Characterization of Nonrandom Chromosomal Gains and Losses in Multiple Myeloma by Comparative Genomic Hybridization

Bone marrow samples were collected from 30 patients, 25 with MM, 4 with MGUS, and 1 with WM. Of these, 17 were ascertained at the University Clinic of Navarra (UCN; Pamplona, Spain), and the remaining 13 were ascertained at the Memorial Sloan-Kettering Cancer Center (MSKCC; New York, NY). The samples were derived from 22 at diagnosis and 8 previously treated patients (Table 1). The ages of the patients ranged from 33 to 77 years, with a median of 59 years. Among the 30 patients, 16 were women and 14 were men. Cytogenetic analysis at UCN included unstimulated short-term and B-cell–stimulated 48-hour cultures,12 whereas cytogenetic analysis at MSKCC was performed on unstimulated short-term cultures of biopsy samples. G-banding by standard procedures was used in both laboratories. Karyotypes of cases no. 16, 17, and 19 were previously reported.6 

Tumor DNA was extracted from bone marrow samples and subjected to CGH analysis essentially as described.11 Briefly, the tumor (test) and normal (reference) DNAs were labeled by nick-translation with fluorescein-12-dUTP and Texas Red-5-dUTP (NEN-DuPont, Boston, MA), respectively. Equal amounts (200 ng) of tumor and normal DNAs were coprecipitated with 10 mg of human Cot-1 DNA (GIBCO/BRL, Gaithersburg, MD) and resuspended in the hybridization mix before in situ hybridization to human metaphase chromosome spreads prepared from phytohemagglutinin-stimulated lymphocytes from normal individuals. After hybridization, the slides were washed and the chromosomes were counterstained with 4,6-diamino-2-phenylindole (DAPI) to enable identification of the chromosomes. Fluorescent hybridization signals and DAPI-staining patterns were captured with a cooled charge-coupled device (CCD) camera (Photometrics, Tuscon, AZ) attached to a Nikon Microphot-SA microscope and processed using an image analysis system (Quips, Vysis, IL). The software performed a calculation of the green (tumor DNA) to red (normal DNA) fluorescent ratios along the length of each chromosome. The average of readings from eight chromosomes were graphed for each chromosome and compared with the profile for the same chromosome in a reference DNA/reference DNA hybridization to set the boundaries of gain and loss. Ratios greater than 1.20 and less than 0.80 were considered to represent chromosomal gain and loss, respectively. These threshold levels were tested as reported previously by us.12 CGH detects DNA sequence copy number relative to the average copy number in the tumor but not the ploidy level of the tumor.10 Therefore, the ploidy level of the tumor in this study was determined by conventional cytogenetics (Table 1). Chromosomal regions near the centromeres of 1, 9, 13-16, 21, and 22 were not scored for CGH analysis because of the highly repeated sequences in these regions. In the present study, control DNA was always obtained from a male donor, and the ratios for gains and losses of sex chromosomes were appropriately adjusted when test DNA was not matched. Recurrence of a change was defined by its presence in 2 or more tumors.

Thirteen of the 30 (43%) patients showed clonal chromosomal abnormalities by G-banding (Table 1). Karyotypes were hyperdiploid (30%), pseudodiploid (47%), or hypodiploid (23%). A 14q+ marker chromosome was detected in 5 cases (38%); in 3 of them, it was derived from the t(11;14)(q13;q32) translocation. Among the remaining chromosomal abnormalities seen, recurrent changes comprised del(11) and del(16) in 3 cases each and −X and +1q in 2 cases each.

The proportion of patients with chromosomal aberrations increased to 21 by CGH (70%; Table 1). The clonally abnormal cases included 18 MM, 2 MGUS, and the single WM cases. The proportion of bone marrow plasma cells in cases that were clonally abnormal by CGH ranged between 16% to 98%. Overall, 10 chromosomal changes were recurrent (Fig 1). The most frequent changes were complete or partial deletions of chromosome 13 and gain of chromosome 19 or 19p (30%). Other frequent losses were complete or partial deletions of 6q (13%) and 16 (17%), with the commonly deleted regions spanning bands 6q21 and 13q14-21 (Fig 1). The partial karyotypes displayed in Fig 2 show representative losses of 13q, 6q, and 16q. None of the cases studied showed any of the recurrent losses detected to be unique aberrations (Table 1). Recurrent gains included partial gains of 11q (20%) and, less frequently (∼10%), gains of 9q, 12q24, 15q23-qter, 17q22-24, and 22q. Among gains, 2 were unique changes: +19 in cases no. 4 and 7 and +11q13-21 in case no. 30. Representative partial karyotypes of cases no. 18 and 26, showing +9, +11q, and +15, are shown in Fig 2.

Comparison of abnormalities detected by G-banding versus CGH in the cases in which both types of data were available (Table 1) showed a variation in the degree of concordance between the two methods of analysis. A subset of 6 cases (1, 2, 3, 17, 26, and 30) showed a high level of concordance between the abnormalities detected by the two techniques; in contrast, the remaining 7 cases (6, 7, 9, 16, 19, 22, and 27) showed discrepancy in almost every single abnormality. In addition, the types of abnormalities detected in different proportions of cases also seemed to depend on the technique used. Thus, rearrangements of 11q were observed by G-banding in 5 of the 6 cases that showed this anomaly by CGH; in contrast, only 1 case with abnormal karyotype by G-banding showed monosomy 13 compared with 9 cases in which this anomaly was detected by CGH.

Cytogenetic analysis of MM and related disorders has so far been hampered by the low proliferative activity of plasma cells, with abnormal karyotypes having been noted in 30% to 50% of cases.2-6,13 Interphase fluorescence in situ hybridization studies using centromeric probes have shown numerical chromosomal aberrations in 80% to 90% of cases, suggesting that clonal chromosomal abnormalities are frequent in these disorders.14,15 Therefore, we reasoned that application of CGH would provide new information on gains and losses of chromosomal regions in MM and related disorders, because this technique is performed using small amounts of DNA rather than cultured cells. In the present study, 70% of tumors showed clonal chromosomal changes. Thus, CGH was more sensitive in detecting chromosomal copy number changes than conventional G-banding. The 6 cases (12, 13, 20, 21, 25, and 29) that did not show clonal chromosomal abnormalities by CGH shared some features: they were obtained at diagnosis, the plasma cell infiltration in them was either less than 30% or they were diagnosed as MGUS, and they showed normal karyotype by G-banding analysis. We have previously described a similar association between absence of cytogenetic aberrations and lack of prior treatment or less than 30% plasma cell infiltration.6 12 A possible explanation for this association, and thus the trend seen in the CGH results, is that at the early stages of the disease, genetic changes may not involve recognizable chromosomal alterations. However, cases no. 10, 11, and 24, which presented a high proportion of plasma cells in the bone marrow, are not compatible with this explanation. These tumors may carry a chromosomal translocation(s) and/or small deletion (<2 kb) that cannot be detected by CGH.

Overall, we observed a variable correlation between the results of G-banding and CGH analyses. Thus, we noted 10 recurrent changes (Fig1), all of which have been previously described by us and others in MM.1-6 13 We have identified 9 cases (30%) that showed total or partial gains of chromosome 19 by CGH. This change has been one of the most frequent in our series and it has been noted as a unique anomaly in 2 cases. However, by G-banding, this change was detected only in 2 cases. The significance of this trisomy in MM development remains to be elucidated. Partial loss of chromosome 13 (also seen in 30% of the cases) was missed by conventional cytogenetic analysis in 8 of the 9 cases in which it was detected by CGH. We found some other discrepancies also in the results of the 13 cases in which both techniques were successful (Table 1). These discrepancies can be explained by the fact that G-banding analysis is based on the study of chromosomes of the clone proliferating in vitro, which may or may not be the predominant tumor clone, and hence may or may not be representative of the CGH result.

The CGH analysis allowed us to narrow the chromosome 13 deletion to a common region that spans the bands 13q14-21. The same region has previously been reported to be frequently deleted in chronic lymphocytic leukemia (CLL) and in 1 case of MM.16-18 Therefore, the candidate tumor-suppressor gene proposed in CLL and other lymphoid malignancies may be involved in the genesis and/or progression of MM as well. In the same way, deletions affecting 6q have been reported in MM and other lymphoid malignancies.19 We have identified the common region of deletion at 6q21, a region that has also been identified in a subset of non-Hodgkin's lymphoma by G-banding and LOH studies.20Finally, 16q loss has been observed in a subset of patients in our series; del(16)(q22) has previously been reported as unique change in MM.6 Therefore, candidate tumor-suppressor genes at 6q21 and 16q22 may be of importance in the pathogenesis of MM.

Using CGH, we have shown that 70% of MM biopsies present recurrent chromosomal gains and losses. These included deletions of 13q and gains of 11q that have previously been associated with a poor outcome.7 8 These abnormalities can be easily detected by CGH, without relevance to G-banding analysis, thus providing a valuable approach to identifying prognostically significant lesions using small amounts of DNA from nondividing cells.