Comparison of peripheral myelin protein 22 (PMP22) expression levels between controls and patients with the leucine 7 frameshift (Leu7fs) mutation with the use of immunoelectron microscopy of skin biopsy specimens. Immunoelectron microscopy studies were performed in skin biopsies from healthy controls and patients with the Leu7fs mutation. The PMP22 was labeled with antibodies conjugated with gold particles. These particles were quantified with Image-Pro Plus software (Media Cybernetics; Bethesda, Maryland) as previously described.11 Mean PMP22 levels (horizontal bars) were decreased in patients with the Leu7fs mutation (161.6 particles/μm2 [SD, 4.5 particles/μm2]; 18 myelinated dermal nerve fibers) compared with control patients (211.4 particles/μm2 [SD, 5.5 particles/μm2]; 17 myelinated dermal nerve fibers).
Changes of compound motor action potential (CMAP) amplitudes by individual patient age.
Li J, Ghandour K, Radovanovic D, Shy RR, Krajewski KM, Shy ME, Nicholson GA. Stoichiometric Alteration of PMP22 Protein Determines the Phenotype of Hereditary Neuropathy With Liability to Pressure Palsies. Arch Neurol. 2007;64(7):974-978. doi:10.1001/archneur.64.7.974
Copyright 2007 American Medical Association. All Rights Reserved. Applicable FARS/DFARS Restrictions Apply to Government Use.2007
Hereditary neuropathy with liability to pressure palsies (HNPP) is caused by a 1.4-megabase deletion at chromosome 17p11.2, which bears the PMP22 gene and other genes. However, whether other genes besides PMP22 contribute to the phenotype is unknown. Whether any mutation within the coding region of the PMP22 gene ultimately causes HNPP by reducing the amount of peripheral myelin protein 22 (PMP22) expressed in myelin is also unknown.
To determine whether affected patients develop a phenotype identical to that found in HNPP and whether the leucine 7 frameshift (Leu7fs) mutation reduces PMP22 levels in myelin.
We evaluated affected family members by neurological examination, electrophysiology, and skin biopsies. We identified a large family with a Leu7fs mutation of PMP22 (11 affected members across 3 generations) that predicts truncation of the protein prematurely and eliminates PMP22 expression from the mutant allele.
We found that PMP22 levels were reduced in peripheral nerve myelin in dermal skin biopsies in patients with an Leu7fs mutation. Through clinical and electrophysiological evaluation, we also found that patients with the Leu7fs mutation were indistinguishable from patients with HNPP caused by deletion. We also found that a length-dependent axonal loss became pronounced in elderly patients with Leu7fs mutations, similar to what has been described in heterozygous knockout mice (pmp22 +/−).
Taken together, these results confirm that the phenotypic expression is identical in patients with Leu7fs mutation and patients with HNPP caused by chromosome 17p11.2 deletion. They also demonstrate that reduction of PMP22 is sufficient to cause the full HNPP phenotype.
Peripheral myelin protein 22 (PMP22) is a glycoprotein expressed by Schwann cells and is localized to compact myelin in the peripheral nervous system.1 Newly synthesized PMP22 is subject to vigorous quality control that results in the degradation of more than 90% of PMP22 in the endoplasmic reticulum, with only a small fraction reaching the Schwann cell membrane.2
The level of PMP22 appears critical to the normal function of peripheral nerve myelin, because alterations of PMP22 gene dosage result in 2 allelic disorders: hereditary neuropathy with liability to pressure palsies (HNPP) and Charcot-Marie-Tooth disease type 1A (CMT1A). Hereditary neuropathy with liability to pressure palsies and CMT1A are caused by deletion or duplication of chromosome 17p11.2, respectively.3- 5 Several point mutations in PMP22 have been reported to cause neuropathies that clinically resemble HNPP and CMT1A.6,7 However, particularly for HNPP, detailed clinical and electrophysiological analysis from most of these mutations has not yet been performed. One thoroughly evaluated family with a frameshift mutation developed certain features similar to HNPP, but markedly slowed nerve conductions and abnormal physical findings8 were atypical of those with HNPP caused by chromosome 17p11.2 deletion.9
Whether any mutation (missense, frameshift, or truncation) within the coding region of the PMP22 gene ultimately causes HNPP by reducing the amount of PMP22 expressed in myelin is unknown. It is also unknown whether other genes or proteins within the 17p11.2 domain contribute to the phenotype of HNPP.
We have evaluated a family with a Leucine 7 frameshift (Leu7fs) mutation of PMP22, which permits us to address both of these issues. The frameshift results in an abnormal amino acid sequence between codon 6 and codon 36, at which point there is a premature termination of the protein; there should be at most 7 normal amino acids expressed from the mutant allele. The mutation has been previously shown to cause a mild neuropathy that is suggestive of an HNPP phenotype.6 To determine whether PMP22 levels are reduced as theoretically predicted from the truncation, we measured and demonstrated reduced PMP22 levels in myelinated dermal nerves of Leu7fs patients. We also used nerve conduction studies to demonstrate an identical pattern of focal abnormalities to those observed in HNPP patients with chromosome 17p11.2 deletion.9,10 Taken together, these results demonstrate that the Leu7fs mutation results in reduced PMP22 levels and that reduction of PMP22 levels is sufficient to produce the full HNPP phenotype. Our results also demonstrated that older patients with the Leu7fs mutation developed a length-dependent axonal neuropathy similar to that found in our older HNPP deletion patients.
Eleven patients from a large family in Sydney, Australia, with the Leu7fs mutation of PMP22 were examined by 2 of us (J.L. and G.A.N.). Their DNA analysis was performed at the Molecular Genetics and Medicine Laboratory at the University of Sydney. The methodology and the results of these DNA studies have been previously published.6 Direct sequence analysis of exon 1 showed a 2–base pair deletion at nucleotide 207, predicting a short peptide largely truncated after amino acid number 7. All patients gave informed consent to participate in the study, which was approved by the ethics committees at Wayne State University and University of Sydney. Results from these patients were compared with those from our published series of HNPP patients with deletion at chromosome 17p11.2.9
Skin biopsies were performed as previously described.11 Briefly, skin from the lateral aspect of the index finger was biopsied. Tissues were fixed in a glutaraldehyde, 0.5%, and paraformaldehyde, 4%, fixative, then embedded in LR white, and sectioned and incubated with antibodies. The primary antibody was anti-PMP22, a rabbit monoclonal antibody (Lab Vision, Fremont, California; 1:200 concentration). The images were imported into Image-Pro Plus software (Media Cybernetics; Bethesda, Maryland). Myelin regions were divided into small squares by the software. Gold particles were counted in any square that had intact myelin. The density of PMP22 was derived as summated particles divided by summated areas of the squares.
The skin biopsies were performed on 2 patients with the Leu7fs mutation and 2 controls who were age and sex matched. The average density of PMP22 from each myelinated dermal nerve fiber served as 1 data point. Seventeen myelinated nerve fibers from patients or controls were analyzed and statistically compared using the t test. A P value < .05 was considered significant.
Motor and sensory nerve conduction studies were performed using conventional methods.10 The distal stimulation distances for motor conduction studies were 7 cm in the arms and 9 cm in the legs.
To determine whether the Leu7fs mutation reduced PMP22 expression, we measured PMP22 levels in myelinated fibers from dermal nerves using quantitative immunoelectron microscopy techniques (Figure 1). The mean PMP22 density from patients with the Leu7fs mutation was 161.6 particles/μm2 (SD, 4.5 particles/μm2; 18 myelinated dermal nerve fibers) compared with control values of 211.4 particles/μm2 (5.5 particles/μm2;17 myelinated dermal nerve fibers). The differences were statistically significant at P < .001. Thus, these data demonstrated a reduction of PMP22 in myelinated fibers of patients with the Leu7fs mutation.
Hereditary neuropathy with liability to pressure palsies has a characteristic phenotype that is distinguished by transient episodes of weakness and sensory loss.9 Characteristic findings on nerve conduction studies are illustrated by prolonged median and peroneal distal motor latencies, and a slowing of ulnar conduction around the elbow and peroneal conduction around the knee.10 Although Leu7fs mutation patients have been previously reported to have mild neuropathies,6 details of the clinical or physiological features were not included. Therefore, we evaluated the clinical and electrophysiological features of 11 affected family members bearing the Leu7fs mutation.
Eleven cases were studied (8 men and 3 women). The mean age of symptomatic onset was 15 years (SD, 7 years) and the mean age at evaluation for this study was 51 years (SD, 27 years). All patients had experienced transient episodes of focal weakness or sensory loss (Table 1). No patient presented with symptoms of symmetrical length-dependent polyneuropathy. Common symptoms (of 38 clearly described events) were focal sensory loss (20), focal weakness (18), or combined focal motor and sensory disturbances (7). The focal motor symptoms were predominantly confined to territories innervated by a single nerve or brachial plexus. However, sensory symptoms were less frequently confined to the distribution of single nerves, though they were focal. Focal symptoms were often triggered by compression, stretching, or repetitive use of the affected extremity. The duration of the focal symptoms varied from minutes to months (Table 1).
The neurological examination in patients with the Leu7fs mutation revealed only mild abnormalities in most patients (9 of 11), similar to those in patients with HNPP caused by chromosome 17p11.2 deletion.9 Most patients (9 of 11) had mild sensory abnormalities symmetrically distributed in their feet. This was particularly pronounced in older patients. The remaining 2 patients had normal neurological examination results. Deep tendon reflexes were abnormal in half of our patients. Foot deformities were seen in only 1 patient.
Seven patients had scapular winging. Interestingly, all were on the right side. Most of the 7 patients did not have a history of arm weakness. We studied 3 of them with needle electromyography, with normal results. Thus, brachial plexopathy is unlikely to be the explanation for their scapular winging. Findings from the neurological examination are summarized in Table 2.
Patients with the Leu7fs mutation had a similar electrophysiological pattern to that in patients with HNPP caused by deletion.10 As in deletion patients, accentuated motor distal slowing occurred at sites subject to compression, such as the median nerves at the wrist (a mean of 118% of the upper limit of normal) and the peroneal nerves at the ankle (a mean of 110% of the upper limit of normal). Alternatively, the distal motor latency was minimally prolonged or normal in ulnar nerves (a mean of 106% of the upper limit of normal; median vs ulnar, P<.01) and tibial nerves (a mean of 86% of the upper limit of normal; peroneal vs tibial, P<.05). Focal slowing was frequent across the elbow for the ulnar nerve (12 of 12 patients) and around the fibular head for the peroneal nerve (5 of 12 patients [42%]), as is the case with patients with the chromosome 17p11.2 deletion.9,10 In contrast, mean conduction velocities at segments not liable to compression were well preserved (median, 51 m/s [SD, 7.9 m/s], normal velocity > 48 m/s; ulnar, 56 m/s [SD, 8.3], normal velocity > 49 m/s; peroneal, 43.1 m/s [SD, 5.5] m/s, normal velocity > 41m/s; tibial, 40 m/s [SD, 6.2 m/s], normal > 41 m/s).
In addition to the clinical and physiological findings described, we noted that all 3 Leu7fs patients older than 65 years had markedly reduced compound muscle action potential (CMAP) amplitudes in both lower and upper extremities. Heterozygous knockout mice (pmp22 +/−), an animal model of HNPP, have been shown to develop axonal loss when they become older.12 We therefore asked whether age-dependent axonal loss was occurring in Leu7fs patients. We examined the CMAP amplitudes in both upper and lower extremities with the ages of the patients (Figure 2). We found that CMAP amplitudes did decline with age in each of the 4 nerves analyzed but that values did not become markedly abnormal until patients were older than 65 years. Consistent with these findings, the 3 oldest patients had distal symmetrical abnormalities on their examinations, such as marked ankle dorsiflexion weakness in 2 patients (a score of 2 of 5 on the Medical Research Council scale), conspicuous calf atrophy of both legs in 2 patients, and unsteady gait in all 3 patients. All 3 elderly patients also had sensory deficits in both hands. Recently we have evaluated an 83-year-old male patient with HNPP caused by the 17p11.2 deletion, the oldest patient we have seen with this deletion. He also had severe axonal loss with peroneal CMAP amplitudes unobtainable at the extensor digitorium brevis, median CMAP amplitudes of 3.9 mV, and ulnar CMAP amplitudes of 5.2 mV. Sural, median, and ulnar sensory nerve action potential amplitudes were unobtainable. Taken together, these results suggest that length-dependent axonal loss may occur in elderly patients with HNPP, similar to what occurs in the pmp22 +/− mice.
This study demonstrates that the Leu7fs mutation reduces PMP22 levels in peripheral nervous system myelin and causes a clinical phenotype that is indistinguishable from HNPP caused by the 17p11.2 deletion. Like typical HNPP patients, Leu7fs patients all present with transient episodes of weakness or sensory loss. Like typical HNPP patients, Leu7fs patients also have prolongations of median and peroneal distal motor latencies, and focal ulnar motor nerve conduction velocity slowing around the elbow and focal peroneal motor NCV slowing around the fibular head.
These findings are significant because, first, they clearly demonstrate that a reduction in PMP22 is sufficient to reproduce all the findings of HNPP deletion. Other genes, in addition to PMP22, are contained within the 17p11.2 segment deleted in patients with HNPP,13 and it has been hypothesized that these genes may serve as modifiers for the HNPP (or CMT1A) phenotype. Our results demonstrate that reduced levels of PMP22 are sufficient to produce the entire HNPP phenotype, independently of effects from any other genes.
Second, the results raise the question of why there was only a 23.6% reduction rather than the predicted 50% reduction of PMP22 in myelinated nerves of patients with HNPP.11,14 One possible explanation is that quantitative immunoelectron microscopy in dermal nerves is not sensitive enough to distinguish between a 50% and 23.6% reduction. Further studies are required to answer this question. Alternatively, however, it is possible that the Schwann cell could have reduced the amount of PMP22 degraded by the endoplasmic reticulum or up-regulated PMP22 transcription/translation to compensate for loss of the Leu7fs allele. Pareek and colleagues2 have demonstrated that most of translated PMP22 is degraded by the endoplasmic reticulum rather than being transported to the myelin sheath. Conceivably, less wild-type PMP22 is degraded when there is a loss-of-function mutation, such as the Leu7fs mutation.
Third, results from our analysis of this large family provide insight into the natural history of HNPP, particularly with respect to the question of axonal loss. Heterozygous knockout mice (pmp22 +/−) develop a slowly progressive axonal loss.12 However, it is unclear whether a progressive axonal loss also occurs in humans, in whom neurophysiological changes have primarily been those suggesting focal slowing.10 Our study demonstrates a marked reduction of CMAP amplitudes in the older Leu7fs patients and a progressive decline of CMAP amplitudes with aging (Figure 2). Taken together, these results suggest a slow progressive axonal loss in patients with HNPP, independent of demyelination, that may be independent of their transient symptoms of weakness or sensory loss.
Finally, direct measurement of PMP22 expression on compact myelin in this study was successfully achieved by human skin biopsies, a minimally invasive technique that we have established recently.11 This technique has revealed the overexpression and reduction of PMP22 in patients with CMT1A and HNPP, disorders caused by the duplication and deletion, respectively, of the 17p11.2 chromosome containing the PMP22 gene.11 We have also used this technique lately to verify an intronic mutation of myelin protein zero gene that predicted alteration in this gene splicing. By using reverse transcription–polymerase chain reaction in skin biopsies, we have successfully confirmed the predicted skipping of exon 4 and downstream frameshift of the mutant protein zero gene in the patients harboring this mutation.15 Our study once again demonstrates the usefulness and effectiveness of human skin biopsy in probing molecular mechanism of inherited neuropathies.
Correspondence: Jun Li, MD, PhD, Department of Neurology, Wayne State University, 4201 St Antoine St, UHC-8D, Detroit, MI 48201 (firstname.lastname@example.org).
Accepted for Publication: June 15, 2006.
Author Contributions:Study concept and design: Li and M. E. Shy. Acquisition of data: Li, Ghandour, Radovanovic, R. R. Shy, Krajewski, M. E. Shy, and Nicholson. Analysis and interpretation of data: Li, Ghandour, Radovanovic, and M. E. Shy. Drafting of the manuscript: Li, Ghandour, and M. E. Shy. Critical revision of the manuscript for important intellectual content: Li, Radovanovic, R. R. Shy, Krajewski, M. E. Shy, and Nicholson. Statistical analysis: Ghandour. Obtained funding: Li and M. E. Shy. Administrative, technical, and material support: Radovanovic, R. R. Shy, Krajewski, and M. E. Shy. Study supervision: Li and Nicholson.
Financial Disclosure: None reported.
Funding/Support: This study is supported by grants K08 NS048204 from the National Institute of Health and MDA 3742 from the Muscular Dystrophy Association (Dr Li).
This article was corrected for error in byline on 8/3/2007, prior to publication of the correction in print.