Conjugation of Arginine–Glycine–Aspartic Acid Peptides to

Thermoreversible N-isopropylacrylamide Polymers

ERIN SMITH, JIANG BAI, CASSIE OXENFORD, JENNIFER YANG, RANJANI SOMAYAJI, HASAN ULUDAG

Department of Biomedical Engineering, and Department of Chemical and Materials Engineering, Faculty of Engineering,
University of Alberta, Edmonton, AB T6G 2G6, Canada

Received 18 June 2003; accepted 21 August 2003

ABSTRACT: Thermoreversible polymeric biomaterials are finding increased acceptance

in tissue engineering applications. One drawback of the polymers is their synthetic
nature, which does not allow direct interaction of mammalian cells with the polymers.
This limitation may be alleviated by grafting arginine– glycine–aspartic acid (RGD)
containing peptides onto the polymer backbone to facilitate interactions with cell-
surface integrins. Toward this goal, N-isopropylacrylamide (NiPAM)-based thermo-
reversible polymers containing amine-reactive N-acryloxysuccinimide (NASI) groups
were synthesized. Conjugation of RGD-containing peptides to polymers was demon-
strated with 1H NMR spectroscopy and reverse-phase high-pressure liquid chromatog-
raphy. The conjugation reaction was optimal at 4 °C and pH of 8.0, and increased with
the increasing NASI content of polymers. With a peptide grafting ratio of 0.25 mol %,
there was no significant change in the lower critical solution temperature of the
polymers. Finally, the NASI-containing polymers, cast as films, on tissue culture
polystyrene, were shown to conjugate to RGD-containing peptides and support C2C12
cell attachment. We conclude that NASI-containing thermoreversible polymers are
amenable for grafting biomimetic peptides to impart cell adhesiveness to the polymers.
© 2003 Wiley Periodicals, Inc. J Polym Sci Part A: Polym Chem 41: 3989 – 4000, 2003
Keywords: functionalization of polymers; stimuli-sensitive polymers; biocompatibil-
ity; biomimetic

INTRODUCTION ture above the solubility temperature of the

polymers. The fact that such a phase separation
Thermoreversible polymers are finding in- is driven only by a temperature change is at-
creased utilization in the tissue engineering tractive because no other exogenous molecules
field because of their potential to act as carriers need to be introduced to the studied system.
in the delivery of therapeutic proteins and cells. The thermoreversible polymers are being ex-
The thermoreversible (i.e., temperature-depen- plored in controlled delivery formulations
dent solubility) character of the polymers al- where the temperature-dependent solubility is
lows one to carry out desired manipulations, exploited to modulate the release rate of thera-
such as mixing and injecting, in a solution state peutic proteins.1–3 The thermoreversible poly-
but eventually enables one to induce a phase mers have also served as extracellular attach-
separation simply by increasing the tempera- ment matrices for cell and tissue culture be-
cause of the feasibility of with a temperature
change to discharge the cells from the attach-
Correspondence to: H. Uludag (E-mail: hasan.uludag@ ment matrix,4 –7 a process that is gentler than
ualberta.ca)
Journal of Polymer Science: Part A: Polymer Chemistry, Vol. 41, 3989 – 4000 (2003)
the proteolytic trypsin-mediated release from
© 2003 Wiley Periodicals, Inc. an attachment matrix.
3989
3990 SMITH ET AL.

The performance of thermoreversible polymers peptides containing the triamino acid sequence,
in a particular application is dependent on their arginine-glycine-aspartic acid (RGD), were used
physicochemical properties. Polymers of N-isopro- as a biomimetic moieties because of their well-
pylacrylamide (NiPAM) have been the commonly demonstrated ability to interact with cell-surface
used thermoreversible polymers and NiPAM integrin receptors.16 The attachment of C2C12
polymers have been systematically modified to cells, a cell line that exhibits bone-depositing os-
suit the needs of particular applications. Among teoblast phenotype under the influence of mor-
the properties optimized for various applications phogenetic proteins, to the RGD-grafted thermor-
are: temperature at which the phase transition eversible polymers was subsequently investi-
occurs (so called, lower critical solution tempera- gated.
ture, LCST), polymer architecture (molecular
weight, structure, etc.) and the presence of spe- EXPERIMENTAL
cific chemical moieties (e.g., fluoride for increased
stability or OCOOH groups for hydrophilicity).8,9 Materials
In our hands, NiPAM polymers with controlled NiPAM, methylmethacrylate (MMA), ethyl-
LCST10 and molecular weight11 were prepared to methacrylate (EMA), N-hydroxysuccinimide
deliver morphogenetic (i.e., tissue inducing) pro- (NHS), benzoylperoxide (BPO), 1,4-dioxane, di-
teins for bone regeneration. The polymers were ethyl ether and tetrahydrofuran (THF) were pur-
designed to localize injected proteins at a site of chased from Aldrich (Milwaukee, WI). NiPAM
administration. This approach prevents the rapid was further purified by crystallization from n-
loss of the administered proteins from the site, hexane. The inhibitors in MMA and EMA were
because of the formation of a thermo-gelling poly- removed by distillation under a high vacuum be-
mer barrier and prevention of the free protein fore polymerization. NASI was prepared by react-
diffusion. Better retention of the proteins is de- ing acryloyl chloride with NHS, as described in
sired to ultimately result in a better bone forma- Uludag et al.12 GYRGDS was custom-synthesized
tion at the administration site. The polymers with by Alberta Peptide Institute (Edmonton, AB) and
lower LCST (12–14 °C; LCST of NiPAM ho- purified to 95% homogeneity by a reverse-phase
mopolymer is ⬇27 °C), polymers with high molec- high-pressure liquid chromatography (RP-
ular weight (⬎400 kD), and polymers containing HPLC). The identity of the peptide was verified by
protein-reactive N-acryloxysuccinimide (NASI) analysis of amino acid composition and molecular
groups were the most effective for protein reten- weight (MW) from electrospray mass spectrome-
tion. With NASI-containing polymers, the pro- try (expected: 654.26 amu, obtained: 654.53 amu).
teins were chemically conjugated to the polymer Other peptides, RGD, GRGDSP, GRGDSPK, and
backbone via the proteins’ amine groups.12 The carboxyl-YVAR were obtained from SIGMA (St
conjugation was achieved without the need for Louis, MO) and BACHEM (Torrense, CA). The
additional coupling agents, and the polymer con- peptide purities were all ⬎95% and their molec-
jugation itself did not exhibit an adverse effect on ular mass were confirmed by electrospray mass
the protein activity.13 spectrum analysis by the manufacturers. The
A significant drawback of the designed poly- phosphate buffer used for conjugation reactions
mers is their inability to support cell attachment was prepared by mixing 0.1 M of Na2HPO4 and
because of their synthetic nature. We desired to 0.1 M of NaH2PO4 䡠 H20 solutions to give a desired
engineer the thermoreversible polymers, not only pH. Hanks Balanced Salt Solution (HBSS),
to control in situ retention of morphogenetic pro- DMEM tissue culture medium and Penicillin/
teins, but also to allow direct attachment of cells Streptomycin antibiotic solution were obtained
responsive to the morphogenetic proteins. Cell from GIBCO (Grand Island, NY). Fetal Bovine
attachment is critical because it influences cellu- Serum (FBS) and 3-(4,5-dimethylthiazol-2-yl)-
lar response to morphogenetic proteins, ulti- 2,5-diphenyltetrazolium bromide (MTT) were
mately leading to a more robust bone deposition from SIGMA. The C2C12 cells were a generous
under the influence of morphogenetic proteins.14 gift from Dr. Civetilli (Washington University, St.
This study was performed to test whether biomi- Louis, MO).
metic peptides could be conjugated to the ther-
moreversible polymers. The amine-reactive Ni- Polymer Synthesis and Characterization
PAM/NASI polymers described in a previous pub- The preparation of NiPAM, NiPAM/NASI, Ni-
lication15 were used as the base polymer and PAM/NASI/MMA, and NiPAM/NASI/EMA poly-
 CONJUGATION OF RGD PEPTIDES 3991

⬇ 3.9 ppm), MMA (OCH3: ␦ ⬇ 3.6 ppm) and NASI

(OCH2OCH2-: ␦ ⬇ 2.8 ppm) were normalized for
the number of Hs and used to calculate the rela-
tive ratio of each monomer. Because of overlap of
chemical shifts for NiPAM and EMA, elemental
analysis was used for the composition of EMA-
containing polymers. A spectroscopic method,
based on aminolysis of NASI groups with 0.1 M
NH4OH,12 was additionally used to determine the
NASI content of the polymers.
The polymer MWs were determined by gel per-
meation chromatography and static light scatter-
ing. A 20 ␮L of polymer solution (20 mg/mL in
THF) was manually injected onto a 7.8 䡠 300 mm
Styragel HMW 6E column (Waters, Inc.; Milford,
MA), eluted with 1 mL/min of THF and the elu-
tion pattern was detected by refractive index/light
scattering detectors (PD2020; Precision Detec-
tors, Andover, MA). The polymer MWs were esti-
mated by Precision MW Analysis software:18
MW ⫽ K 2 䡠 I LS/K 1 䡠 I RI 䡠 (dn/dc), where ILS is the
light scattering signal intensity, IRI is the refrac-
tometer signal intensity, dn/dc is the refractive
index increment, and K2 and K1 are the calibra-
tion constants. The calibration constants were ob-
tained with polystyrene standards.
The polymer LCSTs were determined by spec-
troscopy. The polymers were dissolved at 5–10
mg/mL in 0.1 M phosphate buffer (pH ⫽ 7.4) and
Figure 1. Summary of the chemical scheme used for
the synthesis of NiPAM/NASI/AMA polymers. BPO
1 mL of polymer solution was added into a spec-
served as the free-radical initiator. The conjugation trophotometer cuvette, which equipped with a
reaction between the NASI-containing polymers and water-circulation unit. The water temperature
the RGD peptides (R-NH2, where the terminal amino was adjusted between 10 and 30 °C (in 0.5 °C
group is indicated) is also shown. See individual figure increments every 10 min) with a refrigerated/
legends for the exact nature of R. heated water circulator and the optical density
was determined at 420 nm. Actual temperature of
the samples was measured with a digital ther-
mers was adopted from a reported procedure,17 mometer. The data was fitted to a sigmoidal curve
and was described previously.15 Figure 1 summa- and the LCST was taken as the midpoint of the
rizes the used chemical synthesis scheme. Briefly, inflection point.
desired amounts of NiPAM, NASI, MMA and
EMA monomers were dissolved in dioxane and
Peptide Conjugation to Polymers
0.2 mol % (0.023 g) BPO was added to this solu-
tion. The solution was purged with N2 while stir- In a typical conjugation reaction, GYRGDS and a
ring with a magnetic stirrer and the temperature polymer were separately dissolved in 0.1 M phos-
was increased to 70 °C by an oil bath. After 22 h, phate buffers and mixed at 1:1 ratio to give de-
the polymer was precipitated by hexane and pu- sired concentrations of peptide and polymer. Un-
rified by dissolving the precipitate in THF and less otherwise stated (see figure legends), all re-
precipitating the solution by diethyl ether. The actions were performed at 4 °C for 24 h with a
polymers were dried at 50 °C in vacuo for 1 week. polymer concentration of 4 mg/mL and peptide
The compositions of polymers containing Ni- concentration of 0.1 mg/mL. In an experiment
PAM, MMA, and NASI were determined by 1H where the effect of pH on reaction rate was ex-
NMR spectroscopy. The peak areas corresponding plored, the reaction was carried out in 0.1 M
to unique chemical shifts of NiPAM (OCHO: ␦ phosphate buffers at pH of 6.0, 7.0, and 8.0, and
3992 SMITH ET AL.

0.1 M carbonate buffer at pH of 9.2. At the com- overnight in a humidified environment of 95/5%
pletion of the reaction, the polymer/peptide solu- air/CO2. MTT dissolved in HBSS (5 mg/mL) was
tion was directly injected onto a C-18 column added to the wells on the next day to give a final
(5-␮m particle size, 4.6 mm 䡠 5.0 mm; VYDAC; MTT concentration of 1 mg/mL. After a 2-h incu-
Hesperia, CA). The peptide was eluted with a bation period with MTT, the cells were washed
linear gradient of 0.1% trifluoroacetic acid (TFA) with HBSS and analyzed by light microscopy.
in H2O (A) and 0.1% TFA in 90/10% Acetonitrile/ The conjugation of two RGD peptides to poly-
H2O (B): 0% B for 2 min, 0 –100% B in 12 min, mer films was also investigated with the de-
100 – 0% B in 2 min and 0% B for 2 min. The scribed HPLC methodology. With NiPAM/EMA
disappearance of the peptide peak at 7.5– 8.1 min and NiPAM/NASI/EMA (JB1 and JB2, respec-
was used as a measure of peptide conjugated to tively), polymer films were case as above, and the
the polymer. The polymer eluted as a separate aqueous solutions of two peptides, GRGDSP and
peak at ⬇14 min under the elution conditions so GRGDSPK (0, 0.02, 0.04, and 0.08 mg/mL in 0.1
that a reduction in peptide peak could be used as M phosphate buffer, pH ⫽ 7.4) were incubated
a quantitative measure of conjugation without with the polymer films for 24 h at 37 °C in a
interference from the polymer. The extent of con- humidified environment. The peptide solutions
jugation reaction was expressed as either %pep- were then removed and injected to the HPLC for
tide conjugated or number of peptides conjugated quantitation of the peptide concentration in the
per polymer: {(peptide concentrated 䡠 conjugated supernatant. The reduction in peptide concentra-
fraction) ⫼ MWpeptide} ⫼ {polymer concentrated tions was taken as the amount of peptide conju-
⫼ MWpolymer}, where MWpeptide and MWpolymer gated to the polymer films. To account for the
were the peptide and polymer molecular weights, effect of evaporation during the 24-h incubation
respectively. period, a nonreactive peptide (carboxyl-YVAR,
Peptide conjugation to polymers was also in- where the NASI-reactive N-terminal ONH2 was
vestigated with 1H NMR. For this, NiPAM/NASI carboxylated) were incubated in the reaction me-
(8.4% NASI; 4 mg/mL) was dissolved in distilled/ dium and any changes in the concentration in this
deionized water with: (1) no peptide, (2) 2.5 peptide were used to account for the effect of
mg/mL RGD, (3) 0.4 mg/mL RGD and (4) 0.4 evaporation. The peptide conjugations were ex-
mg/mL RGD ⫹ 0.1 M NH4OH. After 24 h of incu- pressed as the nmol peptide conjugated per cen-
bation, the samples were extensively dialyzed timeter squared film area.
(12–14 kD cutoff) against distilled/deionized wa-
ter to remove unreacted peptides, lyophilized and
dissolved in D2O for analysis with a Bruker RESULTS AND DISCUSSION
AM300 (Billerica, MA). Peptide specific 1H-peaks
were identified by analysis of an RGD solution (1 This study was performed with the long-term pur-
mg/mL) in D2O. pose of designing cell-compatible, synthetic poly-
mers that exhibit thermoreversible solubility un-
der aqueous conditions. We envisioned achieving
Cell Culture on Peptide-Grafted Polymer Films
this by grafting biomimetic peptide sequences
The ability of peptide-grafted polymers to support onto thermoreversible polymer backbones. In-
cell attachment was tested in 24-well tissue cul- stead of using chemical crosslinkers to link the
ture plates. Myoblastic C2C12 cells were used for biomimetic peptides with suitable reactive groups
this assay because of their ability to respond to on polymers, thermoreversible polymers with pre-
bone morphogenetic proteins.19 0.5 mL of a poly- activated NASI groups were prepared so that
mer solution (20 mg/mL in methanol) was added grafting can be performed without additional re-
to wells and the solvent was allowed to evaporate agents. The NHSOesters can undergo hydrolysis
overnight at room temperature in a sterile flow- in aqueous medium, but aminolysis of the ester is
hood. After washing the dry films with HBSS, 0.1 preferred in the presence of primary amines.20
mg/mL GYRGDS in 0.1 M phosphate buffer (pH The aminolysis reaction was enhanced at higher
⫽ 8) was added to wells and allowed to incubate pH values but a high aminolysis:hydrolysis selec-
for 24 h at 37 °C. Trypsinized C2C12 cells tivity could still be achieved at neutral pH values
(⬇60,000 cells/well in 0.5 mL of tissue culture (⬇7.4). Others have shown that small molecular
medium), suspended in DMEM/10% FBS, were amines having a high pKa and lacking bulky side-
then added to the polymer films and incubated groups exhibited relatively higher reaction rates
 CONJUGATION OF RGD PEPTIDES 3993

Table 1. Polymers Used for this Studya

Polymer Monomers Feed Ratio Composition LCST (°C) MW (kD)

B NiPAM 100 100 26.7 212

D NiPAM:NASI 96.5:3.5 97.3:2.7 25.5 175
F NiPAM:NASI 93.5:6.5 94.9:5.1 24.0 183
G NiPAM:NASI 90.9:9.1 92.8:7.2 19.3 191
H NiPAM:NASI 90.9:9.1 91.6:8.4 NDb 383
K NiPAM:NASI:MMA 93.0:2.1:4.9 93.1:1.3:5.6 24.8 622
Q NiPAM:NASI:MMA 88.1:1.9:10.0 87.1:1.3:11.6 22.7 457
S NiPAM:NASI:MMA 81.1:1.8:17.1 78.0:1.1:20.9 19.9 404
U NiPAM:EMA 84.6:15.4 78.5:21:5 24.6 404
JB1 NiPAM:EMA 80:20 76.5:23.5 17.3 594
JB2 NiPAM:NASI:EMA 77.6:19.4:3.0 73.8:23.7:2.5 16.6 495
a
The feed ratios and the final polymer compositions were based on monomer mole ratios.
b
ND: Not determined.

with NHS esters.21 Protein conjugation to acti- were obtained. Moreover, there was no detectable
vated thermoreversible polymers was shown in peptide conjugation when the reaction was per-
our previous studies, as well as reports from the formed with excess NH4OH, indicating that dial-
literature.22,23 One study reported conjugating ysis was effective in removing the unreacted RGD
RGD peptides to NiPAM polymers but this was peptide (not shown).
achieved via a three-step process in which the An RP-HPLC method was subsequently devel-
thermoreversible polymers were sequentially re- oped for routine quantification of conjugation ef-
acted with a poly(ethylene glycol) spacer, a ficiency. The short-sequence peptide, RGD, was
crosslinker [sulfosuccinimidyl 4-(N-maleimidom- not suitable for this method because of its nonre-
ethyl)cyclohexane-1-carboxylate], and finally by tention on C-18 column. A more hydrophobic pep-
the peptide.24 The pre-activated polymers used in tide, GYRGDS, was synthesized for this part of
this study are preferable to eliminate multi-step the study and, as expected, a clear retention peak
conjugation processes. at 7.5– 8.1 min could be obtained under the elu-
tion conditions described in the Experimental sec-
tion. With a set of GYRGDS standards, the peak
RGD-Peptide Conjugation to Polymers
area was correlated (r2 ⫽ 0.9937) with the in-
The properties of the polymers used for this study jected peptide concentration: Area ⫽ 47781 䡠 conc.
are provided in Table 1. An initial study was (in ␮g/mL) ⫹ 408155. As expected, there was no
performed to investigate peptide conjugation by detectable conjugation when the peptide was in-
1
H NMR. The NiPAM/NASI polymer incubated cubated with NiPAM homopolymer lacking NASI,
with RGD (2.5 mg/mL) had the expected arginine or when the peptide was incubated with NiPAM/
OCH2- triplet at ␦⬇3.2 ppm but other 1H peaks NASI polymer in the presence of 0.1 M NH4OH.
characteristic of RGD were not detectable be- With NiPAM/NASI with 8.4 mol % NASI, the
cause of broad polymer peaks at 3.9 ppm and the effects of polymer concentration, temperature and
peaks between 2.5 and 1.0 ppm (Fig. 2). The ar- pH on conjugation efficiency were explored (Fig.
ginine peak was lacking for the control polymer 3). The conjugation efficiency was increased as
(i.e., polymer incubated without RGD, not the concentration of polymer was increased [Fig.
shown). On the basis of arginine and NiPAM spe- 3(A)]. Concentrations beyond 8 mg/mL were not
cific protons, the grafting efficiency was estimated tested because injections at these concentrations
to be 3.8 peptides per polymer. To ensure that this destabilized the HPLC system by causing signif-
was not free RGD (i.e., undialyzed RGD mixed icant pressure fluctuations. To obtain reproduc-
with polymer), a subsequent RGD reaction was ible elution patterns, 4 mg/mL polymer concen-
performed with 0.1 M NH4OH as a competitive tration was used throughout this study. The con-
inhibitor of RGD conjugation. The peptide concen- jugation efficiency was higher at lower
tration was lower for this reaction (0.4 mg/mL) temperatures (4 °C) and gradually decreased as
and, accordingly, only 0.7 peptides per polymer the temperature was increased to 50 °C [Fig.
3994 SMITH ET AL.

Figure 2. 1H NMR spectrum of RGD (A) and RGD reacted NiPAM/NASI (B) in D2O.
Note that arginine triplet 5 was detectable in the conjugate, but other RGD specific
peaks were overshadowed by the polymer peaks (except multiplet 4 that was partly
detectable). S indicates a solvent peak that was also present in the control polymer (i.e.,
polymer incubated with no RGD). Expanded solvent peak and triplet 5 was shown as a
boxed insert in B. NiPAM specific H (indicated by asterisk) was at ␦ ⬇ 3.9 ppm.
 CONJUGATION OF RGD PEPTIDES 3995

Figure 3. Effect of the concentration of polymer dissolved in solution (A), reaction

temperature (B), and pH (C) on the peptide conjugation efficiency. The peptide concen-
tration was 0.1 mg/mL. The polymer used for this study was NiPAM/NASI with NASI
mol % of 8.4 and the peptide was GYRGDS. Except for indicated temperatures in B and
pH values in C, all reactions were performed at 4 °C and pH of 8.0. The standard
deviations (SD) were shown in C but omitted in A and B for clarity. Where the SDs are
not visible, they are smaller than the plotted symbols.

3(B)]. At 37 and 50 °C, the polymer was insoluble versely affected.12,21 A 24-h reaction period was
(i.e., cloudy), remained suspended in a gelled used in all of the above reactions and increasing
state in solution but readily underwent de-gela- the incubation time any further did not influence
tion when temperature was reduced; however, the conjugation efficiency (not shown). On the
this did not appear to prevent the conjugation basis of the data in Figure 3, the grafting effi-
reaction. Unlike this study, a previous study from ciency was estimated to range between 1.5 pep-
our group observed no influence of the tempera- tides per polymer (lowest conjugation at 50 °C)
ture on protein conjugation between 4 and 37 and 8.8 peptides per polymer (highest conjugation
°C.12 However, the assessment method for the at pH ⫽ 8.0).
latter study was an electrophoresis-based (SDS- A critical issue in designing thermoreversible
PAGE) technique, which might not have been as polymers is the influence of polymer composition
robust as the RP-HPLC that was used for peptide on the peptide reactivity. Although one would
quantitation in this study. Reduced peptide con- desire to modify polymer composition and struc-
jugation at higher temperatures might be indica- ture to suit the needs of a particular application,
tive of increased hydrolysis rate over peptide ami- it is essential that polymer reactivity is not com-
nolysis. Although the relative hydrolysis and ami- promised as a result of alterations in polymer
nolysis rates were previously determined with structure. To determine the influence of polymer
proteins and with small amines (ethanolamine),10 properties on conjugation efficiency, GYRGDS
this could not be performed with the peptides in was reacted with NiPAM/NASI of varying NASI
this study because of a very slow reaction rate molar percentage (0 –7.2%) and NiPAM/NASI/
under spectroscopic conditions (not shown). MMA polymers of constant NASI molar percent-
The pH of the reaction medium also influenced age (1.1–1.3%) but varying MMA molar percent-
the conjugation efficiency, higher pH values giv- age (5.5–20.9%). The conjugation efficiency was
ing a higher rate of conjugation [Fig. 3(C)]. The directly proportional to the NASI molar percent-
optimal pH of ⬇8.0 observed for peptide conjuga- age of the polymers [Fig. 4(A)]. With a relatively
tions in this study was similar to the optimal pH constant NASI molar percentage, varying the
(7.4 – 8.1) observed for conjugation of proteins al- MMA molar percentage did not influence the con-
bumin, lactalbumin and IgG.10 Increasing the re- jugation efficiency [Fig. 4(B)]. The polymer with
action pH beyond this value is expected to in- the highest MMA molar percentage had an LCST
crease the hydrolysis rate of NASI to a level of 19.9 °C, indicating that lowering the LCST did
where the peptide conjugation rate will be ad- not appear to significantly influence peptide con-
3996 SMITH ET AL.

attach on the carbonyl of NASI ester,21 the pro-

tonated ⑀-NH2 was likely a bystander in the con-
jugation process.

Peptide Conjugation and Polymer LCST

The LCST of thermoreversible polymers repre-
sents the temperature at which hydrophobic in-
teractions among O(CH3)2 groups of NiPAM
causing insolubility in an aqueous environment is
balanced by H-bonding between water and
Figure 4. Effect of NASI molar percentage (A) and OHNO groups that retains a polymer in solution.
MMA molar percentage (B) on conjugation efficiency. The hydrated polymer collapses to a globular, hy-
The polymers in A were NiPAM/NASI polymers with drophobic state above LCST to form micellar
varying molar percentage of NASI, whereas polymers structures. The LCST of NiPAM polymers is de-
in B were NiPAM/NASI/MMA polymers where the pendent on presence of other groups in the poly-
NASI molar percentage was 1.1–1.4% and MMA molar mer backbone.15,26 Whereas charged groups, such
percentage varied between 5.6 and 20.9% (Table 1). as OCOO⫺ and ONH3⫹, increase the LCST, hy-
Whereas NASI molar percentage increased the peptide drophobic groups, such as OCH3 and OCH2CH3
conjugation, MMA molar percentage did not influence in MMA and EMA decrease the LCST. Conjugat-
the conjugation efficiency.
ing RGD-containing peptides to NiPAM polymers
will introduce at least one positively charged
group (because of arginine) and possibly other
jugation. This was similar to the conjugation re- charged and/or hydrophobic groups depending on
actions performed with the recombinant human
Bone Morphogenetic Protein-2 and the NiPAM/
NASI/AMA; with a range of alkylmethacrylates
(AMAs), ethyl-, butyl- and hexylmethacrylate, to
tailor the LCSTs between 12.9 and 24.7 °C, low-
ering the LCST did not result in an adverse effect
on protein conjugation.10 The results from this
study further indicate that one can tailor the
LCST of thermoreversible polymers without com-
promising their peptide conjugation efficiency.
The peptides RGD and GYRGDS were ex-
pected to undergo conjugation via N-terminal
ONH2 groups, because there were no other amine
groups present in the peptide. To determine if the
peptide conjugation could be enhanced by the
presence of an additional ONH2 group, a lysine-
containing peptide, GRGDSPK, was reacted with
NiPAM/NASI (8.4 mol %) and its conjugation ef-
ficiency was compared to that of GYRGDS. No
significant difference in conjugation efficiency
was observed between the two peptides when a
range of NiPAM/NASI concentrations were used
(Fig. 5). This was indicative of N-terminal ONH2 Figure 5. Conjugating efficiency (mean ⫾ SD) of
GYRGDS and GRGDSPK at various NiPAM/NASI (8.4
groups being primarily responsible for conjuga-
mol % NASI) concentrations. The conjugation efficiency
tion under the experimental conditions. This was
was increased with higher concentrations of polymers
likely due to the lower pKa of the N-terminal dissolved in solution, but there was no apparent differ-
ONH2 (8.0) as compared to that of ⑀-NH2 of lysine ence between the conjugation of the two peptides (the
(⬇11.0). The latter group was expected to be pro- difference at 1 mg/mL polymer was considered an
tonated at the reaction pH of 7.4. Because it is the anomaly). Where the SDs are not visible, they are
un-protonated ONH2 that exhibits nucleophilic smaller than the plotted symbols.
 CONJUGATION OF RGD PEPTIDES 3997

Figure 6. LCST of NiPAM/NASI (8.4 mol % NASI) polymer mixed with GYRGDS (A)
and the same polymer grafted with 8.5 and 0.9 peptides per polymer (B). Although
mixing the polymer (4 mg/mL) and peptide (0.1 mg/mL) did not change the LCST, a
slight but insignificant increase in LCST (⬇0.5 °C) was observed for the peptide grafted
polymers. The difference in the peptide graft ratio did not lead to a difference in the
LCST.

amino acids adjacent to the RGD sequence. The late] moieties typically require ⬎1 mol % to ele-
conjugated peptides should not alter the thermor- vate the LCST of random polymers.15 Incorpora-
eversible character of the chosen polymers. To tion of hydrophobic groups (based on ethyl-, bu-
determine whether peptide conjugation influ- tyl-, and hexylmethacrylates) typically required
ences LCST, GYRGDS was conjugated to NiPAM/ 3– 8 mol % to observe a 1–2 °C reduction in the
NASI (8.4 mol %) at two peptide concentrations LCST. Therefore, a higher peptide incorporation
(0.1 and 0.05 mg/mL) and conjugation efficiency was likely to be needed to alter the LCST of ther-
was determined by RP-HPLC: the conjugates had moreversible polymer. An independent group re-
8.5 and 0.9 peptides per polymer (57.7 and 12.3% ported conjugation of a longer RGD peptide, CG-
conjugation, respectively). As a control, LCST of GNGEPRGDTYRAY, to NiPAM gels;24 only a
the NiPAM/NASI polymer was determined along slight decrease in LCST (⬇1 °C) was observed in
with a polymer/GYRGDS mixture, which was not this study, a change we considered not significant
allowed any reaction time. The LCSTs for the for the use of this polymer in tissue engineering
polymer and its mixture with the peptide were applications. Unfortunately, there was no infor-
identical [27.6 °C, Fig. 6(A)], indicating that the mation available on the peptide grafting effi-
presence of the free peptide in a polymer solution ciency for the longer peptide conjugation, so it
did not influence the phase transition. There was was not possible to directly compare our results
only a slight increase in LCST after polymer con- with this report.
jugation [⬇0.5 °C, Fig. 6(B)], regardless of the
number of peptides conjugated per polymer.
Thermoreversible Polymers and Cell Attachment
Conjugation conditions described in this part
resulted in ⬇0.25 mol % peptide incorporation The ability of conjugated RGD peptides to impart
into the polymer, which was apparently suffi- cell adhesiveness to the thermoreversible poly-
ciently low not to influence the LCST. This value mers was investigated. The polymers were pre-
was indeed low when one considers that nega- pared as a film for cell culture and, accordingly,
tively charged (e.g., acrylic acid) and positively the conjugation reaction was performed on the
charged [e.g., 2-(dimethylamino)ethyl methacry- polymer films. Previous attempts to graft NiPAM
3998 SMITH ET AL.

polymers onto tissue culture polystyrene were in-

tended to induce cell harvesting by a temperature
change (instead of enzymatic discharge); a re-
duced temperature resulted in a hydrophilic sur-
face because of NiPAM hydration, which led to
cell detachment.7,27 In these studies, NiPAM
graft density was 1–7 ␮g/cm2 and, in one study,
increasing the NiPAM graft density beyond 3.0
␮g/cm2 inhibited hepatocyte cell attachment to
polystyrene surfaces.28 Our attempts at coating
tissue culture surfaces below 2.5 mg/cm2 resulted
in uneven attachment of C2C12 cells onto the
surfaces: part of the surface did not support cell
attachment at all, whereas other spots exhibited
dense cell attachment. We attributed this obser-
vation to cell contact with underlying polystyrene
surfaces; apparently, NiPAM coating densities
⬍ 2.5 mg/cm2 were not sufficient for complete
coating of the original tissue culture surface. To
ensure a lack of contact between the C2C12 cells
and the underlying tissue culture surfaces, we
used a coating density of 5 mg/cm2 so that only
the NiPAM surface was available for cell attach-
ment. With these conditions, conjugation of two
peptides, GRGDSP and GRGDSPK, to polymer Figure 7. Conjugating (mean ⫾ SD) of GRGDSPK
films were characterized (Fig. 7). The peptides (top) and GRGDSP (bottom) peptides to the NiPAM/
differed in only the terminal K residue, which EMA and NiPAM/NASI/EMA films. The conjugation
efficiency (nmol/cm2) to NiPAM/NASI/EMA films was
provided an additional ONH2 group in the case of
increased with increasing peptide concentration, but
K-containing peptide. Conjugation to NASI-con- there was no apparent difference between the two pep-
taining polymer was significantly higher for both tides. Some amount of peptide was lost from the incu-
peptides. Some conjugation of the peptides was bating solution for NiPAM/EMA polymers, but this was
observed for the polymer lacking NASI, but this is believed to reflect passive diffusion of the peptide into
believed to represent passive diffusion of the pep- the polymer films and not actual conjugation.
tide into the polymer film, rather than the actual
conjugation per se. Both peptides conjugated to On polymer NiPAM/EMA polymer treated with or
the polymer films to a similar extent within the without the peptide [Fig. 8(B,C)], the cells re-
concentration range tested, indicating that the mained rounded during culture and did not ex-
lysine ONH2 did not participate in the conjuga- hibit the typical spread morphology as seen with
tion reaction, as was the case previously (see Fig. tissue culture polystyrene. Small cellular aggre-
5). These results further confirm the peptide-re- gates were visible on NiPAM/EMA but this was
activity of the NASI-containing polymers when particularly the case with NiPAM/NASI polymers
the polymers were fabricated into a film. The treated with buffer alone [Fig. 8(D)]. These cells
surface peptide concentrations obtained in our remained as aggregates in long-term culture (up
hands (0.5–2.5 nmol/cm2) were comparable to the to 5 days, not shown). The NiPAM/NASI treated
concentrations previously demonstrated to be with GYRGDS, however, allowed cell attachment
suitable for cell attachment.16 and spreading reminiscent of tissue culture poly-
Finally, cell attachment to the polymer films styrene [Fig. 8(E)]. The extent of cell attachment
was microscopically inspected. For this, NiPAM/ was not as high as the tissue culture polystyrene
EMA and NiPAM/NASI (0 and 3.8 mol % NASI, (based on visual observation and MTT quantita-
respectively) were used to form polymer films, tion; to be reported separately), but the spread
which were exposed to phosphate buffer with and morphology of cells was indicative of the ability of
without GYRGDS (24 h at 37 °C). As a positive GYRGDS to support cell attachment.
control, the C2C12 cells attached and spread on Cells on NiPAM polymers were previously
tissue culture polystyrene as expected [Fig. 8(A)]. shown to attach to adsorbed fibronectin,29,30
 CONJUGATION OF RGD PEPTIDES 3999

Figure 8. Cells cultured on tissue culture polystyrene (A), NiPAM/EMA polymer

treated with phosphate buffer (B) or GYRGDS in phosphate buffer (C) and NiPAM/
NASI polymer treated with phosphate buffer (D) or GYRGDS in phosphate buffer (E).
Note that for nonadhesive surfaces (NiPAM/EMA polymer treated with and without the
peptide, and NiPAM/NASI polymer treated with buffer), cells remained rounded,
formed aggregates and exhibited growth in clumps. For adhesive surfaces (tissue
culture polystyrene and NiPAM/NASI polymer treated with the peptide), cells spread
on the surface and grew as a monolayer.

which is a large (340 kD) RGD-containing protein. CONCLUSIONS

The fact that immobilizing a small (0.65 kD)
GYRGDS induced cell adhesiveness was indica-
The results of this study indicated that NASI-
tive of this peptide to effectively substitute for
fibronectin in cell attachment to NiPAM, as was containing thermoreversible polymers could be
shown for other types of surfaces.12 In addition to used for effective conjugation of RGD-containing
the RGD peptides, the growth factor insulin was peptides. The peptide conjugation was optimal at
previously immobilized onto NiPAM surfaces in a pH of ⬇8 and increased with the NASI content
an independent lab to enhance cell growth.31 The of the polymers. The LCST of the NiPAM poly-
chosen cell model (STO fibroblasts) appeared to mers did not influence the peptide conjugation
exhibit accelerated cell division on insulin immo- efficiency. Up to ⬇8 peptides per polymer (⬇0.2
bilized surfaces, but it was not clear whether the mol %) did not alter the LCST behavior of the
observed biological effect was due to the stimula- parent thermoreversible polymer. The grafted
tory effect of insulin on the cells or whether insu- peptides were biologically active, as evident by
lin increased cell adhesion to NiPAM surfaces, increased cell attachment and spreading on Ni-
enabling better cell division and growth on these PAM surfaces conjugated with the peptide. The
surfaces.31 An independent group also reported approach described in this study should allow
incorporation of galactose to NiPAM-containing convenient preparation of peptide-grafted poly-
polymers to attach hepatocytes via the sialoglyco- mers and surfaces. The engineered polymers
protein receptors on cell surfaces.32 The grafted and/or surfaces will be suitable for a variety of
galactose was capable of inducing cell attachment tissue engineering applications, including scaf-
to the designed polymers, indicating ligands other folds for tissue reconstruction and biomaterials
than peptides might also provide alternatives for for protein delivery.
modifying thermoreversible polymers for cell at-
tachment. The use of NASI-containing pre-acti- The authors thank X.-J. Fan for the polymer synthesis
vated polymers will be suitable for conjugation of and the Canadian Health Institutes of Research
cell adhesive molecules other than the RGD pep- (CIHR) and Natural Sciences and Engineering Council
tides, as long as a suitable (ONH2) reactive group of Canada (NSERC) for their financial support. R. So-
is available in the ligand. mayaji was supported by a summer studentship from
4000 SMITH ET AL.

the Alberta Heritage Foundation for Medical Research 10. Uludag, H.; Norrie, B.; Kouisinioris, N.; Gao, T. J.
(AHFMR). Biotechnol Bioeng 2001, 73, 510 –521.
11. Gao, T. J.; Uludag, H. J Biomed Mater Res 2001,
57, 92–100.
12. Uludag, H.; Wong, M.; Man, J. J Appl Polym Sci
NOMENCLATURE 2000, 75, 583–592.
13. Gao T. J.; Kousinioris, N. A.; Wozney, J. M.; Winn,
A Alanine S.; Uludag, H. Tiss Eng 2002, 8, 429 – 440.
D Aspartic acid 14. Grzesik, W. J.; Robey, P. G. J. Bone Miner Res
EMA Ethylmethacrylate 1994, 9, 487– 496.
G Glycine 15. Uludag, H.; Fan, X.-D. In Drug Delivery in the 21st
HBSS Hank’s balanced salt solution Century; Park K.; Msryn R., Eds.; American Chem-
K Lysine ical Society: Washington, DC, 2000; Vol. 752, pp
253–262.
kd Kilodalton
16. Massia, S. P.; Hubbell, J. A. J Cell Biol 1991, 114,
LCST Lower solution critical temperature
1089 –1100.
MMA Methyl methacrylate 17. Feil, H.; Bae, Y. H.; Feijen, J.; Kim, S. W. Macro-
MW Molecular weight molecules 1993, 26, 2496 –2500.
NASI N-Acryloxysuccinimide 18. Wen, J.; Arakawa, T.; Philo, J. S. Anal Biochem
NiPAM N-Isopropylacrylamide 1996, 240, 155–166.
NMR Nuclear magnetic resonance 19. Yang, X.; Matsuura, H.; Fu, Y.; Sugiyama, T.;
P Proline Miura, N. FEBS Lett 2000, 470, 29 –34.
R Arginine 20. Cline, G. W.; Hanna, S. G. 1988. J Org Chem 1988,
RP-HPLC Reverse-phase high-pressure liquid 53, 3583–3586.
chromatography 21. Cline, G. W.; Hanna, S. B. J Am Chem Soc 1987,
109, 3087–3091.
S Serine
22. Chen, J. P.; Hoffman, A. S. Biomaterials 1990, 11,
V Valine
625– 630.
Y Tyrosine 23. Pollack, A.; Blumenfeld, H.; Wax, M.; Gaughn,
R. L.; Whitesides, G. M. J Am Chem Soc 1980, 102,
6324 – 6336.
REFERENCES AND NOTES 24. Stile, R. A.; Healy, K. E. Biomacromol 2001, 2,
185–194.
25. Tournier, E. J. M.; Wallach, J.; Blond, P. Ann Chim
1. Bromberg, L. E.; Ron, E. S. Adv Drug Delivery Rev Acta 1998, 361, 33– 44.
1998, 31, 197–221. 26. Kratz, K.; Hellweg, T.; Eimer, W. Colloids Surf A
2. Kwonand, G. S.; Okano, T. Pharm Res 1999, 16, 2000, 170, 137–149.
597– 600. 27. Kwon, O. H.; Kikuchi, A.; Yamato, M.; Sakurai, Y.;
3. Shah, S. S.; Wertheim, J.; Wang, C. T.; Pitt, C. G. J Okano, T. J Biomed Mater Res 2000, 50, 82– 89.
Controlled Release 1997, 45, 95–101. 28. Sakai, H.; Doi, Y.; Okano, N.; Yamada, N.; Sakurai,
4. Shimizu, S.; Yamazaki, M.; Kubota, S.; Ozasa, T.; Y. In Advanced Biomaterials in Biomedical Engi-
Moriya, H.; Kobayashi, K.; Mikami, M.; Mori, Y.; neering and Drug Delivery Systems; Ogata, N.;
Yamaguchi, S. Art Org 1996, 20, 1232–1237. Kim, S. W.; Feijen, J.; Okano, T., Eds.; Springer:
5. Stile, R. A.; Burghardt, W. R.; Healy, K. E. Macro- Tokyo, 1996; pp 229 –230.
molecules 1999, 32, 7370 –7379. 29. Kushida, A.; Yamato, M.; Konno, C.; Kikuchi, A.;
6. Collier, T. O.; Anderson, J. M.; Kikuchi, A.; Okano, Sakurai, Y.; Okano, T. J. Biomed Mater Res 1999,
T. J Biomed Mater Res 2001, 59, 136 –143. 45, 355–362.
7. von Recum, H. A.; Kim, S. W.; Kikuchi, A.; Oku- 30. Chen, G.; Imanishi, Y.; Ito, Y. J Biomed Mater Res
hara, M.; Sakurai, Y.; Okano, T. J Biomed Mater 1998, 42, 38 – 44.
Res 1998, 40, 631– 639. 31. Chen, G.; Ito, Y.; Imanishi, Y. Biotech Bioeng 1997,
8. Winnik, F. M.; Davisson, A. R.; Hamer, G. K.; 53, 339 –344.
Kitano, H. Macromolecules 1992, 25, 1876 –1880. 32. Goto, M.; Makino, Y.; Kobayashi, K.; Cho, C. S.;
9. Ringsdorf, H.; Simon, J.; Winnik, F. M. Macromol- Akaike, T. J Biomater Sci Polym Ed 2001, 12,
ecules 1992, 25, 7306 –7312. 755–768.