Editor—The genetic causes of mental retardation are still largely unknown so that about 34% of cases of severe to moderate and 80% of mild mental retardation remain unresolved.1 ,2 Consequently, genetic counselling is difficult in these cases. Chromosomal rearrangements are still the most frequent cause of mental retardation and cytogenetic analysis at 400-550 band resolution cannot detect rearrangements smaller than 5 Mb. Therefore, any new technologies to improve cytogenetic analysis would be of great benefit. In recent years, there has been evidence that the cause of 6-7% of mental retardation involves subtle chromosome rearrangements.

Several molecular methodologies have been used successfully to investigate the integrity of telomeres, such as hypervariable polymorphisms (HVPs), FISH, and CGH, while others based on DNA array approaches are being developed.2 The choice of method rests on considerations of feasibility, cost, reproducibility, and sensitivity. FISH and microsatellite analyses are the techniques most commonly used by numerous laboratories and both of them are still being evaluated in terms of their performance and practicability. FISH based analysis with telomeric specific probes provides a simultaneous investigation of all chromosomes ends, resulting in total detection of chromosome rearrangements without requiring the examination of parental chromosomes.3-5 However, this method does not assign parental origin, it loses UPD events, and only gives information regarding abnormalities spanning the probe's DNA region. Although now simplified by being available in kit form, the method not only requires considerable expertise in molecular cytogenetics, but is also expensive and time consuming.

HVP analysis using microsatellites, which is both easy and inexpensive, has two main problems, as it requires both parental DNA samples and high informativeness, established by the number and frequencies of alleles. However, even for highly heterozygous loci ≅0.8, the detection rate for monosomy and trisomy is 0.64 and 0.5, resulting in about 50% of potential carriers for subtelomeric rearrangements remaining undetected.6

In order to underline how microsatellite analysis can give a detection rate matching that obtained using FISH, we performed a comparative pilot investigation on 60 unrelated mentally retarded patients using a collection of microsatellite polymorphisms mapping chromosomal subtelomeric regions and telomere specific probes according to the multiprobe FISH protocol, as previously described by Knightet al.3

From the 60 subjects selected for the study, 30 (group 1) were already part of a larger FISH based screening for cryptic telomeric rearrangements on 200 Italian mentally retarded patients.7The remaining 30 subjects (group 2) were referred by clinical geneticists because their clinical characteristics strongly suggested chromosomal abnormalities. Fundamental clinical criteria for both groups included: (1) a degree of mental retardation ranging from moderate (IQ 50 to 70) to severe (IQ <50), (2) a normal karyotype, (3) absence of diagnosis, and (4) no known consanguinity. Multiple congenital anomalies (MCA) and dysmorphism were present in 40 of the 60 subjects included in the study. FISH analysis with telomere specific probes (Cytocell multiprobe kit) was performed on fixed lymphocytes metaphase chromosomes, prepared directly from peripheral blood, as previously described.3 A total of four subtle telomeric rearrangements were found by this method in 60 subjects. Two de novo deletions (46,XY,6q− and 46,XX,2q−) carried by two probands in group 1 have been described elsewhere.7 In subjects in group 2, FISH analysis showed two different unbalanced derivatives: 46,XY,der(15p)t(15p;17p) and 46,XX,der(4p)t(4p;8p). The clinical presentations will be reported elsewhere. The unbalanced derivatives 4p;8p and 15p;17p were easily shown by FISH in the probands as well as the balanced translocation in the carrier mothers. For the translocation 15p;17p, the rearrangement was confirmed by FISH using specific YAC probes obtained from the Screening YAC Center, since the kit used does not include probes for the p arms of acrocentric chromosomes.

The percentage of anomalies disclosed in our sample by this method was 6.6%, matching the frequency obtained in others studies.4Since all subjects from both groups were studied simultaneously, no bias is represented by the group 1 subjects being involved in another study. For HVP analysis using microsatellites, we initially designed a panel of 41 subtelomeric markers (table 1) chosen on the basis of their informativeness (Het range 0.52-0.93) and telomeric position. Twenty five of the markers had a physical distance of 0-2 Mb from the chromosomal ends. The remainder were genetically mapped and represent those most telomeric and informative. Their distance from the telomere was assigned according to the integrated mapping resource (www.gdb.org). The genetic distribution was spread over a range of <2.5 cM.

Manual genotyping was performed with multiplex PCR on 96 well polycarbonate plates, amplifying three distinct polymorphic loci per single reaction mix simultaneously. For each family trio (proband and parents), 41 PCR were performed, that is, 13.6 per person. It should be noted that for complete use of the microsatellite panel only about 2 μg of DNA is necessary per sample. Detection of allelic segregation pattern was achieved by two alternative methods, the first was that described by Dib et al 8 and the second used the ABI-PRISM 302 (Perkin-Elmer) microsatellite analyser, as recommended by the manufacturer. Detailed protocols are available from Dr M Fichera (mfichera{at}oasi.en.it).

In our sample, microsatellite analysis showed the monosomy 2q and 6q in group 1. The derivatives 4p− and 15p− from subjects with unbalanced translocations (group 2) were not detected.

In 60 probands analysed by this method, given that the detection rate (DR) for monosomy of our panel is 0.59, the expected number of monosomic cases would be 3.4. compared to the four cases disclosed by FISH. Both trisomies from the group 2 patients were missed.

The limitations of a microsatellite panel including only one marker for the telomere prompted us to construct an enriched panel of three polymorphic markers spread over a defined telomeric DNA region. The aim of using three consecutive markers was twofold: (1) to increase the informativeness for monosomy and trisomy and (2) to avoid using parental DNA. A pilot model was performed for four telomeres, 2p, 3q, 18q, and 22q, chosen according to the clone sequences available on the GenBank Data Base. Since 4p and 8q are not present in the data bank, these were not selected. The three polymorphic loci in the four telomeres chosen are situated at a mean distance of 0.6 Mb from the telomere. Nine of these 12 microsatellites have already been physically mapped in GenBank. The remaining three (CA)n, two for telomere 22q and one for telomere 18q, were selected from the sequences available in the GenBank. The criterion for choosing them was the length of the repeat. The heterozygosity value was verified in our patients and was equal to 0.76.

Microsatellite analysis using three consecutive polymorphic loci was evaluated for the proband both with and without parental allelic segregation. The results obtained for the four loci are shown in fig 1. The informativeness was calculated for both anomalies, on the principle that at least one of the three loci was informative, I=1−[(1−i_a)(1−i_b)(1−i_c)] where i_n represents the informativeness of each locus. In this way, the monosomy informativeness rises to 0.93 (compared to 0.60 from the analysis of a single locus) and that of trisomy to 0.83 (doubling the DR obtained with a single locus).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Graphic representation of informativeness for monosomy (grey) and trisomy (black) considering the parental allelic segregation for one, two, and three microsatellites per telomere. New mapped polymorphic markers for telomeres 2p and 22q are located within 700 kb from their telomere.

Analysing the sample without parental allelic segregation, the result for monosomy is interpreted as the exclusion of the anomaly. The probability of exclusion is given by the presence of at least one of the three loci in heterozygosity, therefore e=1−[(1−e_a)(1−e_b)(1−e_c)] where e_n is the heterozygosity of a single locus. Without parental allelic segregation, it is not possible to determine the DR, which depends on the approach chosen to investigate the unresolved cases. Conversely, the DR for trisomy remains the same with or without parental allelic segregation because the presence of three different alleles is obligatory.

No linkage disequilibrium was observed during allelic segregation, guaranteeing the mutual independence of the loci. Nevertheless, if this approach were extended to all the telomeres and linkage disequilibrium were found between three loci, it would be enough to calculate the exclusion of monosomy and the informativeness of trisomy, taking into account the haplotype frequency in the population. In this case both values should be decreased.

This study finally clarifies the limits of resolution of the HVP approach, even when using three aligned loci. Further, an analysis extended to the parents is evidently not necessary because, even if it is impossible to affirm the detection rate for monosomy without the parents, the probability of excluding the anomaly is as high as 98.5%, reducing the number of telomeres to be investigated to 1.5%. For trisomy, however, the missed DR remains 17%.

Even though FISH analysis has a 100% DR, the approach suggested here represents an easy and economical method which gives acceptable results and, in any case, directs the investigation towards a more focused FISH analysis.

In conclusion, considering that all genomic sequences will soon be deposited in GenBank, this could represent a refinement of the whole genome microsatellite screening recently described by Rosenberget al.10