Leukemia (2003) 17, 427–436 2003 Nature Publishing Group All rights reserved 0887-6924/03 $25.00 www.nature.com/leu Chromosome abnormalities clustering and its implications for pathogenesis and prognosis in myeloma CS Debes-Marun1, GW Dewald2, S Bryant3, E Picken1, R Santana-Da´vila1, N Gonza´lez-Paz1, JM Winkler1, RA Kyle1,2, MA Gertz1, TE Witzig1, A Dispenzieri1, MQ Lacy1, SV Rajkumar1, JA Lust1, PR Greipp1 and R Fonseca1 1Mayo Clinic Division of Hematology, Mayo Clinic, Rochester, MN, USA; 2Department of Laboratory Medicine and Pathology, Mayo Clinic, Rochester, MN, USA; 3Division of Biostatistics, Mayo Clinic, Rochester, MN, USA The nonrandom recurrent nature of chromosome abnormalities eral partner chromosomes, and the most common of these in myeloma suggests a role for them in disease pathogenesis. translocations include t(11;14)(q13;q32), t(4;14)(p16.3;q32), We performed a careful cytogenetic analysis of patients with 9,10 1,12,13 abnormal karyotypes (n = 254), to discern patterns of associ- and t(14;16)(q32;q23). Aneuploidy is ubiquitous, and ation, search for novel abnormalities and elucidate clinical monosomy of chromosome 13 is seen in nearly one-half of implications. Patients with karyotypic abnormalities suggestive patients.10,14 Little is known about the role of other chromo- of myelodysplasia/acute leukemia were excluded. In this study some anomalies in myeloma. we compared survival by abnormality only between patients Investigations of the relationship between disease with abnormal karyotypes. Patients with abnormalities were progression/pathogenesis and chromosome evolution have more likely to have features of aggressive disease as compared to all other patients without abnormalities entered into the not been easy because the karyotype of neoplastic cells in myeloma often involves many numeric and structural abnor- myeloma database (lower hemoglobin, higher ␤2-microglobu- lin, labeling-index and plasmocytosis; all P Ͻ 0.0001). Several malities.3,7 In addition, some chromosome abnormalities in groups of patients could be readily identified; hypodiploid myeloma are cryptic to conventional cytogenetic studies (for (22%), pseudodiploid (36%), hyperdiploid (31%) and near-tetra- example t(4;14)(p16.3;q32)), while other chromosome abnor- ploid (11%). Clustering associations were seen among several malities in myeloma are subtle and easily overlooked by con- trisomies and monosomy of chromosome 13 and 14. Several monosomies (−2, −3, −13, −14 and −19), 1p translocations/ ventional cytogenetic studies (for example t(14;16)(q32;q23)). deletions, and hypodiploidy were associated with a signifi- Other abnormalities, such as those of chromosomes 1 are cantly shorter survival. Trisomy of chromosome 13 was rare recurrent and further support the hypothesis that certain recur- (<2%). Even among patients with abnormal karyotypes, specific rent chromosome abnormalities in myeloma are nonran- chromosome abnormalities can impart biologic variability in dom.3,15 Moreover, several investigators suggest that the myeloma, including several monosomies, hypodiploidy and observed clinical heterogeneity among patients with myeloma abnormalities of 1p. Leukemia (2003) 17, 427–436. doi:10.1038/sj.leu.2402797 is likely to be related to the underlying genetic and cyto- 2,16 Keywords: multiple myeloma; prognosis; translocations (genetics); genetic abnormalities and certain specific chromosome chromosome deletion; ploidy abnormalities are known to be associated with adverse out- come.8,17–19 To perform a comprehensive karyotype analysis in mye- Introduction loma, we studied clustering patterns of specific chromosome abnormalities detected by conventional karyotype analysis in As shown by studies using interphase fluorescence in situ a large cohort of patients with myeloma to look further for hybridization (FISH), all patients with multiple myeloma or clues to the pathogenesis of this disorder, and to address their monoclonal gammopathy of undetermined significance have clinical implications. The implications of these abnormalities chromosome abnormalities in the clonal plasma cells.1,2 Stud- were only considered among patients with abnormal meta- ies using conventional chromosome studies almost never phases, thus removing the effect of comparing these patients detect neoplastic metaphase cells in monoclonal gammopathy to those in whom an abnormal karyotype cannot be detected. of undetermined significance, and occasionally will detect abnormalities in myeloma.3–6 Chromosomally abnormal clones can be found in 18% to 35% of myeloma patients at Patients and methods diagnosis, 40% to 60% with aggressive disease and up to 85% of patients with plasma cell leukemia.3–6 Clearly, the potential Patient demographics of conventional cytogenetic studies to detect an abnormal clone in myeloma is associated with aggressiveness of dis- All patients entered in the clinical cytogenetics database of ease.3–7 Studies assessing the clinical implications of chromo- Mayo and who were diagnosed with multiple myeloma were somal abnormalities in myeloma, detected by karyotype included in this study. Patients who had a previous diagnosis analysis, need to emphasize that an abnormal karyotype is of myeloma and were subsequently diagnosed with myelodys- more likely to be encountered in a highly proliferative clone.8 plasia or acute leukemia, and who exhibited the characteristic Several recurrent, nonrandom abnormalities have been chromosome abnormalities for these entities (eg − − detected in myeloma.9,10 Among them, translocations involv- t(1;7),pseudodiploid and 7, or 5 occurring in isolation) were ing the immunoglobulin heavy-chain (IgH) locus have been excluded from further study. Our search yielded a total of 254 implicated in the origin of myeloma11 and are seen in 50– patients with chromosome abnormalities, which form the 60% of patients. Translocations at the IgH locus involve sev- basis for this report. We compared results of these patients with the results of 3483 patients with various plasma cell pro- liferative disorders who had only normal metaphases entered Correspondence: R Fonseca, Mayo Clinic Hematology/ST 6-28, Roch- into the Mayo Clinic Dysproteinemia database. Pertinent clini- ester, MN 55905, USA; Fax 507-266-9277 cal and prognostic features are available for these patients ␤ Received 18 July 2002; accepted 25 September 2002 including among others the plasma cell labeling index, 2- Chromosome clustering in myeloma CS Debes-Marun et al 428 Table 1 Prevalence of chromosomal abnormalities in myeloma Cases Controls P value Patients (n) 254 3483 Age (years) 59 (31–89) 64 (17–96) <0.0001 Males (%) 153 (60%) 2075 (60%) 0.90 Caucasian race (%) 243 (96%) 3363 (97%) 0.38 Serum M-spike (range) 3.58 (0.2–10.4) 2.90 (0.1–15.1) <0.0001 IgG:IgA (ratio) 1.89:1 2.84:1 0.051 ␬␭ (ratio) 1.46:1 1.77:1 0.34 Urine M-spike (present) 0.31 (0–21.13) 0.23 (0–16.99) 0.22 Hemoglobin (g/dl) 10.6 (4.5–15.8) 11.3 (2.7–19.6) <0.0001 Calcium (mg/dl) 9.4 (6.9–17.5) 9.4 (6.6–18.1) 0.54 Creatinine (mg/dl) 1.2 (0.4–11.0) 1.1 (0.4–20.1) 0.02 ␤ 2-microglobulin (mg/dl) 3.95 (1.08–63.8) 3.0 (0.7–82.0) <0.0001 PCLI (%) 0.8 (0–15.6) 0.3 (0–41.0) <0.0001 Bone marrow plasma cells (%) 55.5 (2–99.9) 30 (0–100) <0.0001 Bone lesions (%) 157 (66%) 1862 (57%) 0.04 Numbers represent median and (range). microglobulin, hemoglobin, serum creatinine and bone Clustering analysis: We classified karyotypic abnormali- marrow plasmacytosis. ties in a binomial fashion (present/not), independent of the number of times the abnormality was present. For a specific abnormality to be included it had to be present in at least Statistical analysis seven patients (3%). Clustering analysis was performed on all data using methods of average linkage, single linkage and Descriptive and survival statistics: Descriptive statistics complete linkage using two measures of distance, Kendall’s were used to characterize patients in this study. To test for tau and Jaccard’s coefficient.24 Six cluster trees resulting from association between abnormalities the Fisher’s exact test20 this analysis were compared. was used for nominal variables, and the Wilcoxon rank sum test was used21 for continuous variables. Overall survival was estimated using the methods of Kaplan and Meier.22 The log- Results rank test was used to test for differences in survival between groups.23 Survival time was always calculated from the time Patient population of diagnosis. Patients were assigned to one of four categories; pseudodiploid for those having 45–46 chromosomes, hypodi- We compared results of 254 patients with multiple myeloma ploid if they had 44 or less chromosomes, hyperdiploid if they who had chromosomal anomalies with the results of 3483 had 47 to 74 chromosomes, and near-tetraploid or tetraploid patients with various plasma cell proliferative disorders who if they had 75 or more chromosomes. had only normal metaphases. Clinical and demographic vari- The following clinical prognostic factors were assessed as ables of these patients are summarized in Table 1. Patients continuous (or dichotomized) variables according to previous with detectable chromosomal anomalies were slightly reports: serum calcium (>12 mg/dl), hemoglobin (>12 g/dl), younger (P < 0.0001), had a higher serum monoclonal spike Ն ␤ ␤ plasma cell labeling index (PCLI) (<1 vs 1%), 2- (P < 0.0001), lower hemoglobin (P < 0.0001), higher 2- microglobulin (у2.7, and Ն4 mg/dl), serum creatinine (>2 mg/dl), the presence or absence of bone lesions and the bone marrow plasma cell percentage. Table 2 Prevalence of specific numerical chromosomal abnor- malities Chromosome 1 2345678 Monosomy % 9 9 4 15 10 9 11 21 Trisomy % 4 4 22 4 13 6 15 4 Chromosome 9 10 11 12 13 14 15 16 Monosomy 5 17 12 16 43 28 7 21 Trisomy % 25 2 20 3 2 3 22 1 Chromosome 17 18 19 20 21 22 X Y Monosomy % 19 12 8 17 9 23 22 17 Figure 1 Distribution of total chromosome number among Trisomy % 2 9 22 3 11 3 3 1 patients with karyotype abnormalities.