To study the instability of FMR1 triplet repeats in the general population, we screened a prospective sample of 24,449 anonymized mother-offspring pairs and analyzed transmissions of intermediate-size (45-54 triplets) and premutation-size (55-200 triplets) alleles. We screened all mothers for alleles > or = 45 triplets by Southern blot and studied transmission of 545 maternal alleles to their offspring using polymerase chain reaction. Out of 21,411 maternal samples with conclusive results, we identified 250 carriers of at least one intermediate-size allele and 39 carrying a premutation-size allele. Out of a subsample of 430 transmissions of normal-size alleles (< 45 triplets), we observed four (< 1%) unstable transmissions. There were 6/90 intermediate-size unstable alleles (7%) and 11/25 unstable premutation-size alleles (44%). Two mothers transmitted a typical full mutation. The incidence of fragile X syndrome was thus 1/12,225 newborns (upper limit of 95% confidence interval: 1/4638 newborns), but larger in males (1/6209) than females (none detected in over 12,000 newborn females). Intermediate-size alleles were more unstable than normal-size alleles (p = 0.0027), but more stable (about sixfold) than premutation-size alleles (p < 0.0001). Unstable premutation-size alleles harbored the major fragile X haplotype (T50-T42-T62), and this haplotype appeared to be a good predictor of instability in premutations (p = 0.02). Incidence and instability are important to determine the feasibility and cost effectiveness of putative FMR1 screening programs. Carriers of FMR1 alleles of 55+ triplets with no family history of the disease may have a significant risk of expansion to a full mutation in a single generation.

The fragile X mental retardation 1 gene (FMR1) mutation is strongly correlated with specific and marked neurobehavioral and neuroanatomical abnormalities. The protein product, FMRP, is highly expressed in neurons of the normal mammalian brain, and absent or in low levels in leukocytes from individuals with fragile X (FraX)-associated mental impairment. Inferences which arise from these findings are that FMRP has a critical role in the development and functioning of the brain, and that leukocyte-derived molecular assessments provide a good indicator of FMR1 expression in that organ. This latter conclusion appears true in most cases even though the typical FMR1 mutation is an unstable triplet repeat expansion which demonstrates somatic heterogeneity within and across tissues. Blood to brain correspondence in FraX patients has only rarely been confirmed by the direct study of human brain specimens and, to our knowledge, it has never been studied in living individuals with the FMR1 mutation. In this report, we describe the FMR1 patterns in olfactory neuroblasts (ON) from two living brothers with expansion mutations in their leukocytes who are mentally retarded and autistic. ON were chosen for study because they are accessible neurons closely linked to the brain. In both subjects, the ON genotype was highly, but not perfectly, consistent with that observed in leukocytes. Protein phenotypes across tissues were completely consistent showing the absence of FMRP-immunoreactivity (-ir). These results augment the limited amount of direct evidence which indicates that FMR1 mutation patterns in leukocytes are a good, albeit potentially fallible, reflection of such patterns in the brain. This report further demonstrates the feasibility of using ON samples to evaluate the FMR1 mutation in humans in vivo.

Studies on the population genetics of fragile X syndrome have relied on haplotyping microsatellite markers closely flanking the FMR1 gene on normal and fragile X chromosomes. Strong linkage disequilibrium between the fragile X mutation and a few haplotypes have been reported in several populations, suggesting the occurrence of a limited number of initial mutational events [Chiurazzi et al., 1996a]. The three microsatellite repeats most frequently employed in haplotyping are DXS548 [Riggins et al., 1992], FRAXAC1, and FRAXAC2 [Richards et al., 1991]. Different systems have been used by various groups for identifying the numerous alleles at each of these marker loci, thus impeding reliable comparisons of the results across populations and conclusions on the identity of fragile X "founder chromosomes." In fact, several authors attempted to align the allele distributions of the flanking microsatellites by assuming that the mode(s) in populations of European origin were the same [Chiurazzi et al., 1996a;Macpherson et al., 1994;Rousseau et al., 1995]. However, this "alignment" procedure of different distributions cannot be considered satisfactory and an absolute reference must be introduced at some point. Unfortunately, sizing of polymerase chain reaction (PCR) products by running them next to markers or even sequences (e.g., M13) derived from other loci is not a reliable method at the single base-pair resolution level, because the different GC content and secondary structures make fragments of the same length migrate with slight differences. Thus, only one (or more) allele(s) of the same microsatellite which have been actually sequenced would constitute such an absolute reference. However, sequence information for all three loci obtained from the same samples is not yet available, and considering that sequencing of repeated sequences can be troublesome and that different centers may obtain different results, we first decided to genotype a few lymphoblastoid cell lines for these markers and to start distributing their DNA to all interested laboratories. Therefore, we have established a panel of DNA samples derived from several cell lines maintained in our laboratories and that we are intending to sequence. It is important to realize that, even in the absence of sequence information, the sharing of reference DNAs will allow comparison of the results obtained in the various FMR1 haplotype studies. All DNA samples (seven males and one female for a total of eight) were genotyped for DXS548, FRAXAC1, and FRAXAC2 independently by each participating group in Italy, the United Kingdom, and Canada, respectively. This letter to the editor is intended to call the attention of fellow researchers who wish to replicate our sequencing effort and to inform them that this panel is available freely to all researchers and can be requested at the Internet address http://www.rmga.qc. ca/panel/. Upon request via the WWW site, the closest reference laboratory will provide two or three representative DNA samples from the panel. Ideally, all ...