PCR-based strategy for the diagnosis of hereditary neuropathy with liability to pressure palsies and Charcot-Marie-Tooth disease type 1A

The principal findings in Charcot-Marie-Tooth disease type 1 (CMT1) are distal symmetric muscle weakness, hypo- to areflexia, and bilateral pes cavus.^1,2 Severely slowed conduction velocities of peripheral nerves have often been reported.¹ Nerve biopsy specimens show marked demyelination and remyelinization of peripheral nerves and the appearance of irregular myelin formations. CMT1 can be subdivided into CMT1A, which is linked to chromosome 17,³ and CMT1B, which is linked to chromosome 1.⁴ Most CMT1A cases show a 1.5-megabase (Mb) duplication on chromosome 17p11.2.^5,6 The peripheral myelin protein 22 gene (PMP22) is located⁷ within the duplicated segment. Mutations within the PMP22 gene have also been identified as the cause of CMT1A.^8,9

In hereditary neuropathy with liability to pressure palsies (HNPP), recurrent peripheral nerve palsies (e.g., ulnar nerve, median nerve, and peroneal nerve palsies) occur because of minor compression trauma, whereas foot deformity is less frequently observed than it is in CMT1.¹⁰ Nerve conduction velocities are significantly reduced at compression sites of peripheral nerves. Morphologically, typical sausage-like formations (tomacula) of peripheral myelin can be found in teased-fiber preparations of peripheral nerves of affected individuals.¹¹ In most HNPP patients the same region that is duplicated in CMT1A is deleted.¹² Also, there have been reports that point mutations in PMP22 can cause HNPP.^10,13,14 A gene dosage effect has been proposed^15,16 as the underlying pathomechanism for the development of CMT1A and HNPP.

For molecular genetic testing Southern blotting, pulsed field gel electrophoresis, and fluorescent-in-situ-hybridization have been used.^17,18 In segregation studies, polymorphic markers (e.g., pVAW409R3a and pEW401HE) have been used that are located within the duplicated/deleted region or map to the junctional region(CMT1A-REP) of the duplication/deletion outside the PMP22 locus.^19,20

We developed a rapid PCR-based strategy for the determination of PMP22 gene dose. Using this strategy, the CMT1A duplication as well as the HNPP deletion can be analyzed within 1 day.

Methods. Four HNPP families were analyzed (figure 1). All patients gave their informed consent for genetic analysis. Additionally quantitative PCR results from HNPP patients and CMT1A patients were compared with the results of Southern blot analysis. Fifteen samples from individuals with HNPP, 15 samples from CMT1A patients, and 20 control samples were analyzed. Southern blot analysis was done by the Antwerp group, and quantitative PCR analysis was done by the Münster group in a blinded fashion.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1. HNPP pedigrees in which haloptype analysis and semiquantitative PCR were performed. Filled symbols represent affected individuals. Letters indicate haplotypes as previously described.²¹

Haplotype analysis. Previously identified genetic markers were used for haplotype analysis at the PMP22 gene locus.²¹ Briefly, two DNA fragments (287-bp and 497-bp long) of the PMP22 3′-UTR were PCR-amplified from individual DNA samples. PCR amplification was done as described previously.²¹ Sequencing of the immobilized template was performed with one of the fluorescein-linked primers following the instructions of the AutoRead T7 Sequencing Kit (Pharmacia Biotechnology, Freiburg, Germany). DNA electrophoresis and sequence analysis was performed on an automated laser fluorescent DNA sequencer (A.L.F. DNA Sequencer, Pharmacia Biotechnology).²² Haplotypes were identified as described previously.²¹

PCR analysis of PMP22 gene dose. A segment of PMP22 exon 4 was amplified by PCR with the primers listed in thetable. As a reporter fragment we coamplified a segment from the cholesterol ester transfer protein (CETP) gene, which is located on a different chromosome. To detect the amplification products on an automated sequencer, each set of primers contained one fluorescence-labeled primer. Each reaction tube contained 10 pmol of each primer, 0.1 mmol of each dNTP, 0.1 mg genomic DNA, 0.5 units of SuperTaq (HT Biotechnology Ltd, Cambridge, UK), and buffer (50 mM KCl, 10 mM TRIS-HCl [pH 9.5], 1.5 mM MgCl₂, 0.1% Triton X-100, 0.01% gelatin) in a 25-µl volume. The reaction was done on a Perkin-Elmer PCR System 9600 (Perkin-Elmer, Norwalk, CT). Touch-down PCR was performed using a hot start technique. Initial denaturation at 95 °C for 3 minutes was followed by one cycle of 30 seconds denaturation, 45 seconds annealing at 65 °C, and extension for 45 seconds at 72 °C. The next cycles used identical denaturation and extension conditions, but the annealing temperature was gradually reduced to 57 °C within 21 cycles. PCR fragments were visualized on a 3% Nusieve/1% agarose gel.

Three microliters of the amplified fragment were mixed with an equivalent amount of formamide and 3% dextran blue. This mixture was denatured for 2 minutes at 95 °C. Fragments were then loaded on an 8.5% denaturing polyacrylamide gel. The gel was run on an automated laser fluorescent DNA sequencer (A.L.F. DNA Sequencer, Pharmacia Biotechnology) for 3 hours. Fluorographic signal intensity, representing the amount of amplified fragments as the area under the peak was calculated with the Fragment Manager(Pharmacia Biotechnology). The signal of the PMP22 fragment was normalized in relation to the CETP fragment. These normalized values were compared with the intensity of a PMP22 control that was taken from a nonaffected individual. The assay was run in triplicate.

Southern blot hybridization. DNA was isolated from whole blood samples according to a standard extraction procedure. For RFLP analysis, the digestion was performed with Msp I (Boehringer Research Laboratories, East Sussex, UK), and the fragments were separated in a 0.7% agarose gel. The presence or absence of the 1.5-Mb CMT1A tandem duplication or 1.5-Mb HNPP deletion, or both, in the 50 DNA samples was determined by the previously reported Msp I Southern blot hybridization method with the RFLP markers pVAW409R3a (D17S122) and pEW401HE(D17S61),⁶ PCR analysis of the short tandem repeat markers RM11-GT (D17S122) and AFM191xh12 (D17S921),²³ or with the CMT1A-REP probes (pLR7.8 and pLR6.0) hybridized on Eco RI/Sac I or Eco RI/Sac I/Nsi I Southern blots.²⁰

Results. We analyzed four HNPP families. In pedigree 1, haplotype analysis showed that all affected members were hemozygous for an allele that they inherited from the non-affected parent (seefigure 1), whereas all unaffected members showed a cumulative heterozygosity rate of 80%, which is close to the expected frequency for these markers in the normal population.²¹ These data suggest that the disease-associated allele carries a deletion encompassing the markers within the PMP22 gene we used. A similar situation was observed in pedigrees 2 and 3 (seefigure 1).

The only affected individual in pedigree 4 suffered from HNPP as determined by the presence of clinical symptoms and electrophysiologic examination. Because neither of his parents showed clinical signs of the disease, we investigated the biological relation of this proband with his parents. The family history and a segregation using polymorphic markers at the APOb, CETP, and LPL gene loci confirmed paternity. In addition, we excluded the presence of a mutation at the MPZ locus by gene sequence analysis. Thus, the affected son in this pedigree carries a de novo PMP22 mutation.

In the quantitative PCR, HNPP patients in pedigrees 1 to 4 who were hemizygous at the PMP22 locus showed a reduced fluorescence signal for exon 4 of PMP22 as measured by the area under the peak in the fluorogram (figure 2). The reduction ranged between 34 and 56% (figure 3). Unaffected HNPP family members showed no reduction below 93 percent (see figure 3).

• Download figure
• Open in new tab
• Download powerpoint

Figure 2. Fluorogram (original in four colors) showing the densitometry scans for the amplified PCR fragments. The first four peaks represent the CETP fragment of the four different PCR reaction products. The following peaks represent the PMP22 fragments. The last peak is from the normal PMP22 control. Normalization of the peaks was done by electronic shift making the area under the peaks identical. Percentages indicate the area under the peak of normalized values in relation to the normal PMP22 control. The difference between the signal of the PMP22 control fragment and the CETP fragment results from the differences in amplification efficiency of the two different fragments.

• Download figure
• Open in new tab
• Download powerpoint

Figure 3. Quantitative PCR data obtained in (A) blinded study and (B) four HNPP families. Mean and variance from triplicate analysis. Classification was done by clinical and electrophysiologic examination and Southern blotting (see text).

Southern blot analysis of samples from the probands and controls demonstrated the presence of a deletion in all 15 HNPP patients. A duplication in the 17p11.2-12 area was found in all CMT1 patients, whereas normal signal intensity was observed in all 20 controls.

Quantitative PCR analysis in the 15 CMT1A patients showed signal intensities ranging from 128 to 181% of the mean values obtained for the normal control sample (see figure 3). None of the CMT1A patients was misdiagnosed by the PCR method. Fluorescent analysis of quantitative PCR products detected 14 of 15 HNPP patients. For these 14 patients the signal intensity was between 32 and 70% of the control intensity(see figure 3). One HNPP patient showed a signal intensity of 125% and was thus misdiagnosed between normal and CMT1A. Subsequent reevaluation of this patient's DNA gave results ranging from 54 to 106% (mean 93%). The normal samples ranged from 78 to 120% of the normal standard DNA; the mean was 99.6%.

Apart from the one misdiagnosed sample, there was no overlap in the range of fluorescent intensity observed for HNPP patients, normal controls, and CMT1A patients (see figure 3).

Discussion. It has now been firmly established that deletions at the PMP22 gene locus cause HNPP, that duplications of the same 1.5-Mb-long DNA fragment cause CMT1A, and that these mutations are the most frequent genetic lesion in both disorders.⁹ The determination of the PMP22 gene dosage that we describe in this paper thus allows the diagnosis of the underlying genetic defect in most HNPP and CMT1 cases in a single day.

Our method is faster than previously used techniques, which include Southern blotting, fluorescent-in-situ-hybridization, and pulsed field gel electrophoresis.^17,18 Despite this improvement in the speed of analysis, the method is very reliable as demonstrated by the absence of overlap in the area-under-the-peak values observed for deletions, duplications, and in the normal situation. Validation of the assay in 50 prediagnosed DNA samples showed a sensitivity of 96.6%. Because we were not able to obtain a new DNA sample from the individual, we were unable to confirm or rule out contamination of the DNA sample with PCR product as a possible cause for the observed increase in signal intensity.

The variance in under-the-peak values observed between two different PCR reactions and the variance of experimental conditions (e.g., pipetting error) are reduced or eliminated through the coamplification of a single copy comparison gene. The higher variance in the fluorescence signal observed in CMT1A patients compared with that of HNPP patients may relate to differences in conformation resulting from different copy numbers present in the reaction tubes.

In informative HNPP families, the method can be combined with a previously reported haplotyping system²¹ that allows the diagnosis by the aid of segregation analysis. In the case of normal fluorescent signals, the next step should be direct sequencing of the candidate genes.

In conclusion, we developed a rapid and reliable test system for the genetic diagnosis of inherited peripheral neuropathies. When the reliability of this system is confirmed in further studies, it may make molecular genetic diagnosis of HNPP and CMT1A much faster and easier to perform. The speed of obtaining the molecular genetic diagnosis may lessen the need for sural nerve biopsies in the future.

Acknowledgment

We thank Mrs. Joke Nowitzki for excellent technical assistance.