Clemson University

TigerPrints
All Theses Theses

12-2015

Analysis of Citrate Content in Bone Using High

Performance Liquid Chromatography
Matthew Konstantin Pysh
Clemson University, [email protected]

Follow this and additional works at: https://tigerprints.clemson.edu/all_theses

Part of the Biomedical Engineering and Bioengineering Commons

Recommended Citation
Pysh, Matthew Konstantin, "Analysis of Citrate Content in Bone Using High Performance Liquid Chromatography" (2015). All
Theses. 2297.
https://tigerprints.clemson.edu/all_theses/2297

This Thesis is brought to you for free and open access by the Theses at TigerPrints. It has been accepted for inclusion in All Theses by an authorized
administrator of TigerPrints. For more information, please contact [email protected].
 ANALYSIS OF CITRATE CONTENT IN
BONE USING HIGH PERFORMANCE
LIQUID CHROMATOGRAPHY

A Thesis
Presented to
the Graduate School of
Clemson University

In Partial Fulfillment
of the Requirements for the Degree
Master of Science
Bioengineering

by
Matthew Konstantin Pysh
December 2015

Accepted by:
Melinda K. Harman, Ph.D., Committee Chair
Mark Schlautman, Ph.D.
Katherine Weisensee, Ph.D.
 ABSTRACT

Approximately 1.5 wt. % of bone is comprised of citrate molecules bound to the

surface of apatite crystals. Furthermore, 80 to 90% of physiological citrate is contained

within bone. Recently, interest in citrate has increased due to the potential of a citrate

based method for estimation of postmortem intervals, time since death, of skeletal

remains. The broad objective of this research was to develop and validate a High-

Performance Liquid Chromatography (HPLC) method for quantitative and qualitative

analysis of citrate in bone.

An appropriate HPLC column and operating conditions for the detection of citrate

were selected and subsequently used to analyze standard solutions consisting of sodium

citrate dihydrate in water from 0 to 2.5 mM, a physiological range, for linearity, impact

of storage time, and intra-day test method precision. After testing on simple matrices, an

existing bone processing protocol was optimized by determining the optimal mass of

sample, presence of matrix effects, and most suitable bone type to sample (cortical or

trabecular). Finally, the HPLC method and optimized processing method were combined

to analyze the impact of short term (14 day) storage conditions, intra-bone factors

associated with sampling at different anatomical locations, and applicability to a model

for the determination of postmortem interval (PMI).

The HPLC method developed in this study demonstrated acceptable linearity over

a physiological range of standards (0 to 2.5 mM) in all sets of standards analyzed (n =11).

Also, the method provided repeatable results under normal operating conditions. Analysis

of three bone masses (50, 75, and 100 mg) and different bone types (cortical and

ii
trabecular) demonstrated 75 mg, cortical only bone samples were most suitable for

analysis. Furthermore, standard addition and recovery tests indicated the presence of

matrix effects within the processing protocol of the bones samples. Finally, a total of 36

bone specimens were prepared with the optimized processing protocol and analyzed for

citrate. Chromatograms of each sample were analyzed for peak resolution and separation,

as well as used to calculate the normalized concentration of citrate (wt. %.). Postmortem

intervals were then calculated using a previously published citrate degradation model by

Schwarcz et al. in order to assess precision. The mean concentration of citrate in all

samples analyzed in this thesis was 1.12 ± 0.6 wt. %, which translated into a mean

calculated PMI of 7.93 ± 13.72 years. The results from this thesis demonstrate the need

for a better model of PMI estimation, as well as further research on the distribution of

citrate in bone.

iii
 DEDICATION

This thesis is dedicated to my family for their support and encouragement

throughout my entire academic career. I would not be where I am today without them. In

addition, I would like to dedicate this thesis to all of those who have helped me grow by

challenging me to do my best in everything that I do.

iv
 ACKNOWLEDGMENTS

First, I would like to acknowledge my research advisor, Dr. Melinda Harman, for

her guidance and support throughout this project. I would like to also thank her for

helping me to grow as a scientist and engineer. Second, I would like to thank Dr. Mark

Schlautman for his willingness to meet with me and share his knowledge and ideas on

challenges faced throughout the project. I would also like to thank Dr. Katherine

Weisensee for her input and support throughout this project. Next, I would like to thank

Dr. Guzeliya Korneva for all the assistance she provided in running our HPLC samples

and troubleshooting issues encountered in the testing process. I would also like to thank

Maria Torres for her encouragement throughout this process. Also, thank you to the

faculty and staff of the Department of Bioengineering and CUBEInC, for helping me to

grow in and out of the classroom. Lastly, thank you to the members of the ReMED

Laboratory for their input and support throughout this project.

v
 TABLE OF CONTENTS

Page

TITLE PAGE .................................................................................................................... i

ABSTRACT ..................................................................................................................... ii

DEDICATION ................................................................................................................ iv

ACKNOWLEDGMENTS ............................................................................................... v

LIST OF TABLES ........................................................................................................viii

LIST OF FIGURES ....................................................................................................... xii

CHAPTER

1 Motivation ...................................................................................................... 1

2 Literature Review........................................................................................... 4

2.1 Bone .................................................................................................. 4

2.2 Chemical Composition of Bone ......................................................... 7
2.3 Citrate ............................................................................................... 13
2.4 Citrate in Bone ................................................................................ 16
2.5 Weathering of Skeletal Remains Post Mortem ................................ 26
2.6 Postmortem Interval Estimation ...................................................... 27
2.7 Analytical Techniques Used to Measure Citrate Concentration ..... 36
2.8 High Performance Liquid Chromatography .................................... 39
2.9 Research Objectives ......................................................................... 44

3 Methods and Materials ................................................................................. 46

3.1 Introduction ..................................................................................... 46

3.2 Preparation of Chemical Solutions .................................................. 47
3.3 Execution of Chemical Analysis using High Performance
Liquid Chromatography ............................................................... 60
3.4 Selection and Preparation of Bone Sample Solutions ..................... 73
3.5 Quantitative and Qualitative Analysis of Citrate Concentrations
Using HPLC ................................................................................. 87

vi
Table of Contents (Continued) Page

4 Results ........................................................................................................ 101

4.1 Introduction .................................................................................... 101

4.2 HPLC Method and System Performance ....................................... 101
4.3 Validation of HPLC Method Acceptability with Standards .......... 110
4.4 Analysis of Citrate Peaks in Bone Solution Chromatograms ........ 124
4.5 Applied Test with Citrate HPLC Method ...................................... 131

5 Discussion .................................................................................................. 181

5.1 Introduction .................................................................................... 181

5.2 HPLC Test Method and Validation on Standards (Aim 1) ............ 182
5.3 Selection and Preparation of Bone Specimens (Aim 2)................. 186
5.4 Utilization of Bone Preparation Protocol and HPLC Method
for Qualitative and Quantitative Analysis of Citrate
Content in Bone (Aim 3)............................................................ 198
5.5 General Observations on HPLC Performance ............................... 202
5.6 Overview of Expenses ................................................................... 203
5.7 Future Work ................................................................................... 205

6 Summary and Broad Significance ............................................................. 210

6.1 Summary ........................................................................................ 210

6.2 Broad Significance ......................................................................... 214

APPENDICES ............................................................................................................. 217

A: List of Materials Used Throughout Thesis ................................................ 218

B: Peak Area and Retention Time for All Standard
Solutions Used for Calibration Curve
Linearity Analysis in Chapter 4 ........................................................... 220
C: Mass of Bone, Peak Area, Retention Time, Bone
Type, and Citrate Concentrations (mM and
wt.%) of all Bone Solutions Examined in
Thesis ................................................................................................... 232
D: Standard Addition Calibration Curves Produced in
Aim Three Bone Solutions................................................................... 238

REFERENCES ............................................................................................................ 247

vii
 LIST OF TABLES

Table Page

2.1 Primary Cells Mineralized in Vertebrate Bone

(Adapted from Boskey6) ........................................................................ 12

2.2 List of Reported Citrate Concentrations in

Tissues (adapted from Costello11).......................................................... 16

2.3 Literature Values of Citrate Concentration (wt. %) in

Bone ....................................................................................................... 17

2.4 List of Important Studies on Citrate in Bone ............................................... 19

2.5 List of Relevant Studies on Citrate Degradation

Methods of Detection ............................................................................. 28

2.6 Determination of Error (days) in Exhumed Human

Remains Using Schwarcz Method of PMI
Estimation .............................................................................................. 31

2.7 Citrate Concentrations in Porcine Ribs Analyzed

at Various Time Points Adapted from
Schwarcz et al. ....................................................................................... 31

3.1 Expected Mass (mg) of Citrate in Various Weights

of Bone Powder..................................................................................... 51

3.2 Expected Number of Moles Citrate in Various

Weights of Bone Powder ....................................................................... 52

3.3 Expected Concentrations of Citrate in Bone ................................................ 52

3.4 Preparation Volumes needed for all Concentrations

[0-2.5 mM] in a Set of Standard Citrate Solutions ................................ 54

3.5 List of Chemical Solutions Used throughout Thesis ................................... 60

3.6 List of Selected Columns for HPLC Analysis of Organic

Acids ...................................................................................................... 62

viii
List of Tables (Continued)

Table Page

3.7 Summary of Four Key Attributes in Terms of Similarity

between Animal and Human Bone ........................................................ 75

3.8 Estimated Volumes of 1.0 M KOH (Base) needed to

reach pH 2 After Bone Digestion........................................................... 85

4.1 Peak Areas and Retention Times for Each Concentration

of Standard Set 08A at Time=0 ........................................................... 111

4.2 Determination of Linearity for All Analyses of Citrate

Standards Based on R2 and y-intercept
Percentages of Max response ............................................................... 114

4.3 Set 4A Repeatability Data of 10 injections of

0.5 mM Standard Solution ................................................................... 116

4.4 Set 4B Repeatability Data of 10 injections of

0.5 mM Standard Solution ................................................................... 116

4.5 Average Peak Area for All to (0 weeks) Standards

Analyzed .............................................................................................. 118

4.6 Average Peak Area for all tf (≤4 weeks) Standards

Analyzed .............................................................................................. 118

4.7 p-Values Obtained from paired t-tests of peak areas

of fresh (t=0 weeks) and aged (t = ≤ 4 weeks) ..................................... 119

4.8 Comparison of Slopes Obtained in Calibration Curves

of Fresh and Aged Standards .............................................................. 119

4.9 Slope and R2 Values of Stock Solutions with

pH 9 Prepared in this Thesis ................................................................ 122

4.10 Slope and R2 Values of Stock Solutions with

pH 2 Prepared in this Thesis ................................................................ 122

4.11 Peak Areas of Standard Concentrations Prepared

from pH 2 Stock Solutions ................................................................... 123

ix
List of Tables (Continued)

Table Page

4.12 Peak Areas of Standard Concentrations Prepared

from pH 9 Stock Solutions ................................................................... 123

4.13 Results of Unpaired t-test Analysis for Standards

Prepared with pH 2 and pH 9 Stock Solutions .................................... 123

4.14 Peak Areas of Citrate Peak in 75 mg Cortical Only Bone

Solution Using Different Fits of Integration from
Waters Breeze Software ....................................................................... 128

4.15 List of Citrate Concentrations (wt.%) for Three

Consecutive Injections of Mixed Bone
Specimens ............................................................................................ 130

4.16 Average Peak Areas and Concentrations (mM and wt. %)

of Citrate in 50, 75, and 100 mg Mixed Bone Samples ....................... 136

4.17 Peak Areas for Standard Addition Set One................................................ 140

4.18 Peak Areas for Standard Addition Set Two ............................................... 141

4.19 Peak Areas for Standard Addition Set Three ............................................. 142

4.20 Slope and R2 Values Obtained in All Mixed Bone

Standard Addition Sets Tested ............................................................. 143

4.21 List of Percent Recovery for Each 75mg Mixed

Bones Samples spiked With Citrate in this
study ..................................................................................................... 152

4.22 Percent Recovery of Citrate Based on Adjusted

Concentrations of Citrate in Cortical Only
Standard Recovery Test Solutions ....................................................... 154

4.23 Citrate Concentrations and Estimated PMI for

Trabecular, Cortical, and Mixed Samples
Analyzed With HPLC Test Method ..................................................... 159

x
List of Tables (Continued)

Table Page

4.24 Calculated wt. % Values of Citrate in Bone Solutions

After Short Term (14 days) Storage in Varied
Environments ....................................................................................... 163

4.25 Calculated Normalized Concentrations (wt. %) of

Citrate in Three Anatomical Regions along a
Porcine Rib Bone ................................................................................. 166

4.26 Estimated PMI Values of Bone Samples Using the

Schwarcz Citrate Degradation Model in Three
Anatomical Regions along a Porcine Rib Bone ................................... 167

4.27 Detected Concentration of Citrate (wt.%) Using Three

Different Filtering Methods ................................................................. 171

4.28 Comparison of Concentrations Obtained by Calibration

Curve and Standard Addition Method for Selected
Storage and Location Tests .................................................................. 173

4.29 Comparison of calculated PMI (years) for Standard

Addition and Calibration Curve Methods of
Concentration Determination ............................................................... 175

4.30 Calculated PMI Using Schwarcz Citrate Degradation

Model for all Bone Specimens Tested in the
Storage Condition Study ...................................................................... 177

4.31 Calculated PMI Using Schwarcz Citrate Degradation

Model for all Bone Specimens Tested in the
Location Study ..................................................................................... 178

4.32 Calculated PMI Using Schwarcz Citrate Degradation

Model for all Bone Specimens Tested in the
Filter Method Study ............................................................................. 179

4.33 Mean Calculated PMI and wt. % for all 75mg

Cortical Only Samples Tested in Thesis .............................................. 180

5.1 List of Equipment and Associated Costs Needed

for HPLC Method ................................................................................ 204

xi
 LIST OF FIGURES

Figure Page

2.1 Identification of Cortical (Compact) and Trabecular

(Cancellous) Bone in Porcine Rib Specimen ........................................... 4

2.2 Chemical Structure of Citrate ...................................................................... 14

2.3 Formation of Citrate in the Kreb’s Cycle .................................................... 15

2.4 Degradation of Citrate in the Kreb’s Cycle ............................................... 15

2.5 Citrate content in Porcine Rib Bones Plotted Versus Days ......................... 32

2.6 Plot of all calculated citrate concentration versus known

PMI (days) using the data obtained in the Schwarcz
and Kanz Studies.................................................................................... 35

2.7 Plot of all citrate concentration versus calculated (predicted)

PMI (days) using the Schwarcz model for data obtained
in the Schwarcz and Kanz Studies ......................................................... 35
13
2.8 C NMR Spectra of Bone, of Organic Residues at the
Interface with Apatite, and of 13C-labeled Citrate
in Bone ................................................................................................... 38

2.9 Diagram of a High-Performance Liquid Chromatography

System .................................................................................................... 42

3.1 Waters Breeze 2 HPLC System Used in this Thesis.................................... 61

3.2 Aminex HPX-87H Column Used in this Thesis .......................................... 64

3.3 Autosampler Tray with Vials Placed in a Specific Test Order .................... 66

3.4 Baseline capture after one hour of equilibration with mobile

phase for multiple wavelengths (top) and single
wavelength (bottom) .............................................................................. 68

3.5 Example of peak integration in the Breeze 2 Software

for a 0.5 mM standard solution .............................................................. 69

xii
List of Figures (Continued)

Figure Page

3.6 Image of Rack of Ribs positioned in a Cranial to

Caudal Manner ....................................................................................... 78

3.7 Images of Bone Pre and Post Defleshing ..................................................... 79

3.8 Section of Bone Acquired for Analysis ....................................................... 80

3.9 Images of Bone Pre (A) and Post (B) Grinding with
Mortar and Pestle ................................................................................... 83

3.10 Overview of the Preparation Process of Porcine Rib

Bone Citrate Solutions ........................................................................... 86

3.11 Image of Bone with Midpoints Marked and

Anatomical Regions Identified .............................................................. 96

4.1 Chromatogram of 2.5mM Standard Solution Prior to

Introduction of Bone Solutions into Column ....................................... 103

4.2 Chromatogram of 2.5mM Standard Solution after the

Introduction of Bone Solutions into Column ....................................... 103

4.3 Chromatogram Produced by Blank Standard

through Column [0 mM] ...................................................................... 104

4.4 Chromatogram of 75mg Cortical Only Bone Sample Obtained at PNORM . 106

4.5 Chromatogram of 75mg Cortical Only Bone Sample Obtained at PMIN .. 106

4.6 UV-Vis Spectra of Four Citrate Standards ................................................ 109

4.7 UV-Vis Spectra of Unknown Peak and Citrate Peak ................................ 109

4.8 Chromatograms of Each Standard Solution Concentration

Analyzed in Set 08A ............................................................................ 111

4.9 Calibration Curve produced by Set of Standards 08A

forced through the origin ..................................................................... 112

xiii
List of Figures (Continued)

Figure Page

4.10 Calibration Curve produced by Set of Standards

4A without being forced through the origin ........................................ 112

4.11 Chromatograms of Cortical Only Bone Sample Obtained

from the Dorsal Region of a Porcine Rib Bone both
spiked with 1.63 mM citrate and unspiked .......................................... 125

4.12 Representative Image of Tangential Skim Integration to

Determine Area under Citrate Peak in
Chromatogram ..................................................................................... 126

4.13 Screen captures of Breeze 2 Integration Fits ............................................. 128

4.14 Overlaid Chromatograms of Three Consecutive

Injections of a Mixed Bone Solution ................................................... 129

4.15 Chromatograms of 50, 75, and 100 mg Mixed Bone

Samples from Optimal Mass Group One
Overlaid................................................................................................ 133

4.16 Citrate Peaks of 50, 75, and 100 mg Mixed Bone

Samples from Optimal Mass Group One
Normalized and Overlaid ..................................................................... 134

4.17 Chromatograms of Original Bone Sample and

Three Additions Overlaid .................................................................... 139

4.18 First-order (Straight line) fit of the Peak Areas

Obtained in Standard Addition Set One............................................... 140

4.19 First-order (straight fit) line of the Peak Areas

Obtained in Standard Addition Set Two .............................................. 141

4.20 First-order (Straight line) fit of the Peak Areas

Obtained in Standard Addition Set Three ............................................ 142

4.21 Comparison of Standard Addition Set 1 Curve with

Standards set 05_A (t=0) forecasted forward ...................................... 144

xiv
List of Figures (Continued)

Figure Page

4.22 First-order (Straight line) fits of three Standard Additions

performed on 75 mg Cortical Only Bone Solutions
and Calibration Curve of Set 08 A ....................................................... 145

4.23 First-order (Straight line) fits of three Standard Additions

performed on 75 mg Cortical Only Bone Solutions
adjusted and compared to Calibration Curve of Set
08 A ...................................................................................................... 146

4.24 Non-adjusted Standard Recovery Calibration Curve for

Standard Recovery Set One Solutions Compared to
Standard Calibration Curve.................................................................. 148

4.25 Adjusted Standard Recovery Calibration Curve for

Standard Recovery Set One Solutions Compared
to Standard Calibration Curve ............................................................. 148

4.26 Non-adjusted Standard Recovery Calibration Curve for

Standard Recovery Set Two Solutions Compared
to Standard Calibration Curve ............................................................. 149

4.27 Adjusted Standard Recovery Calibration Curve for

Standard Recovery Set Two Solutions Compared
to Standard Calibration Curve ............................................................. 149

4.28 Non-adjusted Standard Recovery Calibration Curve for

Standard Recovery Set Three Solutions Compared
to Standard Calibration Curve ............................................................. 150

4.29 Adjusted Standard Recovery Calibration Curve for

Standard Recovery Set Three Solutions Compared
to Standard Calibration Curve ............................................................. 150

4.30 Comparison of each adjusted citrate spiked bone

solution to set 06_A calibration curve ................................................. 151

4.31 Non-adjusted Standard Recovery Calibration Curve

for Cortical Only Sample Compared to Standard
Calibration Curve ................................................................................. 153

xv
List of Figures (Continued)

Figure Page

4.32 Adjusted Standard Recovery Calibration Curve for

Cortical Only Sample Compared to Standard
Calibration Curve ................................................................................. 154

4.33 Chromatogram of 75mg Cortical Only Bone Sample

from Dorsal Region of Porcine Rib Bone ............................................ 157

4.34 Chromatogram of 75mg Trabecular Only Bone Sample

from Dorsal Region of Porcine Rib Bone ............................................ 157

4.35 Chromatogram of 75mg Mixed Bone. Solution from

Standard Addition Experiments done Earlier
in Testing ............................................................................................. 157

4.36 Chromatograms of Three Different Storage Conditions

Overlaid...................................................................................................... 162

4.37 Close in View of Citrate Peaks in Different Locations of the

Same Bone ........................................................................................... 165

4.38 Chromatograms from MWCO Group One of 75mg

Bone Overlaid ...................................................................................... 170

4.39 Plot of PMI and Concentration of Bone Samples in

Storage Condition Study ...................................................................... 177

4.40 Plot of PMI and Concentration of Bone Samples in

Location Study ..................................................................................... 178

4.41 Plot of PMI and Concentration of Bone Samples in

Filtration Method Study ....................................................................... 179

xvi
 CHAPTER 1

MOTIVATION

Estimation of the postmortem interval (PMI), or time elapsed since death, can be

critical in cases involving decomposed human remains. Having a reliable and accurate

estimation of PMI can help identify the individual by narrowing the pool of decedents to

which the remains might belong [40]. In cases where homicide is suspected, PMI can be

used to eliminate suspects and corroborate witness testimony. Furthermore, having a

reliable PMI can help to better account for natural and events and environmental forces

that may have occurred, allowing a more complete taphonomic analysis [40]. Methods

for estimation of PMI by forensic anthropologists and pathologists have traditionally

been highly qualitative in nature. Among these techniques is the use of gross observations

of soft tissue decay help to provide a general time estimate. Anthropologists also rely on

their knowledge of the environmental region to help determine a more narrow range of

PMI [40]. Finally, specific benchmarks can he used to establish or verify maximum or

minimum time estimates [40].

While the techniques described above are well established and accepted, an urgent

need for quantitative methods of PMI estimations using skeletal remains has been

recognized by the Scientific Working Group for Forensic Anthropology (SWGANTH)

housed in the National Institute for Standards and Technology and the Technology

Working Group Operational Requirements. In particular, methods that have a practical,

1
innovative, and useful approach for the determination of time since death of an

individual. SWGANTH stress the necessity of systematic studies determining the key

environmental variables on the decomposition process of skeletal remains. In particular,

they identified two needs of interest: (1) studies that rely on skeletal material instead of

the taphonomic environment and (2) critical review of existing forensic anthropology

methods.

Alternative methods using skeletal remains have been investigated as a means to

circumvent the limitations of the tissue decomposition techniques. One of the most

promising and recent techniques of PMI estimation using skeletal remains has sought to

utilize the degradation of citrate [16, 32, 49].

Each of these studies on citrate degradation has primarily utilized absorption

spectroscopy for analysis of citrate concentrations. In order to make this technique work,

specialized kits for citrate detection are necessary. The results from these kits have not

been shown to be repeatable [16]. In addition, the chemistry involved requires a pH level

where hydroxyapatite precipitation is thermodynamically favored as indicated by the

presence of a precipitate in the sample [16, 21]. Overall, a need was apparent for a better

analytical technique to measure citrate concentrations within bone samples.

The overall objective of this thesis was to develop and validate an HPLC method

for quantitative and qualitative analysis of citrate in bone. In order to achieve this

objective, experiments were developed to achieve three main aims. First, an HPLC

method was developed and tested on a physiologically relevant range of citrate standards.

Second, an existing processing method was tested and optimized to ensure the proper

2
preparation and sampling of bone samples using HPLC. Third, bone samples were

prepared and tested to evaluate the qualitative and quantitative analysis provided by the

HPLC method developed in aim one.

Chapter 2 presents a comprehensive overview of the current research on citrate in

bone. A review of the literature demonstrated the need for better testing methods, as most

of the research to date has relied on the unreliable enzymatic assay kits. Chapter 3,

presents a comprehensive list of the methods and materials used throughout the thesis in

order to achieve the aims of this thesis. Chapter 4 presents the results associated with

each of the three aims defined in this thesis. Chapter 5 contains a discussion of the results

and their significance. Finally, Chapter 6 contains the conclusion and brief discussion of

the broad significance of the work performed in this thesis.

3
 CHAPTER 2

LITERATURE REVIEW AND RESEARCH OBJECTIVE

2.1 Bone

Bone comprises the largest percentage of connective tissue mass in the body, and

unlike other connective tissue matrices, is physiologically mineralized and constantly

regenerated as a consequence of bone turnover throughout its life [47]. Lamellar (Mature)

bone is typified by s well-organized arrangement of collagen fibers, lamellae arranged in

circumferential as well as tubular arrangements (osteons), and strong mechanical

properties [33]. Furthermore, lamellar bone can be classified into two primary types:

cortical or compact, and trabecular or cancellous as seen in Figure 2.1 below [33, 51].

Trabecular

Cortical
2 mm

Figure 2.1: Identification of Cortical (Compact) and Trabecular (Cancellous)

Bone in Porcine Rib Specimen

The adult human skeleton is approximately comprised of 80% cortical and 20%

trabecular bone [51]. While this ratio holds true for the overall skeletal structure, some

bones and skeletal sites have different ratios of trabecular to cortical. In the femoral head,

the ratio is 50/50, while in the radial diaphysis the ratio is 5/95. The vertebra has a ratio

4
of 75/25 [9]. Ultimately, the architecture and amount of cortical bone at any site relate to

the function of the bone in that area [26].

2.1.1 Cortical Bone

Cortical or compact bone is a calcified hard, dense bone that forms the outer layer

of bone around the marrow cavity and the trabecular bone [33, 51]. It is porous but it has

a high ratio of solid tissue to space compared to trabecular bone indicating a higher

porosity in trabecular bone compared to cortical [17, 51, DUTTA]. Furthermore, pores in

trabecular bone are easily visible under the microscope, while pores in cortical bone (10-

20 μm) are not visible to the naked eye [17]. Adult cortical bone is comprised of a

collection of cylindrical units known as the osteon or Haversian system that run parallel

to the outer surface of the bone [33, 39, 51]. A central canal containing a neurovascular

bundle is found in each Haversian system and is known as a Haversian canal. These

canals contain one or more blood vessels that carry blood to and from the osteon [39].

Perforating canals, also known as Volkmann’s canals, extend perpendicular to the

surface. Blood vessels in these canals, supply osteons deeper in bone with blood. Small

spaces known as lacunae separate lamellae, and are connected to each other and the

central Haversian canal by channels known as canaliculi [39, 51]. Cytoplasmic processes

are extended into the canaliculi by osteocytes that are found in the lacunae [51].

A series of nested cylinders around the central canal are formed by lamellae of

each osteon. When these sections are cut transversely, a target like pattern is formed by

the concentric lamellae with the central canal as the core [39]. A spiral is formed by the

5
collagen fibers in each of the lamellae which adds both strength and resiliency to the

bone. Cement lines separate osteons, have a high content of inorganic matrix, and

correspond to areas of bone resorption and deposition [51].

2.1.2 Trabecular Bone

Trabecular or cancellous bone consists of a honey-comb like structure that is

commonly found at the ends of long bones and in the middle of bones in the pelvis,

vertebrae, and other flat bones [51]. Within this type of bone, lamellae are not arranged in

osteons. The matrix of the bone form supporting bundles of fibers in a meshwork known

as trabeculae [39]. A network of rod-and-plate-like elements are formed by trabeculae

and act as scaffolding for the marrow cavity, lighten bone, and allow room for blood

vessels and bone marrow [33]. Neither capillaries nor venules are found within the

trabecular matrix. As a result, nutrients diffuse along canaliculi that open onto the surface

of trabecular in order to reach the osteocytes. Furthermore, blood vessels within the tissue

deliver nutrients and remove waste produced by the osteocytes, and red bone marrow is

distributed between the trabeculae [39].

Trabecular bone is typically located in areas of low stress or in places where

stresses arrive from many directions. In addition, trabecular bone can withstand stresses

applied from many directions, and is much lighter compared to cortical bone. From a

functional standpoint, the lighter weight of trabecular bone reduces the weight of the

skeleton and makes it easier for muscles to move bones [39]. Finally, the trabecular

6
framework provides protection and support to the cells of the bone marrow contained

within the bone.

2.2 Chemical Composition of Bone

Bone is a heterogeneous composite material consisting of (in descending order) a

mineral or inorganic phase, an organic phase, and water [7]. The inorganic phase is

comprised of hydroxyapatite (Ca10(PO4)6(OH)2) that is analogous to geologic

hydroxyapatite [7]. The organic phase is comprised primarily of type I collagen,

noncollagenous proteins, and lipids. Furthermore, bone contains proteins in the

extracellular matrix of bone that can be categorized as follows: (a) structural proteins

(collagen and fibronectin) and (b) proteins with specialized functions [7]. Relative

amounts of these constituents have been noted to vary with age, site, gender, ethnicity,

and health status [7]. Properties of bone are defined by the amount and proper

arrangement of each of these components (quality and quantity) [7]. In general, dry bone

is reported to be anywhere from 30 to 40% organic by weight with the rest, 60 to 70%,

being inorganic components [33, 39]. The inorganic or mineral portion gives rigidity to

the bone, while the organic components provide a slight resilience, which accounts for its

strength [58].

7
 2.2.1 Inorganic Matrix

The inorganic matrix is an important factor in the rigidity and hardness of bone

and consists primarily of calcium phosphate [7, 51, 53]. Also, it contains small but

significant amounts of calcium carbonate, as well as very small quantities of ionic

components such as Mg2+, Na+, K+, Cl-, F-, and citrate [53]. However, it must be noted

that citrate is an organic molecule, and is only included under the inorganic matrix as it is

attached to Hydroxyapatite.

A number of techniques including gravimetric analyses, analysis of calcium and

phosphate contents, bone mineral density (BMD), and micro-computed tomography

(micro-CT) have been used to quantify the mineral present in bone (6). Such analyses

have shown that mineral content of bone ranges from approximately 30% dry weight to

98% dry weight. Most bones have approximately 60 to 70% mineral/dry weight

depending upon site, species, and stage of development [10, 12].

X-ray diffraction studies have determined that a proportion of the calcium

phosphate exists in crystalline form with a three-dimensional atomic configuration

(crystal lattice) similar to geological calcium phosphate minerals, commonly known as

the apatites [53]. Bone apatites are truly hydroyapatites having a lattice structure, poor

crystallinity, and smaller crystal size [53]. The ratio of calcium: phosphorous in these

crystals can vary, and contain besides calcium, the other ionic components listed

previously [53]. Due to the small size, apatite crystals have the property of chemisorption

of excess phosphates and other ions, which leads to absorption of water and the forming

of a hydration shell around mineral crystals [53]. Cation exchange across the hydration

8
shell between Ca2+ ions of the apatite crystal surface and Na+ ions of the extracellular

fluid is believed to account for the significant sodium content in bone [53].

2.2.2 Organic Matrix

The organic matrix is primarily comprised of Type I Collagen that consists of a

repeated tripepetide sequence, which form a left handed helix [33]. These collagen fibers

constitute about 95% of the organic matrix and approximately 40% of the total body

collagen [53]. Bone collagen fibers increase in diameter and become more closely packed

with increasing age [53]. Glycine comprises one-third of the total amino acid content of

bone collagen, while alanine, proline, and hydroxyproline together constitutes one-third

combined [53]. In addition, hydroxylysine, a unique amino acid, is significant within

collagen and is believed to play an important role in ensuring stable interchain linkage

between tropocollagen moieties in bone.

One of the more distinct physical properties of collagen in bone is the extreme

insolubility in reagents which easily solubilize collagen from other sources [53]. A key

example of this can be seen in mineral solutions, as less than 0.5% of the total collagen of

bone is soluble in neutral salts [24]. Furthermore, less than 1% is dissolved in dilute acid

solutions, 0.5 M acetic acid [24, 41]. Although, it has been demonstrated that as much as

30% of bone collagen can be dissolved by repeated freezing and thawing in 3% acetic

acid [24, 53].

The remainder of the organic matrix is comprised of small quantities of complex

protein-polysaccharide moieties as interfibrillar ground substance that constitute 0.5% of

9
bone tissue by weight and are extremely complex in chemical structure and physical

properties [53]. Glycosaminoglycan, sialic acid, etc. are contained within the

carbohydrate portion of these protein polysaccharides while the protein portion contains

40% acidic amino acids. Chondroitin 4 sulphate, chondroitin 6 sulphate and keratin

sulphate are included in the other muco-polysaccharids of bone [53]. The exact

biochemical role of the compounds is not clear, but they are believed to be of significance

in the dynamic functions of bone such as mineralization of the matrix, bone resorption,

and regulation of bone remodeling and morphogenesis of bone [53].

2.2.3 Mineralization of Bone Matrix

Calcification, the hardening of organic tissue due to the deposition of calcium

salts within the tissue substance, of the bone matrix is a distinct characteristic of bone.

Currently, several hypothesis have been developed to explain the calcifiability of bone,

but none seem to be satisfactory [53]. However, the following three facts are well

understood and worth stressing. First, bone mineral crystals are arranged parallel to the

long axis of bone collagen fibrils, suggesting specific structural elements that occur

repeatedly along the collagen fibrils, initiate mineral nucleation and crystal formation.

Second, optimal concentrations of the main ionic components of bone mineral should

occur locally in matrix before mineral nucleation and crystallization occur. Third,

glycolytic enzymes and ATP are essential for mineralization to occur [53].

These facts demonstrate that mineralization of bone matrix is a complex physic-

chemical process initiated by interactions between structural elements in collagen

10
molecules, optimal ionic environment, and ATP. The requirement of ATP appears to

suggest that mineral nucleation is an energy consuming process. This mechanism remains

unknown, however, once this process is initiated, crystal growth proceeds forming the

characteristic hydroxyapatite crystals along the collagen fibrils. These crystals then give

rigidity and strength to bone.

2.2.4 Cellular Regulation of Mineralization Process

The organic matrix that is mineralized is produced by cells, which also control the

flux of ions into the extracellular matrix and register signals that mark the starting and

stopping of the mineralization process. The extracellular matrix surrounding these cells

define both the sites where mineralization will commence and the size to which the

crystals should grow, in addition to providing an oriented surface for mineral deposition

[6]. The size and organization of the collagen fibrils limits the dimensions mineral

crystals can obtain, but without the noncollagenous proteins, mineralization will not

occur in a measurable time period [6].

The main cells involved in the mineralization process of bone are the following:

osteoblasts, osteocytes, and osteoclasts. A summary of the properties and functions of

each of these cell types is provided below and is followed by a more thorough discussion

of each of the cell types (Table 2.1).

11
 Table 2.1 Primary Cells Mineralized in Vertebrate Bone (adapted from Boskey6)

Cell Type Properties and Function

Osteoblast Orchestrates coupling of bone formation and bone remodeling and

synthesizes matrix; round or flat-bone forming cell
Osteocyte Linked to other similar cells by canaliculae; osteoblast surrounded by
mineral
Osteoclast Binds to bone surface and releases acid and enzymes that respectively
remove mineral and matrix in response to signals; multinucleated large-
bone resorbing cell

Ostoeoblasts are the differentiated bone-forming cells that secrete bone matrix

and are versatile secretory cells while retaining the ability to divide. These cells secrete

the collagen and ground substance that comprise the initial unmineralized bone or

osteoid. Osteoblasts are also responsible for calcification of the matrix [48]. Calcification

of the matrix is believed to be initiated by osteoblasts through the secretion of small, 50-

250 nm, membrane-limited matrix vesicles into the matrix. The vesicles are rich in

alkaline phosphatase and are only secreted during the periods in which cells are

producing the bone matrix [48].

Osteocytes are the mature bone cells that are enclosed by the bone matrix that it

previously secreted as an osteoblast. These cells have the ability to synthesize matrix, as

well as resorb it to a limited extent, which is important in the homeostasis of blood

calcium [48]. Death of osteocytes causes a resorption of the bone matrix by osteoclast

activity, followed by remodeling or repair of the bone tissue by osteoclasts. Osteocytes

are smaller than their osteoblast precursors due to the reduced perinuclear cytoplasm

[48].

12
 Osteoclasts function to resorb bone and are large multinucleated cells that when

active rest directly on the surface of bone where resorption is to occur. The portion of

these cells in direct contact with bone is commonly divided into two parts, the ruffled

border and clear zone. The ruffled border is the central region that contains numerous

plasma membrane infoldings forming micro-villous type structures. The clear zone is a

ring like perimeter of cytoplasm that serves to demarcate the limits of the bone area being

resorbed [48]. In addition, osteoclasts contain vesicles that contain hydrolytic enzymes

such as collagenase, which digest the organic components of the bone matrix. Prior to the

start of this process, the bone matrix must be decalcified [48]. The prevailing theory is

that dissolution of the calcium salts occurs through the secretion of organic acids by the

membranes of the ruffled border, which is also aided by low pH favoring the action of

acid hydrolyses [48]. This process creates a local acidic environment in the clear zone

adjacent to the ruffled border that creates a compartment at the site of the ruffled border

where focal decalcification and degradation of the matrix occurs [48].

2.3 Citrate

In general, citrate is any anionic form, salt, or ester of citric acid with a chemical

formula of C6H5O73-, and a structure as seen in Figure 2.2. It is historically known as an

important metabolic byproduct of the Kreb’s Cycle [22].

13
 Figure 2.2 Chemical Structure of Citrate

2.3.1 Kreb’s Cycle

The Kreb’s Cycle, also known as the citric acid cycle or tricarboxylic acid (TCA)

cycle, is a series of reactions used by all aerobic organisms to completely oxidize glucose

derivatives to carbon dioxide. Its overall function is to harvest high-energy electrons from

carbon fuels, but it does not generate a large amount of ATP or oxygen as a reactant [3].

Overall, it involves a number of oxidation-reduction reactions that result in the oxidation

of an acetyl group into two molecules of carbon dioxide. In eukaryotes, these reactions

occur within mitochondria. The following chemical reaction (Eq 2.1) summarizes the

cycle [46]:

𝑎𝑐𝑒𝑡𝑦𝑙 𝐶𝑜𝐴 + 3𝑁𝐴𝐷 + 𝐹𝐴𝐷 + 𝐴𝐷𝑃 + 𝐻𝑃𝑂4−2 → 2 𝐶𝑂2 + 𝐶𝑜𝐴 + 3𝑁𝐴𝐷𝐻 + + 𝐹𝐴𝐷𝐻 + + 𝐴𝑇𝑃 Eq 2.1

Ultimately, the citric acid cycle serves as a final common pathway for the

oxidation of fuel molecules including amino acids, fatty acids, and carbohydrates. It also

serves as an important source for precursors for many building blocks of molecules

including amino acids, nucleotide bases, cholesterol and porphyrin [3].

14
 2.3.2. Citrate in the Kreb’s Cycle Pathway

Citrate is historically known as one of the early intermediates in the Kreb’s cycle

[11]. In the Kreb’s Cycle, Acetyl-CoA undergoes condensation with oxaloacetate giving

up its acetyl group and forming the six carbon molecule citrate. Citrate is then undergoes

a reversible isomerization to form isocitrate. The formation reaction (Figure 2.3) is

catalyzed by citrate synthase, while the reversible degradation reaction (Figure 2.4) is

catalyzed by aconitase.

Figure 2.3 Formation of Citrate in the Kreb’s Cycle. Reaction is catalyzed by citrate
synthase, which causes condensation of acetyl-CoA and oxaloacetate

Figure 2.4 Degradation of Citrate in the Kreb’s Cycle. Reaction is catalyzed by

Aconitase which allows for the reversible isomerization of citrate and isocitrate.

15
 All mammalian cells produce citrate, as it is important within the Kreb’s Cycle

[11]. The sum amount of citrate produced consists of the existing concentration of citrate

plus the citrate that was utilized by cells, commonly referred to as “gross citrate

production”. Another useful measurement is “net citrate production” which is the

component of synthesized citrate not utilized by cells, remaining as accumulated citrate.

In normal cellular metabolism, citrate produced is usually utilized in order to avoid a high

degree of citrate accumulation [11].

2.4 Citrate in Bone

Citrate levels in bone and teeth are roughly 100 to 400 times greater than soft

tissues and plasma, with the only exception being the prostate. As seen in Table 2.2

below, citrate levels in bone and teeth are generally accepted to be in the 20 to 80

μmols/gram range.

Table 2.2 Citrate Concentrations in Tissues (adapted from Costello11). Calculated

Concentrations in wt.% are based around 75 mg sample with the average value for citrate
in the range when provided.

Average Calculated
Location Citrate (μmols/gram)
Concentration (wt.%)
Bone / Teeth ~20-80 0.946
Cartilage ~1 0.019
Blood Plasma ~0.2 0.004
Other Soft Tissues ~0.2-0.4 0.006
Prostate Tissue ~10-15 0.236
Prostate Fluid ~50-150 1.891

16
 Approximately 80 to 90% of all citrate within the body is contained within bone.

Furthermore, a number of studies have reported citrate concentrations in bone from 1.09

to 2.0 dry wt. % of bone [14, 16, 20, 21, 28, 34, 49]. A table is provided with a list of

these studies and the weight percent obtained, as well as the types of bone used (Table

2.3).

Table 2.3 Literature Values of Citrate Concentration (wt. %) in Bone. The measurement
methods and species are further clarified in Tables 2.4 and 2.5.

Citrate wt. Bone

Author Year
% Type
Dickens14 1941 1.60 Cortical
Knuuttilla34 1985 1.09 Cortical
Gibbs21 1991 1.50 Trabecular
Schwarcz49 2010 2.00 Cortical
Hu28 2010 1.20 Cortical
Dunphy16 2014 1.69 Cortical
Froome20 2015 1.12 Cortical
Mean 1.457
SD 0.313

17
 2.4.1 Early studies on citrate in bone

The presence of citrate in vertebrate bone was first reported almost 75 years ago

[11, 14]. Initially, there was a great deal of research in this area, but starting in the mid-

1970s, interest in the topic waned and the topic was rarely studied [11, 49]. Recently,

interest in citrate has resurfaced thanks to technological advances which have allowed for

a better understanding of its importance in the bone matrix. Early literature on the topic is

still important and has proven critical in driving current research. Below is an overview

of some of the important early studies on citrate in regards to bone (Table 2.4).

18
 Table 2.4 List of Important Studies on Citrate in Bone.
Primary Tissue or
Year Source Measurement Method Result
Author (s) Solution
Citrate, Calcium, Citrate, calcium, and phosphate
Kupyer35 1938 N/A Van Slyke manometric apparatus form an insoluble complex
and Phosphate

Steamed Bone Colorimetric pentabromoacetone Reported 1.61 % of dry, fat, and protein
Dickens14 1941 Bone free weight; bone contains approximately
Meal method of Pucher
70% of total citrate in body
Dixon and
1953 Bone Rats Colorimetric Enzyme Reaction Amount of citrate in bone regions
Perkins15
are in inverse ratio to metabolic activities

Reported mean citrate concentration of 1.09

Human Iliac wt.%;
Knuttila34 1985 Cortical Bone Enzymatic Assay Kit
Crest carbonate concentration dependent on citrate
metabolism

Cow and Citrate is critical for the stabilization of the

Hu28 2010 Cortical Bone Nuclear Magnetic Resonance (NMR)
Chicken apatite nanocrystals in bone

Osteoblasts and Citrate is produced by primary human

Franklin19 2015 Human Fluoroenzymatic Method
MSC osteoblasts

Rabbit and Citrate anions help to bridge mineral platelets

Davies13 2014 Bone Nuclear Magnetic Resonance (NMR)
Horse and bone

19
 2.4.1.1 Precipitation of Citrate in Presence of Phosphate and Calcium

Kupyer (1938) was attempting to identify information on the nature of the

unidentified constituents of urine through an analysis of its titration curve [35].

Throughout this process, he determined that it was advantageous to remove citrate from

the urine samples by pretreating the solutions. In aqueous solutions, the precipitation was

found to be incomplete, but in urine it was found to be more complete [35]. This more

complete precipitation was later discovered to be attributed to phosphate present in urine.

Subsequent experimentation determined that citrate, calcium, and phosphate formed a

precipitate under alkaline conditions (base was added to turn litmus blue; pH > 8.3).

Furthermore, this study demonstrated that the concentrations of calcium and of phosphate

in the solutions determine the amount of citric acid precipitated. This was determined by

producing solutions which contained various amounts of calcium chloride, potassium

acid phosphate, and citric acid in centrifuge tubes. Citrate was measured by determining

the amount of CO2 liberated when the solutions were oxidized with KMnO4 in the Van

Slyke manometric apparatus. The study also demonstrated that when no phosphate was

present, practically no citrate was precipitated, while increased amounts of phosphate

corresponded to increased precipitation all the way until complete precipitation. Calcium

was also shown to be necessary in concentrations higher than the amount needed to react

with phosphate and citrate [35]. This study was the first to indicate the potential

importance of citrate in the mineralization of bone.

20
 2.4.1.2 High Levels of Citrate in Bone

Dickens followed Kupyer’s findings by being the first to report extremely high

levels of citrate in bone [14]. In Dickens’ study, citric acid was estimated by the

colorimetric pentabromoacetone method of Pucher et al. with dioxan instead of pyridine

as the final diluent [14, 45]. Both ox bone and steamed bone meal were analyzed in the

study for citrate content, but ox bone was believed to be underestimated due to

difficulties in the crushing of the bone [14]. Steamed bone meal was shown to be

analogous to bone, as the overall mineral composition closely resembled typical values of

human bone. Analysis of the bone meal showed citrate concentrations of 1.61% of dry,

fat, and protein free weight in the bone meal.

Ultimately, this study provided a resource on citrate concentrations in a variety of

solid tissues. Furthermore, Dickens determined that the hard substance of bone contains

citric acid, and hypothesized that it may constitute approximately 70% of the total

concentration of citrate in the body. In addition, he documented lower citrate

concentrations in the marrow and cartilage, as well as variations believed to be caused by

hormonal and dietary conditions. [14].

2.4.1.3 Citrate in Circulating Fluids and at Site of Calcification

Eventually, early research in the field gave way to a study by Dixon and Perkins

in 1952 in which they studied the influence of citrate concentrations present in circulating

fluids and at the site of calcification on calcification mechanisms [15]. Bone of the

enzyme systems concerned with the Kreb’s Cycle were analyzed to determine the role of

21
citric acid in the tissues. Thus, enzymes contributing to the formation and conversion of

citric acid, namely citrogenase, aonistase, and isocitric dehydrogenase, were analyzed in

different regions. The impact on calcification in tibia slices from achitic rats was also

studied [15].

Tibiae from rats were cut into thin sections and used as the animal model in this

study. Citrogenase activity was estimated by using the enzyme system which forms

citrate from oxaloacetate and acetate based upon the technique of Stern and Ochoa [50].

The reaction involves the conversion of acetate to acetyl phosphate, the transfer of the

acetyl group to form acetyl-coenzyme A, and the final condensation of the acetyl group

with oxaloacetic acid to form citric acid. Citric acid was then determined in TCA extracts

by the colorimetric method of Weil-Malherbe and Bone [59]. A standard curve was used

and a standard included with each determination.

The authors determined that a mechanism for high local concentrations of citric

acid within bone existed due to activities of citrogenase and aconitase much greater than

isocitric dehydrogenase when compared to other tissues. Also, citric acid may be co-

precipitated with calcium phosphate during its deposition, and thus may hold calcium in a

complex form [44]. Furthermore, it was determined that high citrate concentrations may

help to reverse the process of calcification and solubilize bone salt already produced.

Finally, the authors suggested that citrate concentrations were in an inverse ratio to the

metabolic activities of the bone region [44]. This research would suggest that citrate

concentrations would vary based on the metabolic activity of the region where bone

samples would occur. This would be problematic for a consistent estimation of

22
postmortem intervals, as there would be uncertainty associated with the sampling of

bone, as bone taken from regions with low metabolic activity would have higher citrate

concentrations and those taken from regions with high metabolic activity would have

lower citrate concentrations.

2.4.1.4 Relation between Citrate and Carbonate in Human Cortical Bone

Knuuttila investigated the correlation between citrate and carbonate in bone

specimens from the iliac crest whose death had been caused by acute coronary disease or

accidents [34]. Cortical bone was separated using a dental drill and analyzed for citrate

content using a citrate enzymatic kit. Carbonate in the samples were also analyzed using

a microdiffusion method. Analysis of the bones yielded a mean citrate concentration of

1.09 wt% (10.9 ±3.1 mg/g) in the 128 samples tested and was found to not depend on

subject age. Further analysis of the correlation between citrate and carbon dioxide yielded

a statistically significant negative correlation, confirmed by regression analysis.

Ultimately, this study suggested citrate had an impact on the binding of carbonate to

apatite structures or that carbonate concentration is dependent on citrate metabolism [34].

2.4.1.5 Citrate Stabilization of Apatite Nanocrystals in Bone

Hu et al. (2010) demonstrated that citrate molecules are strongly bound to and

cover the surface of apatite nanocrystals and are highly conserved across fish, avian, and

mammalian bone [28]. Nanocrystals of apatitic calcium phosphate impart the organic-

inorganic nanocomposite of bone with advantageous mechanical properties. The small

23
thickness of the Nanocrystals (approximately 3-nm) provides the advantageous

mechanical properties and likely help to prevent crack propagation. The authors sought to

obtain a better understanding of the factors controlling the nanocrystals in bone in order

to prevent and treat diseases including osteoporosis that lead to millions of fractures each

year. Nuclear Magnetic Resonance (NMR) was used to show that surfaces of the apatite

crystals in bone are studded with strongly bound citrate molecules. Analysis of the data

produced in this study revealed that bound citrate comprises 5.5 wt% of organic matter in

bone, which allows it to bind more calcium than all noncollagenous proteins in bone due

to more COO- groups. They were further able to demonstrate that the area density of

citrate on bone apatite was around 1/(2 nm)2 [28]. Also, it was shown that bound apatite

interferes with nanocrystal thickening as seen in a wide range of vertebrate orders,

indicating its importance in stabilizing nanocrystal size at the thickness for advantageous

mechanical properties and for fast resorption during bone remodeling.

2.4.1.6 Osteoblasts produce Citrate in Bone

Franklin et al. (2014), sought to demonstrate that osteoblasts are specialized cells

that produce citrate which the bone incorporates into its structure, in contrast to the

prevailing view that citrate in bone is derived from blood [19]. The authors argued that

this was a specialized metabolic capability not commonly occurring in other cells as they

use citrate via the Kreb’s cycle for bioenergetics / metabolic requirements. In osteoblasts,

net citrate production demands mitochondrial citrate oxidation is inhibited so that citrate

can be exported to cytosol and ultimately exported outside of the cell into the

24
extracellular environment. Furthermore, it was believed that osteogenic differentiation of

mesenchyme stem cells (differentiation of stem cells into osteoblasts) was responsible for

osteoblasts capable of producing citrate [19].

In the study, human mesenchymal stem cells (MSC) and osteoblasts were

obtained and studied. Citrate production by the cells was determined by using a

specialized cell culture method followed by a fluoroenzymatic method. Ultimately, the

results demonstrated that citrate is produced by primary human osteoblasts, but not by

undifferentiated mesenchyme cells. The authors determined that osteogenic

differentiation was responsible for citrate producing osteoblasts [19]. Plasma was

eliminated as a source of citrate, making this the first study to demonstrate that

osteoblasts are capable of producing citrate that is incorporated into the bone matrix [19].

In summary, this study brought about the new view that citration by citrate producing

osteoblasts occurs at the same time as mineralization throughout the process of bone

formation.

2.4.1.7 Citrate Bridges between Mineral Platelets in Bone

Davies et al. determined that citrate anions help to bridge mineral platelets and

bone [13]. They further hypothesized that the presence of the anions serves to separate

platelets that have disordered regions in between them rather than stepped

transformations into more ordered regions of mineral. Octacalcium phosphate citrate was

used as a model for the citrate bridging between the layers of calcium phosphate mineral

[13].

25
 Using a number of techniques including NMR, X-ray diffraction, and electronic

structure calculations, the authors worked to establish a quantitative structure for the

material. The model was compared with fresh rabbit and horse bone to validate the OCP-

Citrate model. Overall, the authors were able to compare the models, and felt that the

model of citrate ions incorporated as hydrated layers into the mineral structure was fairly

analogous to the results obtained from the bone samples. Furthermore, the authors

believed that the incorporation of citrate into an OCP-like structure could act to restrict

the growth of mineral crystals and produce a plate-like morphology similar to what is

seen with bone. This architecture works to prevent the formation of large crystals that

would have deleterious mechanical properties; therefore, the authors claim that citrate in

the bone mineral may explain changes in crystallinity for individuals with metabolic

diseases, while also making sense of molecular level bone mechanics [13].

2.5 Weathering of Skeletal Remains Post Mortem

Later stages in decomposition of human remains include skeletonization followed

by extreme skeletonization [30]. At these stages, little or no soft tissues are present on the

skeletal remains. The exposure of these remains to an outdoor context initiates changes in

physical and chemical features through a weathering cascade. Key parameters of bone

weathering are similar to those found in body decomposition, and ultimately lead to loss

of collagen and dissolution of bone mineral [57]. Weathering of bone is best described as

a cascade during which collagen is degraded by microorganisms in the present of water

and oxygen independent of soil ions and pH. Subsequently, reactions with hydroxyapatite

26
and calcium ions in bone mineral that are highly dependent on acidic and calcareous soils

occur [36, 61].

2.6 Postmortem Interval Estimation

Estimation of the postmortem interval (PMI), or time elapsed since death, can be

critical in cases involving decomposed human remains. Having a reliable and accurate

estimation of PMI can help identify the individual by narrowing the pool of decedents to

which the remains might belong [40]. In cases where homicide is suspected, the PMI can

be used to eliminate suspects and corroborate witness testimony. Furthermore, having a

reliable PMI can help to better account for natural and events and environmental forces

that may have occurred, allowing a more complete taphonomic analysis [40]. As

discussed in Chapter 1, PMI estimations by forensic anthropologists and pathologists

have traditionally been highly qualitative in nature, but a need for a more quantitative

approach has been well documented.

This section will discuss studies on the newly emerging quantitative technique for

estimation of PMI based upon degradation of citrate as first described by Schwarz. A

summary of each of the studies discussed is provided below (Table 2.5).

27
 Table 2.5 List of Relevant Studies on Citrate Degradation Methods of Detection

Primary Measurement
Year Tissue or Solution Source Result
Author (s) Method

Enzymatic Assay No difference in citrate content between

Gibbs21 1991 Trabecular Bone Human
Kit sexes; 1.5 wt %

Enzymatic Assay
Schwarcz49 2010 Cortical Bone Porcine Limit of 100 years with 1% error; 2 wt. %
Kit

Accuracy of PMI was "unsatisfyingly low";

32 Enzymatic Assay
Kanz 2013 Cortical Bone Human Absolute difference between known and
Kit
calculated PMI was 19.5 years

Enzymatic Assay High standard deviations and variance in

Dunphy16 2014 Cortical Bone Porcine
Kit samples; 1.69 wt.%

U-V Vis Assay 1.12 wt.% with HPLC; 1.02 wt.% with UV-
Froome20 2014 Cortical Bone Porcine
and HPLC Vis Assay

28
 2.6.1 Investigation of Differences in Citrate Content for Determination of Sex

Citrate content of trabecular bone was analyzed in both archaeological and

modern specimens of unknown origin to determine if variation in time of citrate content

could be used to determine sex of human skeletal remains [21]. Differences in citrate

content of trabecular bone samples were analyzed with U.V-enzymatic citrate lyase

spectrometry at a pH of approximately 7. Results demonstrated that differences in

trabecular citrate content in bone were not due to the testing of different parts of the

body. It was also shown that no significant differences in citrate content between sexes

was obtained to justify the use of this method as a means to determine sex of skeletal

remains. The analysis of citrate content in trabecular bone of recently embalmed

individuals had citrate contents of 1.5 ± 0.1 wt. % [21]. Furthermore, the formation of a

precipitate of some bone in all samples analyzed was reported as the pH approached 7

[21].

2.6.2 Potential of Citrate Degradation Model for Estimation of Post Mortem Interval

Schwartz et al. took the work of Gibbs and applied it to demonstrate a correlation

between citrate degradation and PMI. Portions of rib bone, either pig or human, were

analyzed for citrate content with a citrate lyase method in the form of a colorimetric

enzyme assay kit (Xygen Diagnostics, Burgesssville, ON). This study found that citrate

concentration of bone decreased with remarkable regularity in a number of diverse

environments [49]. Furthermore, results indicated that storage environment had minimal

impact on the rates of citrate loss with the exception of no observed loss at temperatures

29
below 0° C. Results from this study also indicated that citrate levels would reach zero

around 95 years from time of death. Schwarz et al. go on to clarify that this lifetime is

only applicable to samples stored at temperatures greater than 0° C. Samples stored

below this temperature, T<0° C, would be expected to have longer lifetimes. Ultimately,

the authors argue that storage conditions make it hard to predict very long-term behavior

of citrate, but they seem to believe that the rate of decline may demonstrate some increase

[49].

When analyzing the study, it appears the authors believe their data demonstrates a

high degree of correlation between citrate content and PMI once soft tissues had fully

decayed. The method was believed to have an error of approximately 1% of age based on

the only error being an analytical error in the detection of citrate (±0.003%). An equation

was developed to determine the PMI interval based on the calculated citrate level as seen

below (Eq. 2.2).

C(t) = -0.673 x log(t) +3.021 Eq. 2.2

In this equation, C(t) is the citrate content in wt% and t is the time since death in

days [49]. The equation was determined to have an r2 value of 0.986. However, when this

technique was applied to three exhumed human remains to compare calculated values

with known values of PMI, each of the samples exceeded the reported precision of one

month (Table 2.6). Furthermore, one sample had a much higher citrate level than

expected, for which no obvious explanation was available.

30
 Table 2.6 Determination of Error (days and % of age) in Exhumed Human Remains
Using Schwarcz Method of PMI estimation49

Sample Citrate Known PMI Calculated Error Error

(wt.%) (days) PMI (days) (days) (% of age)
52-04 1.12 556 692 + 136 24.5
57-05 1.48 1010 198 - 812 80.4
85-06 1.38 324 280 - 44 13.6

Another potential problem with this study was the use of six different porcine ribs

for determination of citrate degradation in porcine rib samples. When the concentrations

obtained in the study (Table 2.7) is plotted (Figure 2.5) it shows a fairly high degree of

linearity (r2= 0.981), but there is no evidence provide that this is not variation between

bones in each individual rack.

Table 2.7 Citrate Concentrations in Porcine Ribs Analyzed at Various Time Points
Adapted from Schwarcz et al.49

Sample Time (days) Citrate (wt.%)

Tennessee 0 0 1.974
Tennessee 1 30 1.945
Tennessee 2 60 1.821
Tennessee 3 90 1.745
Tennessee 4 120 1.702
Tennessee 5 150 1.590
Tennessee 6 180 1.470

31
 2.5

y = -0.0028x + 2.0004
2 R² = 0.9807
Citrate Content (wt. %)

1.5

0.5

0
0 20 40 60 80 100 120 140 160 180 200
Time (days)

Figure 2.5 Citrate content in Porcine Rib Bones Plotted Versus Days. (Adapted from
Schwarcz49)

Schwarz et al. documented the use of an enzyme based colorimetric method for

the determination of citrate concentration in bone. Also, the authors believed their study

demonstrated the ability to use citrate content of bone to estimate PMI up to about 100

years with a 1% error in the age determination. The error in estimation of the three

samples demonstrated the need for a better model to ensure more precise PMI

estimations. Furthermore, it appeared that this study was limited by a small sample size

and a limited timeframe of decay, indicating the need for further study [49].

32
 2.6.3 Attempt to Validate PMI Estimation with Citrate Method Developed by Schwarcz

Kanz et al. produced a study in 2013 that utilized the citrate method developed

and identified by Schwarz et al. to evaluate the citrate concentrations of human remains.

In the study, temporal bones and femora of 20 individuals buried in wooden coffins and

body bags were subjected to the citrate detection method described by Schwarcz [32].

Namely, enzymatic assay kits were obtained from R-Biopharm AG (Darmstadt,

Germany) that utilized a citrate lyase method. In this method citrate lyase converts citrate

to oxaloacetate and pyruvate; the mixtures react with nicotinamide-adenine dinucleotide

(NADH) to form L-malate and L-lactate. The amount of citrate consumed is equivalent to

the amount of NADH consumed and can be measured at 340nm with a photometer [32].

The use of the buried individuals allowed a direct comparison of the PMI estimate

produced by citrate degradation methods to known PMI.

The results in the study demonstrated a marked difference between the temporal

and femur bone samples of the same individuals from the body bag test group, while

differences were insignificant for those buried in wooden coffins [32]. Furthermore,

similar levels of underestimated in PMI were seen with the analysis of the femora for

both the body bag and wooden coffin test groups. Finally, femurs showed slight

differences between body bag and wooden coffin degradation rates of citrate when

compared to the temporal bone population. From this, it was surmised that cortical bone

from long bones should be preferentially used in citrate-based PMI estimations [32].

In comparing the results of this study with those in Schwarcz, Kanz et al.

observed the following. First, the average absolute difference between known and

33
calculated PMI was 19.5 years. Thus, the accuracy obtained from the Schwarcz equation

was “unsatisfyingly low”. Second, Schwarz et al. plot the data in their study on a log time

scale, then concludes a linear relationship between citrate concentration (wt. %) and PMI

(days). When this data is plotted in a more straightforward manner, citrate concentration

vs. PMI (days) (Figures 2.6 – 2.7) with raw data from both studies included, the

relationship is more difficult to see. The data demonstrate decay in the first 180 days that

is measurable, but less useful over longer burials.

34
 2.5
Citrate Concentration (wt.%)

Schwarcz Porcine Rib

1.5
Schwarcz Human
Kanz Body Bad FEM
1
Kanz Body Bag TEM
Kanz Coffin FEM
0.5 Kanz Coffin TEM

0
0 5000 10000 15000 20000 25000
Time (Days)

Figure 2.6 Plot of all citrate concentration versus known PMI (days) using the data
obtained in the Schwarcz and Kanz Studies (Schwarcz49 and Kanz32).

2.5

Schwarcz Porcine Rib

2 Schwarcz Human
Citrate Concentration (wt. %)

Kanz Body Bag FEM

Kanz Body Bag TEM
1.5 Kanz Coffin FEM
Kanz Coffin TEM

0.5

0
0 2000 4000 6000 8000 10000 12000
Time (Days)

Figure 2.7 Plot of all citrate concentration versus calculated (predicted) PMI (days)
using the Schwarcz model for data obtained in the Schwarcz and Kanz Studies
(Schwarcz49 and Kanz32).

35
 2.7 Analytical Techniques Used to Measure Citrate Concentration

Analysis of techniques used to measure citrate concentrations demonstrated that

the most prevalent technique is the enzymatic assay kit method as used in both the

Schwarcz et al. and Kanz et al. studies [32, 49]. Literature further indicated the need for

alternative methods of citrate detection due to precipitation and reliability issues with the

method [32, 49]. This section will briefly introduce the use of the enzymatic assay kit and

issues associated with the method presented by Dunphy [16]. Then a brief survey of

analytical techniques commonly used to measure citrate concentrations will be provided

in order to demonstrate three alternative techniques to enzymatic assays that would be

suitable for the measurement of citrate. The three methods discussed are: Capillary

Electrophoresis, Nuclear Magnetic Resonance, and High-Performance Liquid

Chromatography.

2.7.1 Evaluation of Enzymatic Assay Kit Performance

In this study, Dunphy set out to further evaluate the potential of enzymatic based

citrate detection using absorption spectroscopy and high-performance liquid

chromatography [16]. Bone solutions were prepared using the method developed by

Schwarcz et al. and analyzed using both the enzymatic assay kits and HPLC. Citrate

concentrations of 1.26 and 1.39 wt. % were reported with the colorimetric assay kits,

while a concentration of 1.69 wt. % was found using HPLC. These values, when

compared to the 2010 study Schwarz et al., were not as high with the best recovery, 85%,

being seen with HPLC. Furthermore, the results demonstrated high standard deviations

36
and intra-sample variance indicating the need for a more thorough investigation of the

sample preparation techniques, as well as the need for critical evaluation of the enzymatic

assay kits compared to other techniques to quantify citrate concentrations.

2.7.2 Alternative Methods for Measurement of Citrate

The first alternative method investigated was the use of capillary electrophoresis.

In this technique anions are injected into a thin capillary glass tubing in the form of a

small plug that moves rapidly toward an anode generated by a negative power supply.

Capillary electrophoresis is sensitive, reproducible, and rapid [26]. In a study by Holmes

et al. citrate concentrations in urine were investigated using capillary electrophoresis;

however, this analytical technique has not been used in regards to citrate in bone.

The second technique identified was Nuclear Magnetic Resonance (NMR)

spectroscopy. NMR spectroscopy uses electromagnetic radiation to cause transitions

between energy levels. The range of NMR transition frequencies is typically anywhere

from 10-1000 MHz. NMR signal intensities depend upon differences in the populations

of energy levels [27].

Hu et al used NMR to investigate citrate bonded to apatite surfaces. According to

the study, NMR was able to unambiguously identify the signals generated by citrate in

bone. This study was able to determine that citrate accounts for 5.5 wt% of the organic

matter in bone [28]. An example NMR spectra of bone obtained in the study is provided

below (Figure 2.8).

37
 Figure 2.8 13C NMR Spectra of Bone, of Organic Residues at the Interface with
Apatite, and of 13C-labeled Citrate in Bone. (A) Fish bone (B) avian bone (C) Bovine
Bone (D) Spectra of 13C near 31P in bovine bone. (Image taken from Hu 201028,
PNAS)

2.7.3 High Performance Liquid Chromatography

The third and final analytical technique investigated was the use of High-

performance liquid chromatography (HPLC). HPLC is an analytical technique that

utilizes a high pressure to force solvent through closed columns containing very fine

particles that give high-resolutions separations [23].

Dunphy, et al. demonstrated that HPLC had a higher recovery rate when detecting

citrate in bone compared to absorption spectroscopy. The study also demonstrated that

HPLC was able to detect citrate levels that were closer to those reported in literature by

Schwarz and Gibbs [16, 21, 49].

38
 2.7.4 Method Comparison

A thorough evaluation of these techniques showed HPLC to be the best of the

aforementioned analytical methods for this particular purpose. This was confirmed in

literature, in particular Dunphy who highlighted a number of advantages with HPLC for

citrate detection in bone compared to absorption spectroscopy [16]. Furthermore, it was

determined to be a widely available analytical tool, and one commonly used in forensic

investigations. In addition, HPLC equipment was available for use in this thesis, while

both NMR and capillary electrophoresis equipment were unavailable. As such, HPLC

was selected for use throughout this thesis as an alternative analytical technique for the

determination of citrate content in bone.

2.8 High-Performance Liquid Chromatography

In general, chromatography has been accepted to universally cover the science of

separations. Essentially, the term is used to describe techniques that enable samples of

chemical mixtures to be separated by exploiting differences in physical or chemical

properties [42]. These differences govern the rate of migration of the components within

the mixture that are being influenced by a moving fluid a “bed” of stationary phase [42].

Finely ground solids or liquid coating on the solids may form this stationary phase and

are contained within a tube of metal or glass which is known as the chromatographic

column [42].

Column chromatography deals with the separation of a mixture of components by

establishing conditions under which individual components flow at different rates

39
through the packed column. A mobile phase is used to flow the sample through the

column, and is known as elution of the sample from the column [42].

Differential rates of elution arise from the interaction between the components of

the sample and the material or coating used to pack the column. In general, there are four

principles by which components of a sample are selectively retained [42]. The first of

these is liquid-liquid chromatography, which exploits differences in the partition

coefficients. The second is liquid-solid chromatography that utilizes absorption effects on

surfaces such as silica gel. The third principle is ion-exchange chromatography, which

takes advantage of dissociation of weak or strong electrolytes. The final principle is steric

exclusion chromatography, which is based on molecular size or shape [42].

Retention is defined to be the interaction of a sample with the column packing. In

any chromatographic system, the degree of retention of a compound is a characteristic of

that sample as it depends on the size, solubility, absorption, and ionization characteristics

of that compound in the specific environment in which the system is set up [42].

HPLC is the most employed chromatographic technique due to its importance in

the pharmaceutical industry. It is estimated that HPLC accounts for 60% of the

worldwide separation science market, while Gas Chromatography (GC) accounts for 35%

[38]. Overall, HPLC is the most powerful of all the chromatographic techniques as it is

able to perform analyses and separations that would be challenging or impossible for

other forms of chromatography [37].

40
 2.8.1 High Performance Liquid Chromatography Systems

A typical HPLC system consists of a solvent delivery system, an autosampler

with an autoinjector, a detector, a high-pressure column, and a computer control system

with integrated software to control the system and data output as seen below (Figure 2.9).

An oven to control the column temperature is also included with many commercial

HPLC systems [23].

41
 HPLC Column

Injector

AutoSampler

Sample Manager

Computer Data Station

Solvent
Sample
(Mobile Phase)
Detector
Reservoir
Pump

Solvent Manager Waste

Solvent Delivery System

Figure 2.9 Diagram of a High-Performance Liquid Chromatography System. The main elements of HPLC systems are
included namely the pump, column, detector, autoinjector, as well as the computer data station to control test parameters.

42
 2.8.2 High-Performance Liquid Chromatography Method Development

Prior to carrying out quantitative experiments on an analyte in a set of samples, it

is necessary to develop a suitable method. The problem addressed by the method needs to

be clear defined, then all available information regarding that problem need to be

collected. Information that is typically gathered includes the details of the physiochemical

properties of the analyte, the physiochemical properties of other components in the

matrix, and the features both good and bad of the available methods [38].

In some ways, method development is considered an art that can only be mastered

through years of experience, but in reality, a simple thought process could be used [38].

In the case of citrate, this process revealed the need for an ion-exchange column.

Following the identification of a column type, testing parameters for the HPLC need to

be selected [38].

After the column and testing conditions have been determined, manipulation of

experimental variables relating to mobile phase is typically done to optimize the method.

Variables in this instance may include the nature and percentage of organic components,

pH, ionic strength, nature and concentration of mobile phase additives, and temperature

[38].

43
 2.8.3 Citrate Detection Using HPLC

A review of the literature indicated the use of HPLC for the detection of sildenafil

citrate, a common pharmaceutical [2]. This study was more focused on sildenafil rather

than citrate, so the method provided was not determined to be suitable for this study.

Dunphy utilized HPLC to evaluate citrate concentrations in bone, but the testing

parameters of the HPLC and the column used were not clearly defined. A survey of

columns available for detection of citric acid was done to help get some starting

parameters from literature provided by the company.

2.9 Research Objectives

The overall objective of this thesis was to develop and validate an HPLC

method for quantitative and qualitative analysis of citrate in bone. In order to achieve this

objective, experiments were developed to achieve three following three aims.

(1) An HPLC method was developed and tested on a physiologically relevant

range of citrate standards

(2) An existing bone specimen processing method was optimized for use with

HPLC to ensure the proper suitable bone samples were used with the HPLC

method

44
 (3) Bone samples were prepared and analyzed both qualitatively and

quantitatively using the HPLC method developed in Aim one and the Bone

Processing protocol developed in Aim two.

Aim 1 was accomplished through an array of experiments performed on a

physiologically relevant range of citrate standards (0 to 2.5 mM). In particular, linearity,

storage time of standards, stock solution pH, and intra-day test method precision

(repeatability) were examined. Aim 2 was accomplished by designing experiments to

qualitatively and quantitatively analyze citrate detection in three different bone masses

(50, 75, and 100 mg). Also, matrix effects were also analyzed through the design and

execution of standard addition and standard recovery tests. Finally, different types of

bone (cortical, trabecular, and mixed) were analyzed to determine the optimal type of

bone and limitations with the HPLC test method. Aim 3 was accomplished by analyzing

different short term (14 day) storage conditions (hydrated, dehydrated, and room

temperature), intra-bone factors associated with sampling at different anatomical

locations (dorsal, central, and ventral), and an analysis of the calculated PMI produced by

the method described by Schwarcz et al. for all 75 mg cortical only bone samples

analyzed throughout this thesis [49].

45
 CHAPTER 3

MATERIALS AND METHODS

3.1 Introduction

The methods established and used throughout this thesis fall into four general

subsections, namely the preparation of chemical solutions, quantitative and qualitative

analysis of simple citrate matrices with HPLC, selection and processing of bone

specimens, and the quantitative and qualitative analysis of citrate in bone with HPLC.

Subsection 3.2 “Preparation of Chemical Solutions” includes methods for the

generation of standard solutions, mobile phase for HPLC, and other chemical solutions

needed throughout the thesis. Subsection 3.3 “Quantitative and Qualitative Analysis of

Simple Citrate Matrices using HPLC” discusses methods used in the analysis of standard

solutions comprised of citrate in water including storage time studies, HPLC test

parameters, and HPLC performance. Subsection 3.4 “Selection and Processing of Bone

Samples”, provides methods used throughout the selection and preparation of bone

specimens. This section also includes justification for the use of a porcine model in this

thesis as well as a discussion of simple quantitative and qualitative methods used to

optimize the processing protocol of bone specimens. Finally, subsection 3.5 “Quantitative

and Qualitative Analysis of Citrate Concentration in Bone Using HPLC” discusses

methods used for the detection and quantitation of citrate in bone specimens. In addition,

46
qualitative and quantitative methods used to further optimize the sample selection and

analytical technique are discussed.

3.2 Preparation of Chemical Solutions

In the section below, methods and materials needed for the chemical

solutions required in this thesis are provided. Furthermore, the methods and rationale

used for the preparation of the stock citrate solutions used for standard preparation is

provided. Chemicals specific to the processing protocol will be listed in section 3.4.

3.2.1 Chemicals

Laboratory grade Trisodium Citrate Dihydrate (Na3C6H5O7.2H2O) in granular

form (S279-500, Thermo Fisher Scientific, Waltham, MA) was used to generate

calibration standards. Furthermore, it was used throughout standard addition and standard

recovery tests to spike samples. After use, the container provided by the manufacturer

was sealed by tightly securing the cap on the container. The container was then left on the

counter at room temperature (22ºC).

Laboratory grade 1N (0.5M) Sulfuric Acid (H2SO4) (SA212-1, Thermo Fisher

Scientific) was selected for use in this thesis for all solutions requiring sulfuric acid. The

main uses of this chemical were pH reduction of standard solutions to prevent

microbiological activity during storage, and the preparation of the 5mM H2SO4 mobile

phase used in HPLC analysis. Laboratory grade 1N (1.0 M) Hydrochloric Acid (HCl)

(SA48-1, Thermo Fisher Scientific) was selected for use in this thesis for all solutions

47
requiring hydrochloric acid. The main uses of this chemical were demineralization of

bone specimens in order to release citrate into solution. After use, the lid was secured

tightly onto each bottle provided by the manufacture and stored in an acid storage cabinet

at room temperature (22ºC) within the lab.

Unless otherwise designated, all chemical solutions requiring water were prepared

using ultrapure deionized water (18.2 MOhm-cm resistivity at 25°C, 2ppb TOC) from a

designated water filtration system (MilliQ, EMD Millipore, Billerica, MA) hereafter

referred to as deionized water. Fresh water was obtained from the system for each

preparation of standards as well as all other uses of water throughout this thesis.

3.2.2 Stock Solution Preparation

In most instances, 100mL of 0.1M stock citrate solution was prepared for each

run to ensure adequate solution for all experiments. Later on in the study, it became clear

that 100mL was not necessary, so calculations were adjusted accordingly for 25 and

50mL volumes. The following calculations are based on 100mL, as this was the most

often prepared volume.

3.2.2.1 Calculations Used to Prepare Stock Solution

Stock solutions of 0.1M citrate were prepared by dissolving granular form

Trisodium Citrate Dihydrate in deionized water. The appropriate mass of Trisodium

Citrate Dihydrate to be added was calculated using its molar mass of 294.1 g/mole as

demonstrated in the following equation (Eq. 3.1):

48
 𝑔 𝑚𝑜𝑙 𝑔
294.1 𝑚𝑜𝑙 ∗ 0.1 = 29.41 𝐿 Eq. 3.1
𝐿

The subsequent calculation determined the amount of Trisodium Citrate

Dihydrate needed for 100 mL of solution (Eq. 3.2)

𝑔 1𝐿
29.41 ∗ ∗ 100𝑚𝐿 = 2.941 𝑔 Eq. 3.2
𝐿 1000 𝑚𝐿

Based on these calculations, 2.941g of Sodium Citrate Dihydrate powder was

weighed on a precision (0.0001 grams) balance (Adventurer Pro, Ohaus, Parsippany,

New Jersey) and placed into a 50mL glass beaker (02-540G, Thermo Fisher Scientific)

containing approximately 20mL of deionized water. The solution was mixed for 15

minutes by inserting a clean magnetic stir bar into the solution and placing the beaker

onto a magnetic stirrer plate (6795-220, Corning).

Once mixed, the solution was carefully transferred into a 100mL volumetric flask

with a precision of ±0.1mL (89000-404, VWR). Four of the stock solutions prepared in

this thesis (1, 2, 6, and 10) were prepared by adding only deionized water to the mark on

the flask, resulting in a basic stock solution with a pH of approximately 9. The remaining

stock solutions (3, 4, 5, 7, 8, 9, and 11) were prepared at a pH of 2 in order to match the

pH of the mobile phase. This was achieved by adding 34mL of 0.5M H2SO4 into a

volumetric flask containing the dissolved citrate solution. Finally, deionized water was

49
added to the volumetric flask until the meniscus was level with the 100mL line of the

flask using a sterile polypropylene transfer pipette (414004-008, VWR, Radnor, PA).

A cap was then placed onto the volumetric flask and the solution was further

mixed by gentle shaking for approximately two minutes. Immediately after, the 0.1M

stock solution was transferred for storage into a Pyrex 100mL glass container with screw

top cap (1396-100, Corning, Inc.) and stored at 5°C until use.

3.2.3 Preparation of Citrate Standards

This section introduces the methods and materials used to prepare the citrate

standards used throughout the thesis. In addition, a discussion of the rationale for the

range of standards used is provided.

3.2.3.1 Concentration Range

Prior studies have suggested maximum citrate concentrations of approximately

1.6 wt. % in trabecular bone and approximately 2.0 wt. % in cortical bone [11, 21, 49].

An average of these two values was used to generate a theoretical concentration of 1.8

wt. % within a complete bone specimen that combines equal parts trabecular and cortical

bone. A simple calculation was then performed to determine the expected mass of citrate

within a bone specimens containing the maximum wt. % for a given type and mass of

bone (Eq. 3.3).

𝑀𝑎𝑠𝑠 𝑜𝑓 𝐵𝑜𝑛𝑒 (𝑚𝑔) ∗ 𝑤𝑡% = 𝐸𝑥𝑝𝑒𝑐𝑡𝑒𝑑 𝑚𝑎𝑠𝑠 (𝑚𝑔) Eq. 3.3

50
 Using Equation 3.3, the maximum expected mass of citrate in bone samples

based on literature values for cortical and trabecular [21, 49] with weights varied from 50

mg to 100 mg were calculated. A maximum value of 1.5 wt.% for trabecular bone and a

maximum value of 2.0 wt.% for cortical bone were used. These values were averaged to

obtain a maximum value of 1.75 wt.% for combined (cortical and trabecular).

Table 3.1: Maximum Expected Mass (mg) of Citrate in Various Weights of Bone Powder

Bone Type 50 mg 75 mg 100 mg

Trabecular 0.750 mg 1.125 mg 1.500 mg

Cortical 1.000 mg 1.500 mg 2.000 mg

Combined Trabecular & Cortical 0.875 mg 1.313 mg 1.750 mg

Similarly, the expected number of moles of citrate within the different bone

sample weights and types were calculated using the molar mass of Citrate (189.1 g/mol)

and the following equation (Eq. 3.4).

10000 µmol
* Expected mass citrate (mg)=Expected moles citrate (µmol) Eq. 3.4
1891 mg

The expected number of moles of citrate in bone samples were calculated for the

different types and masses of bones to be used in this thesis (Table 3.2).

51
 Table 3.2: Expected Number of Moles Citrate in Various Weights of Bone Powder
Bone Type 50 mg 75 mg 100 mg
Trabecular 3.97 µmols 5.95 µmols 7.93 µmols
Cortical 5.29 µmols 7.93 µmols 10.58 µmols
Both 4.63 µmols 6.94 µmols 9.25 µmols

The expected number of moles from Table 3.2 were subsequently used to

determine molar concentrations based on an expected final solution volume of 5 mL as

seen in the following equation (Eq. 3.5) and applied below (Table 3.3).

𝐸𝑥𝑝𝑒𝑐𝑡𝑒𝑑 𝑀𝑜𝑙𝑒𝑠 (𝑢𝑚𝑜𝑙𝑠)

= 𝐶𝑜𝑛𝑐𝑒𝑛𝑡𝑟𝑎𝑡𝑖𝑜𝑛 (𝑚𝑀) Eq. 3.5
𝐸𝑥𝑝𝑒𝑐𝑡𝑒𝑑 𝐹𝑖𝑛𝑎𝑙 𝑣𝑜𝑙𝑢𝑚𝑒 (𝑚𝐿)

Table 3.3 Expected Concentrations of Citrate (mM) in Bone

Type of Bone 50mg 75mg 100mg

Trabecular 0.79 mM 1.19 mM 1.59 mM

Cortical 1.06 mM 1.59 mM 2.12 mM
Both 0.93 mM 1.39 mM 1.85 mM

Overall, a citrate concentration ranging from 0 to 2.5mM was determined to be

sufficient enough to cover the expected concentrations of citrate in bone based on the

previously reported concentration of citrate in bone discussed above [11, 21, 49].

52
 3.2.3.2 Preparation of Standards

Seven concentrations of citrate standards were selected and prepared for the

generation of calibration curves. They were as follows: 0 mM, 0.1 mM, 0.5 mM, 1.0 mM,

1.5 mM, 2.0 mM, and 2.5 mM, which are hereafter referred to as a set of standard

solutions. In later work, the 0.1mM solution was not included in some sets as it appeared

to have no significant impact on the linearity of the calibration curve obtained. However,

it must be noted that removing the 0.1 mM standard would have an impact on regression

statistics.

The required volume of 0.1M-citrate stock solution required for each

concentration was calculated using the following process. First the following equation

(Eq. 3.6) was solved for V1 based on the desired concentration of the standard.

𝐶1 𝑉1 = 𝐶2 𝑉2 Eq. 3.6

The following values were plugged into the equation to determine V1, the volume

of stock solution to be added for each concentration. C1 was known to be 0.1M as it was

the concentration of the stock solution described in section 3.2.2 above. V2, the volume of

standard solution being prepared, was also known to be 10mL. Inserting these values

produced Eq 3.7 below.

𝐶2 ∗10𝑚𝐿
𝑉1 = Eq. 3.7
0.1𝑀

53
 C2 was the concentration of the standard being prepared from 0 - 2.5 mM. Each of

these concentrations were plugged into an equation (Eq. 3.7) to determine V1 values. This

process produced the values in the following table (Table 3.4).

Table 3.4 Preparation Volumes Needed for all Concentrations [0-2.5 mM] in a Set of
Standard Citrate Solutions

Concentration Total 0.1M Stock Deionized Water

(mM) Volume (mL) Volume (µL) Volume (mL)
0.0 10 0 10.00
0.1 10 10 9.990
0.5 10 50 9.950
1.0 10 100 9.900
1.5 10 150 9.850
2.0 10 200 9.800
2.5 10 250 9.750

The calculated volumes of 0.1M-citrate stock solution were then carefully

pipetted (89133-288, VWR ) from the stock storage container into seven, or six in cases

when the 0.1 mM solution was not included, individual 10mL glass volumetric flasks

having a precision of ±0.02mL (89000-398, VWR). Pipette tips were discarded after a

single use to prevent cross contamination. Deionized water was subsequently added to

each volumetric flask until the meniscus was level with the mark on the flask.

Each of the concentrations were transferred from their volumetric flasks to 15mL

polypropylene centrifuge tubes (89401-566, VWR). Caps were securely fastened onto the

54
tubes which were then stored at room temperature (22 C) until use. After all analysis

was completed, standards were stored in the centrifuge tubes at room temperature (22ºC).

3.2.4 Mobile Phase Preparation

A 5 mM (0.005M) H2SO4 mobile phase used was used for all HPLC analysis

throughout this thesis. Fresh mobile phase was prepared prior to flushing of the column

on the day of testing using laboratory grade 0.5M sulfuric acid (SA212-1, Thermo Fisher

Scientific). For the most part, 1L of mobile phase was prepared for each run, as such all

calculations were for this volume.

First, filtered, deionized water was degassed by pulling vacuum on a 1L Pyrex

square base glass bottle (1396-1L, Corning) containing a magnetic stir bar. Each bottle

was placed on a magnetic stirrer plate set at a speed of approximately 2 for at least 24

hours before the mobile phase preparation.

After degassing overnight, the water was carefully transferred into a 1000 mL

glass graduated cylinder with a precision of ±5 mL (89001-686, VWR). The volume of

water remaining after degassing was measured to account for volume loss. A calculation

was used to determine the amount of 1N (0.5 M) H2SO4 solution needed to reach a

desired solution concentration of 5mM (0.005 M).

A sample calculation for the preparation of 1000mL of 5 mM H2SO4 solution is

demonstrated below:

𝐶1 ∗ 𝑉1 = 𝐶2 ∗ 𝑉2 Eq. 3.8

55
 In this instance, C1, C2, and V2 are all known values. C1 is the original

concentration of the H2SO4 solution or 1N (0.5M). C2 is the desired concentration of the

final solution or 5 mM (0.005M) H2SO4. V2 is the final volume of the 5 mM H2SO4

solution, which in this example is 1000 mL. V1, the amount of sulfuric acid to add, was

determined from this calculation.

A sample calculation using this method is seen below. First, the known values

were inserted to simplify the equation. Then simple algebra was used to solve the

equation for V1. As demonstrated below in Equations 3.9- 3.11 below, to prepare 1L of

mobile phase, 10 mL of 1N H2SO4 was needed.

0.5𝑀 ∗ 𝑉1 = 0.005𝑀 ∗ 1000 𝑚𝐿 Eq. 3.9

0.5𝑀 𝑉1 = 5 𝑀 ∗ 𝑚𝐿 Eq. 3.10

𝑉1 = 10 𝑚𝐿 Eq. 3.11

Once V1 had been obtained from the equation, a corresponding amount of water

was taken out of the 1000 mL graduated cylinder and discarded using an Eppendorf

Research Plus 10mL Pipettor (89131-970, VWR) with disposable tip. Immediately after

this step, the corresponding amount of 1N sulfuric acid was added in order to create a 5

mM H2SO4 solution.

56
 Runtimes varied due to the number of samples tested, as such, a calculation was

required to ensure adequate mobile phase had been prepared. As discussed later, for the

most part each sample was run for approximately 15 minutes at a 0.6 mL/min flow rate.

In order to calculate the amount of mobile phase needed, a simple method is described

below.

First, the number of samples was multiplied by the run time for each sample to

determine the overall run time, TT. In almost all experimental runs the number of samples

were approximately 30 and sample run time was usually 15 minutes per sample.

# 𝑠𝑎𝑚𝑝𝑙𝑒𝑠 ∗ 𝑠𝑎𝑚𝑝𝑙𝑒 𝑟𝑢𝑛 𝑡𝑖𝑚𝑒 = 𝑇𝑇 Eq. 3.12

Then the total overall run time was multiplied by the flow rate, 0.6 mL/min for

almost all samples to determine the theoretical volume needed (VT) for the HPLC run.

𝑇𝑇 ∗ 𝑓𝑙𝑜𝑤 𝑟𝑎𝑡𝑒 = 𝑉𝑇 Eq. 3.13

Finally a safety volume (VS), typically 250 mL, was added to this number to

ensure an adequate amount of mobile phase was available for the total time of the run.

𝑉𝑇 (𝑚𝐿) + 𝑉𝑆 (𝑚𝐿) = 𝐴𝑐𝑡𝑢𝑎𝑙 𝑉𝑜𝑙𝑢𝑚𝑒 𝑡𝑜 𝑃𝑟𝑒𝑝𝑎𝑟𝑒 (𝑚𝐿) Eq. 3.14

57
 Inserting typical values for number of samples, 35, and sample run time, 15

minutes, yields a TT = 525 minutes. A flow rate of 0.6 mL/min gives a VT= 315 mL. Then

using a VS=250 mL, the actual volume one should prepare is 565 mL. This process does

not account for the mobile phase needed to prepare the column which is typically another

200 mL. Overall, a good estimate for amount of mobile phase needed for an average

number of samples is 1L.

3.2.5 Preparation of 1.0M Potassium Hydroxide Solution

A 1.0M Potassium Hydroxide (KOH) solution was needed for the bone

processing protocol discussed later. For all KOH solutions prepared in this thesis,

Potassium Hydroxide in pellet form (P251-500, Thermo Fisher Scientific) was dissolved

in deionized water.

A molar calculation was done to determine the weight of KOH pellets needed to

prepare 100 mL of a 1.0M solution. For all solutions prepared in this thesis, 5.611g of

KOH pellets were weighed on a precision (0.0001 grams) balance (Adventurer Pro,

Ohaus) and placed into a 100mL beaker containing approximately 20mL of deionized

water. A clean magnetic stir bar was then inserted into the beaker and subsequently

placed onto a magnetic stirrer plate. After 15 minutes of mixing, the solution was

carefully transferred into a 100mL volumetric flask with an error of ±0.1mL (89000-404,

VWR). Deionized water was added to the volumetric flask until the meniscus was level

with the 100mL line of the flask.

58
 The KOH solutions were then transferred into two polypropylene 50mL

centrifuge tubes (89401-562, VWR). Each of the tubes were capped and stored at room

temperature (22ºC). Each tube was able to provide enough volume for roughly 30 sample

preparations.

In total, six main chemical solutions were prepared and utilized throughout the

entirety of this thesis. In order to better clarify the reference names of the solutions and

their compositions, a table was prepared. In the table below, the reference name of the

solution in this thesis is provided, along with their concentration or concentration range,

the solute, solvent, and any chemical used to adjust pH (Table 3.5).

59
 Table 3.5 List of Chemical Solutions Used throughout Thesis. Maximum concentrations
listed for bone solutions are the concentrations determined in Table 3.3.

Reference Name Concentrations Solute Solvent pH Adjustment

Stock Solution 0.1 M Citrate Powder DI H2O 0.5 M H2SO4

Set of Standards 0 – 2.5 mM Citrate Powder DI H2O 0.5 M H2SO4

Mobile Phase 5.0 mM 0.5 M H2SO4 DI H2O None

Cortical Bone Ground cortical

0 - 2.12 mM 1.0 M HCl 1.0 M KOH
Solution bone

Trabecular Bone Ground trabecular

0 – 1.59 mM 1.0 M HCl 1.0 M KOH
Solution bone

Mixed Bone Ground cortical and

0 – 1.85 mM 1.0 M HCl 1.0 M KOH
Solution trabecular bone

3.3 Execution of Chemical Analysis using High Performance Liquid Chromatography

This section will introduce the methods and materials used to analyze samples

with HPLC. Emphasis will be placed on the HPLC system and software, as well as test

method parameters. The items discussed in this section will be applicable to all HPLC

analysis performed in this thesis

3.3.1 High Performance Liquid Chromatography System

HPLC analyses were carried out using a Waters Breeze 2 HPLC system (Waters

Corp., Milford, MA). It was comprised of the following components: a model 1525

Binary HPLC Pump, model 2707 Autosampler, and model 2998 Photodiode Array

(PDA) Detector (Figure 3.1).

60
 B A
C

Figure 3.1: Waters Breeze 2 HPLC System Used in this Thesis. In this picture the main
components of the system are labelled as follows: (A) model 2707 Autosampler, (B)
model 1525 Binary HPLC Pump, and (C) model 2998 Photodiode Array (PDA)
Detector

61
 3.3.2 Column Selection

Columns were investigated in order to identify an appropriate column for the

specific application of detecting citrate in bone. An appropriate column was defined to be

one capable of detecting citric acid at approximately pH of 2. A search of available

columns indicated that a column dedicated to the detection of organic acids would

provide the best results for this application. Table 3.6 contains a list of five manufacturers

and their respective columns that were found to meet the defined conditions.

Table 3.6: List of Selected Columns for HPLC Analysis of Organic Acids

Manufacturer Column Part Number

Bio-Rad Aminex HPX-87H 1250140
Laboratories

Thermo Scientific Acclaim OA, 5 µm 062902

Analytical (4 x 250
mm)

Agilent Agilent Hi-Plex H, PL1170-6830

Technologies 7.7 × 300 mm, 8 µm

Restek Corporation Allure® Organic 9165585

Acids, 5 µm, 300
mm x 4.6 mm

Grace Davison Prevail Organic 88740

Discovery Sciences Acid 150 x 4.6mm
5µm

62
 The Aminex HPX-87H by Bio-Rad Laboratories (Hercules, CA) was selected due

to its successful use in a preliminary study conducted by Dunphy [16]. Furthermore, the

column advertised a fairly low retention time for citric acid (8.20 minutes), providing for

a relatively low run time for large sample volumes, which is a practical concern for

translation to forensic labs [4, 5].

3.3.3 Aminex HPX-87H Organic Acid Column

All analyses in this study were performed on a single Aminex HPX-87H Ion

Exchange column obtained from Bio-Rad (Bio-Rad, Hercules, CA) (Figure 3.2). In order

to minimize costs, an organic analysis kit containing the Aminex HPX-87H column, two

Cation H+ Micro-Guard Cartridges, and organic acid standard was purchased (125-034,

Bio-Rad). A standard cartridge holder was also obtained in order to hold and connect

guard columns to the system (125-0131, Bio-Rad).

Literature from the manufacturer stated the column has a maximum flow rate

between 0.5 and 0.7 mL/min, a maximum operating temperature of 65° C, a pH range of

1-3, and a maximum operating pressure of 1,500 psi [4, 5]. Furthermore, the shipping

solvent of the column was 5 mM (0.005 M) H2SO4.

63
 Figure 3.2 Aminex HPX-87H Column Used in this Thesis

3.3.4 Software

The built in software of the Waters Breeze 2 System was used to perform data

acquisition and peak integration. Furthermore, raw data was exported from the Breeze

software for further analysis using Microsoft Excel (Microsoft, Redmond, WA). Excel

was also used to analyze the peak areas determined by the system software. Peak

integration will be discussed further in Section 3.3.5 below.

3.3.5 Test Method Parameters

A testing method was established and saved in the Breeze 2 system software. The

parameters of the method were as follows: 35° C, flow rate of 0.6 mL/min, detection at a

time constant standard 210 nm, wavelength scan from 200-400 nm, and an alarm set at

1400 psi to ensure max pressure was not passed.

Pressure was monitored closely as the column was connected in the system to

ensure the machine was performing adequately. Throughout the majority of this thesis,

the operating pressure of the column was approximately 750 psi.

64
 As stated in section 3.3.2, the mobile phase (solvent moving through the column)

used in this study was a 5 mM (0.005M) H2SO4 solution, as recommended by the

manufacturer. The stationary phase (one that stays in place in the column) was sulfur

trioxide, SO3 –, bound to polystyrene according to the manufacturer (Bio-Rad). As cost

was a key consideration in the method development, mobile phase was prepared by

diluting a stronger concentration rather than purchased directly in the more dilute form.

3.3.6 HPLC sample preparation and disposal

Glass HPLC vials with pre-slit Teflon septa (Waters Corporation, Milford, MA)

were used to hold samples for testing. Initial testing was done with 2 mL clear glass vials

(186000307C, Waters), while later testing was done with amber vials (186007194C,

Waters) due to concerns of possible photodegradation. Each vial was labeled with

specimen name including concentration and content, as well as the date of preparation.

Samples were transferred from 15 mL centrifuge tubes to HPLC vials by a Becton

Dickinson Tuberculin syringe (14-829-10D, Thermo Fisher Scientific) with the needle tip

removed for standards, and attached for bone solutions due to smaller volumes of sample.

Nylon membrane Acrodisc syringe filters from Pall life Sciences (28143-985,

VWR) were used to filter samples prior to testing. Each syringe tip filter was primed by

passing a small amount of sample solution, approximately 0.5 mL if possible, through the

filter into a waste container. Immediately after, the syringe tip filter was placed in the

opening of the correctly labeled glass HPLC vial. Approximately 1.5 mL of sample

solution was drawn into the syringe and passed through the syringe tip filter into the vial.

65
 After the solution had been transferred, a pre-slit cap was screwed onto the vial.

The syringe and syringe tip filter were discarded, while the vials were loaded onto an

auto sampler tray and placed into the Waters auto sampler for testing. An example of a

filled tray can be seen below (Figure 3.3).

Sample sets were created and saved with the Breeze 2 software to reflect the order

of the samples placed in the tray by the user. Samples were tested in the order dictated in

the Breeze 2 software. After each run, samples were kept for two weeks in the event

further analysis was needed. Beyond this time period, HPLC samples were emptied into

an appropriate waste container and glass vials were placed into a glass waste container.

Figure 3.3: Autosampler tray with vials placed in a specific test order. Each location
corresponds to a letter and number which is represented in the system software. Test
orders can be altered in the software to reflect the arrangement present in the tray.

66
 3.3.7 Column Preparation and System Equilibration

Prior to analyzing the samples, the Aminex-HPX87H column was preconditioned

as advised by the manufacturer. Approximately 20 mL of degassed mobile phase was

passed through the column at a flow rate of 0.2 mL/min at room temperature. After this

step, the column was placed into the column heater attached to the Breeze system and

allowed to slowly heat to the operating temperature of 35ºC at the same flow rate. Once

heated, the flow rate was increased to 0.6 mL/min over 5 minutes with incremental

changes taking place. Once the column had a steady pressure with no apparent

fluctuation, the lamp of the PDA detector was turned on in order to allow the detector to

equilibrate.

Mobile phase was allowed to move through the column as the lamp was warming

up, and a test method saved with the name “citric acid equilibrium” was used to monitor

the baseline produced by the system. The baseline was monitored in order to determine

when to begin analysis of the samples. Once the baseline appeared to be stable with little

fluctuation, the equilibration method was stopped, and analysis of the samples was

begun. A typical stable baseline obtained at the end of the equilibration process is shown

below (Figure 3.4).

67
 Figure 3.4: Baseline capture after one hour of equilibration with mobile phase for
multiple wavelengths (top) and single wavelength (bottom). These were both considered
stable as the fluctuation range was less than .001 absorbance units for both baselines.

3.3.8 Analysis of HPLC Data

HPLC results were analyzed using the Waters Breeze 2 Software integrated into

the Waters Breeze 2 system. The citrate peak in each chromatogram produced was

integrated using the built in integration capabilities of the Waters Breeze 2 software. A

straight line was drawn from the start of the citrate peak as indicated by an increase of the

absorbance units from the baseline level until a return to baseline was evident in the

chromatogram. When necessary, the zoom functions of the software were used to ensure

accurate start and stop points for the range of integration. An example of peak integration

is provided below (Figure 3.5).

68
Figure 3.5: Example of peak integration in the Breeze 2 Software for a 0.5 mM standard
solution. Retention time and Peak area are highlighted in black, while the red lines under
each peak represent the beginning and end of the peak integration.

Peak area and retention time values were then transferred by copying the data

from the breeze software and subsequently pasting the data into a Microsoft Excel

spreadsheet. Each spreadsheet was named with the run date and saved for further

analysis. Chromatograms were also acquired by either taking screen captures of the

Breeze 2 Software or by copying the chromatogram from the software and pasting it into

a Microsoft Word document (Microsoft).

An export function in the Breeze software was used to export raw data for

analysis. Each file was exported in .awf format, which was then opened with Microsoft

Excel. The .awf files were then used to prepare chromatograms for qualitative analysis of

each sample. Each chromatogram had time on the x-axis and absorbance units, the signal

response of the HPLC, on the y-axis.

69
 3.3.9 Calibration Curve Tests

Eleven sets of standard solutions containing all seven concentrations from 0.0 to

2.5 mM were prepared for this experiment. All 11 sets of standards were made from

0.1M stock solutions prepared on the same day as the standard solutions. Five of the sets

of standards were prepared at a pH of approximately 9, while the other six were prepared

at a pH of approximately 2 by adding sulfuric acid to the stock solutions, as described in

section 3.2.3 above.

The first six sets were analyzed at two time points to determine the impact of

storage time on the concentrations of citrate in the simple matrices. In general, a time

zero measurement was taken followed by a subsequent analysis two to four weeks after

depending on availability of the HPLC. All sets were stored in a cardboard box at room

temperature (22°C) with the caps tightly secured on the 15mL centrifuge tubes. The

concentrations and slopes obtained between time zero analyses and the aged analyses

were compared to analyze variance over storage time. Furthermore, the concentrations

and slopes of the different pH values were compared to determine if the stock pH had any

impact on the observed concentrations of citrate. Standard sets 7-11 were analyzed only

at time zero for use with bone solution tests.

70
 3.3.10 Linearity

Linearity is a measure of how well a calibration curve follows a straight line [23].

Based on the expected citrate concentrations in bone determined in section 3.4, a

concentration range of 0 to 2.5 mM was tested. After analysis, the square of correlation

coefficient (R2) was calculated to measure linearity [23]. The equation used to find R2 is

shown below (Eq. 3.15).

[∑(𝑥𝑖 −𝑥̅ )(𝑦𝑖 −𝑦̅)]2

𝑅2 = Eq. 3. 15
∑(𝑥𝑖 −𝑥̅ )2 ∑(𝑦𝑖 −𝑦̅)2

In this equation, x̅ is the mean of all x values and y̅ is the mean of all y values.

The RSQ function in Microsoft excel was used to find all R2 values. RSQ required the

known x values, concentrations, and the known y values, peak areas. The function

returned the value of R2.

R2 needs to be very close to 1 to represent a linear fit [23]. As the concentrations

of citrate in bone are very low (less than 2.0 wt.%) in comparison to the main

components of bone, linearity boundaries were determined by treating citrate as an

impurity in the sample. As such, an R2 value of 0.98 was deemed acceptable for the range

of 0.1 to 2 wt.% [23].

Another common metric for linearity is the y-intercept of the curve after the blank

concentration is subtracted. This method should produce a y-intercept close to 0 [23].

Treating the method as an impurity assay means that an acceptable y-intercept should be

≤ 10% of the response for the 2.5 mM standard. When the y-intercept was negative, the

magnitude was used to determine the percentage.

71
 3.3.11 Instrument Precision Tests

Instrument precision or injection precision is the reproducibility observed when

the same quantity of one sample is repeatedly injected into an instrument (≥ 10 times).

This is also referred to as intra-day test method precision as the results are obtained under

the same conditions over a short time interval [52]. Relative standard deviation (RSD)

was used to assess the results of repeatability tests in this thesis. RSD was determined by

using the following equation (Eq. 3.16) where s was the standard deviation of the sample

and 𝑥̅ was the mean of the sample.

100∗𝑠
Eq. 3. 16
𝑥̅

An acceptable precision for our test method was defined to be an RSD ≤ 5% near

the limit of quantitation [52]. As instrument precision was considered early in the

development process, 0.5 mM standards were used in this test. In particular, two different

0.5mM standard solutions, one from set 04_B and the other from set 04_A, were

prepared for repeatability tests. Ten injections from each vial were performed and

subsequently analyzed for peak areas of citrate peaks. The standard deviation and mean

of the injection were obtained for each standard and analyzed to determine if the machine

had acceptable instrument precision with the test method.

72
 3.3.12 Statistical Methods

Throughout this thesis, the following statistical tests were used: One way

ANOVA, Repeated Measures ANOVA, unpaired t-test, and paired t-tests. Each test is

specifically stated, as is the obtained p-value. When stated, one way ANOVA tests were

used to compare the differences in group means and were followed by a post-hoc test

(Tukey’s HSD) for pair-wise comparisons. Repeated measure ANOVA tests were used

when populations were related. Two-tailed, unpaired t-tests were done to compare group

means as stated. Two-tailed, paired t- tests were done when data could be paired. An

alpha value of 0.05 was employed in all statistical calculations. All statistical analysis in

this thesis were accomplished with Minitab Statistical Software version 17(Minitab Inc.,

State College, PA).

3.4 Selection and Preparation of Bone Samples Solutions

This section addresses the key methods, materials, and rationale used throughout

the selection and preparation of the bone samples used throughout this thesis. In general,

this section will address the rationale of the porcine animal model and the methods and

materials used throughout the preparation of samples.

73
 3.4.1 Animal Model

Different animal models were investigated for use in this thesis to ensure results

would be applicable in a forensic anthropology context. Furthermore, the results obtained

in this thesis needed to have a high degree of translation to human bone. As such, a set of

criteria was established to ensure the selected model would be acceptable.

Animal species were evaluated for the following five criteria to determine the

optimal model for use in this thesis. First, the bone needed to have similar biochemical

and mechanical properties as human bone. Second, the bones needed to be easily

acquired and widely available to many researchers to ensure results could be validated by

others. Third, the bones used needed to be suitable for a lab environment [8]. Namely,

bones needed to be easy to handle and store in a lab environment, as well as not pose any

unnecessary risks to researchers. Fourth, the overall gross morphology of bone was

considered to ensure the size of the bones was easy to work with in this application.

Finally, the bones needed to have been validated by use in other studies on citrate in

bone, and in particular studies on citrate in a forensic context.

The correlation between the biochemical and mechanical properties of human

bone and the animal models being considered for use was critical, and as such was the

first criterion used to screen possible options. The results of this analysis provided two

suitable species, porcine and canine as seen in Table 3.7 [43].

74
 Table 3.7. Summary of Four Key Attributes in Terms of Similarity between Animal and
Human Bone. [Adapted from Pearce43].

Canine Sheep/Goat Pig Rabbit

Bone Composition
Bone Remodeling
least similar, moderately similar, most similar.

Porcine morphologic and anatomic characteristics were determined to be more

comparable to humans than those of canine origin due to bone regeneration rates [56].

Porcine bone was also found to be more similar to human bone in mineral density and

concentrations when compared to canine species [1]. In order to further validate the

results of the first screening process, the second criterion discussed, ease of use within a

lab, was evaluated for the porcine and canine animal models. This analysis yielded the

following: (1) a porcine model was more readily available in large quantities and (2) was

more suitable for a lab environment due to ease of use and minimal risk posed to

researchers in comparison to a canine model. A porcine model was further substantiated

by an evaluation of porcine bone gross morphology that demonstrated an easy to work

with size for the application in this thesis. Finally, porcine bone models been used in

similar studies on bone decomposition [1, 29, 49, 60]. As such, it appeared that using a

porcine model would allow us to produce data that was translational to human bone, as

well as develop a method that could be applied to real world forensic applications.

75
 3.4.2 Selection of Porcine Ribs

Once an adequate animal species had been selected, a set of criteria was

established to determine the appropriate anatomical location of bone to sample. In

general, this analysis was done to ensure adequate access to samples, easy to work with,

and also able to meet our experimental needs.

Based on our overall objectives, an optimal anatomical bone model was defined

to be one that was: (1) accessible, (2) easy to section, and (3) translational to human

bone. Accessible in this thesis was defined as a bone source that was readily available

and easily obtained for use. This criterion narrowed down the possible anatomical

locations to ribs, feet, or loin as they are common cuts of meat available in a meat section

at a grocery. The second criterion narrowed the decision to one location, as porcine rib

was the only location that was deemed to be easy to section due to adequate bone and an

easy to work with gross morphology. Finally, this selection was validated by similar

studies that also selected to work with porcine rib bones as a sufficient model of human

bone [16, 49].

3.4.3 Sectioning Rib Bones

In total, six fresh racks (complete set of ribs) were obtained from a local grocery

story for all bone analysis in this thesis. Each rack consisted of 12 to 13 porcine ribs, and

in some instances a sternum. All bone analyses were performed on the fresh porcine rib

bones either immediately after purchase or after storage in a freezer at approximately -

18° C for no longer than 8 weeks. Frozen ribs were thawed in the refrigerator at

76
approximately 5°C overnight or at room temperature approximately one hour before

sample preparation. The racks were numbered in the order they were obtained with Rack

1 being the first rack and Rack 6 being the last. Each bone specimen was labeled with the

rack from which is was obtained by including an R with the rack number immediately

following (R#).

Bones in each rack were numbered from 1 to 12 or 13 depending on the number

of ribs present. Numbers were assigned by placing the rack in a cranial to caudal manner

with the cranial most rib on top (Figure 3.6). Cranial to caudal manner was determined by

the curvature and gross morphology of each rack of ribs. Each rack contained a very

small bone that was in all cases the caudal most rib present, and as such was placed on

the bottom of each set. A profound curve to each rack could be observed to further

confirm the appropriate direction of each rack of ribs. Numbers were assigned in an

ascending fashion, such that the small rib on the bottom had the highest number assigned.

Care was taken to ensure that each rib number was appropriately documented in all

samples prepared and was recorded for each sample by placing a B with the bone number

immediately following the rack and rack number (R#B#).

Each specimen cut from a bone was given a number to help identify the location

of the specimen along a bone. In general, a 2 mm section was taken off of the dorsal end

of the bone and removed as it was exposed during the processing of the ribs by the

grocer. Next, 2-4 mm sections were taken along the bone with each section being given a

number in an ascending fashion such that the first specimen was number one and the last

was the total number of specimens cut from each bone. Each specimen was then

77
identified by the rack number, bone number and specimen number by placing an S and

specimen number immediately after the bone number (R#B#S#). This labeling system

was maintained throughout the entirety of the thesis.

Cranial

Caudal
100 mm

Figure 3.6 Image of Rack of Ribs positioned in a Cranial to Caudal manner.

3.4.4 Defleshing of Porcine Rib Bones

Bones were defleshed by removing the soft tissue surrounding the porcine ribs

through mechanical separation with a stainless steel surgical scalpel blade (Figure 3.7). In

some instances, a large stainless steel kitchen knife was necessary to remove large

sections of tissue that were too thick to mechanically shear with the scalpel. After the

majority of the soft tissues had been removed, forceps were used to remove the final layer

of soft tissue on the bones. Care was taken during the process to preserve the periosteum

78
on each bone, but in some instances, the removal of the cartilaginous regions on the

ventral side of the bone caused the periosteal layer to detach from the bone.

10 mm

(A)

10 mm

(B)

Figure 3.7 Images of Bone Pre and Post Defleshing. A stainless steel scalpel was used to
mechanically remove most of the soft tissues on the bone.

3.4.5 Bone Sectioning

Following defleshing, porcine rib bones were sectioned into 2-4 mm specimens

by a water-cooled band saw (Microedge, Dorn and Hart, Villa Park, IL). An initial 2 mm

section was taken off the dorsal side of the bone in order to ensure sample analyses was

not impacted by the bone exposed during the processing of the ribs by the grocer.

Immediately after, continuous 2-4 mm sections were cut for each bone, until the bone

79
became too short to safely introduce to the blade without posing a risk to the researcher.

At this point, a pair of stainless steel forceps were used to hold the ventral end of the

bone and slowly and carefully introduce the bone to the blade. After processing, the band

saw was cleaned with bleach and tap water to ensure no residual citrate would be present

in subsequent sections. The remaining parts of bone not sectioned were placed into a

polypropylene storage container and placed into a freezer at at approximately -18° C.

2 mm

Figure 3.8 Section of Bone Acquired for Analysis. The average section was 2-4 mm in
width and was obtained using a water-cooled bandsaw.

3.4.6 Bone defatting

Bone sections were defatted in a mixture of chloroform and ethanol (1:1). For

each specimen, a 50 mL Pyrex glass beaker (13912-149, VWR) was obtained and placed

into a fume hood [49]. Immediately after, one mL of reagent grade chloroform (S25248,

Fisher Scientific) was added into the beaker under a fume hood. Next, one mL of

laboratory grade ethanol (S25309B, Fisher Scientific) was added to the container.

80
Specimens were then placed into the glass container and allowed to sit in the mixture for

one hour. Finally, the samples were allowed to air dry for 10 minutes under the hood.

3.4.7 Bone Type Preparation Steps

After defatting, bone specimens were transferred to a lab bench and allowed to

dry for an additional 10 minutes. Bone marrow was preserved in all samples to prevent

potential loss of citrate in processing. When specimens containing both cortical and

trabecular bone types were desired, specimens skipped the first step and proceeded to the

bone grinding method. Cortical only specimens were prepared by manually removing

trabecular bone with a stainless steel scalpel. Trabecular specimens were prepared by

saving the removed trabecular bone from the above step in a 15mL polypropylene

centrifuge tube.

3.4.8 Bone Grinding

All bone specimens were ground using a mortar and pestle. In this process, each

specimen was cooled with liquid nitrogen and pulverized by hand using a mortar and

pestle in order to generate a powder with a particle size <20 um. Bones were ground until

an adequate amount (approximately half) of the bone was a fine powder (Figure 3.9). All

steps in this process were done under a hood to prevent unnecessary exposure to the bone

powder to both researchers and others using the lab. Liquid nitrogen was carefully poured

onto the bone from an acceptable container, and subsequently left undisturbed for 15

seconds. The mortar and pestle were then used to grind the bone into a fine powder. On

average, the grinding process took 10 minutes per bone specimen with the manual

81
grinding. Once an adequate amount of bone was finely ground, all powders were allowed

to air dry for at least five minutes prior to weighing. Next, the bone powder was weighed

to ensure enough of the bone specimen had been obtained for all testing, and

subsequently transferred into a clean 15 mL centrifuge tube until further testing was

needed. In cases where subsequent processing would not occur for over six hours,

centrifuge tubes were placed into a refrigerator at 5ºC until one hour before desired use.

82
 2 mm

(A)

1 mm

(B)

Figure 3.9 Images of Bone Pre (A) and Post (B) Grinding with Mortar and Pestle. Bone
sections were ground with the mortar and pestle until an adequate amount for sampling
(approximately 100 mg) of the bone was obtained for testing.

83
 3.4.9 Bone Digestion

The finely ground bone powder was transferred from the centrifuge tube to a

clean polystyrene weigh boat (89106-764, VWR). Then 75 mg except where noted was

weighed and transferred into a clean 15mL centrifuge tube. Care was taken to ensure all

of the weighed powder was transferred into the tube. In addition, the tube was capped and

gently tapped to ensure the powder was on the bottom of the tube. Next, 2mL of 1.0M

HCl was added to each centrifuge tube, which was then placed into a 60°C water bath for

four hours to demineralize the bone.

3.4.10 Final Processing Steps

After digestion of the bone, each solution was neutralized with a 1.0 M potassium

hydroxide solution as described in Section 3.2. Potassium Hydroxide was added until the

bone solution was determined to have a pH of 2 as indicated by pH test strips with a

range from pH 0 -2.5 and a 0.3 resolution (Millipore MColorpHast pH-indicator strips

(non-bleeding), VWR) The volumes of KOH necessary were estimated using Visual

MINTEQ equilibrium software as seen below in Table 3.8. The values in the table

assume 2mL of strong acid (1M HCl), pH adjustments with a strong base (1M KOH), and

bone consisting of approximately 60 or 70% hydroxyapatite mineral.

84
 Table 3.8: Estimated Volumes of 1.0 M KOH (Base) needed to reach pH 2 After Bone
Digestion

50 mg bone 100 mg bone 150 mg bone

digested digested digested
60% 70% 60% 70% 60% 70%
HA HA HA HA HA HA
Final estimated
pH after bone 0.254 0.285 0.480 0.587 0.890 1.225
digestion
1.45 1.40 1.00 0.85 0.55 0.35
Milliliters (mL)
(1450 (1400 (1000 (850 (550 (350
of strong base
µL) µL) µL) µL) µL) µL)
added

Immediately after neutralization, sample tubes were centrifuged at 1200xg for 5

minutes (Sorvall ST 40R, Thermo Scientific). The supernatant was decanted and stored in

a clean 15mL centrifuge tube. The remaining pellet, consisting of collagen, was saved in

a labeled tube and stored in a freezer at -18°C. The centrifuge tube containing the

supernatant was labelled and stored in a refrigerator at approximately 5°C until HPLC

analyses were performed. An overview of the entire preparation process for porcine rib

bone solutions (from sections 3.4.3 to 3.4.10) is provided below (Figure 3.10).

85
 Deflesh bones mechanically using a scalpel then obtain porcine rib bone section using water
cooled diamond blade band saw

Defat bone specimens in a 1:1 Chloroform to Ethanol Mixture in a 50 mL Beaker for one hour
and allow to air dry for 10 minutes

Grind bone into a fine powder using a mortar and pestle cooled with liquid nitrogen. Allow to air
dry for 5 minutes.

Measure out 75 mg of bone into a 15 mL centrifuge tube and add in 2 mL of 1.0M HCL

Demineralize each sample in a 60°C water bath for 4 hours

Neutralize samples to pH 2 with approximately 1.4 mL of 1.0M KOH

Centrifuge each sample for five minutes at 1200 x g, then carefully decant supernatant into a clean
15mL tube

Prepare each sample for analysis with HPLC

Figure 3.10: Overview of the preparation process of porcine rib bone citrate solutions.
Modified procedure for the preparation of porcine rib bone specimens for HPLC analysis
based on protocols by Schwarcz et al. and Kanz et. al [32, 49]

86
3.5 Quantitative and Qualitative Analysis of Citrate Concentration in Bone Using HPLC

A series of 8 tests were developed and executed to optimize translation of the

HPLC test method to bone samples. In particular, optimal mass of samples were studied

to determine the best mass of bone to sample. Second, standard recovery and standard

addition tests were performed to determine potential losses of citrate during the bone

processing steps as well as the presence of matrix effects [23]. Third, different types of

bones were analyzed qualitatively and quantitatively to determine the optimal bone type

with the method. Fourth, different storage environments were evaluated to determine if

where samples are stored would have an impact on the citrate concentration. Fifth,

locations along a bone were sampled to determine if the location of a specimen along a

bone had an impact on the measured concentration of citrate. Sixth, samples filtered with

molecular weight cutoff centrifuge tubes and unfiltered samples were qualitatively

analyzed to determine if the extra filtration had a noticeable impact on the peaks observed

in chromatograms. Seventh, different methods used to obtain concentrations in samples,

standard addition and calibration curve methods, were compared to determine if any

significant difference in concentrations was seen between the two populations. Finally,

different integration fits in the breeze software were qualitatively analyzed to determine

the optimal method of peak integration for the method.

87
 3.5.1 Optimal Mass Test

Prior studies on citrate with porcine rib bone all utilized 50 mg samples

throughout sample preparation [16, 21, 49]. Each of these studies were done with

enzymatic assay kits in microwell plates requiring small volumes of sample for analysis.

As such, large masses of bone were not considered, as they offered no obvious advantage

with the analytical technique. With HPLC, larger volumes of samples can be prepared

and analyzed as the only limitation to sample volume is the volume of the vial containing

the solution.

As discussed in Section 3.2, 50, 75, and 100 mg samples have expected

concentrations of 1.06 mM, 1.59 mM, and 2.00 mM respectively for cortical only

solutions (Table 3.3). Using a 50mg sample with the calibration curve range selected

would cause all results to be in the lower range of concentrations, which would not be

desired. The larger masses should produce larger peaks, and also provide concentrations

in the upper range of the calibration curve. Overall, each of these masses were tested to

determine if the HPLC method was impacted by the mass of bone in the solution

prepared and analyzed. In order to determine whether or not the mass of bone prepared

for analysis has an impact on the concentration of citrate detected, three different weights

of bone (50mg, 75mg, and 100mg) were processed and analyzed using HPLC.

The method used for this analysis is as follows. Two porcine rib bones (R3B5 and

R3B6) were prepared and sectioned into 18 bone sections, 12 sections from R3B5 and 6

sections from R3B6, as outlined in section 3.4. The 18 bone sections were subsequently

divided into three different groups (groups 1, 2, and 3) of bone each containing six of the

88
sections. Group 1 was comprised of sections 1-6 from R3B5. Group 2 was comprised of

sections 7-12 from R3B5. Group 3 was comprised of sections 1-6 from R3B6. The 6

sections for each group were then ground together into a large pile of bone powder using

the method described in section 3.4.8. The piles for each of the 3 groups of bone were

then used to prepare a solution for each of the three different masses (50, 75, and 100

mg). Solutions were then labeled based on the mass of bone present and the group from

which the bone was obtained. For example, the 50 mg sample from group 1 was labeled

as bone solution OM 50-1.

The bone samples were then run through the final steps of the sample preparation

method listed in section 3.4.10 with two main differences between mass groups. First, the

amount of 1.0M KOH solution added after digestion was varied, as the higher masses

required less base to reach a pH of 2 due to higher initial pH. Second, 50 mg samples

were digested for one hour, 75 mg samples were digested for two hours, and 100 mg

samples were digested for four hours in order to ensure adequate demineralization

occurred. The volumes added to each mass can be found in Section 3.4.10 (Table 3.8).

Samples were then analyzed qualitatively and quantitatively to determine the

impact of mass on citrate detection. Qualitatively, chromatograms were analyzed for peak

resolution and separation. Quantitatively, peak areas were obtained as described in

Section 3.3.8. A calibration curve was then used to determine the molar concentration of

the samples that were normalized into wt. % and compared using the statistical

techniques described in section 3.3.8.

89
 3.5.2 Standard Addition Tests

Standard addition tests were designed and executed to screen the digested, pH2-

adjusted bone solutions for matrix effects at the analytical stage. Matrix effects were

defined to be any signal change caused by anything in the sample other than analyte [23].

In this test, three bone solutions were prepared using the method defined above and

subsequently prepared as outlined below.

First, a single bone solution was transferred into a 10mL graduated cylinder and

the volume of the sample was recorded. As each specimen was prepared in the same

manner, it was assumed that this volume was equivalent in all samples prepared. In

general, the volume of recovered solution was approximately 3 mL. This volume was

subsequently divided by the number of additions being prepared, typically 3 additions, in

order to determine the amount of sample to transfer to a microcentrifuge tube. In general,

800 µL was transferred into each of the microcentrifuge tubes. It was critical to ensure at

least 500 µL was transferred into each tube, as the HPLC autoinjector required this

volume in each injection for complete analysis of the sample.

Once the solution had been appropriately divided into the microcentrifuge tubes,

calculations were performed in a manner similar to section 3.2.3.2 in order to determine

the appropriate volume of the 0.1M stock solution to add to each microcentrifuge tube to

achieve a specific spiked concentration of citrate.

90
 3.5.3 Standard Recovery Tests

Standard recovery tests (also referred to commonly as fortification tests) were

done to evaluate the impact of the bone processing method on citrate concentrations in

bone samples. Prior to demineralization of the bone samples, known amounts of citrate

were added to each individual sample in order to calculate the recovery of additional

citrate in each sample. A range of concentrations was determined for the study based

upon the theoretical citrate concentrations established in Table 3.2. The selected

concentrations were 0mM, 0.5mM, 1.0mM, 1.5mM, and 2.0mM. Each of these

concentrations was converted into an expected mass value based upon an assumed an

approximate final volume of bone solution of 5mL. The concentrations of citrate in each

sample were then compared to the calibration curve concentrations to determine the %

recovery of the sample. This was calculated as seen in the equation below (Eq. 3.17) [23].

𝐶𝑠𝑝𝑖𝑘𝑒𝑑 𝑠𝑎𝑚𝑝𝑙𝑒 −𝐶𝑢𝑛𝑠𝑝𝑖𝑘𝑒𝑑 𝑠𝑎𝑚𝑝𝑙𝑒

% 𝑟𝑒𝑐𝑜𝑣𝑒𝑟𝑦 = Eq. 3.17
𝐶𝑎𝑑𝑑𝑒𝑑

Percent recovery was used to determine if citrate was being lost during bone

demineralization and/or subsequent processing steps. A high percentage of recovery

implied no losses or matrix effects were present, while a lower percent recovery indicated

the loss of citrate throughout the processing of the samples or analytical technique.

Furthermore, percent recovery was expected to be approximately the same in each

concentration as each of the samples were processed under the same conditions. Failure

91
to have a similar percent recovery in all samples was taken to indicate the presence of

matrix effects in the samples.

In order to test standard recovery, ten bone samples from varying locations and

different porcine ribs were then combined into one pile of bone powder to minimize the

variance in sample testing. Both trabecular and cortical bone were tested in this

experiment in order to ensure the most complex matrix. It was assumed that trabecular

and cortical matrices would be less complex than both combined. Presence of matrix

effects in the mixed sample was taken to mean matrix effects were present in cortical

only or trabecular only samples as well, and as such as strategy would need to be adopted

to overcome the interferences.

75 mg of bone powder was weighed using a precision scale (0.0001 mg)

(Adventurer Pro, Ohaus) and subsequently transferred into a clean 15mL centrifuge tube.

Trisodium Citrate Dihydrate in granular form was then added based on the calculated

values corresponding to the desired concentrations of citrate to be examined. The bone

processing protocol then proceeded as normal. This process was repeated three times for

and each set was analyzed to determine percent recovery.

3.5.4 Analysis of Citrate Concentrations in Various Bone Types

Previous studies on citrate content in bone have used either cortical or trabecular

only samples of porcine rib bone [21, 49]. No direct comparison between the types of

bone and their impact on the analytical technique or obtained concentrations appeared to

be available in literature. As such, this study was performed to gather data on qualitative

92
and quantitative differences between types of bone (cortical, trabecular, and a mixture of

both) using HPLC. The main objectives of this experiment were to determine limitations

of the HPLC method developed in regards to bone type, as well as analyze the

concentrations observed in the different bone types. Overall, it was believed that each

bone type analyzed would have no significant impact on the output of the HPLC analysis.

The methods used for this experiment are as follows. A total of six 2-4 mm

sections of bone were obtained from two fresh porcine rib bones prepared using the

method described in section 3.4.3. After sectioning the bones were defatted using the

chloroform and ethanol solution described in section 3.4.6. Immediately after defatting,

trabecular and cortical bone from the specimens were differentiated by both the color and

structure of bone. Trabecular bone was subsequently removed with a stainless steel

scalpel and placed into a clean 15 mL polypropylene centrifuge tube. The remaining

cortical bone was inspected to insure the trabecular bone was completely removed by

visual inspection with and without magnification, then placed into a separated 15 mL

centrifuge tube. The mixed cortical and trabecular samples analyzed were prepared as

part of the optimal mass testing protocol and were included only for comparison

purposes. Each of the cortical, trabecular, and combined samples were prepared for

HPLC analysis using the preparation method described in section 3.3.6.

Samples were then analyzed qualitatively and quantitatively to determine the

impact of mass on citrate detection. Qualitatively, chromatograms were compared for

peak resolution and separation. Quantitatively, peak areas were normalized into wt. %

and compared using the statistical techniques described in section 3.3.12 when possible.

93
 3.5.5 Analysis of Environmental Storage Conditions on Citrate

Schwarz et al hypothesized that water content has an impact on citrate

concentration within bone, but provided no further justification [49]. As such, exposure

to water in a lab environment was considered to be a potential risk in the preservation of

citrate within samples being prepared for analysis. Thus, this study was done to evaluate

the impact of possible storage environments, in particular a wet environment, within a

forensic lab to which samples may be exposed for a period of up to two weeks. Overall,

three different storage environments (dry, moist, and atmospheric) were evaluated to

determine if storage environment presents a major risk to the measured concentrations of

citrate in bone samples.

The methods used for this experiment are as follows. Three fresh porcine ribs

were prepared and sectioned into three 2-4 mm specimens using the preparation

procedure described in section 3.4.3. One specimen from each bone was then placed

under differing conditions. The first was left at room temperature and atmospheric

conditions. The second specimen was submerged in deionized water and kept under

vacuum. The third specimen was immediately placed in a Fisher IsoTemp 281A vacuum

oven (281A, Fisher) at 110°C for 4 hours, removed, and left at room temperature covered

by a towel. After 14 days, the specimens were immediately prepared for HPLC analysis

as described in section 3.3.6. The water in one of the flasks evaporated after 10 days in

one of the water specimens. As it was later in the process, it was still analyzed as it was

believed to have occurred after adequate exposure of the sample to the water.

94
 Samples were then analyzed qualitatively and quantitatively to determine the

impact of storage conditions on the detection of citrate. Qualitatively, chromatograms

were compared for peak resolution and separation. Quantitatively, peak areas were

normalized into wt. % and compared using the statistical techniques described in section

3.3.12 when possible.

3.5.6 Analysis of Citrate Content in Dorsal, Central, and Ventral Regions

A key concern in using citrate concentration in bone for the estimation of PMI is

if the anatomical location of a sample would have an impact on the concentration of

citrate detected. Prior studies have suggested an even distribution of citrate along a bone,

but no mention of the location of samples tested is provided in these studies [16, 49]. As

such, a need was realized for more information on the concentration of citrate in different

regions along a porcine rib bone. The objective of this experiment was to sample

different anatomical regions (dorsal, central, and ventral) to determine if the location of

the specimen obtained had a significant impact on the concentration of citrate detected.

The methods used for this analysis are as follows. First, three fresh porcine ribs

were defleshed using the preparation procedure described in section 3.4.3. Then, the

length of each bone was measured in mm with a ruler and recorded. Next, each bone was

divided into three equal length sections (ventral, central, and dorsal) based on the total

length of the bone and anatomical landmarks (Figure 3.11). Ventral sections were defined

as the third of the bone where cartilage was attached. Central sections were defined as the

middle third of the bones. Dorsal sections were defined as the remaining third of the

95
bone. The midpoints of each of the three sections were measured with a ruler and marked

on the bone. Each midpoint was then used as a guide to obtain a section of bone from

each of the regions. Cortical only samples were then prepared for HPLC analysis using

the method in Section 3.3.6.

Central

Ventral

Dorsal

Figure 3.11 Image of Bone with Midpoints Marked and Anatomical Regions Identified.

3.5.7 Comparison of Molecular Weight Cutoff Filtered and Unfiltered Samples

Initial chromatograms of bone samples had a substantial unknown peak prior the

elution of citrate. Additional filtering was tested in order to evaluate if the unknown peak

could be removed from the sample. Molecular weight cutoff tubes (MWCO) appeared to

be an appropriate method of filtration for this particular application. As citrate had a

molecular weight of 189.1 Da, the two smallest molecular weight cutoffs available in the

Pall Microsep™ Molecular weight centrifuge cutoff tubes were acquired. The cutoffs

were 3 kDa (89132-004, VWR) and 1kDa (89233-876, VWR). It was hypothesized that

96
the first peak in bone chromatograms was due to a larger biomolecule, in particular a

protein that would be eliminated from the sample when MWCO tubes were utilized to

filter the samples.

The methods used for this analysis are subsequently described. First, three bones

were defleshed as described in Section 3.4.4 then sectioned into 16 2-4 mm bone

specimens as described in section 3.4.4 Three different populations of bone were then

prepared by grinding and mixing five of the specimens together using the process

described in section 3.4.8. Three samples were then prepared, one for each of the filtering

methods (unfiltered, 1K, 3K) from each population of bone specimens by weighing out

75 mg of the cortical only bone and placing it into a new 15 mL centrifuge tube labeled

with the type of filtration. The preparation protocol was then the same for each filtration

method until after the supernatant had been removed after centrifugation of samples for 5

min at 1200 x g. At this point, unfiltered samples were considered ready for the HPLC

sample preparation process, while the MWCO samples required further treatment.

Both the 1K and 3K samples were transferred into the 1K and 3K MWCO tubes

respectively. After transfer, the caps were tightly secured to each of the tubes and placed

into a centrifuge (Sorvall ST 40R, Thermo Scientific). In order to ensure adequate

filtration of the samples, the instructions provided by the manufacturer were analyzed for

ultrafiltration. This analyses demonstrated that setting the centrifuge at 4500 x g for 90

min would be sufficient to ensure ultrafiltration. After centrifugation was complete, the

upper reservoir containing the filter and all of the compounds too large to pass through

the filter membrane were removed from the tubes. The solution in the lower reservoir

97
was then transferred into a clean 15 mL centrifuge tube for storage until preparation for

HPLC analysis. Once the lower reservoir was cleared, the top reservoir was placed back

into the tube and sealed for storage.

Samples were analyzed qualitatively to determine the impact of filtration on the

appearance of the unknown peak in chromatograms. Qualitatively, chromatograms were

compared for peak height of the unknown peaks and separation between the citrate peaks

and unknown peaks.

3.5.8 Integration Fit

Throughout the thesis, citrate peaks did not consistently elute after the completion

of the unknown peak. As such, the method of integrating the area of the citrate peak was

evaluated to determine the optimal fit of integration. Three primary fits were evaluated

(computer, exponential, tangential skim) to determine differences in the calculation of

area between them, and also to determine which appeared to fit the peak most accurately.

Each of the fits was acquired using the built in breeze software data analysis functions.

The computer fit was obtained by using the built in integration function of the computer.

Exponential fits were obtained by setting up an exponential fit using he system software

and having it extend from the beginning of the peak to the end of the peak. Each of the

sections were magnified using the zoom functions of the software and evaluated to

determine the start and stop points. Start and stop time for peak integration could be

programmed into the software to limit the area of peak integrated. Tangential skim fits

were obtained by using the mouse and dragging a line across the bottom of the peak. The

98
line could be made more accurate by using the built in functions to set a start and stop

time for integration. Peak areas obtained by each fit were compared to determine the

most appropriate fit for each of the citrate peaks produced in the study. Furthermore, each

fit was reviewed in a more general sense to determine the most appropriate fit across all

integrations performed in this study.

3.5.9 Standard Addition Compared to Calibration Curve

Two main methods were used to obtain concentrations of citrate in bone samples

throughout this thesis, namely, standard addition and calibration curve methods. From a

practical perspective, the calibration curve technique requires less time and resources and

as such is more desirable for this method. However, calibration methods are prone to

error in the presence of matrix effects in samples [55]. As both methods were adequately

tested in this thesis, a comparison of the concentrations obtained with each method were

performed using results in this thesis, to determine if a significant impact was observed in

the concentrations of citrate detected based on the method of determining concentration.

Statistical tests as described in section 3.3.12 were performed on the location and storage

condition study concentrations with the concentrations obtained from each method to

determine if each method had an impact on the normalized citrate concentration

determined for each of the samples.

99
 3.5.10 Calculation of PMI Using Schwarcz Model

Postmortem interval estimates produced in this thesis were obtained by using the

equation developed by Schwarcz et al. [49]. In this equation, the normalized citrate

concentration (wt. %) were used in the following equation to obtain a PMI estimate in

years (Eq. 3.18).

𝐶(𝑡) = −0.673 ∗ log(𝑡) + 3.021 Eq 3.18

In this equation, C(t) is the detected normalized concentration (wt.%) of citrate in each

sample, and t is the time in days. The number of days was then divided by the typical

number of days in a year, 365, to obtain PMICALC, or the estimated postmortem interval in

years for each bone solution.

100
 CHAPTER 4

RESULTS

4.1 Introduction

The results in this chapter are divided into five general subsections, namely HPLC

Method and System Performance, Validation of Method Acceptability with Standard Solutions,

Analysis of Citrate Peaks in Bone Solution Chromatograms, Applied Tests with Method, and

Performance of HPLC.

Subsection 4.2 “HPLC Method and System Performance” includes results related to the

system parameters observed in this thesis. Subsection 4.3 “Validation of Method Acceptability

with Standard Solutions” provides results related to the linearity of calibration curves, tests on

calibration curve variance in time and stock solution pH, as well as intra-day assay precision

(repeatability) studies. Subsection 4.4 “Analysis of Citrate Peaks in Bone Solution

Chromatograms”, provides results related to bone chromatograms including characteristic peaks

and injection dependence. Subsection 4.5 “Applied Tests with Method” provides results related to

tests done to characterize citrate concentrations in bone. Finally, subsection 4.6 “Performance of

HPLC” provides results related to the general performance of the HPLC method and system over

the thesis.

4.2 HPLC Method and System Performance

This section addresses the overall performance of the HPLC system, as well as

method parameters that varied throughout the thesis. Results are presented to demonstrate

qualitative or quantitative differences in chromatograms or citrate concentrations.

101
 4.2.1 Retention Time

Retention times for citrate standards were carefully monitored throughout the

thesis. Analysis of the test chromatogram provided with the column used for all analysis

in this thesis listed a retention of 7.50 minutes with the same temperature and flow rate

selected for use in the test method. Further information available from the manufacturer

indicated a retention times of citrate of 7.8 and 8.2 minutes [4, 5]. The average retention

times for standards prior to the introduction of bone solutions into the column was 7.86 ±

0.24 minutes. After the introduction of bone samples into the column, average retention

time for standards increased to 8.62 ± 0.14 minutes. Statistical analysis was completed to

determine if the introduction of the bone samples was a factor in the observed retention

times for citrate standards. Introduction of the bone solutions was determined to be a

significant factor in the retention times of citrate standards (unpaired t-test, p < 0.05),

indicating that the performance of the column was altered by the introduction of bone

solutions through the column. Example chromatograms of 2.5 mM standards from set 05,

before the introduction of bone solutions (Figure 4.1), and set 10, after the introduction of

bone solutions (Figure 4.2), are provided below to demonstrate the significant change in

retention times.

102
 0.012

0.010
Absorbance Units

0.008

0.006

0.004

0.002

-0.001
0 2 4 6 8 10 12 14 16 18
Time (Min)

Figure 4.1 Chromatogram of 2.5mM standard solution prior to introduction of bone

solutions into column. Retention time of citrate peak was determined to be 7.97 minutes
[Sample: 2.5mM standard from standards set 05].

0.006

0.005

0.004
Absorbance Units

0.003

0.002

0.001

0.000

0 2 4 6 8 10 12 14 16 18
-0.001
Time (Min)

Figure 4.2 Chromatogram of 2.5mM standard solution after the introduction of bone
solutions into column. Retention time of citrate peak was determined to be 9.55 minutes
[Sample: 2.5mM standard from standards set 10].

103
 4.2.2 Mobile phase

As stated in Chapter 3 (Section 3.3.3), 5mM (.005M) sulfuric acid, pH 2, was

used as the mobile phase in all analyses. Blank sample chromatograms contained no

peaks and demonstrated a stable baseline (Figure 4.3). This was witnessed in all blank

solutions analyzed in this thesis, as such, 5 mM H2SO4 was deemed to be an acceptable

mobile phase for use with this test method and the Waters Breeze 2 HPLC system.

0.0010

0.0005
Absorbance Units

0.0000

-0.0005

-0.0010
0 1 2 3 4 5 6 7 8 9 10 11 12
Time (Min)

Figure 4.3: Chromatogram produced by blank standard through Column [0 mM]. No

peaks and a stable baseline are present.

4.2.3 Operating Pressure

As stated in Chapter 3 (3.3.5), the Aminex HPX-87H column was kept at a

constant temperature of 35ºC by the column heater attached to the Waters HPLC system.

The pressures observed at this temperature were below the maximum operating pressure

of the column (1500psi). The normal operating pressure, PNORM, during analysis was 776

104
psi. The maximum pressure, PMAX, observed in testing was 1176 psi and the minimum

pressure, PMIN, was 565 psi. PMAX was observed in the first run of the column during the

conditioning stages; no samples were analyzed at this pressure. PMIN occurred after a total

of 450 injections comprised of both standards and bone samples. It was hypothesized that

changes in pressure would have no qualitative impact on the quantitative analysis of both

citrate standards and citrate peaks in bone solutions.

Chromatograms of 2.5 mM standards analyzed at PNORM (Figure 4.1) and PMIN

(Figure 4.2) were obtained using the chromatogram processing method described in

Chapter 3 (Section 3.3.8). Peak height, time of elution, and retention times were then

compared to gauge the impact of the change in operating pressure of the column on

citrate standards. This analysis was subsequently repeated on bone solution

chromatograms obtained at PNORM (Figure 4.4) and PMIN (Figure 4.5). As with the 2.5

mM standard solutions, peak height and retention times of citrate peaks were

qualitatively analyzed to evaluate the impact of the change in operating pressure on

citrate peaks. In addition, resolution of the citrate peak in the bone solutions was

analyzed.

105
 0.025

0.020
Absorbance Units

0.015

0.010

0.005

0.000
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14
Time (Min)

Figure 4.4 Chromatogram of 75mg Cortical Only Bone Sample Obtained at PNORM.
(Sample: RT2 from Short Term Storage Study)

0.045

0.040

0.035
Absorbance Units

0.030

0.025

0.020

0.015

0.010

0.005

0.000
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17
Time (Min)

Figure 4.5 Chromatogram of 75mg Cortical Only Bone Sample Obtained at PMIN. This is
from the MWCO data set. 75mg Group 2 Unfiltered. Citrate is the third peak on the
chromatogram (Sample: MW Group 2: Unfiltered).

106
 Analysis of the 2.5 mM standards demonstrated a large decrease in peak height

when the column was operating at PMIN (.005 absorbance units) compared to PNORM (.01

absorbance units). Despite the difference in peak heights, the area under each peak was

approximately the same. This was due to an increase in the elution time of citrate at PMIN

(2 minutes) compared to PNORM (1 minute). Finally, an increase in retention time was seen

at PMIN (9.55 minutes) compared to PNORM (7.97 minutes). Overall, the decreased pressure

impacted the peak height, time of elution, and retention time of the 2.5 mM citrate

standards.

Analysis of bone solutions at PMIN and PNORM demonstrated three main

differences. At PMIN, a total of three peaks were observed, while only two were present at

PNORM. Similar to the trend observed in the 2.5 mM standards, retention time of citrate

was increased at PMIN (9.75 minutes) compared to PNORM (8.75 minutes) In addition, the

citrate peak began to elute approximately one minute later on the PMIN (9.25 minutes)

compared to PNORM (8.25 minutes). Also, the height of the citrate peak at PMIN (0.0025

absorbance units) was approximately 0.5x the peak height of the citrate peak in the PNORM

chromatogram (.005 absorbance units). As these were not the same sample, a direct

quantitative comparison of the citrate peak heights was not appropriate. Overall,

operating pressure did not have a significant impact on the ability to qualitatively and

quantitative analyze citrate peaks in bone sample chromatograms.

107
 4.2.4 Wavelength

As stated in Chapter 3 (3.3.5), all sample analyses were analyzed at both a time

constant standard wavelength of 210 nm and over a range of wavelengths, 190 – 300 nm,

using the 3D wavelength scan capabilities of the HPLC system. 3D data were primarily

used to analyze UV-Vis spectra of chromatogram peaks when identification of citrate

peaks was necessary. When changes in retention time were observed, UV-Vis spectra of

citrate peaks in bone solution were compared to the UV- Vis spectra of citrate standards.

Examples of UV-Vis spectra for four citrate standards [1.0, 1.5, 2.0, 2.5 mM] are

provided below (Figure 4.6). Analysis of the spectra demonstrated maximum response in

absorbance units for the citrate peak at 190 and 210 nm. The time constant standard

channel was set at 210 nm as bone samples contained an unknown peak with a

dominating response prior to 200 nm (Figure 4.7).

108
 2.5mM 2.0mM 1.5mM 1.0mM

0.009
0.008
0.007
Absorbance Units

0.006
0.005
0.004
0.003
0.002
0.001
0
190 200 210 220 230 240 250 260 270 280 290 300 310 320 330 340 350
Wavelength (nm)

Figure 4.6 UV-Vis Spectra of Four Citrate Standards. Include four of the standard
concentrations [1.0, 1.5, 2.0, and 2.5 mM]. The response in absorbance units is shown for
across the wavelength range scanned. The largest responses appear to be at 190nm and
210nm.

Unknown Peak Citrate Peak

0.16
0.14
0.12
Absorbance Units

0.1
0.08
0.06
0.04
0.02
0
200 220 240 260 280 300 320 340 360 380 400
Wavelength (nm)

Figure 4.7 UV-Vis Spectra of Unknown Peak and Citrate Peak. The response in
absorbance units is shown for across the wavelength range scanned. The unknown peak
has a very large response below 200nm.

109
 4.3 Validation of HPLC Method Acceptability with Standards

This section addresses the ability of the HPLC method to detect citrate in simple

matrix standard solutions. Results were analyzed both qualitatively and quantitatively to

ensure the test method provided acceptable linearity over the concentration range selected

for the calibration curve [0 – 2.5 mM]. Furthermore, standards were analyzed to evaluate

degradation in calibration curve response and linearity over time. Finally, intra-day test

method precision (repeatability) was analyzed to evaluate the precision of the test

method.

4.3.1 Linearity and Range

As described in Chapter 3, section 3.3.9, a total of 11 sets of standards were

prepared in this thesis and analyzed for linearity. An R2 > 0.98 and y – intercept value ≤

10% of the 2.5mM response were deemed acceptable degrees of linearity. All analysis

was performed as described in Section 3.2.3.

For complete clarity, an example of the data and results obtained from this

technique is presented below. Chromatograms for each of the standard concentrations

were produced and analyzed (Figure 4.8). The peak area values for each of the

concentrations in Standard Set 08A at time t= 0 weeks (Table 4.1) were then determined

by integration of peaks and entered into Excel in order to generate a first-order (straight

line) fit calibration curve with a forced y intercept at 0 (Figure 4.9). Excel was also used

to find the y-intercept value when the calibration curve was not forced through the origin

(Figure 4.10).

110
Table 4.1 Peak Areas and Retention Times for Each Concentration of Standard Set 08A
at Time=0

Conc. Retention Time Area

(mM) (min) Units
0.00 0.000 0
0.50 8.610 26394
1.00 8.618 52725
1.50 8.611 80585
2.00 8.610 107607
2.50 8.613 131583

0.5 mM 1.0 mM 1.5 mM 2.0 mM 2.5 mM

0.007

0.006

0.005
Absorbance Units

0.004

0.003

0.002

0.001

0.000

-0.001
0 2 4 6 8 10 12
Time (Min)

Figure 4.8 Chromatograms of Each Standard Solution Concentration Analyzed in Set

08A. Peak area and height increase as concentration of citrate is increased.

111
 140000

120000

y = 53162x
100000 R² = 0.9996
Area Units

80000

60000

40000

20000

0
0.00 0.50 1.00 1.50 2.00 2.50 3.00
Concentration (mM)

Figure 4.9: Calibration Curve produced by Set of Standards 08A forced through the
origin

140000
y = 53052x + 200.8
R² = 0.9994
120000

100000
Area Units

80000

60000

40000

20000

0
0.00 0.50 1.00 1.50 2.00 2.50 3.00
Concentration (mM)

Figure 4.10: Calibration Curve produced by Set of Standards 4A without being forced
through the origin. Y-intercept determined to be 200.8 area units.

112
 The first-order (straight line) fit calibration curve of standard set 4A at t=0 had an

R2 of 0.9992 and y-intercept of 246.3 Area Units when the calibration curve was not

forced through the origin. When compared to the response of the 2.5 mM standard, the y-

intercept obtained was determined to be ≤10% of the response, 0.19%. Also, the R2 value

is greater than 0.98. As such, the calibration curve obtained for standard set 4A has an

acceptable degree of linearity for this analysis. This process was repeated for each set of

standards in order to ensure acceptable linearity was obtained. A summary of this data is

presented in the table below (Table 4.2).

113
 Table 4.2: Determination of Linearity for All Analyses of Citrate Standards Based on R2
and y-intercept Percentages of max response

Set of Time y-int % of Acceptable

Stock N Slope R2 y-int
Standards (Weeks) Cmax Areaa Linearity
1 01_A 7 0 47899 0.9863 -5785.50 4.67 yes
1 01_A 5 4 52283 0.9920 8560.50 8.46 yes
2 02_A 7 0 46653 0.9937 2685.50 2.40 yes
2 02_A 5 4 40189 0.9897 2738.00 3.53 yes
3 03_A 7 3 51539 0.9990 -2805.70 2.17 yes
3 03_A 6 6 47338 0.9967 -3040.70 2.55 yes
4 04_A 7 0 50996 0.9991 246.30 0.19 yes
4 04_A 7 3 46203 0.9970 -2473.20 2.07 yes
4 04_B 7 0 46343 0.9935 -4529.30 3.81 yes
5 05_A 7 0 52677 0.9981 1374.70 1.04 yes
5 05_A 7 3 46202 0.9995 160.11 0.14 yes
6 06_A 7 0 56016 0.9933 -4803.80 3.29 yes
6 06_A 5 2 58037 0.9926 -3949.60 2.79 yes
7 07_A 6 0 53188 0.9968 2146.40 1.64 yes
8 08_A 6 0 53162 0.9996 200.80 0.15 yes
9 09_A 6 0 49776 0.9995 -498.20 0.40 yes
10 10_A 6 0 46400 0.9996 279.8 0.24 yes
11 11_A 7 0 49534 0.9985 1675.8 1.34 yes

a
Cmax for all but two of the sets, 01_A, t=4wks, and 02_A, t-4wks, was 2.5mM. Cmax for
the others was 2.0 mM. N = number of concentrations in each set

114
 All sets tested had R2 values greater than 0.98 as well as y-int ≤ 10% of the

maximum concentration tested. Thus, all sets tested in this study had acceptable linearity

based on the criteria defined for linearity. Furthermore, the results indicate that using six

concentrations [0, 0.5, 1.0, 1.5, 2.0, 2.5 mM] of standards is sufficient to provide an

acceptable linear calibration curve over a range of 0 to 2.5 mM.

4.3.2 Analysis of Intra-day Method Precision (Repeatability)

Repeatability (intra-day assay precision) was assessed to evaluate how results

varied over a short time interval under the same conditions [52]. As stated in section 3.5

in Chapter 3, two 0.5mM solutions from two different sets of standards (set 04_A in

study one and set 04_B in study two) prepared from the same stock solution (stock 4)

were injected 10 times on the same day, under the same experimental conditions. The

results of the first repeatability study (Tables 4.3) and the second (Table 4.4) are found

below.

115
Table 4.3: Set 4A Repeatability Data of 10 injections of 0.5 mM Standard Solution

Injection Retention Peak Area

Number Time (min) (Area Units)
1 7.904 23460
2 7.909 22374
3 7.924 24122
4 7.912 23227
5 7.909 22569
6 7.905 22179
7 7.907 22075
8 7.920 23015
9 7.911 23665
10 7.923 22337
Mean 7.9124 22902.3
SD 0.007 664.521
RSD (%) 0.09 2.90

Table 4.4: Set 4B Repeatability Data of 10 injections of 0.5 mM Standard Solution

Injection Retention Peak Area

Number Time (min) (Area Units)
1 7.925 14927
2 7.910 14461
3 7.922 15040
4 7.921 15900
5 7.913 15343
6 7.903 16058
7 7.919 14887
8 7.902 14789
9 7.901 14621
10 7.921 16093
Mean 7.9137 15211.9
SD 0.009 573.675
RSD (%) 0.11 3.77

116
 Relative Standard Deviation (RSD) was used to evaluate the instrument precision

of this method. As the low concentrations of citrate are best approximated by an impurity

model, the precision criterion for the method was set at ≤ 5%. That is, the RSD value

obtained for the retention times and peak areas demonstrated acceptable precision of the

instrument so long as the RSD obtained were ≤ 5%.

In the first set of ten replicate injections, an RSD of 0.09% for retention time and

an RSD of 2.90% for the peak areas were obtained. The second set of replicate injection

provided an RSD of 0.11% for retention time and an RSD of 3.77% for the peak areas.

All of these values were within our acceptable criteria for instrument precision, RSD ≤

5%. This indicated that the test method had a high degree of repeatability under normal

operation at the level of intra-day assay precision.

4.3.3 Concentration Variation over time in standards

As described in section 3.3.9 (Ch. 3), four of the 11 sets of standards prepared in

this thesis were analyzed at two time points (to = 0 weeks and tf = 2-4 weeks) depending

on the availability of the HPLC. Set 03_A was not analyzed at t=0 due to the HPLC

system being offline for maintenance, but was analyzed at t=3 weeks which was included

in the aged population. The number of samples for each standard as well as the mean and

standard deviation for each concentration of both to (Table 4.5) and tf (Table 4.6)

standards are provided below.

117
Table 4.5 Average Peak Area for All to (0 weeks) Standards Analyzed

Concentration Average Peak RSD N= Number

SD
(mM) Area (t=0) (%) of Samples
0.0 0 0 - 4
0.1 4460.67 320.13 7.18 3
0.5 27791 1602.16 5.77 3
1.0 49975 2478.21 4.96 4
1.5 76464.25 3082.83 4.03 4
2.0 100749.7 1659.32 1.65 3
2.5 129685.75 10839.2 8.36 4

Table 4.6 Average Peak Area for all tf (≤4 weeks) Standards Analyzed

Average Peak
RSD N= Number
Concentration Area SD
(%) of Samples
(t=2-4 weeks)
0mM 0 0 - 4
0.1mM 2692.33 288.25 10.71 3
0.5 mM 23419.4 2881.01 12.30 3
1.0 mM 47928.6 2855.04 5.96 4
1.5 mM 68621 6068.84 8.84 4
2.0 mM 97487.2 15440.8 15.84 3
2.5 mM 126389.25 10387.4 8.22 4

118
 Table 4.7 p-Values Obtained from paired t-tests of peak areas of fresh (t=0 weeks) and
aged (t = ≤ 4 weeks). Standards Analyzed with HPLC Test Method (p < 0.05 indicates
significant difference)

t-test p Significant
Concentration Difference
value
0.1 mM 0.00 Yes
0.5 mM 0.37 No
1.0 mM 0.27 No
1.5 mM 0.05 No
2.0 mM 0.05 No
2.5 mM 0.66 No
Slopes 0.15 No

Statistical analysis was performed to assess whether time was a factor in the

concentration of citrate detected in the standards over a two to four week time period.

This analysis showed a significant difference (paired t-test, p < 0.05) only in the 0.1 mM

concentration. Otherwise, all samples had no significant difference in the standard

concentrations between the freshly prepared and aged standards (Table 4.7).

Table 4.8 Comparison of Slopes Obtained in Calibration Curves of Fresh and Aged
Standards

Set of Slope Number of Slope Number of

Sf % of Si
Standards (t=0) Standards Used (t ≤ 4 weeks) Standards Used
02_A 46653 5 40189 5 86
04_A 50996 7 46203 7 90
05_A 52677 7 46202 7 88
06_A 56016 6 58037 5 103

119
 In general, the slopes obtained from the aged standards were 10% lower than the

fresh standards (Sf % of Si, Table 4.8). Set 06_A had a higher slope which was caused by

an overestimation of the areas for the higher concentrations [2.0, 2.5 mM] due to an error

in the preparation of the standards. A paired t-test was performed on the slopes and

showed no significant difference (paired t-test, p=0.15) in the slopes of the fresh and aged

standards.

While no statistical significance was present in the fresh and aged standards, all

subsequent analysis used only fresh standards as slight changes in calibration curve

slopes would impact the observed concentration of citrate in bone samples. Due to the

low concentrations present, a slight change in the obtained concentration would impact

PMI estimation greatly. For example, a bone solution with a low citrate concentration,

peak area of 25000 Area units, would have a molar concentration of 0.49 mM at to and

0.54 mM at tf using the calibration curve slopes for set 04_A at to and tf. Assuming a 75

mg sample, these would become 0.42 and 0.46 wt. % respectively. When these values are

used in the equation provided by Schwarcz, PMI estimates are 20.1 and 17.3 years

respectively. Thus, the slight change in slope causes a difference of 3 years in the PMI

obtained for the same sample. If only aged standards are available, the investigator must

be aware of the calibration curves may impact the calculated PMI in bone samples.

120
 4.3.4 Analysis of Stock Solution pH on Standards

Four of the 11 stocks prepared in this thesis were prepared with only water that

resulted in a pH of 9 (Table 4.9). The remaining six stocks were prepared with sulfuric

acid mixed in to lower the pH to 2. This was done to inhibit microbial growth, as well as

have the standards at the same pH as the bone solutions (Table 4.10). Stock 3 was not

included in this analysis as no time = 0 weeks data was available for analysis.

The different pH populations were qualitatively compared to determine if the pH

of the standards had an impact on the linearity of the calibration curve and quantitatively

compared to determine if the pH had an impact on the peak areas of the each standard

concentration [0, 0.5, 1.0. 1.5, 2.0, 2.5 mM]. It was hypothesized that the pH of the

solution would not impact linearity of the calibration curve. Furthermore, it was

hypothesized that pH would not be a significant factor in the detected peak area for each

of the standard concentrations.

121
 Table 4.9 Slope and R2 Values of Stock Solutions with pH 9 Prepared in this Thesis

Set of Inj Volume Stock Acceptable

Stock Slope R2 pH
Standards (µL) Solution Linearity
1 01_A 47899 0.9863 40 H2O 9 Yes
2 02_A 46653 0.9937 50 H2O 9 Yes
6 06_A 56016 0.9933 50 H2O 9 Yes
10 10_A 46400 0.9996 50 H2O 9 Yes

Table 4.10 Slope and R2 Values of Stock Solutions with pH 2 Prepared in this Thesis

Set of Inj Volume Stock Acceptable

Stock Slope R2 pH
Standards (µL) Solution Linearity
4 04_A 50996 0.9991 50 H2SO4 / H2O 2 Yes
5 05_A 52677 0.9981 50 H2SO4 / H2O 2 Yes
7 07_A 52093 0.9996 50 H2SO4 / H2O 2 Yes
8 08_A 53162 0.9996 50 H2SO4 / H2O 2 Yes
9 09_A 49776 0.9995 50 H2SO4 / H2O 2 Yes
11 11_A 49534 0.9985 50 H2SO4 / H2O 2 Yes

All of the sets analyzed had an acceptable degree of linearity (R2 > 0.98) as shown

in Section 4.3.1. As such, no further analysis was done to evaluate linearity, and all

subsequent analysis focused on the impact on peak areas for each concentration.

Set 01_A standards were injected at a lower volume, 40 µL, compared to all other

samples, but were included in the analysis, as the peak areas were within the range seen

in all other samples. The areas for five of the concentrations [0.5, 1.0, 1.5, 2.0, and 2.5

mM] were used for statistical comparison, as not all populations had sufficient samples of

the 0.1 mM standard for statistical significance (Tables 4.11 and 4.12).

122
Table 4.11 Peak Areas of Standard Concentrations prepared from pH 2 Stock Solutions

Concentration Avg. Peak RSD N= Number

SD
(mM) Area (t=0) (%) of Samples
0.50 26854.17 1842.08 6.86 6
1.00 51941.33 2827.42 5.44 6
1.50 78015.33 3857.43 4.94 6
2.00 102287 3357.21 3.28 6
2.50 128759 3145.75 2.44 6

Table 4.12 Peak Areas of Standard Concentrations prepared from pH 9 Stock Solutions

Avg. Peak
Concentration RSD N= Number
Area (t=2-4 SD
(mM) (%) of Samples
weeks)
0.5 22629 2377.62 10.51 3
1.0 48423 2363.41 4.88 4
1.5 70870.5 5969.71 8.42 4
2.0 95997.33 2801.04 2.92 3
2.5 124387.25 13198.1 10.61 4

Table 4.13: Results of Unpaired t-test Analysis of Citrate Peak Areas for Standards
Prepared with pH 2 and pH 9 Stock Solutions

t-test p
Concentration Significance
value
0.5 mM 0.06 No
1.0 mM 0.10 No
1.5 mM 0.07 No
2.0 mM 0.04 Yes
2.5 mM 0.51 No

123
 As seen above (Table 4.13), a significant difference (unpaired t-test, p< 0.05) was

seen in the peak areas of 2.0 mM standards prepared from stock solutions of different pH.

Also, the standard deviations seen in the pH 9 stock solution standards were much higher

than in the pH 2 populations. Some of this may attributed to a larger population of pH 2

standards, as it was the preferred method of preparing the stock solutions. Another cause

may have been the inclusion of two early sets of standards, 01_A and 02_A, which may

have been impacted by a lack of experience in the preparation of the standards.

4.4 Analysis of Citrate Peaks in Bone Solution Chromatograms

This section will seek to address some of the general observations made

throughout the entirety of testing in regards to the performance of the HPLC with bone

samples. Results will include both qualitative and quantitative analysis when possible.

4.4.1 Chromatogram Peaks

The chromatograms produced by bone solutions were consistent throughout the

entire thesis. In general, each chromatogram contained a large peak preceding citrate with

a retention time in the 6.5 to 7 minute time range. Each chromatogram also contained a

citrate peak with retention times ranging from 8.0 in early studies to 9.25 minutes in later

studies. The increased retention time in later analyses suggested that the column was

altered over the course of the thesis. The citrate peak occurred slightly later in bone

samples than in the simple citrate matrices, but was confirmed using spiked samples. In

each of the chromatograms produced in this study, the citrate peak was easily identified;

124
however, as the study progressed, the separation between the two peaks diminished.

Despite this loss of resolution, the beginning and end of the citrate peak were still easily

identified, and sufficient enough for integration to determine peak area. A typical

chromatogram, as well as a spiked chromatogram of the same sample produced in this

study are demonstrated in Figure 4.11 below.

R4B5D (unspiked) R4B5D (spiked with 1.63 mM)

0.029

0.024

0.019
Absorbance Units

0.014

0.009

0.004

-0.001
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Time (Min)

Figure 4.11: Chromatograms of Cortical Only Bone Sample Obtained from the Dorsal
Region of a Porcine Rib Bone both spiked with 1.63 mM citrate and unspiked. Spiked is
the orange line, unspiked is the purple line. [Specimen: R4B5 Dorsal]

125
 4.4.2 Integration of Citrate Peaks in Bone Samples

In general, citrate peaks in bone solutions analyzed in this study had clearly

defined boundaries. As such, areas under citrate peaks were well approximated with a

simple tangential skim fit using the method described in Section 3.5.8, Chapter 3 (Figure

4.12). Initial points for the integration were obtained through the use of the computer

integration fit, but corrections were necessary in most instances as the computer

overestimated the area. Corrections were made by zooming into the citrate peaks and

determining the point at which the citrate peak began to steadily increase. End points for

integration were determined based upon a return to baseline.

Figure 4.12 Representative Image of Tangential Skim Integration to Determine Area

under Citrate Peak in Chromatogram [Specimen: 75-1 Injection 2]

Over the life of the column, peak resolution decreased in bone chromatograms.

While the boundaries of citrate peaks continued to be apparent, a lack of baseline return

meant that a tangential skim fit for integration would underestimate the area under the

peak due to the incomplete separation of the two peaks. As such, different fits for

126
integration were investigated in order to determine how extensive of an impact column

degradation had on the ability to quantify citrate concentrations in the samples. The fits

used were the computer fit or how the computer determined the best fit, a tangential fit or

straight line drawn from start to bottom, and an exponential fit. An example of each of

the fits on a single sample with an uneven baseline (Figure 4.13). The areas obtained by

each method are listed below (Table 4.14). The computer method overestimated the true

area of the peak, and there was no difference between the exponential and tangential skim

fits. The tangential skim integration fit was utilized in all data analysis as it best

approximated the area under the peaks.

127
 Table. 4.14 Peak Areas of Citrate Peak in 75 mg Cortical Only Bone Solution Using
Different Fits of Integration from Waters Breeze Software. Each fit is for bone solution
[R4B7 Ventral]

Fit Type Area

Computer 209311
Exponential 34929
Tangential Skim 34929

0.008

7.111
0.007

0.006

0.005

9.870
0.004
AU

0.003

13.337
0.002

0.001

0.000

0.00 1.00 2.00 3.00 4.00 5.00 6.00 7.00 8.00 9.00 10.00 11.00 12.00 13.00 14.00
Minutes

(A)
0.008
7.111

0.007

0.006
9.543

0.005

0.004
AU

0.003

13.337
0.002

0.001

0.000

0.00 1.00 2.00 3.00 4.00 5.00 6.00 7.00 8.00 9.00 10.00 11.00 12.00 13.00 14.00
Minutes

(B)
0.008
7.111

0.007

0.006
9.543

0.005

0.004
AU

0.003
13.337

0.002

0.001

0.000

0.00 1.00 2.00 3.00 4.00 5.00 6.00 7.00 8.00 9.00 10.00 11.00 12.00 13.00 14.00
Minutes

Figure 4.13 Screen captures of Breeze 2 Integration Fits. (A) Computer Fit, (B)
Exponential Fit, (C) Tangential Skim Fit Integration [R4B7VEN]

128
 4.4.3 Injection Dependence

The initial bone sample analyzed in each HPLC run produced a larger first,

unknown, peak when compared to subsequent injections (Figure 4.14). The unknown

peak in the first injection had a peak height twice as large as the unknown peak in the

second injection of the same sample; citrate peaks did not exhibit the same trend. Peak

resolution was the same in each of the injections, as was the retention time of the citrate

peaks.

0.18
0.17
0.16
0.15
0.14
0.13
0.12
0.11
Absorbance Units

0.10
0.09 Injection 1
Injection 2
0.08
Injection 3
0.07
0.06
0.05
0.04
0.03
0.02
0.01
0.00
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Time (Min)
Figure 4.14 Overlaid Chromatograms of Three Consecutive Injections of a Mixed Bone
Solution. These are from the optimal mass test group, specimen OM-50 1.

129
 First injections of three 50 mg mixed (cortical and trabecular) bone samples

through the column were compared with second and third injections to determine if the

injection number was a significant factor in the detected concentration of citrate (wt. %).

It was hypothesized that no significant difference would be observed in subsequent

injections. Table 4.15 below lists the values for three consecutive injections.

Table 4.15: List of Citrate Concentrations (wt.%) for Three Consecutive Injections of
Mixed Bone Specimens

Injection 1 Injection 2 Injection 3

Sample wt.% wt.% wt.%
OM-50-1 1.067 0.970 0.943
OM-50-2 0.879 0.917 0.919
OM-50-3 0.954 0.871 0.865
Mean 0.967 0.919 0.909
STD 0.077 0.041 0.033

Statistical analysis was completed to determine if injection number (1, 2, or 3)

was a factor in the measured wt. % concentration of citrate. Injection number was not a

significant factor in the average concentrations measured in the three bone samples

(Repeated Measures ANOVA, p=.26), indicating no significant difference in the detected

citrate concentrations (wt. %). Thus, the larger peak witnessed in the first bone sample

injected through the column in a run has no impact on the detected citrate concentration

(wt. %) in bone samples.

130
 4.5 Applied Tests with Citrate HPLC Method

A series of 8 applied tests were performed to evaluate the HPLC test method

developed in this thesis on the detection of citrate in bone samples. In particular, optimal

mass of samples were studied to determine the best mass of bone to sample. Second,

standard recovery and addition tests were performed to determine the presence of matrix

effects [23]. Third, different types of bones were analyzed qualitatively and quantitatively

to determine the limitations of the method with various bone types. Fourth, different

storage environments were evaluated to determine if where short term storage

environment of samples would have an impact on the citrate concentration. Fifth, three

regions of bone were sampled to determine if the location of a specimen along a bone had

an impact on the measured concentration of citrate. Sixth, samples filtered with molecular

weight cutoff centrifuge tubes and unfiltered samples were qualitatively and

quantitatively analyzed to determine if filtration had a noticeable impact on the peaks

observed in chromatograms and detected citrate concentrations. Seventh, different

methods used to obtain concentrations in samples, standard addition and calibration curve

methods, were compared to determine if any significant difference in concentrations was

seen between the two populations. Finally, the results of each of these studies were

pooled and estimated PMI were calculated using the equation developed by Schwarcz et

al. in order to evaluate the suitability of the model [49].

131
 4.5.1 Optimal Mass Study

As described in section 3.5.1 (Ch. 3), nine bone solutions consisting of three

samples for three different mass groups (50, 75, and 100 mg) were prepared and

analyzed. Each of these samples were qualitatively analyzed for citrate peak height and

resolution and quantitatively analyzed for citrate concentration. In this experiment, it was

hypothesized that the mass of bone analyzed would not have an impact on the normalized

concentration, wt. %, of citrate detected.

Qualitative Analysis

Chromatograms for each of the three mass groups were evaluated to determine the

impact of mass on peak heights and separation. Figure 4.15 below contains the raw

chromatograms overlaid. Citrate peaks were then normalized by setting the first

absorbance measurement to zero by adding or subtracting the measured value. This value

was then used as a constant and added or subtracted from each absorbance measurement.

This process was repeated for the 50, 75, and 100 mg bone solution citrate peaks. The

normalized citrate peaks, as well as the overall chromatograms for each sample were

qualitatively analyzed to ensure peak areas and heights scaled with increased mass. An

example analysis is provided below (Figure 4.16).

132
 0.030

0.025

0.020
Absorbance
Units

0.015

Citrate
0.010

0.005

0.000
5 5.5 6 6.5 7 7.5 8 8.5 9 9.5 10
Time (Min)
100 mg 50 mg 75 mg

Figure 4.15: Chromatograms of 50, 75, and 100 mg Mixed Bone Samples from Optimal
Mass Group One Overlaid. Citrate peak is indicated on chromatogram.

133
 0.0036

0.0031

0.0026
Absorbance Units

0.0021

0.0016

0.0011

0.0006

0.0001

-0.0004
7.81 8.01 8.21 8.41 8.61 8.81 9.01
Time (Min)
50 mg 75 mg 100 mg

Figure 4.16 Citrate Peaks of 50, 75, and 100 mg Mixed Bone Samples from Optimal
Mass Group One Normalized and Overlaid. Citrate peaks were normalized by setting the
first measurement to zero absorbance units and using the difference as a constant that was
either added or subtracted from each subsequent measurement.

134
 Analysis of the raw chromatograms showed clearly defined citrate peaks in each

of the three mass groups occurring between 8 and 9 minutes. Furthermore, each

chromatogram had an unknown peak beginning around 6 minutes and ending at shortly

after 7 minutes (Figure 4.13). The normalized citrate peaks demonstrated increasing areas

as mass was increased (Figure 4.14). Each mass group also produced citrate peaks that

were easily integrated and quantified. No clear distinction was evident in the

chromatograms, so quantitative analysis was necessary to further test the impact of mass

on citrate analysis.

Quantitative Analysis

As stated in Section 3.5.1, a total of 27 analyses, 3 injections for each of the 3

mass populations comprised of three samples each, were performed. The peak processing

method described in section 3.3.8 (Ch. 3) was used to determine citrate concentration in

each of the bone samples. Table 4.16 shows the molar concentration (mM) as well as the

normalized concentration obtained by using the wt. % processing procedure outlined in

section 3.3.8 (Ch. 3).

135
 Table 4.16: Average Peak Areas and Concentrations (mM and wt. %) of Citrate in 50,
75, and 100 mg Mixed Bone Samples. These samples contain both Cortical and
Trabecular Bone

Citrate Citrate
Sample Concentration Concentration
(mM) (wt.%)
50-1 0.78 1.020
50-2 0.71 0.918
50-3 0.74 0.961

75-1 1.20 0.967

75-2 1.22 0.982
75-3 1.22 0.980

100-1 1.86 1.057

100-2 1.82 1.030
100-3 1.68 0.947

Statistical analysis was completed to determine if different bone mass (50, 75, and

100 mg) was a factor in the measured wt. % concentrations. Bone mass was not a

significant factor in the average wt. % concentrations (ANOVA, p=0.48), indicating no

detectable difference between the various bone mass samples (Tukey post-hoc multiple

comparison, p>0.05). Thus, the mass of bone used in the sample preparation process had

no impact on the determined concentration of citrate present in the bone when normalized

into weight percent. All subsequent analysis in this thesis used 75mg bone samples as the

average molar concentration obtained in this study at this mass was 1.21 mM. This

concentration was well within the range of standards [0 – 2.5 mM] used to generate the

calibration curve.

136
 4.5.2 Analysis of Matrix Effects

As described in Chapter 3 (Sections 3.5.2 and 3.5.3), experiments were designed

to evaluate the presence of matrix effects in the analytical method. Results below are

provided for standard addition and recovery tests. It this experiment, it was hypothesized

that matrix effects were not present in the analytical method.

4.5.2.1 Standard Addition Tests

Standard Addition Tests with Mixed Bone Samples

Standard addition tests were performed on a total of three mixed bone samples

using the standard addition method described in Chapter 3 (Section 3.5.2). The additions

were approximately 50% 150%, and 300% times the maximum concentration, 2 wt. %,

listed in literature [49]. Volumes for additions were calculated based on a final total

volume of 5 mL, but each bone sample only produced 3.5 mL. As such, the

concentrations tested were higher than initially calculated. Qualitative analysis was done

on the chromatograms obtained to ensure that the additions of citrate were enlarging the

citrate peak. Quantitative analysis was performed to determine if matrix effects were

present as evidenced by a loss of linearity of the curve produced by the additions

compared to the calibration curve produced by standards.

Standard addition tests were first performed on mixed bone samples as it was

theorized that matrix effects would be most prevalent in the complex matrix of all bone.

Calibration curve results were compared with equally divided bone solutions containing

various additions of 0.1M stock solution. It was hypothesized that acceptable linearity

137
and similar response factors would be observed between the standard addition and

calibration curves.

Qualitative

Prior to quantitative analysis, chromatograms of the additions were overlaid to

ensure that as amount of citrate added increased, the area of the citrate peaks increased

proportionally. An example showing this output for the additions of Standard Addition

Set One is provided below (Figure 4.17).

138
 0 mM 1.64 mM 4.00 mM 7.69 mM
0.030

0.025

0.020
Absorbance Units

0.015

0.010

0.005

0.000
7 7.5 8 8.5 9 9.5 10 10.5 11
Time (Min)

Figure 4.17 Chromatograms of Original Bone Sample and Three Additions Overlaid.
Each curve corresponds to a volume of 0.1M stock solution added to a 600 µL aliquot of
solution. [Standard Addition Set One]

Quantitative

Peak areas for each of the additions were obtained using the post-processing

method described in Chapter 3 (Section 3.3.8). It was hypothesized that the curve

produced by the additions would have acceptable linearity as defined by Chapter 3

(Section 3.3.10). Furthermore, it was hypothesized that no significant difference in

concentration in wt. % would be found between the use of a calibration curve and the

standard addition methodology. The areas and calibration curves obtained for each set of

additions are found below (Table 4.17-4.19, Figures 4.18-4.20).

139
 Table 4.17: Peak Areas for Standard Addition Set One. The R2 and x-intercept are also
provided.

Concentration Peak Area (Area

(mM) Units)
0.00 48161
1.64 146859
4.00 267805
7.69 445347
x-intercept (abs) 1.10 mM
R2 0.9980

500000

450000 y = 51201x + 56414

R² = 0.998
400000

350000
Area Units

300000

250000

200000

150000

100000

50000

0
0.00 1.00 2.00 3.00 4.00 5.00 6.00 7.00 8.00 9.00
Concentration (mM)

Figure 4.18 First-order (Straight line) fit of the Peak Areas Obtained in Standard
Addition Set One. The slope of the standard calibration curve was similar, and as such no
matrix effects appeared to occur.

140
Table 4.18: Peak Areas for Standard Addition Set Two. The R2 and x-intercept are also
provided.

Concentration Peak Area (Area

(mM) Units)
0.00 46696
1.64 136245
4.00 265900
7.69 440006
x-intercept (abs) 1.01 mM
R2 0.9984

500000

450000 y = 51151x + 51753

R² = 0.9984
400000

350000
Area Units

300000

250000

200000

150000

100000

50000

0
0.00 1.00 2.00 3.00 4.00 5.00 6.00 7.00 8.00 9.00
Concentration (mM)

Figure 4.19 First-order (straight fit) line of the Peak Areas Obtained in Standard
Addition Set Two. Slope was similar to that of the standard calibration curve, indicating
no matrix effects in the analytical procedure.

141
Table 4.19: Peak Areas for Standard Addition Set Three. The R2 and x-intercept are also
provided.

Concentration Peak Area (Area

(mM) Units)
0.00 Not enough data
1.64 153037
4.00 272671
7.69 469227
x-intercept (abs) 1.25 mM
R2 0.9998

500000
y = 52351x + 65698
450000 R² = 0.9998
400000

350000
Area Units

300000

250000

200000

150000

100000

50000

0
0.00 1.00 2.00 3.00 4.00 5.00 6.00 7.00 8.00 9.00
Concentration (mM)

Figure 4.20 First-order (Straight line) fit of the Peak Areas Obtained in Standard
Addition Set Three. Not enough sample volume was available of the 0 mM standard, so
no data was obtained in this set.

142
 Table 4.20 Slope and R2 Values Obtained in All Mixed Bone Standard Addition Sets
Tested

x- Uncertainty of
Set Slope R2 intercept X-intercept
(mM) (mM)
SA1 51201 0.9980 1.1 0.17
SA2 51151 0.9984 1.01 0.15
SA3 52351 0.9998 1.25 0.05
Mean 51567.667 0.999 1.120 0.12
SD 554.276 0.001 0.099 0.05

Acceptable linearity for the standard addition curves was an R2 >0.98. All of the

first-order (straight line) fits generated in this analysis had acceptable linearity. The

unspiked aliquot of bone sample in standard 3 did not contain enough volume to test due

to loss of solution in transfer, but the linearity of the curve is unaffected.

The results support the hypothesis that matrix effects are not present in the

samples. Linearity is maintained in all three analysis and as the citrate concentration

increases, the peak areas increase in a linear manner. Furthermore, the slopes of all three

curves are within an acceptable 20% of the slope of the calibration curve obtained in this

experimental run, slope = 56016, set 06_A (t=0 weeks) (Table 4.20, Figure 4.21).

143
 500000

450000 y = 51201x + 56414

R² = 0.998
400000

350000
Area Units

300000

250000
y = 51094x + 1976.6
200000
R² = 0.9968
150000

100000

50000

0
0.00 1.00 2.00 3.00 4.00 5.00 6.00 7.00 8.00 9.00
Concentration (mM)

Figure 4.21 Comparison of Standard Addition Set 1 Curve with Standards set 05_A (t=0)
forecasted forward. The lines appear to be parallel over the range of data indicating an
equivalent slope. The R2 values also both indicate high degrees of linearity.

Standard Addition of Cortical Only Bone Solutions

Standard addition tests were subsequently performed with cortical bone solutions

once a cortical only sampling procedure was adopted. These analyses were performed to

confirm the absence of matrix effects in the analytical procedure as determined with the

mixed bone solutions. Additions in these tests were done at lower concentrations, as a

more accurate final volume of bone solutions was used in calculations for additions. In

this section, three of the addition tests performed throughout the thesis will be provided

(Figure 4.22), however, Appendix D contains a complete database of calibration curves

for all cortical only solutions analyzed using a standard addition method throughout this

thesis.

144
 200000
A
160000
Area Units

120000
80000 y = 39811x + 62004
40000 R² = 0.9995

0
-2 -1.5 -1 -0.5 0 0.5 1 1.5 2 2.5
Concentration (mM)

240000
B 200000
Area Units

160000
120000
80000 y = 49297x + 72639
40000 R² = 0.9997
0
-2.5 -2 -1.5 -1 -0.5 0 0.5 1 1.5 2 2.5
Concentration (mM)

200000
C 160000
Area Units

120000
y = 43184x + 69901
80000
R² = 0.9998
40000

0
-2 -1.5 -1 -0.5 0 0.5 1 1.5 2 2.5
Concentration (mM)

140000
120000 D
100000
Area Units

80000 y = 53162x
R² = 0.9996
60000
40000
20000
0
0.00 0.50 1.00 1.50 2.00 2.50
Concentration (mM)

Figure 4.22 First-order (Straight line) fits of three Standard Additions performed on 75
mg Cortical Only Bone Solutions and Calibration Curve of Set 08 A. (A) Solution: VAC
2, (B) RT 2, (C) H2O 2, (D) Calibration Curve [Set 08A]

145
 H202 Linear (Calibration) Linear (VAC 2) Linear (RT 2) Linear (H202)

140000
y = 53162x
R² = 0.9996

120000 y = 49297x + 760.3

R² = 0.9997
y = 43184x - 275.8
R² = 0.9998
100000
y = 39811x + 820.4
R² = 0.9995

80000
Area Units

60000

40000

20000

0
0.00 0.50 1.00 1.50 2.00 2.50 3.00

-20000
Concentration (mM)

Figure 4.23 First-order (Straight line) fits of three Standard Additions performed on 75
mg Cortical Only Bone Solutions adjusted and compared to Calibration Curve of Set 08
A. Additions were adjusted by subtracting the peak area of the unspiked sample (0 mM)
from the subsequent additions. The calibration curve is the first curve in the figure.

In contrast to the mixed bone solutions, the standard addition tests of cortical only

bone solutions demonstrate the presence of matrix effects in the analytical procedure

(Figure 4.23). Also, the effects appear to vary from each sample which were taken from a

single bone. Overall, the effects appear to best resemble a translational matrix effect, such

that as more citrate is spiked into bone solutions, the peak area response decreases

146
proportionally [55]. The difference between mixed and cortical solutions may be due to

the high concentrations of additions performed in the mixed bone solutions that may have

biased the outcomes of the tests.

4.5.2.2 Standard Recovery

As described in section 3.5.3 (Ch. 3), three standard recovery tests were

performed with combined cortical and trabecular bone solutions. Using a complete bone

specimen ensured that the maximum number of components possible in a sample were

present throughout analysis. Furthermore, it was hypothesized that matrix effects present

in the combined matrix would be present in either cortical only or trabecular only

matrices. One cortical only standard recovery test was performed later in the thesis, when

a cortical only sampling protocol was adopted.

Peak areas were obtained for each of the four concentrations in each standard

recovery set, as well as the unspiked bone solution. Areas were then converted into molar

concentrations using the calibration curve obtained from standards set 06 which were

analyzed under the same operating conditions. First-order (straight line) fits were

generated for both raw and processed bone solution data. A calibration curve for the bone

samples was created by subtracting the peak area from each of the spiked samples. The

standard calibration curve as well as the adjusted and non-adjusted matrix calibration

curves for each set were produced to determine the presence or absence of matrix effects.

Each of the non-adjusted and adjusted calibration curves for bone solutions analyzed

using standard recovery are included below (Figures 4.24 – 4.30).

147
 Linear (SR 1) Linear (Calibration Curve)

250000

200000
y = 52034x + 67157
150000 R² = 0.981
Units
Area

100000

y = 52679x
50000
R² = 0.9976

0
-2 -1.5 -1 -0.5 0 0.5 1 1.5 2 2.5 3
-50000
Concentration (mM)

Figure 4.24 Non-adjusted Standard Recovery Calibration Curve for Standard Recovery
Set One Solutions Compared to Standard Calibration Curve. This curve has the raw data
points for the unspiked and spiked bone solutions. [Sample Set: SR1]

Bone Matrix Standards

160000
140000 y = 51891x + 1445.3
R² = 0.9979
120000
100000
Area Units

80000
y = 52034x - 6423.3
60000 R² = 0.981
40000
20000
0
0 0.5 1 1.5 2 2.5 3
-20000
Concentration (mM)

Figure 4.25 Adjusted Standard Recovery Calibration Curve for Standard Recovery Set
One Solutions Compared to Standard Calibration Curve. This curve has the processed
data points for the spiked bone solutions and the unspiked sample is included as the
origin [Sample: SR1].

148
 Linear (SR 2) Linear (Calibration Curve)

200000
180000 y = 47878x + 58803
160000 R² = 0.9746
140000
Area Units

120000
100000
80000 y = 52679x
60000 R² = 0.9976
40000
20000
0
-1.5 -1 -0.5 0 0.5 1 1.5 2 2.5 3
Concentration (mM)

Figure 4.26 Non-adjusted Standard Recovery Calibration Curve for Standard Recovery
Set Two Solutions Compared to Standard Calibration Curve. This curve has the raw data
points for the unspiked and spiked bone solutions. [Sample Set: SR2]

Bone Matrix Standards

140000
y = 51891x + 1445.3
120000 R² = 0.9979

100000
Area Units

80000
y = 47878x - 2079.4
60000 R² = 0.9746

40000

20000

0
0 0.5 1 1.5 2 2.5 3
Concentration (mM)

Figure 4.27 Adjusted Standard Recovery Calibration Curve for Standard Recovery Set
Two Solutions Compared to Standard Calibration Curve. This curve has the processed
data points for the spiked bone solutions and the unspiked sample is included as the
origin [Sample: SR2].

149
 Linear (SR 3) Linear (Calibration Curve)

200000
y = 45327x + 63774
180000 R² = 0.9711
160000
140000
Area Units

120000
100000
80000
60000 y = 52679x
40000 R² = 0.9976
20000
0
-2 -1.5 -1 -0.5 0 0.5 1 1.5 2 2.5 3
Concentration (mM)

Figure 4.28 Non-adjusted Standard Recovery Calibration Curve for Standard Recovery
Set Three Solutions Compared to Standard Calibration Curve. This curve has the raw
data points for the unspiked and spiked bone solutions. [Sample Set: SR3]

Bone Matrix Standards

140000
y = 51891x + 1445.3
120000 R² = 0.9979
100000
Area Units

80000
y = 45327x + 2853.6
60000 R² = 0.9711

40000

20000

0
0 0.5 1 1.5 2 2.5 3
Concentration (mM)

Figure 4.29 Adjusted Standard Recovery Calibration Curve for Standard Recovery Set
Three Solutions Compared to Standard Calibration Curve. This curve has the processed
data points for the spiked bone solutions and the unspiked sample is included as the
origin [Sample: SR3].

150
 200000

180000

160000

140000

120000
Area Units

100000

80000

60000

40000

20000

0
0.00 0.50 1.00 1.50 2.00 2.50 3.00 3.50 4.00
Concentration (mM)

Set 1 Set 2 Set 3 Linear (Standards)

Figure 4.30 Comparison of each adjusted citrate spiked bone solution to set 06_A
calibration curve

Only one out of the three, set one, bone matrix calibration curves produced had

acceptable linearity, R2 > 0.98. Also, only set one had a bone matrix slope that was

parallel to the standard calibration curve. Despite the slopes being equal, the y- intercepts

of the bone matrix of set one and standard calibration curves were not equal. This is best

characterized as a translational matrix effect, meaning that the loss of citrate or change of

signal was independent of the concentration of the citrate spiked into the bone solutions

[55]. The bone matrix calibration curves obtained in sets two and three are best

characterized by a rotational matrix effect. This type of matrix effect displays as a loss of

citrate or the change of signal proportional to the concentration of the citrate spiked into

the bone samples [55].

151
 Further confirmation of matrix effects was seen by calculating the % recovery of

citrate in each of the spiked bone solutions (Table 4.21). Percentage of citrate recovered

in each of the solutions varied from 59% to 103%.

Table 4.21 List of % Recovery for Each 75mg Mixed Bones Samples spiked With Citrate
in this study

Concentration Concentration %
Set
Added Detected Recovery
1 0.450 0.264 58.737
0.930 0.582 62.617
1.500 1.511 100.714
2.770 2.614 94.362
Mean 79.108
SD 18.617
2 0.590 0.581 98.505
1.290 1.008 78.130
1.640 1.298 79.138
2.350 2.251 95.776
Mean 87.887
SD 9.310
3 0.710 0.666 93.775
1.180 1.006 85.221
1.760 1.826 103.744
2.700 2.237 82.864
Mean 91.401
SD 8.202

152
 Cortical Only Standard Recovery Tests

A single standard recovery test was performed with cortical bone once a cortical

only sampling procedure was adopted (Figures 4.31-4.32). All methods used on mixed

bone solutions were repeated except for the use of a 0.1M citrate stock solution instead of

citrate crystals.

Linear (SR 75C) Linear (Standards)

200000
y = 45749x + 74929
R² = 0.998
180000

160000

140000

120000
Area Units

100000

80000
y = 52093x
R² = 0.9996
60000

40000

20000

0
-2.0 -1.5 -1.0 -0.5 0.0 0.5 1.0 1.5 2.0 2.5 3.0
Concentration (mM)

Figure 4.31 Non-adjusted Standard Recovery Calibration Curve for Cortical Only
Sample Compared to Standard Calibration Curve. This curve has the raw data points for
the unspiked and spiked bone solutions. [Sample: SR 75C]

153
 Linear (SR 75 C) Linear (Standards)

140000

120000 y = 52093x
R² = 0.9996
100000
y = 45749x + 1892.3
Area Units

80000 R² = 0.998

60000

40000

20000

0
0 0.5 1 1.5 2 2.5 3
Concentration (mM)

Figure 4.32 Adjusted Standard Recovery Calibration Curve for Cortical Only Sample
Compared to Standard Calibration Curve. This curve has the processed data points for
the spiked bone solutions and the unspiked sample is included as the origin [Sample: SR
75C].

Table 4.22 Percent Recovery of Citrate Based on Adjusted Concentrations of Citrate in

Cortical Only Standard Recovery Test Solutions

Concentration Concentration %
Added (mM) Detected (mM) Recovery
0.50 0.521 104.30
1.00 0.892 89.150
1.39 1.175 84.582
2.50 2.238 89.532
Mean 91.891
SD 7.424

154
 The curve that was obtained had a linear response indicating that as citrate was

added to the samples, the response of the system increased proportionally, R2 > 0.98. In

addition, the cortical only standard recovery curve demonstrated rotational matrix effects

much likes sets two and three in the mixed bone recovery tests [55]. This confirmed the

hypothesis that matrix effects present in mixed solutions would be translational to all

samples analyzed with this technique.

4.5.2.3 Strategy for Reducing Matrix Effects in Bone Sample Analysis

The presence of matrix effects in the standard recovery tests indicated the need

for a strategy to reduce matrix effects. Three following three general approaches were

considered for all subsequent tests: matrix matched standards, standard addition, or

mathematical compensation [55]. Each of the methods were analyzed for feasibility and

ease of implementation with the HPLC method. This analysis led to the adoption of a

standard addition methodology, as it was determined to be the easiest strategy to

implement and was also feasible with the bone samples. All subsequent quantitative

analysis utilized standard addition strategies to compensate for matrix effects in the bone

solutions.

4.5.3 Comparison of Trabecular, Cortical, and Mixed Bone Samples

One of the key questions to arise from a review of literature is whether citrate

concentrations vary in different types of bone (cortical versus trabecular). Data were

analyzed to determine qualitative differences in chromatograms, as well as quantitative

differences in citrate concentrations between the different bone types. As described in

155
section 3.5.4 (Ch. 3), samples of cortical bone only, trabecular bone only, and mixed

(cortical and trabecular) bone were compared. In total, nine cortical only, six trabecular

only, and three mixed solutions were tested. Each of the cortical only and trabecular only

samples discussed here utilized the standard addition method described in Section 3.5.2

and all were 75mg. The mixed samples analyzed in this experiment were taken from the

75 mg standard addition tests discussed in section 4.4.1 (Ch. 4).

Qualitative Comparison

The chromatograms for all three bone types [cortical, trabecular, and both] were

analyzed for separation of peaks, retention time, and peak height. Figures 4.33-4.35

contain a representative chromatogram for each of the types of bone evaluated.

156
 0.019
Absorbance Units

0.014

0.009

0.004

-0.001
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14
Time (Min)
Figure 4.33 Chromatogram of 75mg Cortical Only Bone Sample from Dorsal Region of
Porcine Rib Bone. [Specimen: R4B6DC]

0.016
0.014
Absorbance Units

0.012
0.010
0.008
0.006
0.004
0.002
0.000
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14

Time (Min)

Figure 4.34 Chromatogram of 75mg Trabecular Only Bone Sample from Dorsal Region
of Porcine Rib Bone. [Specimen: R4B6DT]

0.020
0.018
0.016
0.014
Absorbance Units

0.012
0.010
0.008
0.006
0.004
0.002
0.000
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14
Time (Min)

Figure 4.35 Chromatogram of 75mg Mixed Bone. Solution from Standard Addition
Experiments done Earlier in Testing. [Specimen: SA-1]

157
 Two main peaks, unknown and citrate, appeared in all of the samples, and an

extra peak with a retention time of 12.5 minutes appeared in all of the mixed solutions.

Citrate peaks had normal retention times of 9 minutes in cortical only, 9.5 minutes in

trabecular only samples, and 8.5 minutes in mixed samples. UV-Vis spectra of the citrate

peaks were compared to spectra of standard solutions to verify that peaks were citrate.

Furthermore, the cortical and trabecular citrate peaks elute over a longer period of time,

1.75 minutes compared to the one minute elution time of mixed samples.

The bone samples used to generate the cortical and trabecular chromatograms

were taken from the same rib section and analyzed under the same operating conditions.

Mixed samples were tested at a higher operating pressure of the column, and as such

were not directly compared for peak height or separation with the cortical only and

trabecular only chromatograms. When the cortical and trabecular chromatograms are

compared, the citrate peak in the cortical only sample has a higher peak and better

separation as the unknown peak decreases to .002 absorbance units before the citrate peak

begins. In the trabecular only sample, the citrate peak begins while the unknown peak

still produces a signal of approximately .004 absorbance units. Furthermore, the cortical

only sample has a higher peak height, 0.002 absorbance units, compared to the trabecular

only, 0.001 absorbance units. Although the cortical chromatograms demonstrated better

peak height and separation compared to the trabecular samples, a quantitative analysis

was necessary to determine if there was a difference in the citrate concentrations in the

three types of samples analyzed.

158
 Quantitative Comparison

Four of the six trabecular samples tested did not have peaks that could be

integrated, while all of the cortical only samples and two of the three mixed samples did.

One of the two trabecular only samples that was capable of integration had a very low

citrate concentration, 0.23 wt. %, while the other had a larger, 0.83%, but still somewhat

low concentration compared to the cortical only bone samples from the same specimen

(Table 4.23). The variance in the two trabecular only samples represents a range of

approximately 33 years when using the Schwarcz equation. Mixed samples also had

lower concentrations when compared to the cortical only samples. Cortical only samples

provided the highest concentrations of citrate and also had the lowest PMI estimates, less

than one year, with the Schwarcz equation (Table 4.23).

Table 4.23 Citrate Concentrations and Estimated PMI for Trabecular, Cortical, and
Mixed Samples Analyzed With HPLC Test Method. Cortical and Trabecular Samples
Were taken From Same Specimen.

Type of Concentration Concentration PMICALC

Sample
Bone (mM) (wt.%) (Years)
R4 B5 CEN Trabecular 0.95 0.83 4.9
R4 B6 DORS Trabecular 0.26 0.23 38.4
R4 B5 CEN Cortical 1.87 1.62 0.3
R4 B6 DORS Cortical 1.64 1.42 0.7
SA 1 Mixed 0.86 0.73 6.9
SA 2 Mixed 0.83 0.71 7.4

159
 The inability to consistently detect citrate and the low concentrations obtained in

the two samples that had adequate peaks, demonstrated a limitation of the HPLC

analytical method with trabecular only samples. In addition, the inclusion of trabecular

bone in mixed samples was deemed to be a potential risk with the use of mixed bone

samples. This was further substantiated by the lower concentrations of citrate detected in

the mixed samples. As such, cortical only samples were deemed to be the optimal bone

type for use in this analytical method. All subsequent analysis in this thesis utilized

cortical only samples.

4.5.4 Analysis of Bone Storage Conditions on Citrate Concentrations

This experiment was developed to evaluate whether the storage conditions of a

bone specimen over a short time period, 14 days, would have any significant impact on

the detected concentration of citrate. The methods and materials used for this experiment

are provided in section 3.5.5 (Ch. 3). Results in this section include the qualitative

analysis of chromatograms produced by the bones in the different storage environments

and quantitative analysis comparing the normalized citrate concentrations between the

three test populations. It was hypothesized that storage environment would have no

significant impact on the normalized concentration of citrate in bone over the short time

period of 14 days.

160
 Qualitative Analysis

As described in Section 3.5.5 (Ch. 3), chromatograms of each of the three storage

environments were overlaid and compared (Figure 4.36). The areas under each of the

peaks were similar despite the room temperature sample having a higher peak. Also, the

size of the unknown peak preceding citrate had no impact on the citrate peaks produced,

as the room temperature unknown peak was substantially larger for this bone.

Quantitative analysis was needed to test if storage conditions had an impact on the

measured citrate concentrations.

161
 VAC H20 RT

0.040

0.035

0.030

0.025
Absorbance Units

0.020

0.015

0.010

0.005

-0.001
6 6.5 7 7.5 8 8.5 9 9.5 10 10.5 11
Time (Min)

Figure 4.36 Chromatograms of Three Different Storage Conditions Overlaid. Only six
through eleven minutes are shown as no peaks were observed outside these time points

162
 Quantitative Analysis

Citrate peaks in each of the chromatograms were integrated using the chromatogram

post processing method to determine the peak area and retention time. Areas were then

converted into molar concentrations by using the standard addition method described in

Section 3.5.2 (Ch. 3). Concentrations were the normalized into wt. % concentrations for

statistical analysis (Table 4.24).

Table 4.24 Calculated wt. % Values of Citrate in Bone Solutions after Short Term (14
days) Storage in Varied Environments

Sample H2O (wt. %) VAC (wt. %) RT (wt. %)

1 1.478 1.399 1.171
2 1.361 1.336 1.259
3 1.372 1.302 1.362
Mean 1.40 1.35 1.26
SD 0.05 0.04 0.08

Statistical analysis was completed to determine if the storage environment (H2O,

VAC, and RT) was a factor in the measured wt. % concentrations. Storage environment

was not a significant factor in the average wt. % concentrations (ANOVA, p=0.14),

indicating no detectable difference between the various storage environments (Tukey

post-hoc multiple comparison, p>0.05). This demonstrates that, over the 14 days of

storage in the three environments, the manner in which a bone specimen is stored prior to

analysis does not impact the concentration of citrate detected.

163
 4.5.5 Analysis of Specimen Location on Citrate Concentration

As described in section 3.5.6 (Ch. 3), three bones from a single rack of ribs were

divided into three distinct anatomical regions (Dorsal, Central, and Ventral). The

midpoint of each region was identified and a specimen was sectioned for analysis.

Qualitative analysis was done to determine peak height, retention time, and resolution.

Quantitative analysis was performed to determine if the anatomical location of a

specimen was a factor in the detected concentration of citrate.

Qualitative Analysis

The peaks for the three locations along the length of the bone had two main peaks,

with the largest one preceding the citrate peak. This run was the first run in which the

peak preceding citrate did not return to a baseline prior to the start of the citrate peak.

Thus, the citrate peak began before the unknown peak had finished eluting and could be

considered to be a rider peak [31]. This was constant in all of the chromatograms from

bone tested in this study, as such it was not believed to be specific to a particular location

on the bone. The peak areas did appear to be different, but quantitative analysis was

necessary to determine if the peak areas were significantly different. The chromatogram

regions for the citrate peaks in each location is provided below (Figure 4.37).

164
 Dorsal Central Ventral

0.0027

0.0022
Absorbance Units

0.0017

0.0012

0.0007

0.0002
8 8.2 8.4 8.6 8.8 9 9.2 9.4 9.6 9.8 10 10.2 10.4 10.6 10.8 11
Time (Min)

Figure 4.37 Close in View of Citrate Peaks in Different Locations of the Same Bone. All
Samples are Cortical Only, 75mg samples.

165
 Quantitative Analysis

Wt. % values for each of the samples were then determined using the processing

method described in Section 3.3.8 (Ch. 3). This allowed for statistical analysis using the

One-Way ANOVA method described in Section 3.3.12 (Ch. 3). While the values were

pooled, it must be noted that bone samples from rack six were analyzed as the column

performance began to decline due to the drop in operating pressure discussed earlier in

this chapter. The values for wt. % are seen below (Table 4.25).

Table 4.25 Calculated Normalized Concentrations (wt. %) of Citrate in Three

Anatomical Regions along a Porcine Rib Bone

Bone Dorsal Central Ventral

Sample (wt.%) (wt.%) (wt.%)
R4B5 2.071 2.391 1.183
R4B6 1.598 1.028 0.825
R4B7 1.945 2.457 1.469
R6B4 0.631 0.596 0.315
R6B5 0.588 0.461 0.241
R6B6 0.621 0.364 0.096
Mean 1.242 1.216 0.688
SD 0.645 0.879 0.510
RSD (%) 51.93 72.29 74.13

Statistical analysis was completed to determine if the anatomical location of a

specimen (Dorsal, Central, and Ventral) was a factor in the measured wt. %

concentrations. Anatomical location of a bone specimen was not a significant factor in

the average wt. % concentrations (ANOVA, p=0.39), indicating no detectable difference

between the various anatomical locations of bone samples (Tukey post-hoc multiple

166
comparison, p>0.05). This demonstrates that anatomical location of bone samples tested

has no impact on the detected normalized citrate concentration (wt. %).

The standard deviation when adjusted for the mean, RSD, values obtained for

each of the locations indicates a wide degree of variance in the citrate contents in the six

bone samples. The ventral samples had the highest degree of variance as evidenced by an

RSD of 74.13%, while the dorsal had the lowest with an RSD of 51.93%. Citrate content

across bones appears to be highly variable in each of the three regions. Further analysis

was done to determine the impact of concentration variance on the PMI calculated using

the method described in Section 3.5.10. The Calculated PMI for all samples is provided

below (Table 4.26).

Table 4.26 Estimated PMI Values of Bone Samples Using the Schwarcz Citrate
Degradation Model49 in Three Anatomical Regions along a Porcine Rib Bone

Dorsal Central Ventral

Bone PMIcalc PMIcalc PMIcalc
Sample (Years) (Years) (Years)
R4B5 0.071 0.024 1.475
R4B6 0.357 2.507 5.020
R4B7 0.109 0.019 0.554
R6B4 9.741 10.983 28.747
R6B5 11.292 17.428 37.025
R6B6 10.075 24.341 60.822
Mean 5.27 9.22 22.27
SD 5.12 9.25 22.16
RSD (%) 97.15 100.32 99.51

167
 Subsequent statistical analysis was completed to determine if the anatomical

location of a specimen (Dorsal, Central, and Ventral) was a factor in the PMI estimation

(years) obtained from the Schwarcz equation. Anatomical location of a bone specimen

was not a significant factor in the PMI estimation (ANOVA, p=0.17), indicating no

detectable difference between the various anatomical locations of bone samples (Tukey

post-hoc multiple comparison, p>0.05). This demonstrates that anatomical location of

bone samples tested has no significant impact on the estimated PMI (years).

RSD values of PMI (years) obtained for each of the locations indicates a very

large degree of variance in the estimated PMI for the three regions of the six bone

samples. The central samples had the highest degree of variance as evidenced by an RSD

of 100.32%, while the dorsal had the lowest with an RSD of 97.15%.

4.5.6 Analysis of Additional Filtering on Chromatograms

More rigorous filtering of bone samples was considered towards the end of this

thesis in order to eliminate the unknown peak that begins at 6 minutes. One of the

common views held throughout early testing was that the large peak preceding citrate in

the chromatograms of bone samples was caused by some type of protein remaining in the

samples after using syringe tip filters. As such, it was believed that using a cutoff filter

that allowed low molecular weight elements within the solution through, while filtering

out large molecular weights would help to achieve the desired clean samples. In order to

test this hypothesis, both a 1000 Dalton (Da) and 3000 Da cutoff filter were used in

accordance to the manufacturer instructions in order to allow the low molecular weight

168
citrate molecule (189.1 Da) to pass through while filtering out the larger molecules

believed to be the root cause of the preceding peak.

Chromatograms for the three filtration methods (unfiltered, 1K, 3K) of 75mg

bone samples were overlaid and qualitatively analyzed (Figure 4.38). The peak height of

the unknown peak did not decrease significantly in any of the three filtration methods.

This demonstrated that the compound eluting prior to citrate had a molecular weight

below 1kDa. Thus, the extra filtering had no significant impact on the peak height of the

unknown peak.

169
 Unfiltered 1K Cutoff 3K

0.044

0.039

0.034

0.029
Absorbance Units

0.024

0.019

0.014

0.009

0.004

-0.001
4 4.5 5 5.5 6 6.5 7 7.5 8 8.5 9 9.5 10 10.5 11
Time (Min)

Figure 4.38 Chromatograms from MWCO Group One of 75mg Bone Overlaid. There is
no significant difference in the size of the first peak in the three samples indicating that
the first peak is a low molecular weight compound (< 1kDa).

170
 Quantitative Analysis

Citrate peaks were processed with the peak processing protocol as outlined in

section 3.3.8 (Ch. 3). Peak areas were converted into normalized concentrations of

citrate, wt. %, for each of the three samples (unfiltered, 1K, 3K) in the three groups

(Table 4.27). Concentrations were compared to determine if the filtering process

produced any change in the detected citrate concentration.

Table 4.27 Detected Concentration of Citrate (wt. %) Using Three Different Filtering
Methods. 1K and 3K represent the molecular weight (kDa) of particles that are filtered
from the sample.

Unfiltered 1K 3K
(wt.%) (wt.%) (wt.%)
Group 1 0.614 0.612 0.582
Group 2 0.564 0.466 0.423
Group 3 0.572 0.478 0.492
Mean 0.583 0.519 0.499
SD 0.022 0.066 0.065

Statistical analysis was performed to determine if the type of filtration of a sample

(unfiltered, 1K, and 3K) was a factor in the measured citrate concentration (wt. %). Type

of filtration of a bone sample was not a significant factor in the average wt. %

concentrations (ANOVA, p=0.34), indicating no detectable difference between the

various filtration methods (Tukey post-hoc multiple comparison, p>0.05). This

demonstrates that filtration of bone samples had no impact on the detected normalized

citrate concentration (wt. %).

171
 4.5.7 Comparison of Standard Addition and Calibration Curve Concentrations

Concentrations obtained by the calibration curve and standard addition method

were compared to evaluate if methodology had an impact on the concentration observed

in bone specimens. Uncertainties associated with each Standard Additon measurement

are reported, but were not considered in this analysis as it was intended to be a first order

approximation of the differences between the two methods in order to determine how the

method used to obtain a concentration would impact the obtained value. Two sets of test

data (location and storage condition) were pooled for analysis. It was hypothesized that

methodology of calculating the concentration in a bone sample would not have a

significant impact on the normalized concentration, wt. %, of citrate detected in each

bone sample.

Calibration curve concentrations were obtained by finding the concentration

(mM) associated with the peak area obtained. Standard addition values were obtained by

back extrapolation of the standard addition curve and taking the absolute value of the

concentration on the negative x axis. Uncertainties of x-intercept values were determined

using the following equation (Eq. 4.1) [23].

𝑆𝑦 1 (𝑦̅)2
𝑆𝑡𝑎𝑛𝑑𝑎𝑟𝑑 𝑑𝑒𝑣𝑖𝑎𝑡𝑖𝑜𝑛 𝑜𝑓 𝑥 − 𝑖𝑛𝑡𝑒𝑟𝑐𝑒𝑝𝑡 = |𝑚|
√𝑛 + 𝑚2 ∑(𝑥 −𝑥̅ )2 Eq. 4.1
𝑖

The output of these analyses for selected aim 3 bone solutions can be seen below (Table

4.28).

172
Table 4.28: Comparison of Concentrations Obtained by Calibration Curve and Standard
Addition Method for Selected Storage and Location Tests

Calibration Standard Uncertainty of X

Bone Curve Addition of Standard
Study
Sample Concentration Concentration Addition
(wt.%) (wt.%) (wt. %)
Location R4B5D 1.789 2.071 0.12
Location R4B5C 1.623 2.391 0.08
Location R4B5V 1.016 1.183 0.11
Location R4B6D 1.421 1.598 0.07
Location R4B6C 0.678 1.028 0.31
Location R4B6V 0.657 0.825 0.06
Location R4B7D 1.455 1.945 0.21
Location R4B7C 1.564 2.457 0.16
Location R4B7V 1.166 1.469 0.06
Storage RT1 1.352 1.380 0.01
Storage RT2 1.365 1.473 0.02
Storage RT3 1.393 1.594 0.03
Storage VAC1 1.316 1.399 0.06
Storage VAC2 1.161 1.336 0.03
Storage VAC3 1.182 1.302 0.06
Storage H2O1 1.493 1.478 0.03
Storage H2O2 1.332 1.361 0.01
Storage H2O3 1.331 1.372 0.03
Mean 1.294 1.537 0.08
SD 0.283 0.417 0.07
RSD (%) 21.841 27.153 92.757

173
 Statistical analysis was performed on the obtained concentrations to determine if

the methodology of obtaining the citrate concentration (standard addition or calibration

curve) had a significant impact on the normalized citrate concentration, wt. %, measured

in each sample. Method of obtaining the normalized concentration (calibration curve,

standard addition) was a significant factor in the average concentrations measured (paired

t-test, p<0.05), indicating a significant difference between the concentrations obtained

with the two methods. This demonstrates that the method of obtaining the normalized

concentration, wt. %, was significant in the detected citrate concentration for all samples

when uncertainty of x was not accounted for in standard addition population. A more

complete analysis could be performed using the uncertainty values provided in Table

4.28.

Overall, the concentrations obtained with the standard addition methodology were

higher compared to the calibration curve concentrations in all of the samples except for

one (H2O 1). It was unclear how the difference in the concentrations would factor into

PMI estimates, so further analysis was performed to obtain PMI estimates using the

method described in section 3.5.10 (Table 4.29).

174
Table 4.29 Comparison of calculated PMI (years) for Standard Addition and Calibration
Curve Methods of Concentration Determination

Calibration Standard
Bone
Study Curve PMIcalc Addition
Sample
(years) PMIcalc (years)
Location R4B5D 0.186 0.071
Location R4B5C 0.327 0.024
Location R4B5V 2.608 1.473
Location R4B6D 0.654 0.356
Location R4B6C 8.304 2.504
Location R4B6V 8.916 5.013
Location R4B7D 0.582 0.109
Location R4B7C 0.401 0.019
Location R4B7V 1.562 0.555
Storage RT1 0.827 0.752
Storage RT2 0.791 0.547
Storage RT3 0.719 0.361
Storage VAC1 0.936 0.704
Storage VAC2 1.590 0.874
Storage VAC3 1.480 0.982
Storage H2O1 0.511 0.538
Storage H2O2 0.886 0.802
Storage H2O3 0.889 0.773
Mean 1.787 0.914
SD 2.478 1.148
RSD (%) 138.665 125.595

175
 The standard deviation in both methods was large compared to the mean as

indicated by the large RSD values obtained. The standard addition method provided a

more accurate estimation of the PMI. Statistical analysis was performed to determine if

the differences in PMI estimation were significant. Method of obtaining citrate

(calibration curve or standard addition) was found to be significant (paired t-test, p <

0.05) on the calculated PMI (years). The lower RSD for the standard addition group

indicated the need for a standard addition methodology in the sampling protocol.

4.5.8 Analysis of Calculated PMI for Citrate Concentrations Obtained in Study

As the primary motivation for this thesis was to investigate the potential of a

citrate based method for prediction of PMI, all cortical only samples were analyzed to

determine the calculate PMI (years) obtained by the Schwarz equation (Tables 4.30-4.33;

Figures 4.39-4.41). It was hypothesized that all samples should return a PMI close to 0,

as all samples in this thesis were fresh.

176
 Table 4.30 Calculated PMI Using Schwarcz Citrate Degradation Model49 for all Bone
Specimens Tested in the Storage Condition Study

Storage Condition Study

PMICALC
Sample wt.%
(Years)
VAC1 1.40 0.70
VAC2 1.34 0.87
VAC3 1.30 0.98
H2O1 1.48 0.54
H2O2 1.36 0.80
H2O3 1.37 0.77
RT1 1.17 1.54
RT2 1.26 1.14
RT3 1.36 0.80
Mean 1.34 0.90
SD 0.08 0.27

1.80
1.60
1.40
1.20
1.00
0.80
0.60
0.40
0.20
0.00
VAC1 VAC2 VAC3 H2O1 H2O2 H2O3 RT1 RT2 RT3 Mean SD

wt.% PMICALC (Years)

Figure 4.39 Plot of PMI and Concentration of Bone Samples in Storage Condition Study.
Blue bars represent detected concentration and orange bars represent calculated PMI

177
 Table 4.31 Calculated PMI Using Schwarcz Citrate Degradation Model49 for all Bone
Specimens Tested in the Location Study

Location Study
PMICALC
Sample wt.%
(Years)
R4B5 DOR 2.07 0.07
R4B5 CEN 2.39 0.02
R4B5 VEN 1.18 1.47
R4B6 DOR 1.60 0.36
R4B6 CEN 1.03 2.50
R4B6 VEN 0.83 5.01
R4B7 DOR 1.94 0.11
R4B7 CEN 2.46 0.02
R4B7 VEN 1.47 0.55
R6B4 DOR 0.63 9.74
R6B4 CEN 0.60 10.98
R6B4 VEN 0.31 28.75
R6B5 DOR 0.59 11.29
R6B5 CEN 0.46 17.43
R6B5 VEN 0.24 37.03
R6B6 DOR 0.62 10.07
R6B6 CEN 0.36 24.34
R6B6 VEN 0.10 60.82
Mean 1.05 12.25
SD 0.74 15.93

70.00
60.00
50.00
40.00
30.00
20.00
10.00
0.00

wt.% PMICALC (Years)

Figure 4.40 Plot of PMI and Concentration of Bone Samples in Location Study. Blue bars
represent detected concentration and orange bars represent calculated PMI

178
 Table 4.32 Calculated PMI Using Schwarcz Citrate Degradation Model49 for all Bone
Specimens Tested in the Filter Method Study

Filter Method Study

PMICALC
Sample wt.%
(Years)
MWG1 UN 0.61 10.33
MWG1 1K 0.61 10.40
MWG1 3K 0.58 11.53
MWG2 UN 0.56 12.28
MWG2 1K 0.47 17.17
MWG2 3K 0.42 19.88
MWG3 UN 0.57 11.95
MWG3 1K 0.48 16.45
MWG3 3K 0.49 15.67
Mean 0.53 13.96
SD 0.07 3.22

25.00

20.00

15.00

10.00

5.00

0.00

wt.% PMIcalc (Years)

Figure 4.41 Plot of PMI and Concentration of Bone Samples in Filtration Method Study.
Blue bars represent detected concentration and orange bars represent calculated PMI

179
 Table 4.33 Mean Calculated PMI and wt. % for all 75mg Cortical Only Samples Tested
in Thesis.

Citrate
PMICALC
Sample Concentration
(Years)
(wt.%)
MEAN 1.12 7.93
SD 0.63 13.72
RSD 56.64 173.05

The storage condition study provided the least variation in PMI, 0.90 ± 0.27

years, while the location study had the most 12.25 ± 15.93 years. This indicates that the

model provided by Schwarcz is unable to compensate for variation in citrate content

within bone. The wide variance in all bone samples, 56.64 RSD over the whole thesis,

demonstrates citrate concentrations in porcine rib bone vary. The mean wt % obtained in

this thesis is the same as that reported by Froome et al. (1.12 wt. %) [20]. Also, only three

of the cortical only samples analyzed were ≥ 2.0 wt. %.

180
 CHAPTER 5

DISCUSSION

5.1 Introduction

The overall objective of this thesis was to develop and validate an HPLC method

for quantitative and qualitative analysis of citrate in bone. In order to achieve this

objective, experiments were developed to achieve three main aims. First, an HPLC

method was developed and tested on a physiologically relevant range of citrate standards.

Second, an existing bone specimen processing method was tested and optimized to

ensure the proper preparation and sampling of bone using HPLC. Third, bone samples

were prepared and tested to evaluate the qualitative and quantitative analysis provided by

the HPLC method developed in aim one.

Aim 1 was accomplished through an array of experiments performed on a

physiologically relevant range of citrate standards (0 to 2.5 mM). In particular, linearity,

storage time of standards, stock solution pH, and intra-day test method precision

(repeatability) were examined. Aim 2 was accomplished by designing experiments to

qualitatively and quantitatively analyze citrate detection in three different bone masses

(50, 75, and 100 mg). Also, matrix effects were also analyzed through the design and

execution of standard addition and standard recovery tests. Finally, different types of

bone (cortical, trabecular, and mixed) were analyzed to determine the optimal type of

bone and limitations with the HPLC test method. Aim 3 was accomplished by analyzing

181
different short term (14 day) storage conditions (hydrated, dehydrated, and room

temperature), intra-bone factors associated with sampling at different anatomical

locations (dorsal, central, and ventral), and an analysis of the calculated PMI produced by

the method described by Schwarcz et al. [49] for all 75 mg cortical only bone samples

analyzed throughout this thesis.

This chapter is divided into six subsections. Subsection 5.2, “HPLC Test Method

Development and Validation on Standards (Aim 1)”, discusses experiments related to

Aim one in this thesis. Subsection 5.3, “Selection and Preparation of Bone Specimens

(Aim 2)”, addresses the experiments performed to accomplish aim two. Subsection 5.4,

“Utilization of Bone Preparation Protocol and HPLC Method for Qualitative and

Quantitative Analysis of Citrate Content in Bone (Aim Three)”, addresses the

experimental results obtained to accomplish aim three. Subsection 5.5, “General

Observations on HPLC Performance”, briefly discusses the performance of the HPLC

system throughout this thesis. Subsection 5.6, “Overview of Expenses”, addresses the

costs associated with the HPLC method. Finally, subsection 5.7, “Future Work”,

discusses potential areas of future research.

5.2 HPLC Test Method Development and Validation on Standards (Aim 1)

Aim 1 was accomplished through an array of experiments performed on a

physiologically relevant range of citrate standards (0 to 2.5 mM). In particular, linearity,

storage time of standards, stock solution pH, and intra-day test method precision

(repeatability) were examined.

182
 5.2.1 Mobile Phase

Laboratory grade sulfuric acid was used for mobile phase throughout the entirety

of this thesis. At the outset of the project, it was unclear as to the necessary purity and

grade of chemicals needed to obtain optimum results. Due to a desire to minimize

expenses with this analytical technique in regards to translation to forensic labs, non-

HPLC grade chemicals were used throughout testing to determine impact on sample

resolution and overall quality of the method. No testing was done with HPLC grade

sulfuric acid, and as such it is unclear whether or not chemical grade had an impact on

chromatogram shape or retention times.

5.2.2 HPLC Method Parameters

As with mobile phase, HPLC method parameters [temperature, pressure,

wavelength, injection volume, and flow rate] were based on operating conditions

suggested by the manufacturer. This thesis was focused on practical optimization of

HPLC analysis of bone samples, as such certain aspects of HPLC method were not fully

investigated. For the purposes of this thesis, the conditions provided by the manufacturer

produced acceptable results, and as such were deemed acceptable for this method. This

section will discuss the range of testing on method parameters.

In this thesis, a singular flow rate of 0.6mL/min was used for all analysis in order

to minimize stress on the column. When the column pressure dropped after 400

injections, flow rate was increased to 0.7 mL/min. Chromatograms produced at this flow

rate had shorter retention times and no loss of peak resolution was observed. As such, a

183
flow rate of either 0.6 mL/min or 0.7 mL/min would provide acceptable chromatograms

for qualitative or quantitative analysis with this method.

Initial analysis of standards were done with a 40 µL injection volume based on

the manufacturer’s suggested test method. Chromatograms at this injection volume had

small citrate peaks and as a result low areas of citrate. In subsequent analyses, injection

volume was increased to 50 µL which resulted in larger citrate peaks and peak areas that

were more easily integrated. All subsequent analyses were done with injection volumes

of 50 µL. The HPLC system had a max injection volume in the sample loop of 100 µL. In

samples with in which citrate concentration is expected to be difficult to quantify, larger

injection volumes could be used to increase the response of the system. No need was

evident throughout testing for greater injection volumes with the method as all sample

were fresh and as a result expected to have high concentrations of citrate.

As demonstrated in section 4.2.4 (Ch. 4), analysis of standards and bone samples

was performed with two channels. The first was a time constant standard channel with a

wavelength of 210 nm, which was used in parallel with channel that performed a

wavelength scan from 200-400 nm. Analysis of UV-Vis spectra of citrate standards

demonstrated two wavelengths where citrate had a high response in absorbance units, 190

nm and 209.6 nm (210 nm). Analysis of the unknown peak in bone solution

chromatograms demonstrated a dominant response at 190 nm that was muted at 210 nm,

as such the time constant standard wavelength was fixed at 210 nm. As the main point of

interest of this thesis was citrate, no further analysis was performed on the unknown first

peak. It should be noted that when analysis was performed on the first peak, UV-Vis

184
spectra were consistent in each sample, and may be worth investing the potential of doing

a background subtraction when suspected interference with a citrate peak occurs. The

wavelength scan channel was adjusted to scan from 190 nm to 300 nm to allow for

analysis of citrate at 190 nm if desired.

As demonstrated in section 4.2.3 (Ch. 4), pressure in the column varied

throughout the thesis, but did not have an impact on the ability to quantify citrate.

However, during the final 50 sample injections performed in this thesis, a noticeable drop

in pressure (200 psi) occurred after storing the column for one month at 2ºC. The

manufacturer hypothesized that the drop was due to a change in the chemistry of the

column caused by the high salt content of the bone samples. Column regeneration was

attempted using the protocol suggested by the manufacturer, namely, a 25 mM H2SO4

mobile phase solution was run through the column overnight at a flow rate of 0.1

mL/min. After this treatment, pressure and chromatograms were unchanged from the

lower pressure conditions, and as such regeneration of the column was not successful.

5.2.3 Linearity and Range

All of the standards analyzed in this thesis provided acceptable linearity as

documented in section 4.3.1 (Ch. 4). This indicated that the response (i.e. absorbance) of

the HPLC increased in proportion to the increased citrate concentration in standard

solutions. Furthermore, a physiological citrate concentration range (0 to 2.5 mM) was

linear over the entire range. All citrate measured from bone samples in this thesis were

within the concentration range of the prepared standard solutions. Overall, the HPLC

185
method developed was appropriate for the detection of citrate in standards as well as bone

solutions.

5.2.4 Intra –Day Test Method Precision (Repeatability)

Overall, the HPLC method is suitable for low concentrations of citrate that may

be expected from degraded bone samples in forensic anthropology applications. As

shown in section 4.3.2 (Ch. 4), intra –day test method precision (repeatability) of the

HPLC method and Waters Breeze 2 HPLC system was acceptable based on ten

consecutive injections of the 0.5 mM standard. The 0.5 mM standards used in this test

were analogous to a 75 mg bone solution with a normalized concentration of 0.43 wt. %

which is approximately half of the concentrations seen in fresh bone samples analyzed. It

is believed that a similar result would be obtained were this analysis repeated on bone

solutions based on an analysis of samples for which three injections were performed.

Other aspects of precision (inter day variation, reproducibility) were considered for

testing, but time and resource constraints prevented completion of those experiments.

5.3 Selection and Preparation of Bone Specimens (Aim 2)

The second aim of this thesis was to develop and optimize a specimen processing

protocol for HPLC analysis of bone. The bone specimen processing method described by

Schwartz, et al. was evaluated and modified to allow for more efficient processing of

bone samples. Next, bone samples were analyzed with HPLC in order determine

characteristic peaks and retention times in bone solution chromatograms, injection

186
dependence on the detected citrate concentrations, and the ability of the Breeze system

software to integrate citrate peaks. Furthermore, HPLC analysis were performed on three

different bone masses of bone to determine if mass of bone prepared for analysis was a

factor in the detected citrate concentrations in bone samples. Matrix effects were also

analyzed to determine if the processing method required a strategy for the reduction of

matrix effects. Finally, different bone types (cortical, trabecular, and mixed) were

analyzed to determine if the type of bone prepared in this process was a factor in the

detected citrate concentration.

5.3.1 Modifications to Schwarcz et al. Bone Processing Method

An analytical approach was used to develop both a more efficient and easier to

use processing method for bone analysis in this thesis. Also, changes to the Schwarcz

method were needed in order to optimize the prepared samples for HPLC analysis [49].

This section will address two main changes made to the processing method, namely an

increased digestion time for larger masses of bone and a defined sample neutralization

procedure.

5.3.2 Digestion Times

Overall, the time of digestion of samples in this thesis was increased from one

hour to four hours to ensure adequate time for the demineralization reaction to occur. As

demonstrated in Section 4.5.1, mass of bone prepared and analyzed was not a factor in

the normalized concentration of citrate observed. Thus, the processing protocol adopted

187
in this thesis for subsequent analysis was the use of 75 mg samples. Although mass was

not a factor in the citrate concentration detected, a key concern with the use of larger

masses of bone is ensuring that all of the citrate is extracted from the sample.

As the demineralization reaction is rate limited, the concentration of the acid,

time, and temperature at which demineralization occurs will impact the citrate extracted

from the process [18]. Furthermore, it has been shown that increased concentrations of

acid do not impact the demineralization reaction proportionally [18]. In the optimal mass

study, the three masses were digested for different times (one hour for 50 mg, two hours,

for 75mg, and four hours for 100 mg), thus, it was demonstrated that the increased

digestion times had no negative impact on the detection of citrate. While the two hour

digestion of 75 mg samples in the optimal mass study appeared to be appropriate, it was

unclear whether a 150% increase in bone mass used would be sufficiently mineralized

with only an additional hour of digestion time. As a result, four hours was selected in

order to allow for adequate time of demineralization process to occur. Since the 100 mg

samples at four hours showed no loss in the detected citrate concentrations, four hours

was deemed to be suitable for the 75 mg samples. In all samples processed at four hours,

chromatograms provided adequate peaks for integration. Further research is needed to

optimize this step, in order to better streamline the bone processing procedure.

188
 5.3.3 Sample Neutralization

As described in section 3.4.10 (Ch. 3), samples were neutralized to a pH 2 rather

than a pH 5 as written in the Schwarcz protocol [49]. This was done to ensure

precipitation of citrate did not occur prior to analysis of samples, as observed by Gibbs at

pH 5 [21]. Also, a 1.0 M KOH solution was used rather than a 0.5 M solution. This

change required less volume of KOH solution to neutralize the bone sample to pH 2.

Furthermore, an analytical approach was used to standardize the volume of 1.0 M KOH

solution added to each sample, allowing for more efficient processing in a laboratory

setting when compared to a dropwise addition as written in Schwarcz [49]. The amount

of 1.0M KOH needed to reach pH 2 was equivalent in all bone solutions, as long as the

same source of 1.0M KOH was used. As sample volume was lost in the confirmation of

pH, the preparation of an extra sample for quality control purposes was found to be useful

throughout this process.

5.3.4 Chromatogram Peaks

As demonstrated in section 4.4.1 (Ch. 4), chromatograms produced by bone

samples contained two prominent peaks, a large unknown peak that begins to elute

around 6 minutes and a citrate peak immediately following that begins to elute around 8

minutes. In most analysis, each of these two peaks were clearly defined, with the

unknown peak being 5x as large as the citrate peak in terms of both area and peak height.

Analysis of the test chromatogram provided with the Aminex HPX-87H column showed

an oxalic acid peak that elutes around the 6 minute time point. Oxalic acid in bone was

189
investigated in literature, but it appeared that the concentrations reported in bone were

less than those of citrate [25]. Thus, the peak was not oxalic acid, but rather another

analyte that reacts with the ions in the column. A list of compound retention times

provided by the manufacturer was studied to try and determine the actual compound

eluting in the samples.

Using the retention times of the unknown peak and citrate in reference to lactic

acid, it was determined that at the operating parameters of the method, two compounds

have relative retention times of approximately six minutes. The two compounds and

relative retention times were Glucosamine, 6.10 min, and 1-Hexanesulfonic acid, 6.13

min. Glucosamine appeared to be the more likely compound in this application, as

glucosamine is found in the bone extracellular matrix [54]. Furthermore, Glucosamine

has a molecular weight close to citrate (179.17 g/mol) that would be able to pass through

the 1K and 3K molecular weight cutoff filter tubes as evidenced by the preservation of

the peak in chromatograms (Section 4.4.1,Ch. 4). As the unknown peak did not interfere

with the citrate peak, no further investigation occurred in this thesis; however, the

compound causing the peak should be identified and if possible reduced in order to make

the HPLC method more suitable for bone samples with low, < 0.25 wt. %, citrate

concentrations.

5.3.5 Peak Integration

The built in integration capabilities of the Breeze 2 software were sufficient to

perform all integrations needed in this thesis. Care was taken to ensure a consistent

190
approach when correcting integration in order to prevent the introduction of error in the

processing of the data. Namely, in instances where the beginning and end of peaks were

unclear, peaks were magnified with the software and start and stop points were identified

based on a consistent increase in the baseline (start) and a return to the baseline (end). As

demonstrated in section 4.4.2 (Ch. 4), the tangential skim fit described in section 3.3.8

was the most appropriate fit for all analysis as it allowed for correct approximation of the

area under citrate peaks.

5.3.6 Injection Dependence

As shown in section 4.4.3 (Ch. 4), the first injection of bone solutions through the

HPLC system produced larger unknown peaks when compared to subsequent injections.

Statistical analysis demonstrated that the injection number had no significant impact on

the measured concentration of citrate in bone solutions (Repeated Measure ANOVA,

p=0.26). While the unknown peak size in the first injection had no significant impact on

the detected concentration of citrate, an extra bone sample could be prepared and

introduced into the column prior to the necessary analysis of bone solutions as a quality

control measure.

5.3.7 Optimal Mass

As demonstrated in Section 4.5.1 (Ch. 4), mass of bone prepared for analysis was

not a significant factor in the detected concentrations of citrate. Previous methods of

citrate quantitation utilized 50mg bone samples with acceptable results, but none tested

191
suitability with larger masses of bone [16, 49]. The results shown in section 4.5.1 (Ch. 4)

demonstrated that greater masses of bone caused proportional increases in the area under

the citrate peaks. As such, any mass of bone in the range of 50 to 100 mg could be

analyzed using the HPLC method developed in this thesis as the mass of bone prepared

for analysis (50, 75, and 100 mg) was not a significant factor in the detected

concentration of citrate when normalized (wt.%). As stated in section 4.5.1 (Ch.4), 75 mg

was selected for the HPLC bone sample processing method as the theoretical

concentrations of citrate at this mass were within the middle of the concentration range

used in the calibration curve.

The success of the 100mg mass samples was not important with fresh bone

samples analyzed in this thesis, but would be important with aged samples containing

reduced concentrations of citrate. Forensic samples 50 years or older would be expected

to have low concentrations of citrate that may approach the Limit of Quantitation (LOQ)

and Limit of Detection (LOD) of the HPLC method. The use of a larger mass of bone is

one possible way to generate enough signal for peak integration. The success of this

method at higher mass groups shows promise of being able to overcome the problem of

low concentrations in aged samples.

5.3.8 Matrix Effects

As demonstrated in Section 4.5.2 (Ch. 4), standard addition and standard recovery

tests were performed to determine whether or not matrix effects were present in either the

samples or analytical technique. Overall, matrix effects were seen in the standard

192
recovery tests, but not in the standard addition tests. Section 5.3.8.1 will address what

was demonstrated by the presence of matrix effects in standard recovery, but not standard

addition tests. Section 5.3.8.2 will address the nature and extent of matrix effects in the

standard recovery data. Finally, section 5.3.8.3 will discuss the strategy for reducing

matrix effects adopted in subsequent analysis (Section 4.5.2, Ch. 4).

5.3.8.1 Analysis of Matrix Effects in Standard Recovery and Standard Addition

As seen in section 4.5.2 (Ch. 4), matrix effects were present in standard recovery

tests, but not standard addition tests performed on mixed bone solutions. Analysis of

these results indicated that matrix effects were present in the processing protocol of bone

rather than the analytical method.

Standard addition calibration curves had slopes equivalent to the calibration curve

slopes in all three of the standard addition tests performed with mixed bone solutions.

Also, the maintenance of linearity (R2 > 0.98) indicated that matrix as citrate was spiked

into the bone solutions, the detected citrate concentration increased in proportion. This

indicated that no loss of citrate was occurring in bone solutions spiked with either high or

low concentrations of citrate. As the spiking of the solutions was done after processing of

the specimens, this indicated that no matrix effects exist in the analytical technique

utilized for detection of citrate content in bone solutions. Further testing with lower

spiked concentrations is needed to determine if standard additions on cortical only

solutions with additions in a range of up to 3x the concentration reported in this thesis,

1.12 wt. %, are also not impacted by matrix effects.

193
 As seen in section 4.5.2 (Ch. 4), matrix effects were present in standard recovery

tests of all three mixed bone samples analyzed. Matrix effects were evidenced by a loss in

linearity of two of the three mixed bone solutions tested (R2 < 0.98), as well as both

translational and rotational effects in the standard recovery calibration curves when

compared to the standard calibration curve. This indicated that as citrate was added, the

analyte signal was not increasing proportionally in each of the bone samples. As no

matrix effects were seen in standard addition tests, it was determined that degradation in

the detection of citrate occurred in the final steps of bone sample processing. Only two

steps occurred after citrate was spiked into samples, namely the demineralization process

and sample neutralization. As citrate is known to form a precipitate in the presence of

calcium and phosphate, the demineralization process may be the most likely site of

matrix effects in the bone samples [35]. Another potential explanation for the presence

of matrix effects in standard recovery data but no standard addition data was the

introduction of error in the measurement of citrate spiked into the bone samples. As the

target concentrations of citrate were low [0.5, 1.0, 1.5, and 2.0 mM], small masses of

citrate, in some cases single crystals, were weighed and added into the samples prior to

the demineralization process. If error were introduced by amount of citrate added, the

low concentrations should have been consistently scattered, while the high concentrations

should have demonstrated a high degree of stability. All of the concentrations appeared to

be equally scattered in the three recovery tests, and as such it was concluded that error in

the measurement of citrate crystals added was not a cause of the matrix effects realized in

the standard recovery tests.

194
 5.3.8.2 Nature of Matrix Effects Seen in Standard Recovery Data

As shown in section 4.5.2 (Ch. 4), both translational and rotational matrix effects

were observed in the three sets of standard recovery tests performed in this thesis. It was

unclear why set one was the only to demonstrate a translational matrix effect, meaning

the citrate lost was independent of the amount of citrate added [55]. The other two sets, as

well as the standard recovery test on a cortical only sample, demonstrated a rotational

matrix effect, indicating that as the amount of citrate spiked was increased, a loss of

signal increased. When citrate recovery in set one, 79.2 ± 19%, is compared to sets 2 and

3, 87.9 ± 9% and 91.4 ± 8%, it can be determined that the translational matrix effect

causes a greater loss in citrate within these samples. Further analysis is needed to

determine whether or not both type and extent of matrix effects are varied between

different bones and individuals.

5.3.8.3 Strategy for Reducing Matrix Effects in Bone Sample Analysis

As detailed in section 4.5.2 (Ch. 4), three strategies for reducing matrix effects

were considered: (1) matrix matched standards, (2) standard addition, and (3)

mathematical compensation. As bone composition is expected to vary between bones, a

matrix matched standard was found to be too challenging to implement within a

laboratory setting, as such a matrix matched standards strategy was not utilized. As the

sample size for matrix effects across bone samples is still small, the adoption of a

mathematical compensation strategy was also dismissed, Thus, a standard addition

195
strategy was the easiest to work with and most feasible given the current database of

knowledge.

The adoption of a standard addition strategy was not a small decision. This

method required more sample preparation and subsequent analysis for the determination

of a single value. As a result, both longer run times and increased resource requirements

were experienced compared to a calibration curve methodology. In order to reduce the

overall costs of the standard addition method, only three additions were used for each

bone sample for a total of four measurements. As standard addition requires the back

extrapolation of a first-order (straight line) fit of the data to the x-axis, the uncertainty in

x was calculated using a cortical only bone sample in which standard addition was used.

Standard deviation of the x-intercept in x was calculated using the following equation

(Eq. 5.1).

𝑆𝑦 1 (𝑦̅)2
𝑆𝑡𝑎𝑛𝑑𝑎𝑟𝑑 𝑑𝑒𝑣𝑖𝑎𝑡𝑖𝑜𝑛 𝑜𝑓 𝑥 − 𝑖𝑛𝑡𝑒𝑟𝑐𝑒𝑝𝑡 = |𝑚|
√𝑛 + 𝑚2 ∑(𝑥 −𝑥̅ )2 Eq. 5.1
𝑖

Sy is the standard deviation of y, |m| is the absolute value of the slope, n is the

number of data points for the calibration line, y is the absorbance of the unknown, 𝑦̅ is

the mean value of y (area units) for all of the points on the calibration, xi are the

individual values of x for the points on the calibration line, and 𝑥̅ is the mean value of x

for all points on the calibration line [23].

A calculation was done with sample R4B5 Dorsal, a cortical only sample from the

dorsal region, to evaluate the standard deviation obtained by using only 3 additions. The

196
obtained standard deviation in x was 0.135 mM. Translating this into wt % of 75 mg

samples yields a standard deviation of 0.12 wt. %. Ultimately, the error in the x –

intercept of a sample with four additions translates into an error in the PMI estimation

with the Schwarcz equation, as such, the standard addition method adopted in this thesis

needs to be improved. One possible way to accomplish this is to increase the number of

additions performed for each bone sample. Further analysis is needed to balance the

resource and time requirements of additional concentrations with the required accuracy

and precision of the assay.

5.3.9 Analysis of Citrate Concentration in Bone Types

As demonstrated in Section 4.5.3 (Ch. 4), HPLC analyses were performed on

three bone types (cortical, trabecular, and mixed) to determine the most suitable type of

bone for use with the HPLC method. Cortical chromatograms showed better peak

resolution and larger peak heights compared to the corresponding trabecular samples. In

addition, only two of the six trabecular samples analyzed in this experiment provided

peaks that were quantifiable in the system software. Also, the two that were quantified

had low concentrations, 0.23 wt. % and 0.83 wt. %, compared to the corresponding

cortical bone samples, 1.62 and 1.42 wt. % respectively. The wide variance between

cortical and trabecular citrate concentrations seemingly impacted the mixed samples as

the two analyzed had concentrations of 0.73 and 0.71 wt. %.

The data indicates no advantage to using a mixed bone solution in terms of

detected citrate concentrations. In fact, it appears that the use of a mixed sample may lead

197
to depressed detection of citrate in bone. It was unclear whether this was a limitation of

the HPLC detection method or variation in citrate concentrations between cortical and

trabecular regions. In a prior study by Gibbs, the successful quantitation of citrate in

trabecular bone specimens using an enzymatic assay method for citrate detection was

reported In this study, trabecular samples provided an average concentration of 1.5 wt.%

[21]. All other studies on citrate in bone have utilized cortical only samples, and as such

the selection of cortical only samples in this thesis can be validated through literature [16,

20, 27, 32, 49].

As of now, no apparent disadvantage is known with the use of a cortical only

samples for the detection of citrate in bone. Conceivably, the only reason trabecular bone

analysis would be desired is a more rapid degradation of citrate in cortical bone compared

to trabecular bone within a forensic context. Overall, it appears that the HPLC test

method developed in this thesis was most suitable for cortical only bone sample, not

suitable for trabecular analysis, but was suitable for cortical only sample analysis.

5.4 Utilization of Bone Preparation Protocol and HPLC Method for Qualitative and
Quantitative Analysis of Citrate Content in Bone (Aim 3)
In prior sections, it has been demonstrated that the HPLC method was suitable for

use to qualitatively and quantitatively analyze citrate in standard solutions (Aim 1). Also,

experiments had been completed to optimize the bone sampling protocol for use with

HPLC (Aim 2). This section will address experiments that utilized both the HPLC

method and optimal bone sampling protocol to qualitatively and quantitatively analyze

citrate in bone (Aim 3).

198
 Overall, three experiments were designed to achieve this aim. First, bone samples

were exposed to three different short term (14 day) storage conditions (hydrated,

dehydrated, and room temperature) and subsequently analyzed to determine if short term

storage environment was a factor in the concentration of citrate detected. Next, intra-bone

factors associated with sampling at different anatomical locations (dorsal, central, and

ventral) were evaluated to determine whether or not anatomical location was a significant

factor in the observed concentration of citrate. Finally, the equation produced by

Schwarcz et al. was utilized to determine the calculated PMI (years) for the cortical only

bone samples analyzed throughout this thesis [49].

5.4.1 Storage Condition

As shown in Section 4.5.4 (Ch. 4), bone specimens were exposed to three

different short term (14 day) storage environments (H20, VAC, and RT) in order to

determine whether or not the short term storage environment was a factor in the detected

citrate concentration in bone. In one of the specimens stored in water, the water was

pulled out of the flask by the vacuum. This occurred after 10 days. The sample was

included in the study as this occurred shortly before testing. Overall, no significant

difference was observed in either chromatograms or detected concentrations of citrate

between the three short term storage environments analyzed. The bone specimens left out

at room temperature had the lowest concentrations of the three populations, 1.26 ± 0.08,

while the samples stored in deionized water under vacuum had the highest concentrations

of citrate, 1.40 ± 0.05. These findings were surprising, as literature had suggested

199
exposure of bone specimens to water would lead to degraded levels on citrate in bone

specimens [49]. One possible explanation is that deionized water was used in this study

rather than tap or ground water. As such, it did not contain microbial organisms that may

be found in other water sources. Thus, the degradation of citrate in bone when exposed to

water may be the result of microorganisms rather than hydration of the sample in water.

In general, these three conditions were selected in order to provide a range of storage

conditions available in all laboratories. Based on the data, either the storage of bone

specimens in deionized water or dehydration in a vacuum oven would be better than

leaving the samples at room temperature. Further analysis is needed to determine how

these methods work over longer periods of time.

5.4.2 Intra-bone factors associated with sampling at different anatomical locations

As shown in section 4.5.5 (Ch. 4), the anatomical region from which a bone

specimen was obtained (Dorsal, Central, Ventral) was not a significant factor in the

detected concentration of citrate (wt. %). Six sections representing each region were

analyzed. The ventral regions had the highest relative standard deviation, 72.3%, while

the dorsal region had the lowest, 51.9%. These variances suggest that inter-bone

variations can impact results. The wide range of variances also caused a large variation in

the calculated PMI. Each of the three regions had RSD values greater than 95%

indicating a great deal of uncertainty with the calculated PMI. The ventral specimens in

particular provided satisfying low accuracy with the PMI method, 22.3 ± 22.2 years. This

estimate would not be useful in forensic investigations. Further sampling is needed to

200
determine if these variations are universal or unique to the porcine rib model analyzed in

this thesis.

5.4.3 Standard Addition Compared to Calibration Curve

As documented in section 4.5.7, the method of obtaining the citrate concentration

(calibration curve or standard addition) was a significant factor (p < 0.05) in the detected

citrate concentration (wt.%). The standard addition method had a slightly higher variance,

RSD = 27.2 %, compared to the calibration method, RSD = 21.8%. Furthermore, the

concentrations obtained by the standard addition method provided calculated PMI with

less variance, 0.914 ± 1.1 years, compared to the calibration curve concentrations, 1.79 ±

2.5 years. Although, this data indicates that standard addition methodologies provide

more accurate concentrations, it must be noted that the standard deviations in the x-

intercept (standard addition concentration) were reported but not included in analysis.

Overall, the adoption of a standard addition methodology to overcome matrix effects in

samples appears to be necessary for this HPLC method.

5.4.4 Analysis of Calculated PMI for Citrate Concentrations Obtained in Study

As demonstrated in Section 4.5.8, the concentrations of citrate in 75 mg cortical

only bone samples were converted into PMI estimates. Overall, the average concentration

in all samples was 1.12 ± 0.63 wt. %, which translated into an average calculated PMI of

7.93 ± 13.7 years. The large standard deviation indicates a lack of accuracy with this

model when applied. As all specimens analyzed in this thesis were fresh, the calculation

201
should have provided a calculated PMI near 0 years. Only two of the samples analyzed

had PMI of < 0.05 years. Ultimately, a better model is needed to fit the data, as the

current level of accuracy would not be useful in forensic investigations.

5.5 General Observations on HPLC Performance

Overall, the HPLC performance throughout this thesis was highly reliable with

the standard solutions, but inconsistent with the bone sample analysis. Throughout this

thesis, the Aminex HPX-87H column was exposed to a total of 460 injections. About half

of these injections were standards with the other half being bone samples. Over time, the

performance of the column, in particular, appeared to degrade as evidenced by an

increase in retention time for all samples, including standards, and a decreased operating

pressure in the system. Although no changes, other than increased retention time, were

apparent in standards, citrate peaks in bone sample chromatograms began to elute prior to

a return to baseline. However, qualitative and quantitative analysis of the samples was

still possible.

One potential cause of the change in column performance was the binding of

calcium from the bone samples to the solid phase of the column causing a chemistry

change in the column. Typically, Aminex HPX-87H columns should be able to handle a

minimum of 1000 injections without degradation in performance, so a loss of selectivity

after 500 injections was not expected. A better cleaning method after sample processing,

namely the use of a 25 mM H2SO4 left allowed to run through a reversed column with a

flow rate of 0.1 mL/ min should be investigated for column preservation.

202
 5.6 Overview of Expenses

One of the key considerations throughout this thesis was the translation of any

analytical techniques or methods developed in this paper into a forensics lab. As such,

key attention was given to minimizing costs of the HPLC equipment and consumables

needed, while maximizing the performance of the technique. This section will discuss the

supplies and associated cost as of the writing of this thesis (2015) of materials used

throughout the thesis. Below is a list all of the equipment not including chemicals needed

for analysis, as well as part numbers and estimated costs.

203
 Table 5.1 List of Equipment and Associated Costs Needed for HPLC Method

Approximate Quantity/
Equipment Part Number
Cost Price

125-0234,
Aminex HPX- 87 H Column $1500.00 1
Bio-Rad

125-0129, w/
Guard Column 2
Bio-Rad Column

125-0131,
Standard Cartridge Holder $600.00 1
Bio-Rad

HPLC Glass Vials 186007194C, Waters $60.00 100

Pall Syringe Tip Filters 28143-985, VWR $300.00 100

14-829-10D, Thermo
Syringes $25.00 100
Fisher Scientific

As with most other HPLC applications, the greatest expense in this thesis was the

column needed for sample separation. Having not had previous HPLC equipment, a

column, guard column, and cartridge holder for the guard column were needed. All

equipment was obtained from Bio-Rad Laboratories (Hercules, CA). A review of the

invoices from this process showed an expense of roughly $2000 for these parts.

The cartridge holder for the guard columns has no interaction with the samples

and was considered to be a one-time purchase. The cartridge holder was $600. As the

column and guard column cartridges are directly exposed to bone samples, performance

will degrade over time and as such they will require replacement. An organic acid

204
analysis kit was purchased containing the column, and two guard column cartridges for

approximately $1500.

Overall, the costs associated with this method should not prohibit the use of the

method in an applied setting. When averaged over the total number of injections done in

this thesis, 620, the column and guard cartridge could be averaged out to roughly $2.50

per injection. The cartridge holder is purely an upfront expense and will not be limited by

the number of injections, as such it is not to be accounted in an estimate of the cost per

sample.

Consumables used in HPLC analysis included syringe tip filters, glass HPLC

vials, and syringes. Each of these items were sold in packs of 100 and were consumed

simultaneously with each sample preparation. Analysis of invoices demonstrated that the

most expensive of these items were the syringe tip filters, followed by the HPLC vials,

and finally the syringes.

At the time of this thesis, the syringe tip filters made by Pall life sciences were

approximately $300 per 100 count pack. This was by and far the largest recurring

expense in our method. HPLC vials were purchased from Waters and averaged

approximately $60 per 100 count pack. Syringes averaged about $25 per 100 count pack.

Overall, the recurring expenses were $385 per 100 samples. The total number of samples

in this thesis was approximately 400, so it can be estimated that these supplies needed to

be purchased a total of four times, for a total of approximately $1600.00. On average,

each sample required approximately $4.00 in consumables. The cost per measurement

205
increased to $16.00 when the standard addition methodology was adopted as four of each

consumable were necessary for each bone sample.

When the cost per injection in the column and guard column is added to the cost

of HPLC consumables, the total cost of supplies per injection in this thesis was $6.50. A

sample done with standard addition, 4 injections, would require $26.00 in supplies.

5.7 Future Work

A key limitation in this thesis was the lack of experimentation with the parameters

selected for the HPLC method. There was no available literature to reference in regards

to the method, and as such, this thesis required testing of the method and bone samples.

The results in this thesis provide an extensive reference on the optimal samples for testing

and use with the method described in this test. Using these results, further testing can be

done to better tune the method for citrate detection. The rest of this section will address

some of the areas that may have an impact on the outcomes of the method.

One of the more apparent avenues for investigation is the column. In this study, a

single column was used for all experimentation with mixed results. Early tests showed

clear resolution of the peaks and good separation, while later analysis showed less

resolution and poor separation. As of now, it is unclear whether these results are due to

variances between porcine rib racks, or may be the result of degraded column

performance. Repetition of the test setup in this thesis may allow for the method to be

further optimized with a different column.

206
 Another recommendation for future study is investigating the impact of test

method parameters on the chromatograms obtained for each sample. The results obtained

by the test method described in this thesis appeared to be sufficient for our needs.

Varying temperature, flow rate, and injection volume, would be useful to determine if

any significant changes are observed as a result.

5.7.1 Processing Protocol

Grinding of the bone using a mortar and pestle proved to be the most time

consuming aspect of the processing protocol. Identifying an acceptable alternative for

bone digestion would decrease the time required for processing, and help to make this

method more practical. A few alternative methods were investigated briefly in this thesis

(mechanical grinding, freezing) but none of these showed any promise. A search of

literature at the end of the project, produced a simple method for the demineralization of

bone.

The alternative method being considered only requires the placement of a bone

specimen in 1N HCl, for five days. In theory, this demineralization process should allow

citrate to become suspended in solution once the mineral structure was broken down. A

preliminary study demonstrated that bone dissolved in 1N HCl, with the rubbery collagen

maintaining its structure. No further testing was done in this thesis, as it was unclear how

to proceed once the citrate was in solution. Further investigation into this method may

yield a more practical and less time consuming method of processing the samples, which

will reduce run times, and increase yield with this technique.

207
 Another suggestion for future experiments, is the investigation of bone from other

anatomical locations. As of now, most of the research on postmortem interval estimation

by citrate concentrations have used only porcine cortical rib bone [16, 49]. Future testing

should focus on the use of other types of bones from different anatomical locations.

While these bones will be harder to obtain, the results obtained from these studies would

help to determine the success of these analytical techniques on bone in general rather than

the small subset these porcine rib bones represent. Furthermore, other species should be

evaluated, including humans, to ensure that the results are not being influenced,

negatively or positively, by the use of a porcine rib model.

5.7.2 HPLC Method Applied to Bone

As the majority of the work on this thesis focused on the application of the HPLC

method to bone samples, this section will provide the most important recommendations

for future study. Overall, relatively little research in this field has been done on

quantitative and qualitative analytical techniques, as such much was unknown at the start

of this project. Below are a list of three major recommendations for future studies.

First, matrix effects were evaluated in this study only for the sake of determining

their existence. As of now, it is unclear how matrix effects vary from sample to sample

and individual to individual, as well as from species to species. Variations in bone

chemistry across species and individuals should be expected to have an impact on these

matrix effects, and should be mitigated by the adoption of an analytical technique to

overcome them. Matrix matching in standards was investigated, but it appeared that bone

208
chemistry was too complex to accurately mimic due to the complex biomolecules and

cellular components of the bone matrix. One hypothesis is that the phosphorous and

calcium present in bone would be expected to account for the majority of matrix effects.

This was not tested, but a set of standards with a physiologically ratio of calcium,

phosphorous, and citrate may be one solution to test this hypothesis. In general, these

matrix effects warrant further study to better understand the reactions which cause the

reduction of the analyte signal.

Second, the standard addition method used to compensate for matrix effects

should be explored further to ensure the results seen in this thesis are consistent

throughout all samples. Furthermore, the number of additions performed should be

optimized to minimize over and under estimation of citrate concentrations when the curve

is extrapolated.

Third, all of the experiments in this study used time zero bone, freshly sectioned

and immediately processed and analyzed. The effects of aging through either natural

degradation or artificial exposure in a bioreactor were not evaluated. In order to ensure

this method works on samples similar to those found in a forensic context, aging studies

should be performed to trace the degradation in the samples, as well as the ability of the

HPLC method to provide adequate data for qualitative and quantitative analysis.

209
 CHAPTER 6

SUMMARY AND BROAD SIGNIFICANCE

6.1 Summary

The overall objective of this thesis was to develop and validate an HPLC method

for quantitative and qualitative analysis of citrate in bone. In order to achieve this

objective, experiments were developed to achieve three main aims. First, an HPLC

method was developed and tested on a physiologically relevant range of citrate standards.

Second, an existing processing method was tested and optimized to ensure the proper

preparation and sampling of bone samples using HPLC. Third, bone samples were

prepared and tested to evaluate the qualitative and quantitative analysis provided by the

HPLC method developed in aim one.

Chapter 1 introduces the motivation for the work performed in this thesis. Namely

the need for better quantitative measures for Post Mortem Interval of Skeletal Remains.

Chapter 2 presents a comprehensive overview of the current research on citrate in bone.

A review of the literature demonstrated the need for better testing methods, as most of the

research to date has relied on the unreliable enzymatic assay kits. Chapter 3, presents a

comprehensive list of the methods and materials used throughout the thesis in order to

achieve the aims of this thesis. Chapter 4 presents the results associated with each of the

three aims defined in this thesis. Chapter 5 contains a discussion of the results and their

significance.

210
 Overall, aim one was achieved through an array of experiments performed on a

physiologically relevant range of citrate standards (0 to 2.5 mM). In particular, linearity,

storage time of standards, stock solution pH, and intra-day test method precision

(repeatability) were examined. Aim two was accomplished by designing experiments to

qualitatively and quantitatively analyze citrate detection in three different bone masses

(50, 75, and 100 mg). Also, matrix effects were also analyzed through the design and

execution of standard addition and standard recovery tests. Finally, different types of

bone (cortical, trabecular, and mixed) were analyzed to determine the optimal type of

bone and limitations with the HPLC test method. Aim three was accomplished by

analyzing different short term (14 day) storage conditions (hydrated, dehydrated, and

room temperature), intra-bone factors associated with sampling at different anatomical

locations (dorsal, central, and ventral), and an analysis of the calculated PMI produced by

the method described by Schwarcz et al. [49].

In aim one experiments, simple matrix analysis with HPLC demonstrated

acceptable linearity in all of the standard sets tested. HPLC output acquired from theses

samples allowed for both qualitative and quantitative analysis of citrate in the

concentration range expected in bone. Also, time was also shown to not have a

statistically significant impact on the calibration curves produced by standards.

Furthermore, stock solution pH was shown to have a significant difference in 2.0 mM

standards when compared. Overall, the results from these tests demonstrated a successful

method for the detection of citrate in simple standard matrices, accomplishing aim one.

211
 In aim two experiments, a specimen processing protocol for HPLC analysis of

bone was developed and optimized. The bone specimen processing method described by

Schwartz, et al. was modified by keeping samples at a lower pH, increasing the mass of

bone to 75 mg, and increased digestion time to ensure demineralization was complete in

the sample [49]. Next, bone samples chromatograms were shown to have two dominant

peaks, a unknown peak preceding citrate and the citrate peak. Also, bone solutions were

analyzed to determine if injection order (1, 2, or 3) was a significant factor in the detected

citrate concentration. Furthermore, HPLC statistical analysis of citrate concentrations in

three masses of bone showed that mass of bone prepared for analysis was not a factor in

the detection of citrate, validating the use of 75 mg samples in the processing protocol.

Matrix effects were detected in standard recovery tests, but not standard addition tests

indicating a matrix effect in the demineralization or neutralization steps of the bone

processing protocol. A standard addition strategy was adopted to reduce the impact of

matrix effects on the detected concentration of citrate. Finally, qualitative and

quantitative analysis of different bone types (cortical, trabecular, and mixed) led to the

adoption of citrate only samples in the processing protocol, as trabecular peaks were

unreliable. At the conclusion of these experiments, an optimal sample processing

protocol for the detection of citrate in bone samples was documented, accomplishing aim

two.

In aim three experiments, bone solutions were prepared using the optimized

method from aim two, and the analyzed with the validated HPLC method developed in

aim one. Qualitative and quantitative analysis of bone specimens kept in three different

212
storage conditions (hydrated, dehydrated, and room temperature) over a short time period

(14 days) were performed to determine if storage condition was a factor on citrate content

in bone. Statistical analysis showed that storage conditions over a 14 day time period

were not a factor in the detected concentration of citrate in 75 mg cortical only porcine

rib samples. Furthermore, intra-bone factors associated with sampling at different

anatomical locations (dorsal, central, and ventral) showed no statistical difference in

citrate content along the bone. Finally, citrate concentrations obtained in earlier

experiments were used to estimate PMI (years) using the method described by Schwarcz

et al [49]. This analysis provided an average concentration in all samples of 1.12 ± 0.63

wt. %, which translated into an average calculated PMI of 7.93 ± 13.7 years. Overall, it

appeared that the HPLC method from aim one and the optimal bone processing protocol

developed in aim two were able to provide qualitative and quantitative data on bone

samples, accomplishing aim three.

This thesis represents one of the first analysis of citrate in bone using HPLC for

the analytical method. In general, the outcomes shown in this thesis demonstrate that the

HPLC method developed was shown to be reliable in the qualitative and quantitative

analysis of citrate in bone. While the results obtained throughout this work mark progress

in this field, there still exists a great need for further analysis and optimization of the

techniques and methods described in this thesis.

213
 6. 2 Broad Significance

A major challenge throughout this thesis was knowing what concentrations of

citrate should be expected in bone. The average concentration of citrate in bone obtained

in this thesis was 1.12 ± 0.63 wt. %. Compared to the value listed by Schwarcz, 2.0 wt.

%, our method only detected 56 % of the citrate in the cortical bone samples. When

compared to the average concentration value listed in literature of 1.457 ± 0.31wt. % as

found in Section 2.4 (Ch. 2), the method utilized in this thesis detected 77% of the

expected citrate concentration. The range of concentrations observed was 0.66 wt. % to

1.79 wt. %. In order to have a reliable method for postmortem interval estimation, a

better understanding of the variance of citrate concentrations in bone samples is needed.

Were the values obtained in this thesis to be plugged into the Schwarcz equation for PMI

estimation, the maximum weight percent would return a PMI estimate of 67 days, while

the minimum would yield 8.8 years (3222 days). As both of these samples were time 0

samples, the equation provided by Schwarcz does not appear to work with the wide

variance of citrate concentrations seen in the porcine ribs sampled. One possible

explanation is that the porcine rib bones used were not fully mineralized as the slaughter

age of pigs is analogous to a five year old human.

While research on citrate has been evolving since the 1940’s, a wide gap of

knowledge still exists. Modern analytical methods have provided more insight over the

last five years, but still relatively little is known and understood. Over the last two years,

new evidence has emerged that citrate is produced by osteoblasts rather than derived

from blood as previously believed [19]. This poses some interesting questions in regards

214
to the use of a citrate based technique for the estimation of PMI, as the use of a citrate

model would require a stable citrate concentration in all individuals in order to provide

accurate estimations of PMI. As of now, no studies have been conducted on citrate

concentrations in bone where high osteoblast activity would be expected, such as an

actively remodeling bone or in certain bone diseases. As osteoblasts have been shown to

produce citrate, it is reasonable to assume that an increase in citrate concentration would

be observed where increased osteoblast activity was present.

Another interesting consideration with citrate is its potential use in bone implants.

In a recent study by Guo et al. the potential of citrate based polymers for use as biphasic

scaffolds to fill bone voids was investigated [22]. The goal of this study was to replicate

the architecture of native tissue to improve functionality of the implants. The utilization

of a biomimetic scaffold made from a citrate based polymer to replicate both the

architecture and composition of bone tissue was evaluated for biocompatibility and

integration into the host tissue. The hypothesis was that the citrate based scaffold would

induce faster healing of the bone and degrade as native tissue regrew due to the similar

composition of bone tissue.

The scaffolds demonstrated excellent mechanical and histological attributes,

leading to the realization that scaffolds made of this polymer would be suitable as off the

shelf implants to provide structural support to large bone defects [22]. Overall, this study

showed promise on the ability to utilize citrate within the chemical composition of

biomaterials to induce improved tissue growth in bone. The method developed in this

215
thesis would provide quantitative data on the citrate concentrations to determine if citrate

concentrations increase with these implants.

Overall, citrate is a small but important part of bone biochemistry. Prior to 2010,

interest in citrate within bone had waned after the promising initial wave in the early

1950’s. One of the leading reasons for this appears to have been the need for better

techniques of analysis for both quantitative and qualitative studies. The method

developed in this thesis should be utilized to help improve the understanding of citrate’s

role in bone.

216
APPENDICES

217
 Appendix A
List of Materials Used Throughout Thesis

Chemicals
Quantity or
Material Vendor Part Number List Price
Volume
1N (0.5M)
Sulfuric Acid Fisher Scientific SA212-1 $61.91 1L

1N (1.0M)
Hydrochloric Fisher Scientific SA 48-1 $70.13 1L
Acid

Trisodium
Citrate
Dihydrate Fisher Scientific S279-500 $91.33 500g
(Granular
Form)

Potassium
Hydroxide Fisher Scientific P251-500 $115.21 500g
Pellets

Chloroform Fisher Scientific S25248 $18.86 100 mL

95% Ethanol Fisher Scientific S25309B $31.05 4L

High Performance Liquid Chromatography Equipment

Quantity or
Material Vendor Part Number List Price
Volume
Aminex HPX-
87H Organic Bio-Rad 125-0234 $1485.00 1
Acid Column

Micro-Guard
Cation H Refill Bio-Rad 125-0129 $310.00 2
Cartridges

Standard
Cartridge Bio-Rad 125-0131 $675.00 1
Holder

218
 High Performance Liquid Chromatography Disposables
Quantity or
Material Vendor Part Number List Price
Volume
HPLC Glass
Waters 186007194C $52.00 100
Vials

Pall Syringe Tip

VWR 28143-985 $784.98 100
Filters

Becton
Dickinson Fisher Scientific 14-829-10D $26.76 100
Syringes

219
 Appendix B
Peak Area and Retention Time for All Standard Solutions Used for Calibration Curve
Linearity Analysis in Chapter 4

Citrate Standards Set 1 (02/18/2015)

Conc. Time Retention Area
(mM) (weeks) Time (min) Units
0.00 0.00 0.00 0
0.10 0.00 0.00 0
0.50 0.00 7.12 19772
1.00 0.00 7.20 50718
1.50 0.00 7.24 61773
2.00 0.00 7.29 97903
2.50 0.00 7.33 123816

220
 Citrate Standards Set 2 (02/19/2015)
Conc. Time Retention Time Area
(mM) (weeks) (min) Units
0.00 0.00 0.00 0
0.10 0.00 0.00 0
0.50 0.00 0.00 0
1.00 0.00 7.205 45263
1.50 0.00 7.242 72180
2.00 0.00 7.292 98052
2.50 0.00 7.32 112074

221
 Citrate Standards Set 3 (03/03/2015)
Conc. Time Retention Time Area
(mM) (weeks) (min) Units
0.00 0.00 0.00 0
0.10 0.00 7.972 2423
0.50 0.00 7.96 23022
1.00 0.00 7.964 50934
1.50 0.00 7.964 77578
2.00 0.00 7.969 103165
2.50 0.00 7.97 129515

222
 Citrate Standards Set 4 (03/24/2015)
Conc. Time Retention Time Area
(mM) (weeks) (min) Units
0.00 0.00 0.00 0
0.10 0.00 7.99 4104
0.50 0.00 7.959 27661
1.00 0.00 7.967 51985
1.50 0.00 7.965 74239
2.00 0.00 7.971 101878
2.50 0.00 7.971 128144

223
 Citrate Standards Set 4B (04/14/2015)
Conc. Time Retention Time Area
(mM) (weeks) (min) Units
0.00 0.00 0.00 0
0.10 0.00 7.893 3170
0.50 0.00 7.909 14981
1.00 0.00 7.917 46236
1.50 0.00 7.928 70161
2.00 0.00 7.916 90689
2.50 0.00 7.919 118807

224
 Citrate Standards Set 5 (03/24/2015)
Conc. Time Retention Time Area
(mM) (weeks) (min) Units
0.00 0.00 0.00 0
0.10 0.00 7.965 4970
0.50 0.00 7.97 30119
1.00 0.00 7.963 51964
1.50 0.00 7.966 80949
2.00 0.00 7.971 102319
2.50 0.00 7.968 132500

225
 Citrate Standards Set 6 (05/15/2015)
Conc. Time Retention Time Area
(mM) (weeks) (min) Units
0.00 0.00 0.00 0
0.10 0.00 7.951 4308
0.50 0.00 7.966 25593
1.00 0.00 7.976 50688
1.50 0.00 7.97 78489
2.50 0.00 7.967 146025

226
 Citrate Standards Set 7 (05/28/2015)
Conc. Time Retention Time Area
(mM) (weeks) (min) Units
0.00 0.00 0.00 0
0.50 0.00 8.476 24669
1.00 0.00 8.457 57071
1.50 0.00 8.443 83423
2.00 0.00 8.441 104901
2.50 0.00 8.454 130797

227
 Citrate Standards Set 8 (06/05/2015)
Conc. Time Retention Time Area
(mM) (weeks) (min) Units
0.00 0.00 0.00 0
0.50 0.00 8.61 26394
1.00 0.00 8.618 52725
1.50 0.00 8.611 80585
2.00 0.00 8.61 107607
2.50 0.00 8.613 131583

228
 Citrate Standards Set 9 (06/17/2015)
Conc. Time Retention Time Area
(mM) (weeks) (min) Units
0.00 0.00 0.00 0
0.50 0.00 8.819 24907
1.00 0.00 8.801 47965
1.50 0.00 8.8 76047
2.00 0.00 8.8 99881
2.50 0.00 8.801 124069

229
 Citrate Standards Set 10 (09/14/2015)
Conc. Time Retention Time Area
(mM) (weeks) (min) Units
0.00 0.00 0.00 0
0.50 0.00 9.551 22522
1.00 0.00 9.549 47023
1.50 0.00 9.546 71040
2.00 0.00 9.553 92037
2.50 0.00 9.545 115634

230
 Citrate Standards Set 11 (10/28/2015)
Conc. Time Retention Time Area
(mM) (weeks) (min) Units
0.00 0.00 0.00 0
0.50 0.00 8.18 27375
1.00 0.00 8.165 50208
1.50 0.00 8.178 72849
2.00 0.00 8.169 97136
2.50 0.00 8.17 125461

231
 Appendix C
Mass of Bone, Peak Area, Retention Time, Bone Type, and Citrate Concentrations (mM
and wt.%) of all Bone Solutions Examined in Thesis
Aim Two Experiments
Optimal Mass Tests
Retention
Bone Mass Area Concentration
Injection Time wt. %
Sample (mg) Units (mM)
(min)
50-1 49.9 1 8.089 42793 0.81 1.067
50-1 49.9 2 8.076 38916 0.74 0.970
50-1 49.9 3 8.076 37824 0.72 0.943
50-2 50.2 1 8.076 35470 0.67 0.879
50-2 50.2 2 8.106 36975 0.70 0.917
50-2 50.2 3 8.122 37092 0.70 0.919
50-3 49.9 1 8.125 38262 0.73 0.954
50-3 49.9 2 8.129 34913 0.66 0.871
50-3 49.9 3 8.162 34683 0.66 0.865
75-1 75.1 1 8.166 54431 1.03 0.837
75-1 75.1 2 8.184 52143 0.99 0.801
75-1 75.1 3 8.201 54863 1.04 0.843
75-2 75.2 1 8.214 63142 1.20 0.969
75-2 75.2 2 8.209 64334 1.22 0.987
75-2 75.2 3 8.242 60710 1.15 0.932
75-3 75.3 1 8.235 60309 1.14 0.924
75-3 75.3 2 8.234 62531 1.19 0.958
75-3 75.3 3 8.248 63335 1.20 0.971
100-1 99.8 1 8.263 98482 1.87 1.068
100-1 99.8 2 8.27 90557 1.72 0.982
100-1 99.8 3 8.288 91833 1.74 0.996
100-2 100.2 1 8.291 89692 1.70 0.969
100-2 100.2 2 8.3 93320 1.77 1.008
100-2 100.2 3 8.315 90040 1.71 0.972
100-3 100.6 1 8.311 83347 1.58 0.896
100-3 100.6 2 8.309 84168 1.60 0.905
100-3 100.6 3 8.327 80753 1.53 0.869

232
 Standard Addition Tests
Standard Conc. Mass of Retention Type
Peak Conc. Conc.
Addition Added Bone in Time of
Area (mM) (wt.%)
Set (mM) Sample (min) Bone
SA1 0 75.50 8.54 48161 Mixed 0.914 0.779
SA1 1.64 75.50 8.53 146859 Mixed 2.788 2.374
SA1 4 75.50 8.53 267805 Mixed 5.084 4.329
SA1 7.69 75.50 8.54 445347 Mixed 8.454 7.199
SA2 0 74.90 8.54 46696 Mixed 0.886 0.761
SA2 1.64 74.90 8.53 136245 Mixed 2.586 2.220
SA2 4 74.90 8.53 265900 Mixed 5.048 4.333
SA2 7.69 74.90 8.54 440006 Mixed 8.353 7.170
SA3 0 75.30 - Not enough to test Mixed 0.000 0.000
SA3 1.64 75.30 8.55 153037 Mixed 2.905 2.480
SA3 4 75.30 8.55 272671 Mixed 5.176 4.420
SA3 7.69 75.30 8.56 469227 Mixed 8.907 7.605

233
 Standard Recovery Tests
Standard Conc. Mass of Retention
Peak Type of Conc. Conc.
Recovery Added Bone in Injection Time
Area Bone (mM) (wt.%)
Set (mM) Sample (min)
SR S1 0.00 75.10 1 8.32 47576 Mixed 0.903 0.773
SR S1 0.00 75.10 2 8.35 48538 Mixed 0.921 0.789
SR S1 0.00 75.10 3 8.35 48709 Mixed 0.925 0.792
SR S1 0.45 74.70 1 8.35 87165 Mixed 1.655 1.424
SR S1 0.93 75.30 1 8.36 117368 Mixed 2.228 1.902
SR S1 1.50 76.20 1 8.38 159585 Mixed 3.029 2.556
SR S1 2.77 76.40 1 8.39 208555 Mixed 3.959 3.332
SR S2 0.00 74.20 1 8.39 60555 Mixed 1.150 0.996
SR S2 0.00 74.20 2 8.41 61752 Mixed 1.172 1.016
SR S2 0.00 74.20 3 8.41 60338 Mixed 1.145 0.992
SR S2 0.59 76.20 1 8.42 91498 Mixed 1.737 1.466
SR S2 1.29 75.30 1 8.42 113976 Mixed 2.164 1.847
SR S2 1.64 76.30 1 8.44 129252 Mixed 2.454 2.068
SR S2 2.35 75.90 1 8.44 162893 Mixed 3.092 2.619
SR S3 0.00 75.00 1 8.50 61126 Mixed 1.160 0.995
SR S3 0.00 75.00 2 8.51 60754 Mixed 1.153 0.989
SR S3 0.00 75.00 3 8.51 60881 Mixed 1.156 0.991
SR S3 0.71 75.00 1 8.51 95994 Mixed 1.822 1.562
SR S3 1.18 75.80 1 8.53 113895 Mixed 2.162 1.834
SR S3 1.76 75.90 1 8.53 157107 Mixed 2.982 2.526
SR S3 2.70 76.70 1 8.53 178780 Mixed 3.394 2.845
SR75 0.00 73.10 1 8.70 70941 Cortical 1.334 1.173
SR75 0.50 73.00 1 8.70 98140 Cortical 1.845 1.625
SR75 1.00 73.40 1 8.70 119670 Cortical 2.250 1.971
SR75 1.40 74.10 1 8.70 135919 Cortical 2.555 2.217
SR75 2.40 73.70 1 8.71 196938 Cortical 3.703 3.230

234
Short Term (14 days) Storage Condition Tests [Standard Addition Method Used]

Mass of Addition Retention Conc. Conc.

Bone Peak Type of Rack Conc.
Bone in Conc. Time (wt.%) (wt.%)
Sample Area Bone # (mM)
Sample (mM) (min) CC* SA+
VAC1 75.1 0.000 8.91 69301 Cortical 3 1.316 1.115 1.399
VAC1 75.1 0.815 8.94 108337 Cortical 3
VAC1 75.1 1.630 8.96 137375 Cortical 3
VAC1 75.1 2.445 8.95 175572 Cortical 3
VAC2 74.2 0.000 8.95 61184 Cortical 3 1.161 0.996 1.336
VAC2 74.2 0.815 8.97 95864 Cortical 3
VAC2 74.2 1.630 8.98 126530 Cortical 3
VAC2 74.2 2.445 8.99 159115 Cortical 3
VAC3 75.9 0.000 8.98 62289 Cortical 3 1.182 0.992 1.302
VAC3 75.9 0.815 9.01 90476 Cortical 3
VAC3 75.9 1.630 9.00 127651 Cortical 3
VAC3 75.9 2.445 9.02 156934 Cortical 3
H2O1 75.1 0.000 8.82 78636 Cortical 3 1.493 1.265 1.478
H2O1 75.1 0.815 8.84 113021 Cortical 3
H2O1 75.1 1.630 8.85 152060 Cortical 3
H2O1 75.1 2.445 8.87 187165 Cortical 3
H2O2 75.7 0.000 8.85 70177 Cortical 3 1.332 1.120 1.361
H2O2 75.7 0.815 8.88 105079 Cortical 3
H2O2 75.7 1.630 8.89 139497 Cortical 3
H2O2 75.7 2.445 8.89 176020 Cortical 3
H2O3 74.6 0.000 8.88 70126 Cortical 3 1.331 1.136 1.372
H2O3 74.6 0.815 8.92 103440 Cortical 3
H2O3 74.6 1.630 8.92 141600 Cortical 3
H2O3 74.6 2.445 8.93 175036 Cortical 3
RT1 75 0.000 8.73 71212 Cortical 3 1.352 1.147 1.171
RT1 75 0.815 8.76 113808 Cortical 3
RT1 75 1.630 8.77 155495 Cortical 3
RT1 75 2.445 8.78 156383 Cortical 3
RT2 74.5 0.000 8.77 71879 Cortical 3 1.365 1.166 1.259
RT2 74.5 0.815 8.79 113637 Cortical 3
RT2 74.5 1.630 8.79 153634 Cortical 3
RT2 74.5 2.445 8.80 192471 Cortical 3
RT3 74.5 0.000 8.79 73395 Cortical 3 1.393 1.190 1.362
RT3 74.5 0.815 8.81 108838 Cortical 3
RT3 74.5 1.630 8.83 144660 Cortical 3
RT3 74.5 2.445 8.83 184839 Cortical 3
*CC: Calibration Curve Method, +: Standard Addition Method

235
 Analysis of Location Tests
Mass of Addition Retention Conc. Conc.
Bone Peak Type of Rack Conc.
Location Bone in Conc. Time (wt.%) (wt.%)
Sample Area Bone # (mM)
Sample (mM) (min) CC* SA+
Dorsal R4B5 74.5 0 8.929 108394 Cortical 4 2.039 1.789 2.071
Dorsal R4B5 74.5 0.815 8.975 137331 Cortical 4
Dorsal R4B5 74.5 1.63 8.98 183376 Cortical 4
Dorsal R4B5 74.5 2.445 8.979 214997 Cortical 4
Central R4B5 75.4 0 8.967 99554 Cortical 4 1.873 1.623 2.391
Central R4B5 75.4 0.815 8.994 127977 Cortical 4
Central R4B5 75.4 1.63 9.003 153230 Cortical 4
Central R4B5 75.4 2.445 9.003 186713 Cortical 4
Ventral R4B5 74.5 0 8.977 61590 Cortical 4 1.159 1.016 1.183
Ventral R4B5 74.5 0.815 9.024 87550 Cortical 4
Ventral R4B5 74.5 1.63 9.014 133013 Cortical 4
Ventral R4B5 74.5 2.445 9.017 162818 Cortical 4
Dorsal R4B6 75.5 0 9.003 87252 Cortical 4 1.641 1.421 1.598
Dorsal R4B6 75.5 0.815 9.026 120722 Cortical 4
Dorsal R4B6 75.5 1.63 9.045 155230 Cortical 4
Dorsal R4B6 75.5 2.445 9.056 198725 Cortical 4
Central R4B6 75.1 0 9.076 41407 Cortical 4 0.779 0.678 1.028
Central R4B6 75.1 0.815 9.075 110571 Cortical 4
Central R4B6 75.1 1.63 9.079 139464 Cortical 4
Central R4B6 75.1 2.445 9.083 157432 Cortical 4
Ventral R4B6 75.1 0 9.071 40137 Cortical 4 0.755 0.657 0.825
Ventral R4B6 75.1 0.815 9.111 72077 Cortical 4
Ventral R4B6 75.1 1.63 9.116 112160 Cortical 4
Ventral R4B6 75.1 2.445 9.146 139773 Cortical 4
Dorsal R4B7 74.9 0 9.177 63804 Cortical 4 1.667 1.455 1.945
Dorsal R4B7 74.9 0.815 9.176 91781 Cortical 4
Dorsal R4B7 74.9 1.63 9.21 110302 Cortical 4
Dorsal R4B7 74.9 2.445 9.211 136890 Cortical 4
Central R4B7 75.2 0 9.19 95666 Cortical 4 1.800 1.564 2.457
Central R4B7 75.2 0.815 9.19 113849 Cortical 4
Central R4B7 75.2 1.63 9.213 148440 Cortical 4
Central R4B7 75.2 2.445 9.226 173226 Cortical 4
Ventral R4B7 75.7 0 9.177 71806 Cortical 4 1.351 1.166 1.469
Ventral R4B7 75.7 0.815 9.237 104047 Cortical 4
Ventral R4B7 75.7 1.63 9.27 134642 Cortical 4
Ventral R4B7 75.7 2.445 9.281 174202 Cortical 4
*CC: Calibration Curve Method, +: Standard Addition Method

236
 Mass of Addition Retention Type Conc. Conc.
Bone Peak Rack Conc.
Location Bone in Conc. Time of (wt.%) (wt.%)
Sample Area # (mM)
Sample (mM) (min) Bone CC* SA+
Dorsal R6B4 75.5 0.00 8.29 30992 Cortical 6 0.626 0.542 0.631
Dorsal R6B4 75.5 0.82 8.30 62110 Cortical 6
Dorsal R6B4 75.5 1.63 8.30 92397 Cortical 6
Dorsal R6B4 75.5 2.45 8.31 129882 Cortical 6
Central R6B4 74.9 0.00 8.31 20354 Cortical 6 0.411 0.359 0.596
Central R6B4 74.9 0.82 8.31 50322 Cortical 6
Central R6B4 74.9 1.63 8.31 77492 Cortical 6
Central R6B4 74.9 2.45 8.30 98686 Cortical 6
Ventral R6B4 74.75 0.00 8.32 9713 Cortical 6 0.196 0.171 0.315
Ventral R6B4 74.75 0.82 8.32 36599 Cortical 6
Ventral R6B4 74.75 1.63 8.35 62133 Cortical 6
Ventral R6B4 74.75 2.45 8.32 83010 Cortical 6
Dorsal R6B5 74.5 0.00 8.35 24293 Cortical 6 0.490 0.430 0.588
Dorsal R6B5 74.5 0.82 8.36 46046 Cortical 6
Dorsal R6B5 74.5 1.63 8.37 87418 Cortical 6
Dorsal R6B5 74.5 2.45 8.35 103696 Cortical 6
Central R6B5 74.8 0.00 8.36 13911 Cortical 6 0.281 0.245 0.461
Central R6B5 74.8 0.82 8.36 39194 Cortical 6
Central R6B5 74.8 1.63 8.40 65367 Cortical 6
Central R6B5 74.8 2.45 8.39 82873 Cortical 6
Ventral R6B5 74.8 0.00 8.42 7768 Cortical 6 0.157 0.137 0.241
Ventral R6B5 74.8 0.82 8.36 27911 Cortical 6
Ventral R6B5 74.8 1.63 8.42 51838 Cortical 6
Ventral R6B5 74.8 2.45 8.40 71730 Cortical 6
Dorsal R6B6 75.4 0.00 8.43 21719 Cortical 6 0.438 0.380 0.621
Dorsal R6B6 75.4 0.82 8.45 47970 Cortical 6
Dorsal R6B6 75.4 1.63 8.45 74654 Cortical 6
Dorsal R6B6 75.4 2.45 8.46 96636 Cortical 6
Central R6B6 74.5 0.00 8.43 15062 Cortical 6 0.304 0.267 0.364
Central R6B6 74.5 0.82 8.47 33842 Cortical 6
Central R6B6 74.5 1.63 8.47 58842 Cortical 6
Central R6B6 74.5 2.45 8.48 87253 Cortical 6
Ventral R6B6 75.9 0.00 8.48 5746 Cortical 6 0.116 0.100 0.096
Ventral R6B6 75.9 0.82 8.46 27341 Cortical 6
Ventral R6B6 75.9 1.63 8.49 52316 Cortical 6
Ventral R6B6 75.9 2.45 8.48 83067 Cortical 6
*CC: Calibration Curve Method, +: Standard Addition Method

237
 Appendix D

Standard Addition Calibration Curves Produced in Aim Three Bone Solutions

Short Term (14 Days) Storage Condition Tests

Compare to Calibration Curve of Standard Set 08

Hydrated (H20) Samples

H20 1
200000

150000
Area Units

100000
y = 44739x + 78027
50000
R² = 0.9994
0
-2 -1.5 -1 -0.5 0 0.5 1 1.5 2 2.5 3
Concentration (mM)

H20 2
200000

150000
Area Units

100000

50000 y = 43184x + 69901

R² = 0.9998
0
-2 -1.5 -1 -0.5 0 0.5 1 1.5 2 2.5 3
-50000
Concentration (mM)

H20 3
200000

150000
Area Units

100000

50000 y = 43299x + 69617

R² = 0.9993
0
-2 -1.5 -1 -0.5 0 0.5 1 1.5 2 2.5 3
-50000
Concentration (mM)

238
 Dehydrated (VAC) Samples
VAC 1
200000

160000
Area Units

120000

80000
y = 42681x + 70469
40000
R² = 0.9969
0
-2 -1.5 -1 -0.5 0 0.5 1 1.5 2 2.5 3
Concentration (mM)

VAC 2
200000

150000
Area Units

100000

50000 y = 39811x + 62004

R² = 0.9995
0
-2 -1.5 -1 -0.5 0 0.5 1 1.5 2 2.5 3
-50000
Concentration (mM)

VAC 3
200000

150000
Area Units

100000

50000 y = 39400x + 61171

R² = 0.9972
0
-2 -1.5 -1 -0.5 0 0.5 1 1.5 2 2.5 3
-50000
Concentration (mM)

239
 Room Temperature (RT) Samples
RT 1
200000
Area Units

100000 y = 51707x + 71364

R² = 1
0
-1.5 -1 -0.5 0 0.5 1 1.5 2
Concentration (mM)

200000
Area Units

100000 y = 36466x + 79645

R² = 0.8957
0
-2 -1.5 -1 -0.5 0 0.5 1 1.5 2 2.5 3
Concentration (mM)

*Top contains curve without 2.445mM addition, as an error was made in preparation

RT 2
250000
200000
Area Units

150000
100000
y = 49297x + 72639
50000
R² = 0.9997
0
-2.5 -2 -1.5 -1 -0.5 0 0.5 1 1.5 2 2.5 3

Concentration (mM)
RT 3
200000

150000
Area Units

100000

50000 y = 45418x + 72410

R² = 0.9991
0
-2.5 -2 -1.5 -1 -0.5 0 0.5 1 1.5 2 2.5 3
Concentration (mM)

240
 Location Test Set 1 Sample Standard Addition Calibration Curves
Compare to Standard Set 09 Calibration Curve
R4B5 Dorsal

R4B5 Central

R4B5 Ventral

241
R4B6 Dorsal

R4B6 Central

R4B6 Ventral

242
R4B7 Dorsal

R4B7 Central

R4B7 Ventral

243
 Location Test Set 2 Sample Standard Addition Calibration Curves
Compare to Standard Set 10 Calibration Curve

R6B4 Dorsal

R6B4 Central

R6B4 Ventral

244
R6B5 Dorsal

R6B5 Central

R6B5 Ventral

245
R6B6 Dorsal

R6B6 Central

R6B6 Ventral

246
 REFERENCES

[1] Aerssens, J., Boonen, S., Lowet, G., & Dequeker, J. (1998). Interspecies Differences
in Bone Composition, Density, and Quality: Potential Implications for in vivo Bone
Research 1. Endocrinology, 139(2), 663-670.
[2] Atipairin, A., Woradechakul, C., Chee, K. S., Sawatdee, S., & Yoon, A. S. (2014).
METHOD VALIDATION FOR DETERMINATION OF SILDENAFIL CITRATE IN
EXTEMPORANEOUS ORAL SUSPENSION. International Journal of Pharmacy and
Pharmaceutical Sciences, 6(7).
[3] Berg, J. M., Tymoczko, J. L., & Stryer, L. (2002). The Citric Acid Cycle Oxidizes
Two-Carbon Units.
[4] Bio-Rad. Guidelines for Use and Care of Aminex® Resin-Based Columns. N.p.: Bio-
Rad, n.d. Print.
[5] Bio-Rad Laboratories. Aminex HPX-87H Column: Industry Standard for HPLC
Analysis of Small Molecules. N.p.: Bio-Rad Laboratories, n.d. Print.
[6] Boskey, A. L. (2007). Mineralization of bones and teeth. Elements, 3(6), 385-391.
[7] Boskey, A. L. (2013). Bone composition: relationship to bone fragility and
antiosteoporotic drug effects. BoneKey reports, 2.
[8] Chow, P. K., Ng, R. T., & Ogden, B. E. (2008). Using animal models in biomedical
research: a primer for the investigator. World Scientific.
[9] Clarke, B. (2008). Normal bone anatomy and physiology. Clinical journal of the
American Society of Nephrology, 3(Supplement 3), S131-S139.
[10] Cohen, L., Dean, M., Shipov, A., Atkins, A., Monsonego-Ornan, E., & Shahar, R.
(2012). Comparison of structural, architectural and mechanical aspects of cellular and
acellular bone in two teleost fish. The Journal of experimental biology, 215(11), 1983-
1993.
[11] Costello, L. C., Franklin, R. B., Reynolds, M. A., & Chellaiah, M. (2012). The
important role of osteoblasts and citrate production in bone formation:“osteoblast
citration” as a new concept for an old relationship. The Open bone journal, 4.
[12] Currey, J. D. (2002). Bones: structure and mechanics. Princeton University Press.
[13] Davies, E., Müller, K. H., Wong, W. C., Pickard, C. J., Reid, D. G., Skepper, J. N.,
& Duer, M. J. (2014). Citrate bridges between mineral platelets in bone.Proceedings of
the National Academy of Sciences, 111(14), E1354-E1363.
[14] Dickens, F. (1941). The citric acid content of animal tissues, with reference to its
occurrence in bone and tumour. Biochemical Journal, 35(8-9), 1011.

247
[15] Dixon, T. F., & Perkins, H. R. (1952). Citric acid and bone metabolism.Biochemical
Journal, 52(2), 260.
[16] Dunphy M. (2014). An Engineering Approach to Forensic Methods: The Citrate
Method for Postmortem Interval Determination. Master’s Thesis, Dept. of
Bioengineering, Clemson University.
[17] Dutta, P. K., & Dutta, J. (Eds.). (2013). Multifaceted Development and Application
of Biopolymers for Biology, Biomedicine and Nanotechnology (Vol. 254). Springer.

[18] Figueiredo, M., Cunha, S., Martins, G., Freitas, J., Judas, F., & Figueiredo, H.
(2011). Influence of hydrochloric acid concentration on the demineralization of cortical
bone. Chemical Engineering Research and Design, 89(1), 116–124.
doi:10.1016/j.cherd.2010.04.013
[19] Franklin, R. B., Chellaiah, M., Zou, J., Reynolds, M. A., & Costello, L. C. (2014).
Evidence that osteoblasts are specialized citrate-producing cells that provide the citrate
for incorporation into the structure of bone. The Open bone journal, 6, 1.
[20] Froome, C. (2014). Citrate Content of Bone: A Potential Measure of Postmortem
Interval. Proceedings of the National Conference on Undergraduate Research (NCUR)
(1100-1108). University of Kentucky, Lexington, KY.
[21] Gibbs, L. (1991). Citrate, Sex and Skeletal Remains (Doctoral dissertation, MA
Dissertation, McMaster University, Ontario, Canada).
[22] Guo, Y., Tran, R. T., Xie, D., Wang, Y., Nguyen, D. Y., Gerhard, E., & Yang, J.
(2015). Citrate‐based biphasic scaffolds for the repair of large segmental bone
defects. Journal of Biomedical Materials Research Part A,103(2), 772-781.
[23] Harris, D. C. (2007). Quantitative chemical analysis. Macmillan.
[24] Herring, G. M. (2012). The organic matrix of bone. The biochemistry and physiology
of bone, 1, 127-189.
[25] Hodgkinson, A. (1970). Determination of oxalic acid in biological material.Clinical
Chemistry, 16(7), 547-557.
[26] Holmes, R. P. (1995). Measurement of urinary oxalate and citrate by capillary
electrophoresis and indirect ultraviolet absorbance. Clinical chemistry, 41(9), 1297-1301.
[27] Hore, P., Hore, P., Stephen, J. J. W., Jones, J., & Wimperis, S. (2015). Nmr: the
Toolkit: How Pulse Sequences Work. Oxford University Press, USA.
[28] Hu, Y. Y., Rawal, A., & Schmidt-Rohr, K. (2010). Strongly bound citrate stabilizes
the apatite nanocrystals in bone. Proceedings of the national academy of
sciences, 107(52), 22425-22429.

248
[29] Huculak, M. A., & Rogers, T. L. (2009). Reconstructing the Sequence of Events
Surrounding Body Disposition Based on Color Staining of Bone*.Journal of forensic
sciences, 54(5), 979-984.
[30] Iscan, M. Y., & Steyn, M. (2013). The human skeleton in forensic medicine. Charles
C Thomas Publisher.
[31] Jurt, S., Schär, M., & Meyer, V. R. (2001). Procedure for the quantification of rider
peaks. Journal of Chromatography A, 929(1), 165-168.
[32] Kanz, F., Reiter, C., & Risser, D. U. (2014). Citrate content of bone for time since
death estimation: results from burials with different physical characteristics and known
PMI. Journal of forensic sciences, 59(3), 613-620.
[33] Khurana, J. S., & Fitzpatrick, L. A. (2009). Osteoporosis and metabolic bone
disease. In Bone Pathology (pp. 217-237). Humana Press.
[34] Knuuttila, M., Lappalainen, R., Alakuijala, P., & Lammi, S. (1985). Statistical
evidence for the relation between citrate and carbonate in human cortical bone.Calcified
tissue international, 37(4), 363-366.
[35] Kuyper, A. C. (1938). The quantitative precipitation of citric acid. Journal of
Biological Chemistry, 123(2), 405-407.
[36] Lambert JB, Simpson SV, Weiner SG, Buikstra JE. Induced metal-ion exchange in
excavated human bone. J Archaeological Sci, 12(2):85-92, 1985.
[37] Lindsay, S. (1991). HPLC by open learning. Johan wiley and SDons, London, 30-45.
[38] Lough, W. J., & Wainer, I. W. (Eds.). (1995). High performance liquid
chromatography: fundamental principles and practice. CRC Press
[39] Martini, F., & Nath, J. L. (2009). Fundamentals of anatomy & physiology. San
Francisco: Pearson/Benjamin Cummings.
[40] Megyesi, M. S., Nawrocki, S. P., & Haskell, N. H. (2005). Using accumulated
degree-days to estimate the postmortem interval from decomposed human remains. J
Forensic Sci, 50(3), 618-26.
[41] Miller, E. J., & Martin, G. R. (1968). The collagen of bone. Clinical orthopaedics
and related research, 59, 195-222.
[42] Parris, N. A. (2000). Instrumental liquid chromatography: a practical manual on
high-performance liquid chromatographic methods. Elsevier.
[43] Pearce, A. I., Richards, R. G., Milz, S., Schneider, E., & Pearce, S. G. (2007).
Animal models for implant biomaterial research in bone: a review. Eur Cell Mater, 13(1),
1-10.

249
[44] Perkins, H. R., & Dixon, T. F. (1953). Parathyroid and bone
citrogenase.Science, 118(3057), 139-140.
[45] Pucher, G. W., Vickery, H. B., & Wakeman, A. J. (1934). Determination of malic
acid in plant tissue. Industrial & Engineering Chemistry Analytical Edition, 6(4), 288-
291.
[46] Purves, W. K., Sadava, D. E., Orians, G. H., & Heller, H. C. (2000). Life: The
Science of Biology. Macmillan.
[47] Robey, P. G., & Boskey, A. L. (2009). The composition of bone. Primer on the
Metabolic Bone Diseases and Disorders of Mineral Metabolism, 7, 32-38.
[48] Ross, M. H., Kaye, G. I., & Pawlina, W. (2003). Histology: a text and atlas: with
cell and molecular biology. Lippincott Williams & Wilkins.
[49] Schwarcz, H. P., Agur, K., & Jantz, L. M. (2010). A New Method for Determination
of Postmortem Interval: Citrate Content of Bone*. Journal of forensic sciences, 55(6),
1516-1522.
[50] Stern, J. R., & Ochoa, S. (1951). Enzymatic Synthesis of Citric Acid I. Synthesis
With Soluble Enzymes. Journal of Biological Chemistry,191(1), 161-172.
[51] Stevenson, J. C., & Marsh, M. S. (2007). An atlas of osteoporosis. CRC Press.
[52] Shabir, G. A. (2004). A practical approach to validation of HPLC methods under
current good manufacturing practices. Journal of validation technology, 10, 210-218.
[53] Talwar, G. P., & Srivastava, L. M. (2002). Textbook of biochemistry and human
biology. PHI Learning Pvt. Ltd..
[54] Tat, S. K., Pelletier, J. P., Vergés, J., Lajeunesse, D., Montell, E., Fahmi, H., ... &
Martel-Pelletier, J. (2007). Chondroitin and glucosamine sulfate in combination decrease
the pro-resorptive properties of human osteoarthritis subchondral bone osteoblasts: a
basic science study. Arthritis Research and Therapy, 9(6), R117.
[55] Thompson, M., & Ellison, S. L. (2005). A review of interference effects and their
correction in chemical analysis with special reference to uncertainty.Accreditation and
quality assurance, 10(3), 82-97.
[56] Thorwarth, M., Schultze-Mosgau, S., Kessler, P., Wiltfang, J., & Schlegel, K. A.
(2005). Bone regeneration in osseous defects using a resorbable nanoparticular
hydroxyapatite. Journal of oral and maxillofacial surgery, 63(11), 1626-1633.
[57] Vass, A. A., Bass, W. M., Wolt, J. D., Foss, J. E., & Ammons, J. T. (1992). Time
since death determinations of human cadavers using soil solution.Journal of forensic
sciences, 37(5), 1236-1253.

250
[58] Viguet-Carrin, S., Garnero, P., & Delmas, P. D. (2006). The role of collagen in bone
strength. Osteoporosis International, 17(3), 319-336.
[59] Weil-Malherbe, H., & Bone, A. D. (1949). The micro-estimation of citric
acid.Biochemical Journal, 45(4), 377.
[60] Weitzel, M. A. (2005). A report of decomposition rates of a special burial type in
Edmonton, Alberta from an experimental field study. Journal of forensic sciences, 50(3),
641-647.
[61] White EM, Hannus LA. Chemical weathering of bone in archaeological soils. Am
Antiquity, 48(2):316-322, 1983.

251