Cyclin D1 Amplification Is Independent of P16 Inactivation in Head And

Cyclin D1 ampli®cation is independent of p16 inactivation in head and neck squamous cell carcinoma

K Okami1,3, AL Reed1, P Cairns1, WM Koch1, WH Westra2, S Wehage1, J Jen1 and D Sidransky*,1

1Department of OtolaryngologyÐHead & Neck Surgery, Division of Head and Neck Cancer Research, Johns Hopkins University School of Medicine, 818 Ross Research Building, 720 Rutland Avenue, Baltimore, Maryland, MD 21205-2196, USA; 2Department of Pathology, 7181 Meyer Building, Johns Hopkins Hospital, 600 North Wolfe Street, Baltimore, Maryland, MD 21287, USA; 3Department of Otolaryngology, Yamaguchi University, Ube 755, Japan

Progression through the G1 phase of the cell cycle is kinase (cdk) 4 complexes. This cdk4 phosphorylation mediated by phosphorylation of the retinoblastoma activity is negatively regulated by the tumor suppressor protein (pRb) resulting in the release of essential gene p16 (a CDK inhibitor), and positively by cyclin transcription factors such as E2F-1. The phosphorylation D1. Each individual component of this p16-cyclin D1- of pRb is regulated positively by cyclin D1/CDK4 and Rb pathway is commonly targeted in various malig- negatively by CDK inhibitors, such as p16 (CDKN2/ nancies (Weinberg, 1995; Sherr, 1996). We previously

MTS-1/INK4A). The p16 /cyclin D1/Rb pathway plays reported a high frequency (83%) of p16 inactivation in a critical role in tumorigenesis and many tumor types human HNSCC (Reed et al., 1996) and less frequent display a high frequency of inactivation of at least one inactivation (13%) of pRb (Yoo et al., 1994). Cyclin D1 component of this pathway. In order to determine the ampli®cation has been reported in 22 ± 64% of HNSCC overall contribution of these three components to based on several dierent genetic or IHC methods progression of head and neck squamous cell carcinoma (Jares et al., 1994; Kyomoto et al., 1997; Olshan et al.,

(HNSCC), we examined p16 inactivation, cyclin D1 1997; Michalides et al., 1995). However, few reports ampli®cation, and pRb expression in 23 primary have simultaneously examined all components of this HNSCC tumors and ®ve cell lines. p16 inactivation critical pathway in HNSCC (Olshan et al., 1997; Lukas was detected in 19/23 (83%) primary tumors by detailed et al., 1995). genetic analysis and was con®rmed by immunohisto- p16 and pRb inactivation have been reported to be chemistry (IHC). Absence of Rb protein expression inversely correlated in some tumor types (Shapiro et indicative of pRb inactivation was identi®ed in 2/23 (9%) al., 1995; Parry et al., 1995; Sakaguchi et al., 1996; tumors. In this set of tumors, there was a perfect inverse Andl et al., 1998). It has thus been assumed that each correlation between p16 and pRb inactivation. Using component of this pathway is an alternative target for

¯uorescence in situ hybridization (FISH) cyclin D1 abrogation of a common pathway and that alterations ampli®cation was identi®ed in 4/5 (80%) cell lines and of multiple components in the same pathway is not 4/11 (36%) primary tumors. However, 2/4 cell lines and necessary for tumor progression (Serrano et al., 1996). all four primary tumors with cyclin D1 ampli®cation We thus undertook a comprehensive study of all contained a concomitant alteration of p16. Therefore 21/ potential alterations within the p16-cyclin D1-Rb 23 (91%) of primary HNSCC contained at least one pathway in cell lines and primary HNSCC. We found alteration in the p16/cyclin D1/Rb pathway. Although a high frequency of p16 inactivation and cyclin D1 p16 and Rb alteration are apparently exclusive, cyclin D1 ampli®cation in cell lines and primary HNSCC. ampli®cation occurs concomitantly with the loss of p16 Moreover, a signi®cant proportion of primary tumors suggesting an additional role for this ampli®cation in contained concomitant alterations of p16 and cyclin D1 HNSCC. suggesting that unlike p16 and pRb, alterations of these two components are not mutually exclusive.

Keywords: cyclin D1; p16; head and neck cancer

Results and dscussion Introduction p16 status

The G1 phase of the cell cycle is controlled by The p16 gene status in these tumors has been reported phosphorylation of the retinoblastoma protein (pRb) previously (Reed et al., 1996). p16 inactivation was and mediated in part by cyclin D-cyclin dependent detected by IHC in 19 out of 23 (83%) cases (Table 1). The IHC results were con®rmed by genetic analysis as previously reported (12 had HD at the p16 locus, ®ve contained methylation of the promoter region of p16, and one had a frameshift mutation in exon 1; Reed et al., 1996). IHC analysis of p16 appears to be a reliable and practical method to determine p16 *Correspondence: D Sidransky Received 20 October 1998; revised 8 January 1999; accepted 5 March inactivation in fresh tissue as reported (Reed et al., 1999 1996). p16, Cyclin D1 and Rb in head and neck cancer KOkamiet al 3542 the BAC probe on chromosome 11q13 (data not pRb IHC shown). In order to determine the incidence of cyclin

pRb IHC showed strong nuclear staining in both D1 gene ampli®cation, FISH analysis was carried out tumor and normal cells (Figure 1a). pRb inactivation on both HNSCC cell lines and primary tumors. We was observed in two cases out of 23 (9%) exhibiting a identi®ed gene ampli®cation in four out of ®ve (80%) good contrast between the negative staining tumor area cell lines (Table 1). The FaDu cell line contained four

and the positive staining normal areas (Figure 1b). The control signals and more than 20 cyclin D1 signals incidence of pRb inactivation was consistent with our (Figure 1a). We found two typical appearances of previous study on primary HNSCC (Yoo et al., 1994; gene ampli®cation in cell lines 011, 012 and 022. Cell

Andl et al., 1998). line 011 had 4 ± 6 control signals and 10 ± 12 cyclin D1 signals, and in the metaphase spreads of chromosome 11, it exhibited a homogeneously staining region Cyclin D1 ampli®cation by FISH (HSR, Brodeur and Hogarty, 1998) of the cyclin D1 FISH was ®rst performed on normal lymphocyte gene (Figure 2b). Cell line 012 and 022 metaphase

spreads which con®rmed the chromosomal location of spreads displayed 4 ± 6 cyclin D1 signals due to a duplication of the long arm of chromosome 11 seen in each nucleus (Figure 2c). In the primary tumors, we found four cases of gene ampli®cation out of 11 (36%) cases available (Table 1). Figure 2d shows a primary tumor with four control signals and 12 cyclin

D1 signals. Previously, ampli®cation of cyclin D1 was reported in 22 ± 64% of HNSCC (Jares et al., 1994; Kyomoto et al., 1997; Olshan et al., 1997; Michalides et al., 1995). These studies used various methods to detect the ampli®cation including Southern blot analysis, Northern blot hybridization, dierential PCR, or IHC. Although the frequency of amplification detected by these various approaches is somewhat

Table 1 Aberrations of p16, cyclin D1 and pRb in HNSCC primary tumors and cell lines Primary Cyclin D1 Fold tumors/ p16 p16 pRB ampli®- ampli®- Cell line statusa IHC IHC cationb cationc 1 Methylated 7 + N/A 2 HD 7 + N/A 3 Methylated 7 + N/A 4 Methylated 7 + 7 5 Methylated 7 + + 3 6 HD 7 + N/A 7 HD 7 + N/A 8 HD 7 + N/A 9 HD 7 + N/A 10 HD 7 + N/A 11 HD 7 + 7 12 HD 7 + 7 13 HD 7 + N/A 14 Retained + + 7 15 LOH + 7 7 16 Frame shift in 7 + + 3 exon 1 17 HD 7 + + 2 18 HD 7 + N/A 19 HD 7 + N/A 20 LOH 7 + N/A 21 Retained + + 7 22 LOH + 7 7 23 Methylated 7 + + 5 FaDu Point mutation 7 + + 6 Cell line 011 wt + + + 3 Cell line 012 Methylated 7 + + 2 Cell line 022 wt + + + 2 Cell line 029 HD 7 7 7 ap16 status was assessed by microsatellite analysis, sequencing, and promoter methylation assay (see text; Cairns et al. (1994)). Methylated, promoter region was methylated; HD, homozygous Figure 1 pRb IHC in primary HNSCC. (a) Rb-positive tumor b showing retention of nuclear pRb staining in neoplastic cells deletion; LOH, loss of heterozygosity; wt, wildtype. Cyclin D1 ampli®cation was assessed by FISH. N/A, fresh specimen not (arrow) and the normal interstitial tissue (arrow head). (b) Rb- c negative tumor demonstrating absence of nuclear staining in available for FISH. Fold ampli®cation compared to control cancer cells (arrows) with retention of staining in normal region centromeric probe+weak staining, protein expression con®rmed by (arrow heads) Western blot (Liggett et al., 1996) p16, Cyclin D1 and Rb in head and neck cancer KOkamiet al 3543 consistent, there is the limitation of quantitation by all Relationship between p16, cyclin D and pRb of these methods. FISH is the most reliable method to 1 analyse gene ampli®cation (Brodeur and Hogarty, Table 2 shows the correlation between the p16, pRb

1998) because it can demonstrate the minimal region and cyclin D1 status. As reported previously (Shapiro et of ampli®cation in situ and can detect the relative gain al., 1995; Parry et al., 1995; Sakaguchi et al., 1996; of gene copy number compared to a control. One cell Andl et al., 1998), inactivation of p16 and pRB had an line (029) in our study had four signals for both inverse correlation which was signi®cant by Fisher's chromosome 11 and cyclin D1 indicative of a simple exact test (P=0.0397). In our series, we found duplication of the whole chromosome, rather than a concomitant alteration of p16 and cyclin D1 in four true ampli®cation. Furthermore, the appearance of of 11 primary tumors and two of ®ve cell lines. HSRs and apparent duplication, as shown here, These results suggest that these two components are provide very strong evidence of gene ampli®cation not mutually exclusive and can both be altered in a (Brodeur and Hogarty, 1998). signi®cant proportion of primary tumors. These results

Figure 2 Cyclin D1 FISH on tumor cell lines and a primary tumor. (a) Cyclin D1 ampli®cation in cell line FaDu interphase appearance showing four chromosome 11 control signals (green; represented as bright gray) and more than 20 cyclin D1 signals (red; represented as dark gray). (b) Cyclin D1 ampli®cation in cell line 011 metaphase spread exhibiting a homogeneously staining region (HSR) in one chromosome 11 (arrow) with two normal appearing chromosomes (arrow heads). (c) Cell line 022 metaphase exhibited cyclin D1 duplication with 4 ± 6 cyclin D1 signals on the duplicated long arm of chromosome 11 (arrows) and two normal appearing chromosomes (arrow heads). (d) Primary HNSCC (#16) showing 4 control signals and 12 cyclin D1 signals (bright gray) p16, Cyclin D1 and Rb in head and neck cancer KOkamiet al 3544 Table 2 Correlation between the three components of p16/cyclin chemically on archival formalin-®xed and paran-embedded D1/Rb pathway in HNSCC primary tumors and cell lines tissue using a monoclonal antibody, G 3-245 (PharMingen, San Cyclin D1 Diego, CA, USA) at 1 : 100 ± 200 (Sakaguchi et al., 1996; P Cyclin D1 not P Geradts et al., 1994). Brie¯y, 5 mm sections were cut from pRb+ pRb7 valuea ampli®cation ampli®ed valuea blocks and deparanized. After blocking, slides were treated p16+ 4 2 2 4 and microwaved in Citra (BioGenex, San Ramon, CA, USA) p167 22 0 0.0397 6 4 0.6084 for 20 min. Slides were incubated with normal serum followed a by the primary antibody for 2 h at room temperature and (+) present, wt; (7) absent, inactivated. P value by Fischer's exact washed in PBS. Immunohistochemical staining was performed test with the use of the Vectastain Elite ABC kit (Burlongame, CA, USA), 3,3'-diaminobenzidine tetrahydrochloride development, and counterstained with methylgreen. pRb expression was must also be interpreted cautiously, however, because considered normal if de®nite nuclear staining could be of our small sample size. Using more de®nitive identi®ed in all areas of the tumor. Aberrant expression was methods, our ®ndings are consistent with previous de®ned if the tumor area showed no nuclear staining and if results in cell lines (Lukas et al., 1995) and primary de®nite nuclear staining was seen in immediately adjacent non- neoplastic cells. tumors (Olshan et al., 1997). Strikingly, all but two tumors (21/23, 91%) contained an alteration in either p16 or pRb. This observation con®rms the critical Genetic analysis

importance of the p16/cyclin D1/Rb pathway in the Homozygous deletion (HD) of p16 was detected by detailed progression of HNSCC. microsatellite analysis (Reed et al., 1996). The criteria for loss At least one in vivo study (Lukas et al., 1995) of heterozygosity (LOH) and HD was described previously (Reed et al., 1996; Cairns et al., 1995; Okami et al., 1997). showed that cyclin D1 ampli®cation and lack of p16 expression cooperatively provided greater growth The methylation status of the 5' CpG island in p16 was advantage to the tumor than either of the two determined by Southern blot analysis (Merlo et al., 1995; abnormalities alone. Recently, two reports (Zwijen et Cairns et al., 1997) and sequence analysis of exons 1 and 2 of p16 was carried out as described (Cairns et al., 1994). al., 1997; Neuman et al., 1997) showed that cyclin D1 can bind and activate the estrogen receptor indicative FISH of novel role for cyclin D1 as a CDK-independent activator of the estrogen receptor. This observation A genomic DNA probe for FISH analysis was obtained by

implies a novel function of cyclin D1 in a critical screening a genomic library, Down to the Well system transcriptional pathway, perhaps involving differentia- (GenomeSystems, St. Louis, MO, USA), with primers for the

tion responses (Jacks and Weinberg, 1998). 3' untranslated region of the cyclin D1 gene (sense, 5'- The most frequent target of inactivation, p16, is lost GGAAAGCTTCATTCTCCTTGTTG-3', anti-sense, 5'- in most HNSCC and appears critical to the initiation TCTAGGTAAACCTCTGAGGTCC-3'). The BAC (bacterial of these neoplasms (van der Riet et al., 1994; Mao et arti®cial chromosome) clone #10842 containing both the 5' and 3' ends of the gene was labeled with digoxigenein-11-dUTP al., 1996). There is evidence to suggest that both pRb (Boehringer Manheim, Indianapolis, IN, USA) by nick and cyclin D1 ampli®cation occur somewhat later in translation. The FISH procedure used on cell line preparations tumor progression (Yoo et al., 1994; Jares et al., 1994). and frozen sections was described previously (Okami et al.,

Our results suggest that cyclin D1 ampli®cation may 1997; Cairns et al., 1995; Lichter and Cremer, 1992). A biotin- confer an additional advantage to tumor cells in labeled chromosome 11 a satellite probe (Oncor, Gaithersburg

addition to p16/Rb inactivation. This advantage may MD, USA) was co-hybridized with the cyclin D1 probe, and both probes were detected using the in situ hybridization kit explain the association of cyclin D1 ampli®cation with more advanced disease and a poorer outcome (Jares et (Oncor, Gaithersburg MD, USA) according to the manufac- al., 1994; Kyomoto et al., 1997; Michalides et al., turer's instruction. Signals from at least 100 nuclei were counted 1995). under a ¯uorescence microscope equipped with a triple-pass ®lter. In normal epithelia or lymphocytes, the percentage of

nuclei with two signals of cyclin D1 was 52 ± 67%. Three or more Materials and methods signals were observed in only 5 ± 6% of the nuclei in normal areas, suggesting a low incidence of non-speci®c hybridization Cell line and primary tumor samples for this probe. On the basis of the evaluation of normal non malignant areas, we set the criteria for determination of cyclin Five HNSCC cell lines were derived from patients at the D gene ampli®cation as follows. If a nucleus contained; (a) Johns Hopkins Hospital in the Department of Otolaryngol- 1 more than three signals of both cyclin D1 and chromosome 11; ogy (011, 012, 022, and 029) or obtained from the American and (b) if the ratio of cyclin D /chromosome 11 was over 1.1, the Type Culture Collection (FaDu). These cell lines were 1 nucleus was considered to have cyclin D1 ampli®cation. cultured and harvested as previously described (Liggett et Furthermore, if the percentage of the nuclei with amplification al., 1996). The p16 and pRb status of these cell lines were exceeded 20%, the sample was interpreted as having cyclin D1 de®ned previously (Lukas et al., 1995; Liggett et al., 1996). ampli®cation (Jenkins et al., 1997; Simpson et al., 1997). For Twenty-three randomly selected primary HNSCC tumors documentation, images were captured by a CCD camera were collected following surgical resection with prior consent (Photometrics, Tucson, AZ, USA) and processed using the from the Johns Hopkins Hospital patients. DNA from tumor Oncor Image analyzing system (Oncor). samples and blood as a normal control were extracted as described (van der Riet et al., 1994).

IHC Abbreviations pRb, retinoblastoma protein; cdk, cyclin dependent kinase; IHC for p16 was processed on fresh frozen sections as described p16, CDKN2/MTS-1/INK4A; HNSCC, Head and Neck (Reed et al., 1996) using a monoclonal antibody, DCS-50 (Lab Squamous Cell Carcinoma; IHC, immunohistochemistry; Vision, Fremont, CA, USA). pRb was stained immunohisto- FISH, ¯uorescence in situ hybridization. p16, Cyclin D1 and Rb in head and neck cancer KOkamiet al 3545 References

Andl T, Kahn T, Pfuhl A, Nicola T, Erber R, Conradt C, Michalides R, van Veelen N, Hart A, Loftus B, Wientjens E KleinW,HelbigM,DietzA,WeidauerHandBoschFX. and Balm A. (1995) Cancer Res., 55, 975 ± 978. (1998). Cancer Res., 58, 5±13. Neuman E, Ladha MH, Lin N, Upton TM, Miller SJ, Brodeur GM and Hogarty MD. (1998). The Genetic Basis of DiRenzoJ,PestellRG,HindsPW,DowdySF,BrownM Human Cancer. Vogelstein B and Kinzler KW. (eds). and Ewen ME. (1997) Moll. Cell. Biol., 17, 5338 ± 5347. McGraw Hill: New York, pp.161 ± 172. Okami K, Cairns P, Westra WH, Linn JF, Ahrendt SA, Wu Cairns P, Mao L, Merlo A, Lee DJ, Schwab D, Eby Y, L, Sidransky D and Jen J. (1997) Int. J. Cancer, 74, 588 ± Tokino K, van der Riet P, Blaugrund JE and Sidransky D. 592. (1994). Science (Washington DC), 265, 415 ± 417. Olshan AF, Weissler MC, Pei H, Conway K, Anderson S, Cairns P, Polascik TJ, Eby Y, Tokino K, Califano J, Merlo Fried DB and Yarbrough WG. (1997) Oncogene, 14, 811 ± A, Mao L, Herath J, Jenkins R, Westra W, Rutter JL, 818. BucklerA,GabrielsonE,TockmanM,Cho,KR,Hedrick Parry D, Bates S, Mann DJ and Peters G. (1995) EMBO J., L,BovaGS,IsaacsW,KochW,SchwabDandSidransky 14, 503 ± 511. D. (1995). Nat. Genet., 11, 210 ± 212. Reed AL, Califano J, Cairns P, Westra WH, Jones RM, Cairns P, Okami K, Halachmi S, Halachmi N, Esteller M, Koch M, Ahrendt S, Eby Y, Sewell D, Nawroz H, Bartek J Herman, JG, Jen J, Isaacs, WB, Bova GS and Sidransky and Sidransky D. (1996) Cancer Res., 56, 3630 ± 3633. D. (1997). Cancer Res., 57, 4497 ± 5000. SakaguchiM,FujiiY,HirabayashiH,YoonHE,KomotoY, Geradts J, Hu SX, Lincoln CE, Benedict WF and Xu HJ. Oue T, Kusafuka T, Okada A and Matsuda H. (1996) Int. (1994) Int. J. Cancer, 58, 161 ± 167. J. Cancer, 65, 442 ± 445. Jacks T and Weinberg RA. (1998). Science (Washington Serrano M, Lee HW, Chin L, Cordon-Cardo C, Beach D and DC), 280, 1035 ± 1036. DePinho RA. (1996) Cell, 85, 27 ± 37. Jares P, Fernandez PL, Campo E, Nadal A, Bosch F, Aiza G, Shapiro GI, Edwards CD, Kobzik L, Godleski J, Richards Nayach I, Traserra J and Cardesa A. (1994). Cancer Res., W, Sugarbaker DJ and Rollins BJ. (1995) Cancer Res., 55, 54, 4813 ± 4817 505 ± 509. Jenkins RB, Qian J, Lieber MM and Bostwick DG. (1997). Sherr CJ. (1996). Science (Washington DC), 274, 1672 ± Cancer Res., 57, 524 ± 531. 1677. Kyomoto R, Kumazawa H, Toda Y, Sakaida N, Okamura Simpson JF, Quan DE, O'Malley F, Odom Maryon T and A, Iwanaga M, Shintaku M, YamashitaT, Hiai H and Clarke PE. (1997) Am.J.Pathol.,151, 161 ± 168. Fukumoto M. (1997) Int. J. Cancer, 74, 576 ± 581. van der Riet P, Nawroz H, Hruben RH, Corio R, Tokino K, Lichter P. and Cremer T. (1992). Human Cytogenetics vol. 1, Koch W and Sidransky D. (1994) Cancer Res., 54, 1156 ± a practical approach (2nd ed.). Rooney DE and 1158. Czepulkowski BW. (eds). Oxford IRL, pp.157 ± 192. Weinberg RA. (1995) Cell, 81, 323 ± 330. Liggett Jr WH, Sewell DA, Rocco J, Ahrendt SA, Koch W YooGH,XuHJ,BrennanJA,WestraW,HrubanRH,Koch and Sidransky D. (1996) Cancer Res., 56, 4119 ± 4123. W, Benedict WF and Sidransky D. (1994) Cancer Res., 54, Lukas J, Aagaard L, Strauss M and Bartek J. (1995). Cancer 4603 ± 4606. Res., 55, 4818 ± 4823. Zwijen RML, Wientjens E, Klompmaker R, van der Sman J, MaoL,LeeJS,FanYH,RoJY,BatsakisJG,LippmanS, Bernards R and Michalides RJ. (1997) Cell, 88, 405 ± 415. Hittelman W and Hong WK. (1996) Nat. Med., 2, 682 ± 685. MerloA,HermanJG,MaoL,LeeDJ,GabrielsonE,Burger PC, Baylin SB and Sidransky D. (1995) Nat. Med., 1, 686 ± 692.