Correspondence 2533 Disruption of INK4/CDK/Rb cell cycle pathway by gene hypermethylation in multiple myeloma and MGUS

Leukemia (2003) 17, 2533–2535. doi:10.1038/sj.leu.2403133 examination, skeletal survey, serum and urine protein electro- Published online 11 September 2003 phoresis, serum immunoglobulin (IgG, IgA and IgM) levels, renal function tests and serum calcium level. Monoclonal proteins were identified by cellulose acetate or agarose-gel TO THE EDITOR electrophoresis; if there was an abnormal band or equivocal Pathogenesis of cancer involves a stepwise accumulation of pattern, immunoelectrophoresis or immunofixation was per- genetic alterations that disrupt either the normal machinery formed. Of the 34 patients with MM, 15 (44.1%) were female of cell proliferation or its regulation by the activation of with a median age of 61.5 years (range 25–87). There were 21 cellular oncogenes or inactivation of tumor suppressor genes.1 (61.8%), seven (20.6%), five (14.7%) and one (2.9%) with IgG, Cellular proliferation is mediated by progression through the IgA, light chain and IgD MM. Five (14.7%), 10 (29.4%) and 19 cell cycle, where two cell cycle checkpoints are located at (55.9%) patients belonged to Durie–Salmon stage 1, stage 2 and G1S and G2M. Quiescent cells in G0 phase contain hypopho- stage 3 disease, respectively. Eight (23.5%) had impaired renal sphorylated RB, which sequesters the transcription factor E2F. function at diagnosis. The median paraprotein level at diagnosis Upon activation by mitogens, upregulation of D-type cyclins was 29.8 g/l (range 10–78). Of the six patients with MGUS, there results in activation of cyclin-dependent kinases (CDK), namely were three each of male and female patients. The median age CDK4 and CDK6, resulting in phosphorylation of RB. was 82 years (range 44–85). All had monoclonal gammopathy Hyperphosphorylated RB results in release of E2F, which of IgG subtype (three each of IgG/kappa and IgG/lamda). None activates transcription of S1-specific genes, and irreversible of the MGUS patients progressed to myeloma, amyloidosis or commitment to cell cycle progression. Negative regulatory other lymphoproliferative disease after a median follow-up time elements to the cell cycle progression include the INK4 (p15, of 3 years (range 1–7 years). p16, p18 and p19) and the CIP/KIP (p21CIP, p27KIP1 and p57KIP2) High molecular weight genomic DNA was isolated by families of proteins that give rise to cell cycle arrest.1 P15, p16, standard protocols from bone marrow aspirates of 34 patients p18 and p19 are coined INK4 family CDK inhibitors (CKI) with MM, six patients with MGUS and AF10 cell line (kindly as they share similar functional domains (ankyrin repeats provided by Professor Zelig Eshhar, Department of Immunology, for protein–protein interactions) that render them capable of Weizmann Institute of Science, Israel). The methylation-specific competing with cyclin D for CDK4 and 6, and thus inhibition of polymerase chain reaction (MSP) for gene promoter methylation CDKs 4 and 6. was performed as described in detail previously.3 Briefly, Multiple myeloma (MM) is characterized by neoplastic treatment of DNA with bisulfite for conversion of unmethylated, proliferation of monoclonal plasma cells. These neoplastic but not methylated, cytosine to uracil was performed with a plasma cells are believed to arise from a postgerminal center commercially available kit (CpGenome DNA modification kit, B cell, which migrates to the bone marrow, adheres to the Intergen, New York, USA) according to the manufacturer’s marrow stroma, and triggers subsequent bone resorption and a instructions. The primers for the methylated (M-MSP) and 2 paracrine cytokine loop. The natural course of disease may unmethylated (U-MSP) gene promoter regions are shown in progress through monoclonal gammopathy of undetermined Table 1. DNA from eight normal donors was used as negative significance (MGUS) to MM. MGUS is found in approximately control, while methylated control DNA (CpGenome Universal 3% of persons older than 70 years and in 1% of those 50 years or Methylated DNA, Intergen) was used as positive control in all older. During long-term follow-up, approximately one-fourth of the experiments. MSP was performed in a thermal cycler (9700, patients develop MM, amyloidosis, macroglobulinemia, or other PE Biosystems, Foster City, CA, USA). The PCR mixture malignant lymphoproliferative disorders with a median interval contained 50 ng of bisulfite-treated DNA, 0.2 mM dNTPs, 2 mM of 10 years.2 MgCl2, 10 pmol of each primer, 1 Â PCR Buffer II and 2.5 U Aberrant gene promoter methylation is a mechanism whereby AmpliTaq Gold (Perkin-Elmer Biosystems) in a final volume of 3 gene expression is abrogated. It is therefore potentially 50 ml. The identity of the methylated and unmethylated involved in silencing of tumor suppressor genes during 3 sequences was confirmed by automated DNA sequencing. carcinogenesis. We studied if disruption of the INK4/cyclin PCR products were gel purified, sequenced bidirectionally (ABI D-CDK/RB pathway is a common mechanism in the pathogen- Prism dRhodamine Terminator Cycle Sequencing Kit, PE esis of MM and MGUS. Biosystem), and analyzed on an automated DNA sequence Diagnosis of MGUS is made in an asymptomatic patient with analyser (377 ABI Prism, PE Biosystem). monoclonal gammopathy with a low level of M-component None of the four genes tested were methylated in eight (IgG p3.5 g/dl, IgA p2 g/dl, and Bence Jones protein p1g/ normal bone marrow samples. Normal marrow DNA showed 24 h), less than 10% plasma cells in the bone marrow and unmethylated status by positive amplification in U-MSP but lack absence of lytic bone lesions. Diagnosis of MM was made of amplification in M-MSP, and methylated positive control according to standard criteria with marrow plasmacytosis, DNA showed complete methylation by positive amplification in presence of paraprotein and osteolytic bone lesions. Complete M-MSP but not by U-MSP. The specificity of the MSP was staging work-up for both MM and MGUS includes bone marrow verified by DNA sequencing of methylated positive control. DNA sequence of methylated control DNA is aligned and Correspondence: CS Chim, University Department of Medicine, compared with germline sequence of the wild-type DNA (WT). Queen Mary Hospital, University of Hong Kong, Hong Kong. Fax: (852)29741165; E-mail: [email protected] Methylated cytosine residues in CpG dinucleotide remained as Received 29 April 2003; accepted 8 August 2003; Published online 11 ‘C’, whereas unmethylated cytosine read as ‘T’ after bisulfite September 2003 conversion.

Leukemia Correspondence 2534 Table 1 Methylation-specific polymerase chain reaction: primer sequences and reaction conditions

Gene Forward primer Reverse primer Tm/cycles Size (bp) location

p15 M-MSP GCG TTC GTA TTT TGC GGT T CGT ACA ATA ACC GAA CGA CCG A 631C/35 148 9p21 U-MSP TGT GAT GTG TTT GTA TTT TGT GGT T CCA TAC AAT AAC CAA ACA ACC AA 154 p16 M-MSP TTA TTA GAG GGT GGG GCG GAT CGC GAC CCC GAA CCG CGA CCG TAA 651C/35 150 9p21 U-MSP TTA TTA GAG GGT GGG GTG GAT TGT CAA CCC CAA ACC ACA ACC ATA A 151 p18 M-MSP TTATCGAATTGTTATTTTCGTTCG CGTCTCGCCGAAAAAATAATC 641C/35 93 1p32 U-MSP GGGTTATTGAATTGTTATTTTTGTTTG CATCCATCTCACCAAAAAAATAATC 95 RB M-MSP GGGAGTTTCGCGGACGTGAC ACGTCGAAACACGCCCCG 661C/35 152 13q14 U-MSP GGGAGTTTTGTGGATGTGAT ACATCAAAACACACCCCA 152 Tm: annealing temperature; M-MSP: methylation-specific polymerase chain reaction for the methylated allele; U-MSP: MSP for the unmethylated allele.

At diagnosis, p16 hypermethylation occurred in 18 A similar frequency of p15 hypermethylation in both (52.9%) MM and none of the MGUS (P ¼ 0.02). Hypermethyla- MGUS and MM suggested p15 methylation to be an early tion of p15 occurred in 12 (35.3%) MM patients and three event. On the other hand, in contrast to the frequent (50%) MGUS patients (P ¼ 0.68). Concurrent p15 and methylation of p16 in MM,4 there was absence of methylation p16 methylation occurred in six (17.6%) of patients with of p16 in MGUS (P ¼ 0.02). Therefore, p16 methylation MM. None of the patients with MM and MGUS had methy- might be an event associated with progression to MM. lation of p18 or RB. AF10 had completely unmethylated Indeed, this notion was supported by a study of p16 methylation p15, p18 and RB but biallelic methylation of p16. In some in a large cohort of patients with MGUS (N ¼ 40), MM (N ¼ 98) patients with methylated DNA, the presence of amplification by and plasma cell leukemia (N ¼ 5),6 which showed frequent U-MSP might be due to contamination by other cells in the p16 methylation in 40% of MM and 80% of plasma cell marrow biopsy. However, hemizygous methylation cannot be leukemia, but not in any of the patients with MGUS. excluded. Indeed, the absence of unmethylated band in the Furthermore, methylation studies in myeloma, acute leukemia homogeneous cell line indicates biallelic methylation of p16 and esophageal cancer also suggested p16 methylation to be gene. associated with clonal evolution and disease progression. Ideally, MSP status of the neoplastic plasma cells will only be Therefore, the role of p16 methylation as a marker of elucidated if marrow cells have been sorted for CD38- or progression to MM warrants further study. CD138-positive cells. Here, MSP of the four genes has been first p18 is important in the terminal differentiation of B validated in eight normal control DNA by demonstration of the lymphocytes into plasma cell by the induction of cell cycle lack of methylation, and further verified by sequencing. Given arrest.7 Although frequent homozygous deletion (38%) of p18 that methylation detected by MSP is a positive signal with a high genes has been reported in myeloma cell lines, deletion of the sensitivity (up to 1 Â 105 for p15 gene),3 our results are still valid p18 has been identified in only 6% of patients with MM.7 There without sorting of the marrow for plasma cells. is as yet no data on p18 gene methylation in MM, and our study Our study showed concurrent hypermethylation of p15 showed absence of p18 hypermethylation in primary MM, and p16 in a significant number of MM patients. This concurred MGUS and AF10 myeloma cell line. with another study that reported frequent concurrent hyper- RB, located in 13q14, has also been shown to be involved methylation of both genes in myeloma patients.4 This may in an autocrine circuitry of IL-6-mediated myeloma cell imply that p15 and p16 methylation cooperated to disrupt growth.8 Moreover, hemizygous RB deletion, due to 13q the INK4/cyclin D1/RB cell cycle pathway. Indeed, one study deletion, occurs frequently in MM.8 However, in contrast in mantle cell lymphoma has shown disruption of this to other studies that showed hypermethylation to co- pathway either by concomitant hypermethylation of both operate with hemizygous gene deletion for biallelic inacti- p15 and p16, or concurrent p15 and p16 deletion. Similarly, vation of tumor suppressor genes,8 our data did not show another study showed frequent p16 hypermethylation together an important role of hypermethylation for complete suppression with either RB deletions or cyclin D1 overexpression in of RB. myeloma. Nonetheless, despite that both p15 and p16 inhibit In summary, both p15 and p16, but not p18 and RB, were cyclin-dependent kinases 4/6 by competing with cyclin D, frequently methylated. p15, but not p16, p18 and RB, was previous studies suggest p15, but not p16, is upregulated frequently methylated in MGUS. p16 hypermethylation might in a TGFb-dependent manner.4 Therefore, concurrent inactiva- be a marker of progression to MM. tion of p15 and p16 by hypermethylation, although members of the same cell cycle control circuitry, might represent inhibition of cell cycle pathway mediated by biologically distinct signals. Secondly, p15 gene is frequently codeleted Acknowledgements in leukemic cells with p16 deletion because of their close proximity, as p15 and p16 are located within 20 kb We thank Professor Eshhar, Department of Immunology, Weiz- of each other.5 This might also account for the high fre- mann Institute of Science, Israel, for provision of AF10 cell line; quency of concurrent methylation of both p15 and p16 in and Department of Pathology, Queen Mary Hospital, for the myeloma. diagnoses; and Kadoorie Foundation for the funding support.

Leukemia Correspondence 2535 1 CS Chim 1University Department of Medicine, Queen inhibitor gene in multiple human cancers. Nature 1994; 368: 1 TK Fung Mary Hospital, University of Hong Kong, 753–756. R Liang1 Hong Kong 6 Mateos MV, Garcia-Sanz R, Lopez-Perez R, Balanzategui A, Gonzalez MI, Fernandez-Calvo J et al. P16/INK4a gene References inactivation by hypermethylation is associated with aggressive variants of monoclonal gammopathies. Hematol J 2001; 2: 1 Malumbres M, Barbacid M. To cycle or not to cycle: a critical 146–149. decision in cancer. Nat Cancer Rev 2001; 1: 222–231. 7 Kulkarni MS, Daggett JL, Bender TP, Kuehl WM, Bergsagel PL, 2 Kuehl WM, Bergsagel PL. Multiple myeloma: evolving genetic Williams ME. Frequent inactivation of the cyclin dependent kinase events and host interactions. Nat Rev Cancer 2002; 2: 175–187. inhibitor p18 by homozygous deletion in multiple myeloma cell 3 Chim CS, Liang R, Tam C, Kwong YL. P15 and P16 promoter lines: ectopic p18 expression inhibits growth and induces apoptosis. methylation in acute promyelocytic leukemia. J Clin Oncol 2001; Leukemia 2002; 16: 127–134. 19: 2033–2040. 8 Fonseca R, Harrington D, Oken MM, Dewald GW, Bailey RJ, 4 Ng MH, Chung YF, Lo KW, Wickham NW, Lee JC, Huang DP. Van Wier SA et al. Biological and prognostic significance Frequent hypermethylation of p16 and p15 genes in multiple of interphase fluorescence in situ hybridization detection of myeloma. Blood 1997; 89: 2500–2506. chromosome 13 abnormalities (delta13) in multiple myeloma: 5 Nobori T, Miura K, Wu DJ, Lois A, Takabayashi K, an eastern cooperative oncology group study. Cancer Res 2002; Carson DA. Deletions of the cyclin-dependent kinase-4 62: 715–720.

High incidence of trisomies 1q, 9q, and 11q in multiple myeloma: results from a comprehensive molecular cytogenetic analysis

Leukemia (2003) 17, 2535–2537. doi:10.1038/sj.leu.2403153 From August 1998 to November 2002, bone marrow aspirates Published online 2 October 2003 from 357 consecutive patients diagnosed with MM were obtained for molecular cytogenetic analysis during routine TO THE EDITOR diagnostic procedures. Of the 357 patients, 290 (81%) were entered on two national treatment trials and were newly Genomic abnormalities are emerging as crucial prognostic diagnosed. The remaining patients had previous therapy. parameters in multiple myeloma (MM). Deletion of chromo- Informed consent was obtained from all patients. There were some arm 13q (13qÀ) is present in 40–60% of tumors and has 141 female (39.5%) and 216 male patients (60.5%). Almost all been recognized as an independent predictor of a dismal patients had advanced stage myeloma (Salmon–Durie stage: II – prognosis. Using chromosome banding analysis, only about a 21%, III – 76%; others: 3%). At the time of specimen collection, third of myelomas exhibit an abnormal karyotype. Moreover, the ages of the patients ranged from 28 to 76 years (median, 56 evaluable metaphases are missing in up to one-fourth of cases, years). Mononuclear cells were separated by density-gradient as shown in large cytogenetic studies. By molecular cytogenetic centrifugation and cytospin preparations were performed. The techniques, for example, comparative genomic hybridization quality of the cells and the proportion of tumor cells in the bone (CGH) or fluorescence in situ hybridization (FISH), genomic marrow samples were assessed by morphology on a May– changes can be identified in more than 90% of plasma cells Gru¨nwaldFGiemsa-stained cytospin preparation. Plasma cell primary tumors. While banding analyses fail to identify certain enrichment using immunomagnetic beads directed against the chromosomal aberrations, for example, small deletions or CD138 antigen was performed in 29 specimens with a low translocations with breakpoints located near the telomere, such tumor cell fraction. Tumor samples of insufficient cell quality (eg abnormalities can be reliably detected by FISH. Therefore, FISH due to prolonged shipment to our central reference laboratory) has become the most widely used tool for molecular diagnostics or with o1% tumor cells after positive selection of plasma cells in MM today. were not analyzed. Depending on the tumor cell light chain Analogous to the acute and chronic leukemias, genomic type, overnight incubation with the respective AMCA-labeled aberrations most likely represent the biological basis of the light chain antibody was used for the detection of myeloma cells clinical heterogeneity, that is also characteristic for MM. While during fluorescence microscopy. At least 100 AMCA-positive our understanding of the implication of various translocations in plasma cells were evaluated. MM has gradually increased,1,2 little is known about chromo- Based on our CGH study that resulted in the identification of somal imbalances in this disease. Even the exact incidence of genomic regions commonly gained or deleted in MM,3 a most chromosomal gains and losses in MM are unknown to comprehensive set of DNA clones for the detection of the most date. To address this issue, we developed a comprehensive, frequent chromosomal imbalances was assorted. To achieve disease-specific DNA probe set for the evaluation of chromo- high accuracy in diagnosis, probes were selected from the somal imbalances in plasma cell malignancies and determined critical regions of these imbalances as delineated by CGH. This the incidence of genomic gains and losses in a large prospective probe panel comprised probes mapping to chromosome bands series of myelomas. 1p22, 1q21, 6q21, 8p11, 9q34, and 11q25. DNA clones from chromosome bands 13q14 (locus D13S272), 17p13 (p53), and 22q11 were added to this panel, as deletions involving these Correspondence: Dr H Do¨hner, Department of Internal Medicine III, genomic loci have been previously reported to provide University of Ulm, Robert-Koch-Str. 8, Ulm 89081, Germany; Fax: þ 49 731 500 24493 prognostic information in MM. Received 20 May 2003; accepted 7 August 2003; Published online 2 The following DNA clones were used: RPCI-11 71-L-20 October 2003 (1q21.2), RPCI-11 210-F-08 (1q43), RPCI-11 284-J-03 (8p11.3),

Leukemia