Material and methods

The sample was collected in four Italian Genetic Centres. Patients were selected on the basis of the following criteria: (1) the presence of mental retardation that was classified as mild, moderate, or severe according to DSM IV3; (2) patients under 1 year of age, too young for psychological assessment, were selected for the presence of developmental delay; (3) exclusion of pre- or perinatal distress through an accurate evaluation of the pre-, peri-, and postnatal patient history; and (4) exclusion of Mendelian syndromes and of genomic disorders4 ,5 for which a specific diagnostic test is available.

The essential elements of evaluation also included family history, a complete physical and neurological examination of the patients with particular attention to the presence of mental retardation and multiple congenital anomalies, and assessment of the behavioural phenotype. Electroencephalograms, brain CT scan, and MRI were performed in specific situations.

Abnormal methylation and expansion at FRAXA and FRAXE6 were excluded in 52 and 50 males and in 37 and 32 females, respectively.

Routine cytogenetic analysis at the 400-550 band level was performed in all the patients. In those patients in whom a cryptic subtelomeric rearrangement was identified by FISH, prometaphase chromosomes were also analysed to determine if the rearrangement could be detected in retrospect by cytogenetic analysis. Chromosome preparations from peripheral blood or from lymphoblastoid cell lines were used for FISH analysis. The Chromoprobe-T kit with telomere specific clones7 was used according to the supplier's instructions (Cytocell, UK) with minor modifications. To establish the origin of each rearrangement, FISH and microsatellite analysis with subtelomeric probes were performed in the parents of the patients. FISH experiments with different YACs from each rearranged chromosomal region were performed to define its size.

Results

Among our 200 patients (table 1), 44 had mild mental retardation (IQ 50-70), 62 were moderately retarded (IQ 50-35), and 55 were severely retarded (IQ 35-20). A total of 39 patients had mental retardation not otherwise specified. Mental retardation was familial in 53 and sporadic in 147 subjects; it was isolated in 23 subjects, associated with dysmorphism and/or congenital anomalies in 160 subjects, and unknown in 17.

We identified 13 rearrangements (table 2). Seven rearrangements were de novo deletions. Six rearrangements were derivative chromosomes inherited by a balanced parent in the five cases where the parents were available. Twelve rearrangements were present in patients with moderate (7/62, 11.2%) or severe (5/55, 9%) mental retardation, whereas one rearrangement was found in a 3 month old patient, too young to assess the degree of his psychomotor retardation. No rearrangements were found among the 44 patients with mild mental retardation.

In all the 13 subjects with rearrangements, mental retardation was associated with dysmorphism and, excluding case 12, with at least one serious congenital anomaly (table 3). Cases 1-6 have unbalanced translocations. In the two families (cases 1 and 2) in which we could examine other relatives, we found that the translocation was present in several members, both in a balanced and unbalanced state. In family 1, in which an X linked dominant trait had been suspected,8the two translocations' derivatives were present in subjects with different phenotypes and different degrees of mental retardation, whereas in family 2 only one derivative was identified. In this family, the proband's mother received amniocentesis at 19 weeks' gestation because of fetoplacental hydrops. The presence of two paternal uncles with severe mental retardation and multiple congenital anomalies but with a normal karyotype made counselling difficult. After normal cytogenetic results the pregnancy continued. At birth the child showed congenital chylothorax, facial dysmorphism, and bone malformations prompting the counsellor to request chromosome specific telomere testing. The test was then extended to the mentally retarded uncles, who showed the same unbalanced translocation. Also in family 3, the mother received amniocentesis at 18 weeks' gestation after the detection of omphalocele. The cytogenetic results were normal and the pregnancy continued. A previous pregnancy had resulted in a child who died immediately after birth. At necropsy, an atrial septal defect and accessory spleen were recorded. Since accessory spleen has been reported in association with 9p duplication,9 one could speculate that this child was also unbalanced for the maternal translocation, having the der(13) instead of the der(9) found in the proband. In family 4, a paternal cousin was referred with mental retardation. Cases 7-13 have deletions and all of them are de novo. The origin of the deletion was paternal in cases 7 and 11 and maternal in case 9.

We identified by FISH and PCR analysis the breakpoints in most of the rearrangements. The size of the rearrangements ranged from less than 2 cM to 31 cM (table 4). Some of the rearrangements involving large chromosome regions have not been detected cytogenetically because the translocated region has the same size and overlapping banding pattern of the original region. In case 5, when we reanalysed the chromosomes with high resolution banding to verify if the rearrangement could be detected in retrospect, we discovered that the abnormal chromosome 5 appeared to have a longer short arm. In the mother, the translocation was also evident at the 400-550 band level owing to the small size of the der(4). Three rearrangements (cases 1, 5, and 7) involve chromosomal regions whose monosomy is associated with specific syndromes. In case 1, the 9p deletion spans 27 cM. The main anomalies found in the patient at birth were omphalocele, trigonocephaly, and genitourinary abnormalities (right cryptorchidism with micropenis), all of them reported in 9p22 deletions.10-13 However, based on London Dysmorphology Database suggestions,14 C syndrome (trigonocephaly, metopic synostosis, strabismus, cryptorchidism) was suspected. These data indicate that some cases diagnosed as C syndrome could have chromosome 9p microdeletion as already hypothesised.15 We cannot, however, exclude that the trigonocephaly is the result of the 13q duplication since it has been reported in at least three patients with 13q22-qter duplication.15-18 In case 4, the 18 cM 5p deletion covers the region of the characteristic cat-like cry (which has not been reported by the mother) but not that of the distinct facial features associated with the cri du chat syndrome.19 In case 7, the Albright-like phenotype associated with distal 2q deletion20 was in fact present. The breakpoint of the deletion spans the polymorphic markers D2S2253 and D2S338, placing between KW and RA cases of AHO-like syndrome reported by Wilsonet al. 20 We recently narrowed the AHO-like critical region to 10 cM in the subterminal distal 2q (from D2S338 to D2S2253).21

In conclusion, we show that cryptic unbalanced translocations and deletions are the cause of mental retardation and phenotypic anomalies in 13 out of 200 subjects without a previous diagnosis. In five of the cases, the finding of a balanced translocation in one parent makes it possible to monitor future pregnancies for the presence/absence of chromosome anomaly. For the seven de novo deletion cases, although the risk of recurrence is low, prenatal diagnosis of future pregnancies can monitor for the presence/absence of the deletion in case of parental germinal mosaicism.

The finding that cases 2 and 3 were cytogenetically investigated prenatally after the finding of echographic abnormalities and a suspicion of chromosome anomalies suggests extending the telomere test to these situations.

Discussion

The results of our study largely overlap those obtained by Knight et al. 1 Putting together the two sets of data one can conclude that 6.5-7.4% of the children with moderate/severe mental retardation have a cryptic telomeric rearrangement, whereas the probability of finding these anomalies among subjects with mild mental retardation is less than 0.4% (1 out 182 children in the English sample, 0 out 44 in the Italian one). In both studies half of the rearrangements are de novo deletions and half are unbalanced translocations inherited in most of the cases. In six out of the 15 families with cryptic translocations, both subjects with one derivative chromosome and subjects with the other one were present. In families with more members affected by mental retardation, differences in the phenotypes associated with the two derivatives should help the clinician to encourage diagnostic laboratories to search for a chromosome anomaly. Knight et al 1 concluded that abnormalities that include the ends of chromosomes are the commonest cause of mental retardation in children with undiagnosed moderate to severe mental retardation. In fact, in our patients, subtelomeric rearrangements were always associated not only with moderate or severe mental retardation but also with dysmorphic features and congenital anomalies. Therefore, our conclusion is that once a well defined aetiological cause has been excluded, screening with subtelomeric probes should be applied in subjects with a “chromosomal phenotype”. In fact, as stressed by Knight and Flint2 in their review of subtelomeric probes and their use in clinical diagnosis, all telomere positive cases reported to date have had physical anomalies in addition to mental retardation. Recently, a whole genome microsatellite screening22 has led to identification of two cases of interstitial aneusomy among 11 children (18%) with multiple congenital anomalies and mental retardation, normal for subtelomeric chromosome analysis. This indicates that among our 147 subjects with mental retardation associated with an abnormal phenotype and with a normal subtelomeric test, roughly 26 of them should have an interstitial rearrangement.

The detection of cryptic telomeric rearrangements through FISH with subtelomeric probes raised the question of whether most of these rearrangements could have been detectable by appropriate high resolution banding,23 thus saving time and money. The same question had been raised when FISH was introduced to detect the 15q11-q13 deletion in Prader-Willi/Angelman syndrome. The demonstration that both false negative and false positive results occur with high resolution banding24-26 led to the conclusion that this technique is insufficient for deletion detection.27 The same can be extended to telomeric or interstitial rearrangements falling into a size critical for their detection even on high quality chromosome preparations. The setting up of appropriate microarray tests will probably bypass this problem. For the moment, FISH telomere analysis is an excellent support to the clinician faced with a malformed, mentally retarded subject in whom known syndromes have been excluded.