Fine mapping and expression analysis of chromosome 1 with the aim of defining critically deregulated genes important in the pathogenesis of myeloma

Deletions on chromosome 1p and gains on chromosome 1q are common alterations in Myeloma (MM). We analyzed 78 cases using Affymetrix 500K SNP and U133 Plus 2 expression arrays to characterize abnormalities of this chromosome. Copy number, loss of heterozygosity (LOH) analysis and supervised hierarchical clustering (HC) were performed using dChipSNP. LOH and deletion of 1p were found in 22/78 cases (28%). Of these 6 had deletion of the whole of 1p, 9 cases had an interstitial deletion (1p12–p21.3), median size 47.2 Mb (0.2–50.5), and 7 cases had more than one interstitial deletion, median size 5.3 Mb (0.5–55.6). Supervised HC of expression data identified 324 differentially expressed genes comparing deletions of 1p with normal cases. Of these 12 genes were located at 1p12–p21.3, all of which were underexpressed in the samples with deletion. We found one small homozygous microdeletion located at 1p32.3 containing only one transcript CDKN2C . Homozygous deletions by definition contain genes which are completely inactivated and as such are likely to be relevant to the MM pathogenesis. We extended this analysis looking for other alterations at this region and homozygous deletions, mean 0.41 Mb (0.03–1.74), were seen in 4 samples at 1p32.3 (5%). These cases together with another 6 (13%) with deletion at 1p32.3 define a minimally altered region within which CDKN2C is located and in all cases is underexpressed. This data suggests that inactivation of the INK4a/ARF pathway may be important. We looked at other genes in this pathway but did not find homozygous deletions in any of them, however, we did identify another 5 cases with hemizygous deletion. The TC classification utilises data on cyclin D expression and as CDKN2C directly regulates these complexes we looked at how these changes were distributed relative to expression of cyclin Ds. We did not find a consistent pattern as alterations were found in both the elevated D1 and D2 groups. Further analysis of the expression patterns in this pathway and how they relate to genomic changes detected with the SNP array is ongoing. Gains of 1q were identified in 25/78 cases (32%), with gain of the whole of 1q being responsible for the majority. Supervised HC between samples with and without 1q gains identified 659 differentially expressed genes. Of these, 83 genes were located in 1q21–q43, all of which were overexpressed in 1q+. We identified a single case with a small defined region of gain (up to 4 copies) covering 17.5 Mb at 1q21.1–q23.3. Looking within this region following HC we found 42 genes overexpressed (1.3–32.4 fold), including MCL1 and NTRK1, compared to other 1q+ cases. We analyzed our dataset for previously described genes altered in 1q. While we found upregulation of CKS1B , BCL9 and PDZK1 we could not distinguish these from other genes located within this amplicon. Gain of 1q has been associated with poor outcome and we sought other regions co-segregating with 1q+; del13 was the most frequently associated change and all cases with both alterations had an IGH translocation, which may contribute to the poor prognosis of this subgroup. We confirm the importance of changes on chromosome 1 in the MM pathogenesis and provide data supporting the critical deregulation of the G1S cell cycle transition as being important and frequently affected by deletions of 1p affecting CDKN2C .