EDITORIAL

The Death Panel for

Charcot-Marie-Tooth Panels

C harcot-Marie-Tooth (CMT) disease is a category of

hereditary neuropathy. Rather than 1 disease, CMT
is a syndrome of several clinically and genetically distinct
ing. In the United States, this has led to the use of
large genetic panels that can be ordered through com-
mercial laboratories. Unfortunately, these panels are
disorders (Table and for updated listings: http//www. often misused and lead to unnecessary and expensive
molgen.ua.ac.be/CMTMutations/Mutations/MutByGene. testing. Guidelines have been published to help direct
cfm).1 The various subtypes of CMT have been tradi- which genetic tests should be ordered based on inheri-
tance pattern and MNCVs.2 Nevertheless, many
tionally classified according to the nerve conduction
patients still have complete panels performed, includ-
velocities and predominant pathology (eg, demyelination
ing all CMT1-related genes in obvious CMT2
or axonal degeneration), inheritance pattern (autosomal
patients and vice versa. This problem is especially
dominant, recessive, or X-linked), and the specific seen in the United States, but in many European and
mutated genes, of which there have been >35 identified. other countries, where individual gene diagnostic tests
Type 1 CMT or CMT1 refers to inherited demyelinating are available in noncommercial laboratories, it is not
sensorimotor neuropathies, whereas the axonal sensori- uncommon for inappropriate tests to be ordered espe-
motor neuropathies are classified as CMT2. By defini- cially testing for the chromosome 17 duplication in
tion, motor nerve conduction velocities (MNCVs) in the CMT2 patients.
arms are slowed to <38m/s in CMT1 and are >38m/s In this issue of Annals of Neurology, Dr Shy’s
in CMT2. However, most cases of CMT1 are associated group report the results of genetic testing on their
with MNCVs that are much slower. CMT1 and CMT2 very large cohort of presumed CMT patients (787
usually begin in childhood or early adult life; however, patients, of whom 527 were genotyped) and propose a
onset later in life can occur, particularly in CMT2. Both series of algorithms for genetic testing based on age of
onset, inheritance pattern, and MNCVs.3 Although
are most commonly associated with autosomal dominant
there has been a recently published Practice Parameters
(AD) inheritance, although X-linked inheritance and suggesting a similar approach,2 this article takes this a
autosomal recessive (AR) inheritance are also seen. The step further, offering a more refined approach backed
traditional classification of CMT into types 1 and 2 is by the results of their huge cohort of patients. One of
widely used, but additional subtypes of CMT3 and the most important findings in this paper, especially in
CMT4 are also used to varying degrees. CMT3 refers to view of the expense involved in extensive CMT gene
HMSNIII (hereditary motor and sensory neuropathy testing, is the finding that almost 92% of the 527 ge-
type III) in older classifications, and is characterized as a netically defined CMT patients in their cohort had
severe demyelinating or hypomyelinating neuropathy, mutations in 1 of only 4 genes (PMP22, MPZ, GJB1,
with most patients having de novo dominant mutations and MFN2). This finding is the basis of the useful
in the common causative genes for AD CMT1 (PMP22, algorithms they have devised. The authors propose ini-
tially testing only for CMT1A caused by PMP-22
MPZ, and EGR2). CMT4 is an AR demyelinating
duplications in patients with a classical CMT pheno-
polyneuropathy that typically begins in childhood or
type with slow MNCV (15 < and
35m/s). Only if
early adult life, although some classifications refer to this is negative would they next screen for GJB1 muta-
these subtypes as AR CMT1 and the AR forms of tions (CMT1X) or MPZ (CMT1B), if there is clear
CMT2 as AR CMT2. Intermediate forms of CMT with male to male transmission. If these are negative, they
median or ulnar MNCVs between 25 and 45m/s are also suggest screening for point mutations in PMP22, SIM-
present (see Table). PLE, and EGR2. In patients with CMT and severely
Given the large number of genes that can poten- slow MNCVs (
15m/s), they recommend screening
tially cause CMT, the diagnostic approach to patients for both the PMP22 duplication and MPZ mutations
with a possible hereditary neuropathy can be daunt- in those patients who begin to walk after 15 months

V
C 2011 American Neurological Association 1
ANNALS of Neurology

TABLE : Classification of Charcot-Marie-Tooth Disease

Name Inheritance Gene Location Gene
CMT1
CMT1A AD 17p11.2 PMP22 duplication
CMT1B AD 1q21-23 MPZ
CMT1C AD 16p13.1-p12.3 LITAF
CMT1D AD 10q21.1-22.1 ERG2
CMT1E AD 17p11.2 Point mutations in
PMP22 gene
CMT1F AD 8p13-21 NEFL
CMT1X X-linked dominant Xq13 GJB1
HNPP AD 17p11.2 PMP22
CMT2
CMT2A2 (allelic to HMSN AD 1p36.2 MFN2
VI with optic atrophy)
CMT2B AD 3q13-q22 RAB7
CMT2B1 (allelic to LGMD 1B) AR 1q21.2 LMNA
CMT2B2 AD 19q13 Unknown
CMT2C (with vocal cord and AD 12q23-24 TRPV4
diaphragm paralysis)
CMT2D (allelic to distal SMA5) AD 7p14 GARS
CMT2E (allelic to CMT 1F) AD 8p21 NEFL
CMT2F AD 7q11-q21 HSPB1
CMT2G (may be allelic to CMT4H) AD 12q12-q13 Unknown (may be FGD4)
CMT2H AD 8q21.3 Unknown (may be GDAP1)
CMT2I (allelic to CMT1B) AD 1q22 MPZ
CMT2J (allelic to CMT1B) AD 1q22 MPZ
CMT2K (allelic to CMT4A) AD 8q13-q21 GDAP1
CMT2L (allelic to distal hereditary AD 12q24 HSPB8
motor neuropathy type 2)
CMT2M AD 16q22 AARS
CMT2X X-linked Xq22-24 PRPS1
CMT3 (Dejerine-Sottas disease, AD 17p11.2 PMP22
congenital hypomyelinating neuropathy)
AD 1q21-23 MPZ
AR 10q21.1-22.1 ERG2
AR 19q13 PRX
CMT4 (AR CMT1)
CMT4A AR 8q13-21.1 GDAP1
CMT4B1 AR 11q23 MTMR2
CMT4B2 AR 11p15 MTMR13

2 Volume 69, No. 1

 Amato and Reilly: Death Panel for CMT Panels

TABLE : (Continued)

Name Inheritance Gene Location Gene

CMT4C AR 5q23-33 SH3TC2
CMT4D (HMSN-Lom) AR 8q24 NDRG1
CMT4E (congenital AR — Probably includes PMP22,
hypomyelinating neuropathy) MPZ, and ERG2
CMT4F AR 19q13.1-13.3 PRX
CMT4G AR 10q22.2 HK1
CMT4H AR 12q12-q13 FGD4
CMT4J SR 6q21 FIG4
DI-CMT
DI-CMTA AD 10q24.1-25.1 Unknown
DI-CMTB AD 19p12-p13.2 DNM2
DI-CMTC AD 1p35 YARS
Modified from Amato AA, Russell J. Neuromuscular disease. New York, NY: McGraw-Hill, 2008.
CMT ¼ Charcot-Marie-Tooth; AD ¼ autosomal dominant; PMP22 ¼ peripheral myelin protein-22; MPZ ¼ myelin protein zero
protein; LITAF ¼ lipopolysaccharide-induced tumor necrosis factor-alpha factor; ERG2 ¼ early growth response-2 protein; NEFL
¼ neurofilament light chain; GJB1 ¼ gap junction associated protein B1; HNNP ¼ hereditary neuropathy with liability to pres-
sure palsies; HMSN ¼ hereditary motor and sensory neuropathy; MFN2 ¼ mitochondrial fusion protein mitofusin 2 gene;
LGMD ¼ limb girdle muscular dystrophy; AR ¼ autosomal recessive; TRPV4 ¼ transient receptor potential cation channel, sub-
family V, member 4; SMA ¼ spinal muscular atrophy; GARS ¼ glycl-tRNA synthetase; HSPB1 ¼ small heat shock protein B1;
FGD4 ¼ FGD1-related F actin binding protein; GDAP1 ¼ ganglioside-induced differentiation-associated protein-1; HSPB8 ¼
small heat shock protein B8; AARS ¼ alanyl-tRNA synthetase; PRPS1 ¼ phosphoribosyl pyrophosphate synthetase 1; MTMR2 ¼
myotubularin-related protein-2; MTMR13 ¼ myotubularin-related protein-13; SH3TC2 ¼ SH3 domain and tetraticopeptide
repeats 2; NDRG1 ¼ N-myc downstream regulated 1; HK1 ¼ hexokinase 1; FIG4 ¼ SAC domain-containing inositol phosphates
3; DI-CMT ¼ dominant intermediate CMT; DNM2 ¼ dynamin 2; YARS ¼ tyrosyl-tRNA synthetase.

of age, but only for the PMP22 duplication in those transmission in the pedigree, in which case only MPZ
that walk before 15 months of age. If there is no screening is necessary.
PMP22 duplication or MPZ mutation, they suggest The next generation of sequencing techniques
sequencing PMP22. including the use of gene chips, exome sequencing, and
Patients with CMT and intermediate MNCV 35 whole genome sequencing, may serve as cheaper alterna-
< and
45m/s) usually have CMT1X or CMT1B. For tives to more efficiently screen for mutations in terms
patients with no male to male transmission, intermedi- of cost and time in the future. These techniques are
ate MNCVs, and a classical phenotype, the authors rec- not widely available at present and would be too expen-
ommend first screening for GJB1 mutations.3 If this sive for routine use, so the algorithms proposed in this
testing is negative, testing should proceed to MPZ paper are an excellent guide to rationale genetic testing
mutations. Alternatively, if there is male to male trans- of CMT patients currently. In the future, however,
mission, patients should be first screened for CMT1B. these newer techniques may be useful in identifying
As no patients with CMT1A had intermediate novel CMT-associated genes, particularly in CMT2, in
MNCVs, testing for a PMP22 duplication would not which only about 30% of cases can be genotyped at
be warranted. If testing for MPZ and GJB1 is negative, present.
then patients could be screened for mutations in rare A hotly debated area is whether and when patients
genes associated with dominant intermediate forms of should have genetic testing. The authors touched on this,
CMT including DNM2 (DI-CMTB) and YARS (DI- noting that testing can aid in prognosis and genetic
CMTC). In patients with severe CMT2 in childhood counseling.4,5 Further arguments for seeking a genetic di-
(normal or only mildly slow MNCVs, if obtainable), agnosis include the avoidance of unnecessary invasive
screening should begin with mutations in MFN2, the tests, such as nerve biopsies, and in rare cases avoiding
cause of CMT2A. If this is negative, testing for MPZ unnecessary trials of immunotherapy (eg, when there is
and GJB1 would be reasonable, as it would be initially diagnostic consideration of chronic inflammatory demye-
for late onset patients, unless there was male to male linating polyneuropathy). However, one could play devil’s

January 2011 3
ANNALS of Neurology

advocate and suggest that such testing from a clinical

and pragmatic point of view often is not going to change Potential Conflicts of Interest
management or help in the prognosis of an individual
patient. Unfortunately, there are no specific treatments Nothing to report.
available for any of the subtypes of CMT. In terms of
prognosis, there have been natural history studies done
on some CMT subtypes with the primary aim of assess- Anthony A. Amato, MD
ing outcome measures that may be useful in future clini- Department of Neurology
Brigham and Women’s Hospital and Harvard Medical School
cal trials.4,5 But what are you going to tell a patient if
Boston, MA
you genetically confirm they have CMT1A—that their
impairment will increases by an average of 0.686 points/ Mary M. Reilly, MD
yr on the CMT Neuropathy Score (CMTNS) and by an
Department of Molecular Neurosciences
average of 1.368 points/yr on the Neuropathy Impair- MRC Centre for Neuromuscular Diseases
ment Score4 or, in the case of CMT1X, that the UCL Institute of Neurology
CMTNS increased an average of 2.89 points/decade?5 London, United Kingdom
What do these numbers mean for an individual patient
with CMT? Further, should we test for very rare muta-
tions that occur in 1% or less of CMT patients? All these References
issues need to be discussed with patients and their fami- 1. Amato AA, Russell J. Neuromuscular disease. New York, NY:
lies prior to ordering any genetic testing. Nevertheless, a McGraw-Hill, 2008.

strong argument for genetic testing and carefully follow- 2. England JD, Gronseth GS, Franklin G, et al. Evaluation of distal
ing up phenotyped patients is to learn more about the symmetric polyneuropathy: the role of laboratory and genetic test-
ing (an evidence-based review). Muscle Nerve 2009;39:116–125.
natural history of all types of CMT, including the rare
types, so that we can give more individualized prognoses 3. Saporta ASD, Sottile SL, Miller LJ, et al. Charcot-Marie-Tooth sub-
types and genetic testing strategies. Ann Neurol 2011;69:22–33.
in the future. We do not disagree with the authors’
approach; these are just issues that we struggle with on a 4. Shy ME, Chen L, Swan ER, et al. Neuropathy progression in Char-
cot-Marie-Tooth disease type 1A. Neurology 2008;70:378–383.
day to day basis, particularly given the costs of these
studies. We fear it will only be more of a struggle in the 5. Shy ME, Siskind C, Swan ER, et al. CMT1X phenotypes represent
loss of GJB1 gene function. Neurology 2007;68:849–855.
future, but adopting a rational approach to genetic diag-
nosis as outlined in Dr Shy’s paper should help. DOI: 10.1002/ana.22272

4 Volume 69, No. 1