Our major focus has been to identify and characterize translocations to the IgH locus (chromosome 14q32) in multiple myeloma (MM) tumors. We assembled a panel of 36 EBV negative MM cell lines, and find that: 1) Ig translocations are present in all 36 MM cell lines (HMCL), including IgH (33/36 = 92%), Iglambda (5/23 = 23%), and Igkappa (0/21); 2) the location of cloned IgH breakpoints is consistent with errors of B cell specific mechanisms (switch mostly but perhaps somatic hypermutation in a small fraction) in most cases; 3) cloned breakpoints are scattered over a large region, as far as 1 Mb from the dysregulated, overexpressed oncogene; 4) at least 15 of 30 (50%) lines have two (10) or three (5) independent IgH translocations; 5) there are a minimum of 18 (6 recurrent) translocation partners identified by ourselves and others in primary MM tumors; 6) apart from c-myc at 8q24 (see below) five chromosomal loci (cyclin D1 at 11q13; cyclin D3 at 6p21; FGFR3 tyrosine kinase receptor and MM.SET at 4p16.3; and the c-maf or mafB basic zip transcription factors at 16q23 and 20q11,respectively) are recurrent; ; 7) in a panel of 50 advanced tumors, translocations are somewhat less frequent (IgH in 58%, 2 independent IgH in 15% 3 independent IgH in none, Iglambda in 16%, Igkappa in 2%, and no Ig translocation in 35%). Analyses of premalignant MGUS and smoldering MM tumor by others show that IgH translocations are present in 47% of pre-malignant tumors. Our working hypothesis is that primary translocations to Ig loci provide one of the initial immortalizing events in the molecular pathogenesis of myeloma in about 50% of tumors, and usually occur during plasma cell development in germinal centers. In addition, secondary translocations involving one of the Ig loci - but lacking the hallmarks of a process mediated by a B cell specific recombination mechanism - occur as a tumor progression events in tumors that do or do not have primary translocations.. It appears that primary and secondary translocations involve different chromosomal partners (oncogenes), although there might be some overlap. A second focus is to clarify the significance of our finding that there is selective expression of L-myc or one c-myc allele in 9 informative HMCL despite the apparent absence of a translocation, rearrangement, or amplification involving the c-myc locus. From a combination of FISH and SKY analyses, we have evidence for karyotypic abnormalities of L-myc (one HMCL) or c-myc locus in 28/32 (88%) HMCL that we have examined. Thus it seems clear that the selective expression of one c-myc allele is a consequence of a tumor specific, complex structural abnormality (complex translocation, insertion, duplication, inversion, with frequent involvement of 3 different chromosomes but not always an Ig locus) that alters the chromosomal context of one of the two L-myc or c-myc alleles. In all informative cases, it is clear that the myc structural abnormality was present in the primary tumor as well as in the HMCL. The incidence of c-myc abnormalities appears to be much lower (45%) in advanced, primary tumor samples. Some primary tumors show heterogeneity of the karyotypic abnormalities of c-myc, and one tumor had a karyotypic abnormality of N-myc. We have hypothesized that the complex karyotypic abnormalites that appear to dysregulate c-myc rarely - if ever- occur as an early event in tumorigenesis. Instead it appears that the dysregulation of c-myc occurs as a very late progression event, most often a complex translocation that is not mediated by B cell specific DNA modification processes. A third focus is to define other kinds of genetic and phenotypic abnormalities in MM. First, we have shown that K- or N-RAS mutations are present in 17/36(45%) HMCL and approximately 30% of untreated primary MM tumors (29/99), consistent with other studies suggesting that RAS mutations are associated with the transition from MGUS to MM, with some RAS mutations occurring during progression of MM. Second, for HMCL and primary MM tumors that overexpress FGFR3 due to the t(4;14) translocation, we find about 10% with mutations of FGFR3 and about 45% with mutations of RAS, but none with mutations in both. Similar to activated RAS, we have shown that activated FGFR3 can transform NIH3T3 cells. Third, we have identified bi-allelic deletion of p18INK4c in 32% of HMCL but only about 2-3% of unselected tumors. In one case bi-allelic deletion was present in relapsed tumor but not early in tumorigenesis, indicating that this is a very late progression event. Fourth, we have investigated p16INK4A expression by Taqman assays and also assessed methylation of the p16INK4A gene, with our preliminary results suggesting that p16INK4A expression is absent or very low in approximately 70% of untreated primary MM tumors despite the fact that methylation of the gene is present in only about 25% of the untreated primary MM tumors that we analyzed. A fourth focus is to analyze expression arrays of MM cell lines and primary tumors to gain further insight into mechanisms of pathogenesis. These studies, which have been done collaboratively with Drs. P.L. Bergsagel, J. Shaughnessy, and L. Staudt, have enabled us to propose that a unifying molecular event in the pathogenesis of most MGUS and MM tumors is dysregulation of a cyclin D gene despite the very low proliferative index of these tumors. Each of the recurrent translocations, which together appear to be initiating events in about 40% of tumors, are associated with dysregulation of one of the three cyclin D genes: 11q13 and 6p21 translocations directly dysregulate cyclin D1 and cyclin D3, respectively; 16q23 and 20q11 translocations indirectly dysregulate cyclin D2 by the action of the c-maf and mafB transcription factors that are dysregulated by these translocations; the 4p16 translocation is associated with up-regulation of cyclin D2, although the mechanism is unclear. Although cyclin D1 is not expressed at significant levels in normal B cells, about 40% of MM or MGUS tumors ectopically express cyclin D1 despite the absence of translocations or other structural abnormalities of the cyclin D1 locus; the ectopic expression of cyclin D1 in the absence of an Ig translocation does not appear to occur in a large panel of lymphoma tumors. Curiously, the tumors that ectopically express cyclin D1 in the absence of a translocation are not represented in our panel of HMCL. Most of the remaining tumors express high levels of cyclin D2. We are currently focusing our efforts to determine the causes and consequences of the various kinds of cyclin D dysregulation in MM tumors. A final focus is to develop a pathogenic classification of myeloma. There appear to be two pathways involved in the pathogenesis of pre-malignant MGUS and MM. About half of tumors are non-hyperdiploid, and mostly have five recurrent IgH translocations that dysregulate oncogenes on 11q13(CYCLIN D1), 6p21 (CYCLIN D3), 16q23(c-MAF), 20q12(MAF b), and 4p16(FGFR3 and MMSET); compared to hyperdiploid tumors, the non-hyperdiploid tumors also have a higher incidence of other chromosomal structural abnormalities and of chromosome 13 monosomy. Hyperdiploid tumors have multiple trisomies involving chromosomes 3,5,7,9,11,15,19,and 21, and infrequently have one of the five recurrent translocations. Expresson arrays show that virtually all MM tumors have dysregulated and/or increased expression of CYCLIN D1, D2, or D3, providing an apparent early, unifying event in pathogenesis. The patterns of Translocations and CYCLIN D expression enabled us to propose a novel TC classification that includes seven groups: 11q; 6p; MAF; 4p, D1; D1+D2; and D2. The hyperdiploid D1 group is absent in extramedullary MM and HMCL, suggesting a particularly strong dependence on interaction with the bone marrow microenvironment.
Showing the most recent 10 out of 19 publications
