We screened 42 sporadic gastric tumors and found that seven of them had significant microsatellite instability. These were then studied at 26 microsatellite loci, comprising di-, tri- and tetranucleotide repeats. The instability level of individual microsatellites in the tumors was found to be positively correlated with the population average heterozygosity and variance of repeat number of the microsatellite loci, as predicted by the stepwise mutation model. Moreover, as is known to occur in human populations, instability was strongly correlated with the number of repeats at each microsatellite locus and with the perfection of the reiterated sequence. These results demonstrate that microsatellite mutations in unstable tumors show similarities to germline mutations and suggest that their study may be useful in understanding the mechanisms that generate microsatellite variability in human populations. We used this model to test the claims that the microsatellite mutation process is biased towards increased size and heterozygosity with wide differences in allele sizes. These assertions were not confirmed.

Many factors have been implicated in influencing the rate of microsatellite mutations, including the length and base composition of the repeat motif, number of repeats, base composition of flanking sequences and, perhaps most importantly, degree of perfection of the repeats. The latter is of clinical relevance, since in both spino-cerebellar ataxia and fragile X syndrome, alleles with imperfect repeats appear to be much more stable than perfect ones. As yet, the relative importance of increased replication slippage and decreased mismatch repair efficiency in the preference of mutations to occur within perfect repeats has not been fully determined. D13S308E is an asymmetric trinucleotide repeat microsatellite with the sequence (CAT)3CAC(CAT)CAC(CAT)2CAC(CAT)CAC(CAT)15, thus containing two parts: an 11-repeat imperfect portion (underlined above) and a 15-repeat perfect one (bold). We sequenced eight new mutant alleles of D13S308E from three human gastric tumors with instability in this and other microsatellites. In all mutations the size variation occurred exclusively in the perfect part of the microsatellite. These results constitute direct evidence that the molecular basis of microsatellite alterations seen in normal cells is similar to those that occur in human tumors with extensive microsatellite instability. The investigation of mechanisms involved in microsatellite mutations has been handicapped by the fact that they are rare events. The microsatellite instability observed in malignant tumors provides us with a useful general system to study these mechanisms.

No abstract

We have developed a non-isotopic technique based on methylation-specific PCR (MSP) of the FMR1 promoter for the identification of fragile X full mutations among men. DNA samples were first treated with sodium bisulfite to convert unmethylated, but not methylated, cytosines to uracil, followed by PCR amplification with oligonucleotide primers specific for methylated versus unmethylated DNA. We designed two primer pairs: one produced a 142-bp fragment from the bisulfite-treated methylated CpG island, while the other generated an 84-bp product from the treated non-methylated promoter. In normal males only the 84-bp fragment was seen, while the diagnosis of FRAXA was doubly indicated by the appearance of a 142-bp product together with absence or weak visualization of the 84-bp band. As an indispensable internal control for the efficiency of the sodium bisulfite treatment, we used a primer pair specific for the imprinted maternal methylated version of the CpG island of the SNRPN gene on human chromosome 15. Using the methylation-specific PCR we identified with 100% sensitivity and accuracy eight previously diagnosed FRAXA male patients mixed with 42 normal controls. Because of its simplicity and high efficiency, methylation-specific PCR may become the method of choice for the diagnosis of the fragile X syndrome in mentally retarded males.

To screen for monosomy X in spontaneous fetal losses we explored a simple molecular strategy based on loss of heterozygosity at highly polymorphic X-linked loci. We developed a multiplex fluorescent procedure that allows the simultaneous amplification of five dinucleotide repeat polymorphisms in a large low-recombination region in the long arm of the X chromosome. Analysis was performed by computer-assisted laser densitometry. We did not find any instances of homozygosity at all five loci in 30 normal females tested, nor among 37 women whose typing data were retrieved from the Fondation Jean Dausset - CEPH genotype database. In addition, all cases of monosomy X previously diagnosed by conventional cytogenetics presented the anticipated loss of heterozygosity at all loci. We studied 19 spontaneously aborted female fetuses and we found four samples homozygous for the five loci (21%), in good agreement with the expected rate of monosomy X in first trimester spontaneous abortions. We conclude that the loci have high diversity and high efficiency in PCR-amplification and that our multiplex procedure constitutes a simple and useful molecular screening test for monosomy X in abortions and stillbirths.