For molecular diagnostic testing of CMT1 we choose two highly polymorphic dinucleotide repeat markers D17S921 and D17S122, that are closely linked to the CMT1 region. To optimize the assay method for fluorescent detection of alleles, we tested different PCR cycle numbers, buffer compositions and template concentrations. As control samples in these experiments we used DNA probes from clinically and electrophysiologically healthy subjects, DNA probes from patients with preliminary data for CMT1A duplications having two or three alleles for the markers under investigation [the analyses were performed by A.J. in the Laboratory of Neurogenetics, University of Antwerpen, The Netherlands, using an ABI PRISMO automatic analyzer (Applied BioSystems, Foster City, CA, USA)] and null controls for amplifications. The criteria for allele intensity were measurement of the peak area. All data were compared between probes with known genetic defects and normal controls.

To quantify the PCR amplification we examined the PCR product between cycles 22 and 34. Figure 1 shows that before the 25^th cycle the product is insufficient. After the 27^th cycle a plateau effect is present and no significant statistical difference between normal controls and those with duplications is observed. We chose 26 cycles because that is still within the exponential phase of amplification, and the differences between normal controls and probes with duplications having two peaks with different intensities, were most evident.

Various buffer compositions were tested. The concen­tration of MgCl₂ was particularly important for accurate peak ratio. The best results were obtained at a concentration of 2.0 mM MgCl_2. The quantitation of PCR products was found to be linear with respect to the initial template concentration in the range 50-500 ng, and 200 ng was chosen. Our observations are consistent with previously reported data [9]. All other parameters of the PCR reaction were as described by Cudrey et al. [10] for D17S921 and by Lupski et al. [11] for D17S122.

We tested patients with already known CMT1A dupli­cations using previously described PCR conditions. The observed electrophoretic mobility patterns for both markers are summarized into four groups in Figure 2: a) two peaks corresponding to three CMT alleles, with the higher intensity of the fast migrating one; b) two peaks, corresponding to three CMT alleles, with the higher intensity of the slow migrating one; c) one peak; d) three peaks (representing three alleles at the locus under investigation).Peak areas were used as described in the Material and Methods section, to calculate CMT allele number, and they were compared with those of normal controls. The results are presented in Table 1. As can be seen, three peaks are observed only among the affected individuals with CMT1A duplications and are characteristic for this gene defect. In case of one peak only (two or three overlapping alleles) the polymorphic marker is uninformative for the patient himself. There is no statistical difference between this pattern in normal controls and patients with duplications/deletions. To define the CMT status, a combination of two or more different markers should be used, and their inheritance among two or more generations in the pedigree have to be investigated. The electrophoretic pattern with two peaks, where those with a slower mobility have a larger area, is also characteristic for patients with a CMT1A duplication, since it has not been described in normal controls.

The most difficult pattern for interpretation is two peaks, where the fast migrating one has a higher intensity. Such mobility is observed in samples with duplications, but also in normal individuals. That is because the allele with a shorter length is amplified more efficiently than the other one and it therefore has a higher intensity. This is a common situation in normal controls but can also be seen among patients with duplications. Here, the co-migration of two alleles with the same length results in different intensities of the peaks. The measurement of the area-under-peak ratio could be useful for discriminating both cases. In Table 1, it is shown that for marker D17S921 the ratios in the control group vary between 1.04-1.74, while in patients with duplications they are doubled (2.05-3.72). There is no overlap between the groups. The Fischer exact test showed statistically significant differences between samples with duplications and normal controls having one and the same pattern; for marker D17S921 (p = 0.5 x 10^–8) and for marker D17S122 (p = 0.5 x 10^–6). All results were reproducible in serial and in day-to-day experiments. Our findings suggest that observation of electrophoretic pattern with two peaks, with the higher intensity of the faster migrating one, where the peak ratio is over 2, is characteristic for duplication.

The application of CMT1 polymorphic analysis is illustrated in Figure 3. Here a three-generation CMT1 family (No. 57) was analyzed. For marker D17S122, we have detected three alleles in all affected members, suggesting a CMT1A duplication.

For marker D17S921, affected individuals No. 1 and No. 3 have electrophoretic patterns characteristic for probes with duplications. Individuals No. 2 and No. 4 have a similar mobility, but No. 2 has two alleles with an area-under-peak ratio above the critical 2.0 value, that is consistent with duplication status. At the same time, alleles of the normal individual (No. 4) show intensity of the two peaks consistent with normal status.

Using the described strategy, we have analyzed 57 Bulgarian CMT1 families with an autosomal-dominant trait of inheritance for markers D17S921 and D17S122. This approach was informative in 88% of the cases. CMT1A duplication was detected in 12 cases, HNPP in three cases, and duplication/deletion was excluded in 38 of them. These data were compared with those from Southern blot analysis and no false-positive or false-negative results were observed.

We would like to point out that polymorphic analysis is informative only in nuclear HNPP families where segregation data indicate inheritance of a null allele from one of the parents. For single patients alternative diagnostic methods should be used.

In conclusion, we have developed a rapid and reliable approach for CMT1A/HNPP genetic diagnosis that dis­criminates efficiently between affected subjects and controls. Dosimetric measurement combined with statistical evaluation allows a straightforward interpretation of the results. This gives us a real possibility to perform correct and fast molecular diagnosis of CMT1A and HNPP in routine clinical laboratory practice.