Leukemia (2002) 16, 127–134  2002 Nature Publishing Group All rights reserved 0887-6924/02 $25.00 www.nature.com/leu Frequent inactivation of the -dependent inhibitor p18 by homozygous deletion in multiple myeloma cell lines: ectopic p18 expression inhibits growth and induces MS Kulkarni1, JL Daggett2, TP Bender2, WM Kuehl3, PL Bergsagel4 and ME Williams1

1Department of Internal Medicine, University of Virginia School of Medicine, Charlottesville, VA, USA; 2Department of Microbiology, University of Virginia School of Medicine, Charlottesville, VA, USA; 3Navy Medical Oncology Branch, National Cancer Institute, Bethesda, MD, USA; and 4Department of Internal Medicine, New York Hospital–Cornell Medical Center, New York, NY, USA

Multiple myeloma (MM) is a clonal neoplasm of plasma cells unit and phosphorylates the tumor suppressor pRb which offers an excellent model to study multistep molecular and related , releasing sequestered and leading to oncogenesis. In 20–25% of primary tumors and cell lines exam- ined, cyclin D1 is overexpressed due to the translocation the activation of required for DNA replication and S t(11;14)(q13;q32). We have characterized cyclin-dependent kin- phase progression. The activity of cyclin D1-CDK4 and cyclin ase inhibitor p15 (CDKN2B), (CDKN2A) and p18 (CDKN2C) D1-CDK6 holoenzymes is regulated at multiple levels includ- deletions in cyclin D1-expressing and non-expressing MM cell ing cyclin abundance, cyclin D1-CDK4 and cyclin D1-CDK6 lines. p18 was found to be frequently deleted (38%); in some assembly, dephosphorylation of inhibitory sites, CAK/cyclin cases p18 deletions coexisted with hemizygous p16 deletion. H1 activation of the complex byphosphorylationof the con- To examine the function of p18 as a putative tumor suppressor in myeloma cells, a zinc-inducible p18 construct was stably served threonine-160 residue, and 1:1 stoichiometric interac- 5–8 transfected into KMS12, a MM cell line with biallelic p18 and tion of CDKs with the INK4 familyCDKIs. The INK4 family monoallelic p16 deletions as well as cyclin D1 overexpression. members are tumor suppressor genes comprised of the pro- Ectopic expression of p18 caused 40–45% growth suppression teins p16 (CDKN2A), p15 (CDKN2B), p18 (CDKN2C), and as determined by trypan blue exclusion and MTS assays. p18 p19 (CDKN2D) that negativelyregulate CDK4 and CDK6 induction also resulted in apoptosis, suggesting that inhibition activity.9–12 The functional similarityof all four genes is of the cyclin D1/CDK/pRb pathway in these tumor cells could be a crucial step toward the induction of tumor regression via reflected in their structure; each contains highlycon- apoptotic cell death. This pathway is thus frequently served ankyrin domains suggesting all may have a role in mutated and provides a potentially novel target for thera- regulating proliferation. peutic or pharmacologic approaches to human myeloma. p18 predominantlybinds to CDK6 and, to a lesser extent, Leukemia (2002) 16, 127–134. DOI: 10.1038/sj/leu/2402328 to CDK4 and inactivates their kinase activity.9,10 Furthermore, Keywords: multiple myeloma; cyclin D1; apoptosis; cell cycle; p18; p18 can efficientlyblock CDK6 phosphorylationby mantle cell lymphoma CAK/cyclin H1.8 IL-6, a cytokine implicated in myeloma pathogenesis, induces p18 expression in IgG-bearing B lym- phocytes and correlates with cell cycle arrest, differentiation Introduction and cell death.13 Together, these findings suggest CDKI p18 status mayprovide useful insights into the molecular Multiple myeloma (MM) is a clonal plasma cell neoplasm that mechanisms of myeloma pathogenesis. provides an excellent model to studymultistep molecular In order to assess cell cycle regulatory mutations in human oncogenesis. A chromosomal translocation involving the IgH myeloma, we have analyzed MM-derived tumor cell lines locus, most frequentlyinto a switch region, is a nearlyuniver- with or without cyclin D1 overexpression for CDKI p15, p16 1 sal event in MM cell lines and primarytumor samples. Mol- and p18 gene deletions and correlated the association with ecular events contributing to karyotypic instability in a B lym- D1 expression. The KMS12 myeloma cell line, which has phocyte may initially lead to a monoclonal gammopathy of homozygous p18 and hemizygous p16 deletions and overex- undetermined significance (MGUS, ‘benign’ monoclonal presses cyclin D1, was used to analyze complementation of gammopathy) and evolve to smouldering myeloma, overt p18 function. We generated KMS12-derived cell lines trans- MM, and in some cases to extramedullaryprogression or fected with an inducible p18 expression vector, thus re-estab- 1 plasma cell leukemia. lishing the G1 . We report here that ectopic In 20–25% of MM cell lines and tumors examined, cyclin p18 induction causes growth inhibition and apoptosis in the D1 (CCND1) is brought under the influence of immunoglob- KMS12 multiple myeloma cell line, suggesting that this ulin heavychain (IgH) chain enhancer and promoter elements dysregulated cell cycle pathway may serve as a target for bya chromosomal translocation at Bcl-1/CCND1 breakpoints myeloma therapy. via the translocation t(11;14)(q13;q32). This leads to cyclin D1 overexpression and suggests that disruption of cell cycle regulatoryelements maybe pathogeneticallyimportant in this Materials and methods malignancy.2 Cyclin D1 functions as an oncogene in some systems, in part by promoting cyclin D1-CDK phosphorylation Cell culture of pRb.3,4 During the mid- of the cell cycle, the cata- lytic subunit of CDK forms a complex with its regulatory sub- The multiple myeloma cell line KMS12 was provided by Kim- oto and colleagues to one of us (WMK).14,15 The Granta line was derived from a human mantle cell lymphoma and was a Correspondence: MS Kulkarni, Hematology/Oncology Division, Box 16 800716, Universityof Virginia School of Medicine, Jefferson Park kind gift from DM Jadayel, Institute of Cancer Research, Sur- ° Avenue, Charlottesville, VA 22908, USA; Fax: 804–243–6086 rey, UK. Cells were grown at 37 C (5% CO2) in RPMI 1640 Received 6 February2001; accepted 27 July2001 media supplemented with 10% (v/v) heat-inactivated fetal p18 inactivation in multiple myeloma MS Kulkarni et al 128 bovine serum, 2 mM/l L-glutamine, and 100 U/ml penicillin ein I promoter to drive expression of a cloned cDNA.18 The and 100 ␮g/ml streptomycin. Stable KMS12 transfectants were vector includes a neomycin cassette for selection. The vector maintained in growth medium supplemented with 450 ␮g/ml also contains exons III and IV of the human beta-globin gene G418 without penicillin and streptomycin supplementation. which provides splice donor and acceptor sites as well as a polyadenylation signal. The pMth-p18 expression vector was then transfected into exponentiallygrowing KMS12 cells using Southern blot analysis the lipid reagent DMRIE-C (Life-Technologies, Gaithersburg, MD, USA) following the manufacturer’s instructions. Negative Genomic DNA was isolated from cultured cells and clinical control cell lines were generated using the pMth-Neo vector samples using Qiagen Blood and Cell Culture DNA kit alone. Stable clones were selected byG418 resistance, then (Qiagen, Valencia, CA, USA). Ten micrograms of DNA were individuallygrown and analyzedfor inducible expression of digested with EcoRI, fractionated on 0.7% agarose gels, p18 mRNA. Of 79 neomycin-resistant clones examined, five

denatured with 0.5 N NaOH, 1.5 mol/l NaCl, neutralized with (33 a–e) could be induced to express p18 with ZnCl2 when 1 mol/l Tris-HCl pH 7.4, 1.5 mol/l NaCl, and transferred over- added at a final concentration of 120 ␮M to the exponentially night with 20 × SSC (3 mol/l sodium chloride/3 mol/l sodium growing cells. citrate) on to nylon membranes (Stratagene, La Jolla, CA, USA) and UV-crosslinked. cDNA probes for p15 (a gift from Dr David Beach, Cold Spring Harbor, Long Island, NY, USA), and Northern blot analysis p18 (a gift from Dr Y Xiong, UNC, Chapel Hill, NC, USA), were labeled via the random priming technique using the Total RNA was extracted from exponentiallygrowing cells Ready-To-Go kit (Pharmacia Biotech, Piscataway, NJ, USA), using the RNA-STAT 60 System (TEL-TEST, Friendswood, TX, and 50 ␮Ci 32P-CTP (specific activity6000 Ci/mmole, NEN, USA). Ten micrograms of RNA was fractionated on 1% aga- Boston, MA, USA). The membranes were sequentiallyprehy- rose gels containing 0.22 mol/l formaldehyde, and transferred bridized and hybridized with 32P-labeled probes in Quick with 20 × SSC buffer on to dura-Lon-UV membranes, and UV- Hyb solution at 68°C according to the manufacturer’s instruc- cross-linked. Prehybridization and hybridization with alpha- tions (Stratagene). The membranes were washed twice in 2 × 32P CTP-labeled probe, and sequential washes of the mem- SSC buffer and 0.1% (w/v) SDS at room temperature, followed branes were performed as described above for Southern blot bya third wash in 0.1 × SSC buffer and 0.1% (w/v) SDS at analysis. Autoradiography was done using Kodak XAR5 films 60°C. Autoradiographywas performed using Kodak XAR5 at −70°C with intensifying screens, with semiquantitation of films (Eastman Kodak, Rochester, NY, USA) at −70°C with expression bycomparison with positive (Granta cell line) and intensifying screens. Gene deletion was identified by negative (KMS12 cell line) controls. decreased signal intensityrelative to placental control DNA.

Immunoprecipitation and Western blotting PCR analysis Cells were harvested and washed with PBS, left on ice for 30 DNA extracted from MM cell lines was analyzed by PCR to min, then lysed in basic NP-40 lysing buffer supplemented detect p18 gene deletions as previouslydescribed. 17 Primers with protease inhibitor cocktail, as described byXiong and used to amplifyp18 exon 1, exon 2 A and B were as follows: colleagues19 for 30 min at 4°C. Nuclei and debris were P18-exon 1, 5Ј-CCTGATCGTCAGGACCCTAA-3Ј/5Ј-CCAA removed bycentrifugation at 12000 r.p.m. for 10 min in an CCTGCAGCCGAGT-3Ј; p18-exon 2A, 5ЈAGGATTCTACCAT Eppendorf microcentrifuge model 5415 C. Protein concen- TTCTACTT-3Ј/5Ј-CTTCTTTGGCAGCCAAGTGC3Ј; p18-exon tration was determined using the BCA assay(Pierce, Rockford, 2B, 5Ј-GACTTGCTGGAGTTTCAAGCTG3Ј/5ЈTTATTGAAGAT IL, USA). One mg of protein was immunoprecipitated with 10 TTGTGGCTCCCCCA-3Ј. These primers cover the entire p18 ␮g monoclonal p18 Ab-1 (Neomarker, Fremont, CA, USA) at coding region. All PCR reactions were carried out in a final 4°C overnight, and 100 ␮l of prewashed Protein A Agarose volume of 100 ␮l containing 500 ng DNA, 0.5 M each of beads (Life-Technologies) were added for 2 h. The beads were primer, 400 ␮M dNTPs, 1.0 mm Tris-HCl (pH 8.0), 50 mM recovered bycentrifugation, washed with NP-40 lysisbuffer

KCl, 1.5 mM MgCl2, 2.5 U Taq polymerase. PCR reactions four times and resolved byelectrophoresis bySDS-PAGE on were overlayed with two drops of mineral oil. PCR conditions 15% polyacrylamide gels. Proteins were transferred to nitro- used an initial denaturation for 3 min at 94°C, followed by cellulose membrane, and the blots were blocked with 5% 35 cycles of 1 min denaturation at 94°C, 1.5 min annealing nonfat drymilk–0.5% Tween-20 in phosphate buffered saline at 58°C, extension for 2 min at 72°C, and final extension for (PBS) solution with the primaryantibody,monoclonal p18- 10 min. The PCR products were extracted with chloroform, Ab2 (Neomarker), in a 1:100 dilution for 1 h at room tempera- and separated on 1.5% agarose or 10% non-denaturing poly- ture. Following four washes with PBS-Tween 20, immune acrylamide gels. Placental DNA was used in all reactions as complexes were detected using horseradish-conjugated anti- a positive control for p18 amplification. rabbit antibodyat a 1:10000 dilution, and visualized byECL (Amersham, Arlington Heights, IL, USA).

Stable transfections in KMS12 Cell proliferation A full-length p18 cDNA containing the entire coding region of 507 bp (a gift from Dr Yue Xiong) was removed from pUC Cells were plated at a densityof 1 × 105/ml/well in duplicate 118 plasmid backbone bycleavage with BamHI. The cDNA in a 24-well plate. After 24 h of incubation at 37°Cina

fragment was subcloned in the pMth-Neo eukaryotic humidified atmosphere of 5% CO2 and 95% air, cells were ␮ expression vector which uses the zinc-inducible metallothion- treated with or without ZnCl2 for 0, 24, and 48 h. A 100 l

Leukemia p18 inactivation in multiple myeloma MS Kulkarni et al 129 aliquot of cell suspension was taken at the indicated times lines contained homozygous p18 deletions (Table 1, Figure and counted with a hemocytometer by the trypan blue 1a). No p15/p16 deletions were identified bySouthern blot in exclusion test. the eight cyclin D1-negative lines tested. PCR confirmed the frequencyof p18 gene deletions in MM cell lines. PCR ampli- fication of p18 exon 1 generated a clear band in cells with Cell growth MTS assay wild-type p18, whereas homozygous deletions were found in the cell lines KMS12, MM1, MY5, OPM2 (Figure 1b) and Cell viabilitywas also measured byMTS Cell-Titer 96 AQ SKMM2 (data not shown). As p18 deletions are rare in most 22–27 (Promega, Madison, WI, USA), a colorimetric method for human neoplasms and cell lines but have been identified 21,28,29 determining the number of viable cells in proliferation or cyto- in mantle cell lymphoma and some human myelomas, toxicityassays.Cells were plated at a densityof 1 × these frequent deletions in MM cell lines suggest a pathogenic 105/ml/well in triplicate overnight and treated with or without role via cell cycle dysregulation, either alone or in concert ␮ with cyclin D1 overexpression. ZnCl2 for 0, 24, and 48 h. A 100 l aliquot of cell suspension was taken at the indicated times and transferred to 96-well titer plates. Twentymicroliters of MTS solution was added to each well and incubated at 37°C for 2 h. The intensityof the Inducible ectopic expression of p18 in KMS12 soluble brown formazan product, directlyproportional to the number of viable cells, was measured byreading absorbance To examine the relevance of loss of p18 function in MM , we at 490 nm using a Titer Tek microplate reader (ICN Biochemi- re-expressed the p18 gene in the KMS12 cell line, which is cals, Irvine, CA, USA). homozygous p18 null. We generated a full-length p18 cDNA in the eukaryotic expression plasmid pMTH-Neo, which uses a zinc-inducible metallothionein promoter, such that ectopic DNA fragmentation p18 expression is regulated bythe addition of zinc chloride to the growth medium. Asynchronously proliferating KMS12 cells were transfected with the pMth-Neo vectors lacking or × 5 Proliferating cells were plated at a densityof 1 10 /ml in 15 containing the p18 cDNA. Forty-eight hours later, transfec- ml Petri dishes, grown overnight, and treated or not treated tants were transferred into medium supplemented with G418. with ZnCl2 for 0, 18, 36, and 60 h. For the DNA fragmentation 20 Clones resistant to G418 were selected and analyzed for Zn- assay cells were lysed in a buffer containing 1% NP-40 in inducible expression of p18 mRNA byNorthern blot analysis. 20 mM EDTA and 50 mM Tris-HCl (pH 7.5). The supernatant Five clones, 33a–33e, were identified that induciblyexpressed was recovered bycentrifugation at 1600 g for 5 min, brought p18 mRNA (Figure 2a and data not shown). No clones were to 1% SDS and treated for 2 h with 5 ␮g/␮l RNAse A at 56°C. ° identified that expressed detectable amounts of p18 mRNA in Subsequently, Proteinase K was added and incubated at 37 C the absence of ZnCl . The mantle cell lymphoma line for 2 h. The lysate was precipitated using 0.5 volumes of 10 2 − ° GRANTA, which has germline p18 bySouthern blot, was used M ammonium acetate and 2.5 volumes of ethanol at 20 C as a positive control for p18 RNA expression. Differing p18 for 1 h. Precipitated DNA was pelleted bycentrifugation, × transcripts are expressed in various tissues, indicating a tissue- resuspended in 1 TE (10 mM Tris-HCl, pH 7.4 and 1 mM specific regulation.9 These transcript sizes reflect variations in EDTA) and fractionated byelectrophoresis on 1% agarose the 5 prime untranslated region;30 however, all transcripts gels. DNA ladders were visualized byethidium bromide encode the same protein. The endogenous transcripts staining.

Table 1 Analysis of cyclin D1 expression and CDK inhibitor Results p15/p16 and p18 deletions in human multiple myeloma cell lines

CDKI gene deletions in MM cell lines Cell line D1 p15/p16 p18 expression Structural abnormalities of 1 are frequently FLAM 76 + 1DEL 1DEL detected in MM at diagnosis and have been correlated with + 21 H1112 GG more advanced or aggressive disease. The p18 gene has MM.M1 + GG been mapped to human chromosome 1p32. Although p18 KMS12 + 1 DEL DEL deletions have been reported to be rare in human malig- SKMM2 + 1 DEL DEL nancy,22–27 cytogenic analysis has shown abnormalities of KMS11 − GG − 1p32–36 in 16% of human MM as well as 14% of low-grade MM.1 GDEL UTMC2 − GG non-Hodgkin’s lymphoma (NHL) and 12% of diffuse large cell ARK − GG lymphoma. To determine whether p18 may be a critical SKMM1 − GG element of such deletions, to assess the correlation of such H929 − GG deletions, if any, with cyclin D1 translocation and over- MY5 − GDEL expression, and to assess p18 as a candidate MM tumor sup- OPM2 − GDEL − pressor gene, we examined bySouthern blot analysisfive JJN3 ND G TH − ND G cyclin D1-overexpressing and 11 cyclin D1-nonexpressing DELTA − ND G human MM cell lines for p15/p16 and p18 gene deletions. 3/13 6/16 Each cyclin D1-expressing cell line had a karyotypic or cloned (23%) (38%) t(11;14) translocation.2 Three of the five cell lines positive for cyclin D1 expression contained coexisting deletions of G, germline; 1 DEL, hemizygous deletion; DEL, homozygous p15/p16 and p18, whereas three of 11 cyclin D1-negative deletion.

Leukemia p18 inactivation in multiple myeloma MS Kulkarni et al 130

Figure 2 Inducible expression of ectopic p18 mRNA and protein in the KMS12 multiple myeloma cell line clone 33a+. (a) Northern blot analysis of p18 mRNA isolated from KMS12, pMth-Neo (vector- transfected line), the mantle cell lymphoma line Granta (a positive control for endogenous p18 mRNA), and pMth-p18-transfected 33a cell lines, treated with (+) or without (−) zinc chloride for p18 induc- tion. Total RNA was extracted from the cell lines indicated, separated on formaldehyde-agarose gels (10 ␮g/lane), transferred on to nylon membranes and hybridized with a 32P-labeled p18 cDNA probe. Autoradiographic exposure time was 4 h. Equal loading was con- firmed bythe amount of ribosomal RNA (not shown). (b) Immunoprec- ipitation Western blot analysis of p18 protein from KMS12, Granta, pMth-p18 (33a) and pMth-Neo (vector only) cell lines, treated with Figure 1 (a) Deletion of p18 in cyclin D1-expressing multiple (+), and without (−) zinc chloride for p18 induction. Whole cell lysates myeloma cell lines. Representative Southern blot of EcoRI DNA were immunoprecipitated with p18 ab-1, separated by15% SDS- digests seriallyhybridizedwith immunoglobulin JH, chromosome 11q polyacrylamide gel electrophoresis, immunoblotted, and probed with 13 bcl-1 major translocation cluster (MTC) breakpoint and CDKI p15 p18 ab-2. The 18 kDa p18 protein is detected in the Granta cells (lane and p18 probes. (b) PCR of the p18 gene showing deletion of p18 4) and in the zinc-induced 33a cells (lane 5). Lanes 1, protein molecu- exon 1 in four of the 10 multiple myeloma cell lines. Genomic DNA lar weight marker; 2, KMS12−; 3, KMS12+; 4, Granta; 5, 33a+;6, was amplified bymeans of primers spanning exon 1 of p18 (179 bp). 33a−; 7, pMth-Neo+; 8, pMmth-Neo−. Amplification of the 5Ј region of the ␤-globin gene (400 bp) served as a positive amplification control. PCR products were electrophor- esed on 1.5% agarose gels and stained with ethidium bromide. Pla- cental DNA was used as a positive p18 control. in the 33a+ (Zinc-induced) line to be 10 to 20-fold higher than that seen in exponentiallygrowing Granta cells. Expression of p18 protein was examined in the KMS12, pMth-Neo, and 33a (pMth-p18) cell lines with (+) or without − expressed in the Granta cell line are 1.9 and 1.1 kb; the ( ) ZnCl2 byimmunoprecipitation followed byblotting with ectopicallyexpressed p18 mRNA is 1.1 kb size and encodes the p18-Ab1 monoclonal antibody. Whole cell extracts pro- the full-length protein. The pMth-Neo and KMS12 control duced from 33a+ and Granta cell lines expressed detectable lines did not express p18 mRNA, with or without zinc treat- amounts of p18 protein after 18 h of induction with Zn. In ment (Figure 2a). Based upon the results of multiple RNA blot- contrast, KMS12 and the negative control pMTH-Neo did not ting experiments, we estimate the abundance of p18 mRNA contain inducible p18 protein. The ectopicallyinduced p18

Leukemia p18 inactivation in multiple myeloma MS Kulkarni et al

+ 131 from the 33a cell line migrated at the same position as the larger at 48 h (Table 2). When the cells were exposed to ZnCl2 endogenous p18 from Granta cells (Figure 2b). for 5 days, the 33a+ line cell number was reduced to 10–15% of the controls (data not shown). This observed decrease in the rate of cell growth mayrepresent either cell cyclearrest, Ectopic expression of p18 inhibits KMS12 proliferation induction of apoptosis, or both. These results confirm the function of CDKI p18 as a negative growth regulator of cell To test the hypothesis that overexpression of cyclin D1 cycle capable of inducing cell growth inhibition. coupled with the loss of CDKI p18 deletion provides a growth advantage to the cells, we tested the effects of enforced p18 expression on cell growth. 1 × 105 cells/ml were plated and

grown for 24 h followed byZnCl 2 addition to induce p18 ␮ expression. A 100 l aliquot of cell suspension was removed Table 2 Effect of ectopic inducible p18 expression on meta- from the culture plates for cell counts bytrypanblue exclusion bolicallyactive cells measured byMTS color development assay after 24 and 48 h of Zn treatment. After 24 h of p18 induction

with ZnCl2 treatment, cell proliferation was reduced to 40– Cell lines 0 h Metabolically active cells (%) 45% of the uninduced controls (33a−), pMth-Neo and KMS12; the differences remained the same after 48 h (Figure 3). All the 48 h 72 h clones (33a–e) which expressed p18 mRNA in the presence of 5 ZnCl2, exhibited the same growth inhibition patterns in the KMS12− 1 × 10 /ml 100 100 trypan blue cell counting proliferation assay (data not shown). KMS12+ 1 × 105/ml 128 91 Zinc treatment for 24 h had no inhibitoryeffect on prolifer- pMth− 1 × 105/ml 95 80 + × 5 ation of the KMS12 parental cells. Experiments to transfect the pMth 1 10 /ml 85 82 − − − 33a− 1 × 105/ml 86 90 pMth-p18 vector into the cyclin D1 , p18 / OPM2 mye- 33a+ 1 × 105/ml 57 44 loma cell line were unsuccessful due to intrinsic resistance of the wild-type OPM2 cells to high concentrations of G418 and Cells were plated at a density of 1 × 105/ml in a 24-well plate. At a resulting inabilityto select transfected clones. 24 h, ZnCl2 was added to induce p18 expression. Aliquots of 100 We also examined the cell viabilitybythe MTS assayof ␮l cell suspension were analyzed at 24 and 48 h following Zn induc- mitochondrial function. Similar to the cell-counting assay, tion for cell growth by Cell Titer 96 Aqueous One Solution Cell Pro- cells were plated at a densityof 1 × 105 and Zn was added liferation Assay (Promega, Madison, WI, USA). Percentages were after 24 h for p18 induction. A 100 ␮l aliquot of cell suspen- calculated by comparison to the untreated KMS12 line as 100% metabolically active. sion was removed from the culture plates at 24, 48, and 72 KMS12, myeloma cell line; pMth, vector-only transfected KMS12 h of induction and assayed for viable cells. Again, there was − cell line; 33a, pMth-p18 transfected KMS12 cells; , No ZnCl2 + + ␮ a 33% growth suppression in the 33a line. The differences added; , 120 M/l ZnCl2 added. between the induced and uninduced cells were progressively

Figure 3 Effect of ectopic inducible p18 expression on cell growth. KMS12 cells, pMth (vector only) and 33a (pMth-p18 vector transfected) × 5 cells were plated in duplicate at a densityof 1 10 /ml in 24-well plates. ZnCl2 was added to the growth medium after 24 h to induce p18 expression. Following 24 and 48 h of Zn exposure cells were harvested and 100 ␮l aliquots were counted bytrypanblue exclusion for viable − + ␮ cell number. , no ZnCl2 added; , 120 Ml ZnCl2 added.

Leukemia p18 inactivation in multiple myeloma MS Kulkarni et al 132 Apoptosis analysis Table 3 Induction of apoptosis byp18 expression

Cell lines and treatments % Apoptosisa To discern whether the p18-induced growth inhibition was due to G1 phase cell cycle arrest or apoptosis, DNA was KMS12− 7 assayed by flow cytometry after propidium iodine staining. KMS12+ 12 Inducible p18 expression did not cause significant accumu- PMth− 7 lation of cells in G1 phase compared to the negative controls. PMth+ 13 − However, in the presence of ZnCl there was a significant 33a 9 2 33a+ 36 increase in sub 2n cells in the 33a+ line which was more pronounced at 48 h, potentiallyreflecting a significant − + ␮ , no ZnCl2 added; , 120 M/l ZnCl2 added. Cells were plated at number of cells undergoing apoptosis (data not shown). × 5 density of 1 10 /ml. After 24 h of culture ZnCl2 was added to The generation of low molecular weight DNA fragmen- induce ectopic p18 protein expression. Following 24 h of Zn tation is a hallmark of apoptosis. When the p18 gene was exposure cells were washed and resuspended in staining solution induced for 18, 36, and 60 h, DNA isolated from the 33a+ cell for analysis of DNA content by flow cytometry. line exhibited a characteristic apoptotic DNA fragmentation See legend of Table 2 for definition of cell lines. a ladder when resolved by1% agarose gel electrophoresis The percentage of cells undergoing apoptosis, measured as cells containing less than diploid DNA content, was calculated as pro- (Figure 4). As we have found that Flavopiridol (Fp) treatment portion of total cell population. of KMS12 cells at a concentration of 600 nM for 24 h leads to DNA fragmentation (unpublished data), we used Fp-treated KMS12 cells as a positive control for apoptosis and the DNA Discussion ladder. KMS12 and 33a− cell lines showed no appreciable DNA fragmentation under the same conditions. To quantify In a subset of MM (ෂ20–25%), the translocation t(11;14) the extent of apoptosis we examined DNA histograms of (q13;q32) results in cyclin D1 overexpression.15 Several induced p18 as well as appropriate controls byflow cytometry + additional non-random molecular abnormalities have been at 24 h (Table 3). There were 36% apoptotic cells in the 33a described involving BCL2, C-MYC, RAS, , and pRb and cells, whereas in parental KMS12 and the vector-transfected correlated with poor clinical prognosis. However, these mol- pMth-Neo cells (with and without zinc exposure) it ranged ecular changes have not been correlated with various stages from 7 to 13%, suggesting that p18 induction is associated of disease progression. Cyclin D1 overexpression along with with activation of apoptotic pathways. loss-of-function mutations of p53 and pRb genes are consist- ent with the deregulation of normal cell cycle control, parti- cularlyat the G1– transition. Virtuallyall members of the Rb pathwayhave been implicated pathogeneticallyin various forms of hematologic malignancies and other cancers; for example, CDKI p16 is deleted in some ALL, and in some MCL deletions of p16 or p18 coexist along with cyclin D1 overexpression.29,31 In the present study, we have characterized CDKI p15/p16 and p18 deletions in cyclin D1-overexpressing and non- expressing MM cell lines. Homozygous p18 deletion was present in five lines, with a single p18 allele deleted in the FLAM 76 line (data not shown). Some of these deletions coexisted with hemizygous p16 deletion and suggest a selec- tion advantage for concomitant cyclin D1 overexpression and deletion of one or more CDK inhibitors, potentiallycontribu- ting to cell transformation and enhanced cell proliferation. p18 gene deletions and rearrangements have been identified in MCL and MM21,28,29 but not in other neoplasms,22–27 sup- porting a possible role of this CDKI in MCL and myeloma pathogenesis. At this point, it is unknown whether p18 deletions in MM tumor cell lines represent an epiphenomenon of the cell culture process, although there is a report of such deletion in primaryMM and in childhood T cell ALL samples.21,26 It is also unknown whether additional loss-of- function mutations including gene hypermethylation or point mutations are operative in MM or other human malignancies; Figure 4 DNA fragmentation after ectopic p18 induction in the p18 was inactivated bya point mutation in BT-20 human bre- KMS12 multiple myeloma cell line. Exponentially growing 33a cells ast cancer cells rendering it deficient in suppressing cell (pMth-p18 transfected KMS12 cells) were treated (+) or not treated (−) growth.32 with zinc chloride for p18 induction for 18, 36, and 60 h in culture. In view of the high frequencyof p18 deletions in MM cell Cells were harvested and low molecular weight DNA was extracted lines, we investigated the significance of p18 inactivation by and separated by1% agarose gel electrophoresis. KMS12 cells treated replacing its biologic function in a human myeloma cell line. with flavopiridol (FP) served as a positive control for DNA fragmen- tation. Lanes a, blank; b, 33a− 18 h; c, ␾X 174 RF DNA HaeIII marker; Our data demonstrate that inducible ectopic p18 protein is d, 33a− 36 h; e, 3a+ 36 h; f, 33a+ 60 h; g, 33+ 18 h; h, KMS12 with indistinguishable from the constitutivelyexpressed protein FP 600 nM (positive control); i, KMS12 no treatment. and inhibited cell proliferation, consistent with the reported

Leukemia p18 inactivation in multiple myeloma MS Kulkarni et al 133 functions of p18.9,10 We also found that ectopic expression of 8 Kaldis P, Russo AA, Chou HS, Pauletich NP, Soloman MJ. Human p18 in KMS12 cells induced apoptotic cell death and DNA and yeast CDK-activating (CAKS) display distint substrate fragmentation, suggesting that an efficient block of the cyclin specificities. Mol Biol Cell 1998; 9: 2545–2560. 9 Guan KL, Jenkins CW, Li Y, Nichols MA, Wu X, O’Keefe CL, Mat- D1/CDK/pRb pathwayin MM could be a crucial step to era AG, Xiong Y. Growth suppression byp18, a p16 INK4/MTSI induce tumor regression via apoptotic cell death. All five of and p14 INK4/MTS2-related CDK6 inhibitor, correlates with wild- the p18-expressing KMS12 clones generated showed inhi- type pRb function. Genes Dev 1994; 8: 2939–2952. bition of proliferation in vitro despite varying levels of ectopic 10 Hirai H, Roussel MF, Kato JY, Ashmun RA, Sheer CJ. Novel INK4 p18 expression. Clone 33a, which had 10- to 20-fold higher proteins, p18 and p19, are specific inhibitors of the - p18 expression than exponentiallygrowing Granta cells, was dependent kinases CDK4 and CDK6. Mol Cell Biol 1995; 15: proapoptotic. However, this proapoptotic effect maybe non- 2672–2681. 11 Hirama T, Koeffler HP. Role of the cyclin-dependent kinases physiologic given the high levels of induced p18. In IgG-bear- inhibitors in the development of cancer. Blood 1995; 86: 841– ing human lymphoblastoid cells, IL-6-induced p18 expression 854. correlates with cell cycle arrest, terminal differentiation and 12 Serrano M, Lee HW, Chin L, Cordon-Cardo C, Beach D, DePinho cell death.13 P18 is highlyexpressed during myogenicdiffer- RA. Role of INK4a locus in linear suppression and cell mortality. entiation, potentiallymediating accompanyinggrowth arrest Cell 1996; 85: 27–37. byCDK4/CDK6 inhibition. 33 Thus, p18 appears to be a critical 13 Morse L, Chen D, Franklin D, Xiong Y, Chen-KianyS. Induction element in both differentiation and proliferation pathways and of cell cycle arrest and B cell terminal differentiation by CDK inhibitor p18 INK4a and IL-6. Immunity 1997; 6: 47–56. mayserve as a marker for treatment response or as a target 14 Namba MK, Ohtsuki T, Mori M, Togawa A, Wada H, Sugihara T, for new therapeutic approaches in MM. Yawata Y, Kimoto T. Establishment of five human myeloma cell In summary, we report frequent p18 deletions in both cyclin lines. In Vitro Cell Dev Biol 1989; 25: 723–729. D1-positive and -negative MM cell lines. These findings sug- 15 Bergsagal PL, Chesi M, Nardini E, Brents LA, KirbySL, Kuehl MW. gest a growth advantage to the cell, potentiallyleading to Promiscuous translocations into immunoglobulin heavychain enhanced cell proliferation. Information on LOF switch regions in multiple myeloma. Proc Natl Acad Sci USA mutations/deletions of CDKIs and other tumor suppressor 1996; 93: 13931–13936. 16 Jadayel DM, Lucas J, Nacheva E, Bartkova J, Stranks G, DE genes mayprovide molecular markers for disease stage, prog- Schouwer PJJC, Lens D. Dyer MJS, Kruger AR, Catovsky D. Poten- nosis and perhaps therapeutic response. Ectopic p18 tial role for concurrent abnormalities of the cyclin D1, p16 expression suggests that this gene mayplaya vital role in CDKN2 and p15 CDKN2 B genes in certain B cell non-Hodgkin’s inhibiting cell proliferation and promoting cell death in lymphomas. Functional studies in a cell line (Granta 519). Leuke- human myeloma. Furthermore, replacing p18/CDKI function mia 1997; 11: 64–72. via gene therapeutic or pharmacologic techniques maypro- 17 Kodura, PRK, Zariwala M, Soni M, Gong JZ, Xiong Y, Broome vide potentiallynew avenues for therapyof this presently JD. Deletion of cyclin-dependent kinase 4 inhibitor genes in non- Hodgkin’s lymphoma. Blood 1995; 86: 2900–2905. incurable neoplasm. 18 McClinton D, Stafford J, Brents L, Bender TP, Kuehl WM. Differen- tiation of mouse erythroleukemia cells is blocked by late up-regu- lation of a c-myb transgene. Mol Cell Biol 1990; 10: 705–710. Acknowledgements 19 Xiong Y, Zhang H, Beach D. Subunit rearrangement of the cyclin dependent kinases is associated with cellular transformation. The authors thank Dr Lei Leu for assistance in performing the Genes Dev 1993; 7: 1572–1583. Northern blot analyses, and Margaret Rogers and Emma Peho- 20 Hermann M, Lorenz HM, Voll, Grunke M, Woihte W, Kalden JR. A rapid and simple method for the isolation of apoptotic DNA vaz for expert technical assistance. We thank James Lampros fragments. Nucleic Acids Res 1994; 22: 5506–5507. for assistance with manuscript preparation. This work was 21 Tasaka T, Berenson J, Vescio R, Hirama T, Miller CW, Nagai M, supported in part bygrant RO1-GM55985 from the National Takara J, Koeffler HP. Analysis of the p16 INK4A, p15 INK4B and Institute of General Medicine, Bethesda, MD and grants ACS p18 INK4C genes in multiple myeloma. Br J Haematol 1997; 96: FRA-425 from the American Cancer Society, Atlanta, GA, and 98–102. FogartyTWO2297. 22 Tahara H, Smith AP, Gaz RA, Zariwala M, Xiong Y, Arnold A. Parathyroid tumor suppressor on 1 p: analysis of the p18 cyclin dependant kinase inhibitor gene as a candidate. J Bone Mineral Res 1997; 12: 1330–1334. References 23 Kawamata N, Miler CE, Koeffler HP. Molecular analysis of a family of cyclin dependent kinase inhibitor genes (p15/MTS2/INK4b and 1 Halek M, Bergsagel PL, Anderson KL. Multiple myeloma: increas- P18/INK4c) in non-small cell lung cancers. Mol Carcinogen 1995; ing evidence for a multistep transformation process. Blood 1998; 14: 263–268. 91: 3–21. 24 Kawamata N, Seriu T, Koeffler HP, Bartram CR. Molecular analysis 2 Chesi M, Bergsagel PL, Breuts LA, Smith CM, Gerhard DS, Kuehl of the cyclin-dependent kinase inhibitor family: WM. Dysregulation of cyclin D1 by translocation into an IgH p16(CDKN2/MTS1/INK4A), p18(INK4C), and p27(Kip1) genes in gamma switch region in two multiple myeloma lines. Blood 1996; 88: 674–681. neuroblastomas. Cancer 1996; 77: 570–575. 3 Hunter T, Pines J. and cancer II: cyclin D and CDK inhibi- 25 Zariwala M, Liu E, Xiong Y. Mutational analysis of the p16 family tors come of age. Cell 1994; 79: 573–582. cyclin-dependent kinase inhibitors p15INK4b and p18INK4c in 4 Sellers WR, Kaelin WG. Role of the retinoblastoma protein in the tumor-derived cell lines and primarytumors. Oncogene 1996; 12: pathogenesis of human cancer. J Clin Oncol 1997; 15: 3301– 451–455. 3312. 26 Siebert R, Willers CP, Fossa A, Kloke O, Nowrousian MR, Seeber 5 Sherr CJ. G1 phase progression: cycling on cue. Cell 1994; 79: S, Opalka B. Analysis of the novel cyclin-dependent kinase 4 and 551–555. 6 inhibitor gene p18 in lymphoma and leukemia cell lines. Leuke- 6 Sherr CJ, Roberts JF. Inhibitors of mammalian G1 cyclin-depen- mia Res 1996; 20: 197–200. dent-kinases. Genes Dev 1995; 9: 1149–1163. 27 Miller CW, Aslo A, Campbell MJ, Kawamata N, Lampkin BC, 7 Watanabe G, Albanese C, Lee RJ, Reutens A, Vairo G, Henglein Koeffler HP. Alterations of the p15, p16, p18 genes in osteosar- B, Pestell RG. Activityis associated with E2F-mediated inhibitors coma. Cancer Genet Cytogenet 1995; 86: 136–142. of cyclin D1 promoter activity through E2F and Sp2. Mol Cell Biol 28 Williams ME, Rogers MJ, Daggett JL, Bergsagel PL, Keuhl M, 1998; 18: 3212–3222. Bender T. Frequent deletion of the cyclin dependent kinase inhibi-

Leukemia p18 inactivation in multiple myeloma MS Kulkarni et al 134 tor genes p16 and p18 in human multiple myeloma cell lines. 31 Vasef MA, Medeiros LJ, Yospuz LS, Sun NC, McCourtyBrynesRK. Blood 1996; 88 (Suppl. 1): 102a. Cyclin D1 protein in multiple myeloma and plasmacytoma: an 29 Williams ME, Whitefield M, Swerdlow SH. Analysis of the cyclin- immunohistochemical studyusing fixed paraffin-embedded tissue dependent kinase inhibitors p18 and p19 in mantle-cell lym- sections. Mod Pathol 1997; 10: 927–932. phoma and chronic lymphocytic leukemia. Ann Oncol 1997; 8 32 Lapointe J, Lachance Y, Labrie Y, Labrie C. A p18 mutant defective (Suppl. 2): 71–73. in CDK6 binding in human breast cancer cells. Cancer Res 1996; 30 Phelps DE, Hsiao KM, Li Y, Hu N, Franklin DS, Westphal E, Lee 56: 4586–4589. EYHP, Xiong Y. Coupled transcriptional and translational control 33 Franklin DS, Xiong Y. Induction of p18 INK4c and its predominant of cyclin-dependent kinase inhibitor p18INK4C expression during association with CDK4 and CDK6 during myogenic differen- myogenesis. Mol Cell Biol 1998; 18: 2334–2343. tiation. Mol Biol Cell 1996; 7: 1587–1599.

Leukemia