The value of clonality in the diagnosis and follow-up of patients with cutaneous T-cell infiltrates.

The diagnosis of early stages of cutaneous T-cell lymphoma (CTCL) is often difficult, especially for lesions that are at the borderline between reactive and neoplastic skin T-cell infiltrates. T-cell monoclonality in these lesions is considered by some to be an important prognostic factor of neoplastic evolution, whereas others claim that clonality can also be found in benign skin infiltrates and is therefore of limited diagnostic value. To address this controversy, the authors analyzed retrospectively eight patients with lymphocytic skin lesions who progressed to CTCL, and three patients with recurrent T-cell lymphocytic infiltrates who had not developed CTCL. From a total of 65 biopsies of eight progressing patients, 32 were diagnosed as histologically malignant and 33 were diagnosed as benign or borderline. The authors found clonality by either polymerase chain reaction or Southern blot analysis in 88% of malignant and in 79% of nonmalignant lesions. None of the 37 biopsies of non-progressing patients was clonal. These results indicate strongly that the presence of monoclonality in T-cell skin infiltrates is related closely to the risk of developing CTCL. The value of clonality as a marker of malignancy is supported by the absence of T-cell clonal populations in all infiltrates from patients who had not progressed to lymphoma.

Frequent loss of heterozygosity at the DNA mismatch-repair loci hMLH1 and hMSH3 in sporadic breast cancer.

To study the involvement of DNA mismatch-repair genes in sporadic breast cancer, matched normal and tumoral DNA samples of 22 patients were analysed for genetic instability and loss of heterozygosity (LOH) with 42 microsatellites at or linked to hMLH1 (3p21), hMSH2 (2p16), hMSH3 (5q11-q13), hMSH6 (2p16), hPMS1 (2q32) and hPMS2 (7p22) loci. Chromosomal regions 3p21 and 5q11-q13 were found hemizygously deleted in 46% and 23% of patients respectively. Half of the patients deleted at hMLH1 were also deleted at hMSH3. The shortest regions of overlapping (SRO) deletions were delimited by markers D3S1298 and D3S1266 at 3p21 and by D5S647 and D5S418 at 5q11-q13. Currently, the genes hMLH1 (3p21) and hMSH3 (5q11-q13) are the only known candidates located within these regions. The consequence of these allelic losses is still unclear because none of the breast cancers examined displayed microsatellite instability, a hallmark of mismatch-repair defect during replication error correction. We suggest that hMLH1 and hMSH3 could be involved in breast tumorigenesis through cellular functions other than replication error correction.

Allelic losses and DNA methylation at DNA mismatch repair loci in sporadic colorectal cancer.

Normal and tumor DNA samples of 35 patients with sporadic colorectal carcinoma were analyzed for microsatellite alterations at 12 markers linked to mismatch repair loci: hMLH1, hMSH2, hMSH3, hMSH6, hPMS1 and hPMS2. Remarkably, no correlation was observed between the replication error phenotype (RER+) and allelic losses at these loci. Hemizygous deletions, seen in 6/35 (17%) informative cases at hMLH1, 4/27 (15%) at hMSH2/hMSH6 and 6/34 (18%) at hMSH3, were rarely found in RER+ tumors. Since mismatch repair protein components act in molecular complexes of defined stoichiometry we propose that hemizygous deletion of the corresponding loci may be involved in colorectal tumorigenesis through defects in cellular functions other than replication error correction. The analysis of the methylation status of the promoter region of hMLH1 revealed that methylation might be an important mechanism of this locus inactivation in RER+ sporadic colorectal cancer.

Genomic loci susceptible to replication errors in cancer cells.

Microsatellite instability due to a deficiency in DNA mismatch repair is characteristic of a replication error (RER) phenotype. This widespread genomic instability is well documented in hereditary non-polyposis colon cancer (HNPCC) as well as subsets of sporadic carcinomas. Features of the RER phenotype such as the early appearance in tumour development and better prognosis of RER+ colorectal tumours render its examination important for cancer patients. Recently, we identified four loci that were shown to be highly susceptible to RER in cancer cells. Here, we used these loci to detect the RER phenotype in sporadic carcinomas of colon, breast, lung, endometrium and ovary. Replication errors revealed by these four markers followed the same tumour specificity as observed in HNPCC patients. In particular, 24% (6/25) of colorectal, 33% (4/12) of endometrial and 17% (2/12) of ovarian cancers displayed the RER phenotype characterized by an increased allelic mobility, whereas none of the breast (n = 22) and the lung (n = 27) carcinomas were found to be unstable. Assaying RERs sensitive loci provides us with a useful diagnostic tool for HNPCC-like sporadic tumours.

High resolution deletion mapping reveals frequent allelic losses at the DNA mismatch repair loci hMLH1 and hMSH3 in non-small cell lung cancer.

To study the involvement of DNA mismatch repair genes in non-small cell lung cancer, matched normal and tumoral DNA samples from 31 patients were analyzed for both LOH and microsatellite instability with 34 markers at or linked to hMLH1 (3p21), hMSH2 (2p16), hMSH3 (5q11-q13), hMSH6 (2p16), hPMS1 (2q32), and hPMS2 (7p22) loci. Chromosomal regions 3p21 and 5q11-q13 were found to be hemizygously deleted in 55% and 42% of the patients, respectively. Sixty five percent of the patients deleted at hMLH1 were also deleted at hMSH3. The shortest regions of overlap for 3p21 and 5q11-q13 deletions delimited by D3S1561/D3S1612 and D5S2107/D5S624, respectively, were restricted to genetic distances of only 1 cM. Currently, the hMLH1 (3p21) and hMSH3 (5q11-q13) genes are the only known candidates located within these regions. The mutational analysis of hMLH1 and hMSH3 in hemizygously deleted patients led to the detection of 2 new polymorphisms in hMSH3. The consequence of these allelic losses remains unclear, but the lack of inactivating mutation might explain that replication error, the hallmark of mismatch repair genes inactivation in cancer cells, was quasi-absent in tumors. We suggest that hMLH1 and hMSH3 genes could be involved in lung tumorigenesis through dosage effect in cellular functions other than replication error correction.

Overall informativity, OI, in DNA polymorphisms revealed by inter-Alu PCR: detection of genomic rearrangements.

We studied two systems of multilocus markers revealed by PCR using primers directing amplification between Alu repeats in a tail-to-tail orientation. Genomic polymorphisms were detected as the presence or absence of the electrophoretic bands representing DNA fragments of a given length. A total of 104 such fragments segregating as Mendelian markers in a panel of eight CEPH families were analyzed by two-point linkage analysis. Fifty-one of these fragments were localized with respect to CEPH markers; they represented 33 loci, 7 of which were multiallelic. Locus-specific oligonucleotides were developed and used as hybridization probes to identify the mapped loci within a complex pattern of inter-Alu PCR products. A great proportion of inter-Alu PCR polymorphisms represented length variants within amplified DNA segments, while others were presumably due to mutations within the priming sites. To describe the expected number of informative loci per typing experiment we introduced a parameter called overall informativity (OI), which provides a single measure of the multiplex ratio and the informativity of markers contributing to a multilocus system (OI of a single locus is equivalent to its heterozygosity and cannot exceed 0.5 for a biallelic codominant marker). High OI values (5.8 and 11.5) of the two presented systems of inter-Alu PCR markers of random chromosomal distribution render them suitable for mapping genomic rearrangements such as genomic deletions in tumoral tissues. This was illustrated by the detection of loss of heterozygosity in the 9q22-qter region in sporadic colon cancer.

U3 long terminal repeat-mediated induction of intracellular immunity by a murine retrovirus: a novel model of latency for retroviruses.

BL/VL3 radiation leukemia virus (RadLV) is a thymotropic, highly leukemogenic murine leukemia virus (MuLV) which is unable to replicate in vitro in mouse fibroblasts. We have previously reported that the U3 long terminal repeat region of its genome is responsible for this block (E. Rassart, Y. Paquette, and P. Jolicoeur, J. Virol. 62:3840-3848, 1988). By using hybrids of permissive and resistant cells infected with BL/VL3 RadLV or fibrotropic MuLV, we found that the resistant phenotype was dominant. Investigation to determine at which step of the virus cycle the block operates revealed that integration, transcription, and translation of the BL/VL3 viral genome occurred at normal levels in nonpermissive cells. The BL/VL3 RadLV Pr65gag proteins made in nonpermissive cells were also myristylated and located at the membrane, and the levels of their cleaved products were similar to those of fibrotropic MuLV. However, processing of BL/VL3 RadLV Pr85env was impaired in nonpermissive cells. Virions were not released into the culture medium of nonpermissive cells, as measured by reverse transcriptase activity and by content in p30 or gp70 protein and as documented by lower levels of budding particles seen by electron microscopy. These results indicate that BL/VL3 RadLV replication is blocked at a late stage of the virus cycle, i.e., at virion assembly. Interestingly, these BL/VL3 RadLV-infected nonpermissive fibroblasts were resistant to superinfection by fibrotropic Moloney MuLV, and this resistance also occurred at a late step of the Moloney virus cycle. Since this block is dominant, it appears that the U3 long terminal repeat region of the BL/VL3 viral genome has the ability to induce a cellular suppressor factor(s), thus bringing intracellular immunity against itself and against other ecotropic MuLVs.

DNA-binding proteins that interact with the long terminal repeat of radiation leukemia virus.

We used the electrophoretic mobility shift assay to identify the interactions of nuclear proteins with the long terminal repeat of leukemogenic, thymotropic BL/VL3 radiation leukemia virus (RadLV). In the promoter region, we identified a CCAAT box-binding protein (CBP) that has the same binding characteristics as the CCAAT box-binding protein that binds to the Moloney murine sarcoma virus promoter and most likely represents the CP1 factor. In the upstream enhancer region unique to BL/VL3, we detected several sequence-specific complexes, one with T-lymphocyte extracts but not with fibroblast extracts. This U3 region, UEB, may be important for the T-cell specificity of BL/VL3 RadLV. In the enhancer, which has been uniquely rearranged in this virus, we identified three specific protein-binding sites. Two of them showed characteristics of the LVb and core binding sites previously described for other murine retroviruses. But one binding site, identified as Rad-1, is unique to BL/VL3 RadLV and was found downstream, only 1 nucleotide from the core sequence. Rad-1 has a corelike motif on the minus strand, and the factor that binds to it could be competed by a BL/VL3 core-containing fragment. Moreover, the protein-DNA contacts involve the typical three core Gs separated by one T. These results suggest that Rad-1 binds a factor identical or similar to the core-binding factor. Our data suggest that the LVb, core, and Rad-1 motifs may be sufficient for this enhancer, most likely in association with other U3 long terminal repeat sequences, to promote a high rate of transcription of BL/VL3 RadLV in its specific target cells (thymocytes).

Molecular pathology of chronic myelogenous leukemia.

The presence of Philadelphia chromosome t(9:22) is a hallmark of 95% of clinical cases of chronic myelogenous leukemia (CML) as well as 20% of adult acute lymphoblastic leukemia (ALL) and 5% of acute myeloid leukemia (AML). The product of t(9;22) is a fusion protein BCR-ABL. The fusion proteins of CML, ALL and AML have increased tyrosine kinase activity and show a transforming potential in vitro and in animal models. The shorter p190 protein is associated almost only with ALL and AML, while the protein p210 is present in both chronic phase and blast crisis of CML and also in 50% of Philadelphia-positive (Ph1+) ALL. In CML the transition from chronic phase to blast crisis is usually accompanied by additional genetic events, e.g. additional chromosomal abnormalities, and oncogene activation(s). The detailed understanding of molecular basis of CML, and Ph1+ ALL and AML provides highly sensitive molecular and serological methods to complement classical cytogenetics. The advantages and limitations of these techniques are described and discussed below.