CORRESPONDENCE Microsatellite Instability in Myeloid Leukemias

Somatic instability of variable number of tandem repeat (VNTR or microsatellite) sequences has been described recently in colorectal cancer as an indicator of a novel mechanism of carcinogenesis.' The recent report in Blood by Wada et aI2 suggested that microsatellite instability is a frequent feature in the evolution of chronic myeloid leukemia (CML) from chronic phase to blast crisis, and therefore, that defects in replication andor repair mechanisms may be relevant to progression of CML. Our own observations provide no support for this conclusion. We have studied 48 CML blast-crisis patients by comparing constitutional DNA extracted from buccal epithelial cells or chronicphase leukocytes with DNA obtained from blast-crisis leukocytes. Twelve VNTR loci from six different chromosomes were amplified and, after autoradiography, the size as well as the intensity of amplified products were compared between constitutional and blast-crisis samples.3 Occasional instances of loss of heterozygosity (LOH) were found, but we did not find a single case of microsatellite instability, defined as the occurrence of novel alleles at more than one VNTR in a single sample. We have also compared buccal epithelial and blast cell DNA from 17 patients with acute myeloid leukemia of various French-American-British subtypes and similarly failed to find any convincing microsatellite instability (unpublished observations, June 1994). Out of a total of 649 paired amplifications, only 7 instances of novel VNTR bands in 6 acute leukemia samples were found. We believe these novel bands may simply reflect the naturally high mutation rate of VNTR sequences4 The reasons for the discrepancy between these two studies may be partly technical and partly interpretational. We have found that spurious PCR bands may be generated if the amplification conditions are suboptimal or if the DNA sample is impure. In this regard, it may be relevant that Wada et a1 used DNA extracted from Giemsastained slides. We believe that many of the instances of microsatellite instability described by Wada et a1 are unconvincing, especially for the primer pairs Mfd27 and Mfd41 shown in Figs 1 and 2 of their article. Data for the C13-9 and LPL primers was not shown, but microsatellite instability and LOH does appear to be clearer at the deleted in colorectal carcinoma gene (DCC) locus. However, we have also studied the DCC VNTR in 49 paired samples and found no instances of repeat instability or LOH in the 38 cases who were informative. Whether abnormalities in the DCC gene, or genes such as hMLHl and hMSH2 that lead to microsatellite in~tabili ty~.~ are relevant to CML or other leukemias will ultimately be determined by mutational and functional analysis.