Cyclin D 1 ampli cation is independent of p16 inactivation in head and neck squamous cell carcinoma

Size: px

Start display at page:

Download "Cyclin D 1 ampli cation is independent of p16 inactivation in head and neck squamous cell carcinoma"

Anastasia Copeland
6 years ago
Views:

1 Oncogene (1999) 18, 51 ± 55 ã 1999 Stockton Press All rights reserved 0950 ± 9/99 $ Cyclin D 1 is independent of p1 inactivation in head and neck squamous cell carcinoma K Okami 1,, AL Reed 1, P Cairns 1, WM Koch 1, WH Westra, S Wehage 1, J Jen 1 and D Sidransky*,1 1 Department of OtolaryngologyÐHead & Neck Surgery, Division of Head and Neck Cancer Research, Johns Hopkins University School of Medicine, 818 Ross Research Building, 0 Rutland Avenue, Baltimore, Maryland, MD , USA; Department of Pathology, 181 Meyer Building, Johns Hopkins Hospital, 00 North Wolfe Street, Baltimore, Maryland, MD 18, USA; Department of Otolaryngology, Yamaguchi University, Ube 55, Japan Progression through the G 1 phase of the cell cycle is mediated by phosphorylation of the retinoblastoma protein (prb) resulting in the release of essential transcription factors such as EF-1. The phosphorylation of prb is regulated positively by cyclin D 1 /CDK and negatively by CDK inhibitors, such as p1 (CDKN/ MTS-1/INKA). The p1 /cyclin D 1 /Rb pathway plays a critical role in tumorigenesis and many tumor types display a high frequency of inactivation of at least one component of this pathway. In order to determine the overall contribution of these three components to progression of head and neck squamous cell carcinoma (HNSCC), we examined p1 inactivation, cyclin D 1, and prb expression in primary HNSCC tumors and ve cell lines. p1 inactivation was detected in 19/ (8%) primary tumors by detailed genetic analysis and was con rmed by immunohistochemistry (). Absence of Rb protein expression indicative of prb inactivation was identi ed in / (9%) tumors. In this set of tumors, there was a perfect inverse correlation between p1 and prb inactivation. Using uorescence in situ hybridization (FISH) cyclin D 1 was identi ed in /5 (80%) cell lines and /11 (%) primary tumors. However, / cell lines and all four primary tumors with cyclin D 1 contained a concomitant alteration of p1. Therefore 1/ (91%) of primary HNSCC contained at least one alteration in the p1/cyclin D 1 /Rb pathway. Although p1 and Rb alteration are apparently exclusive, cyclin D 1 occurs concomitantly with the loss of p1 suggesting an additional role for this in HNSCC. Keywords: cyclin D 1 ; p1; head and neck cancer Introduction The G 1 phase of the cell cycle is controlled by phosphorylation of the retinoblastoma protein (prb) and mediated in part by cyclin D-cyclin dependent *Correspondence: D Sidransky Received 0 October 1998; revised 8 January 1999; accepted 5 March 1999 kinase (cdk) complexes. This cdk phosphorylation activity is negatively regulated by the tumor suppressor gene p1 (a CDK inhibitor), and positively by cyclin D 1. Each individual component of this p1-cyclin D 1 - Rb pathway is commonly targeted in various malignancies (Weinberg, 1995; Sherr, 199). We previously reported a high frequency (8%) of p1 inactivation in human HNSCC (Reed et al., 199) and less frequent inactivation (1%) of prb (Yoo et al., 199). Cyclin D 1 has been reported in ± % of HNSCC based on several di erent genetic or methods (Jares et al., 199; Kyomoto et al., 199; Olshan et al., 199; Michalides et al., 1995). However, few reports have simultaneously examined all components of this critical pathway in HNSCC (Olshan et al., 199; Lukas et al., 1995). p1 and prb inactivation have been reported to be inversely correlated in some tumor types (Shapiro et al., 1995; Parry et al., 1995; Sakaguchi et al., 199; Andl et al., 1998). It has thus been assumed that each component of this pathway is an alternative target for abrogation of a common pathway and that alterations of multiple components in the same pathway is not necessary for tumor progression (Serrano et al., 199). We thus undertook a comprehensive study of all potential alterations within the p1-cyclin D 1 -Rb pathway in cell lines and primary HNSCC. We found a high frequency of p1 inactivation and cyclin D 1 in cell lines and primary HNSCC. Moreover, a signi cant proportion of primary tumors contained concomitant alterations of p1 and cyclin D 1 suggesting that unlike p1 and prb, alterations of these two components are not mutually exclusive. Results and dscussion p1 status The p1 gene status in these tumors has been reported previously (Reed et al., 199). p1 inactivation was detected by in 19 out of (8%) cases (Table 1). The results were con rmed by genetic analysis as previously reported (1 had at the p1 locus, ve contained methylation of the promoter region of p1, and one had a frameshift mutation in exon 1; Reed et al., 199). analysis of p1 appears to be a reliable and practical method to determine p1 inactivation in fresh tissue as reported (Reed et al., 199).

The incidence of prb inactivation was consistent with our previous study on primary HNSCC (Yoo et al., 199; Andl et al., 1998).

2 5 p1, Cyclin D 1 and Rb in head and neck cancer prb prb showed strong nuclear staining in both tumor and normal cells (Figure 1a). prb inactivation was observed in two cases out of (9%) exhibiting a good contrast between the negative staining tumor area and the positive staining normal areas (Figure 1b). The incidence of prb inactivation was consistent with our previous study on primary HNSCC (Yoo et al., 199; Andl et al., 1998). Cyclin D 1 by FISH FISH was rst performed on normal lymphocyte spreads which con rmed the chromosomal location of the BAC probe on chromosome 11q1 (data not shown). In order to determine the incidence of cyclin D 1 gene, FISH analysis was carried out on both HNSCC cell lines and primary tumors. We identi ed gene in four out of ve (80%) cell lines (Table 1). The FaDu cell line contained four control signals and more than 0 cyclin D 1 signals (Figure 1a). We found two typical appearances of gene in cell lines 011, 01 and 0. Cell line 011 had ± control signals and 10 ± 1 cyclin D 1 signals, and in the metaphase spreads of chromosome 11, it exhibited a homogeneously staining region (HSR, Brodeur and Hogarty, 1998) of the cyclin D 1 gene (Figure b). Cell line 01 and 0 metaphase spreads displayed ± cyclin D 1 signals due to a duplication of the long arm of chromosome 11 seen in each nucleus (Figure c). In the primary tumors, we found four cases of gene out of 11 (%) cases available (Table 1). Figure d shows a primary tumor with four control signals and 1 cyclin D 1 signals. Previously, of cyclin D 1 was reported in ± % of HNSCC (Jares et al., 199; Kyomoto et al., 199; Olshan et al., 199; Michalides et al., 1995). These studies used various methods to detect the including Southern blot analysis, Northern blot hybridization, di erential PCR, or. Although the frequency of amplification detected by these various approaches is somewhat Figure 1 prb in primary HNSCC. (a) Rb-positive tumor showing retention of nuclear prb staining in neoplastic cells (arrow) and the normal interstitial tissue (arrow head). (b) Rbnegative tumor demonstrating absence of nuclear staining in cancer cells (arrows) with retention of staining in normal region (arrow heads) Table 1 Primary tumors/ Cell line FaDu Cell line 011 Cell line 01 Cell line 0 Cell line 09 Aberrations of p1, cyclin D1 and prb in HNSCC primary tumors and cell lines p1 status a Retained Frame shift in exon 1 Retained Point mutation wt wt p1 prb b Fold c a p1 status was assessed by microsatellite analysis, sequencing, and promoter methylation assay (see text; Cairns et al. (199))., promoter region was methylated;, homozygous deletion;, loss of heterozygosity; wt, wildtype. b was assessed by FISH., fresh specimen not available for FISH. c Fold compared to control centromeric probeweak staining, protein expression con rmed by Western blot (Liggett et al., 199) 5

consistent, there is the limitation of quantitation by all of these methods.

a control. One cell line (09) in our study had four signals for both chromosome 11 and cyclin D 1 indicative of a simple duplication of the whole chromosome, rather than a true.

p1, and Rb in head and neck cancer Relationship between p1, cyclin D 1 and prb Table shows the correlation between the p1, prb and cyclin D 1 status. As reported previously (Shapiro et al.

In our series, we found concomitant alteration of p1 and cyclin D 1 in four of 11 primary tumors and two of ve cell lines.

3 consistent, there is the limitation of quantitation by all of these methods. FISH is the most reliable method to analyse gene (Brodeur and Hogarty, 1998) because it can demonstrate the minimal region of in situ and can detect the relative gain of gene copy number compared to a control. One cell line (09) in our study had four signals for both chromosome 11 and cyclin D 1 indicative of a simple duplication of the whole chromosome, rather than a true. Furthermore, the appearance of HSRs and apparent duplication, as shown here, provide very strong evidence of gene (Brodeur and Hogarty, 1998). p1, and Rb in head and neck cancer Relationship between p1, cyclin D 1 and prb Table shows the correlation between the p1, prb and cyclin D 1 status. As reported previously (Shapiro et al., 1995; Parry et al., 1995; Sakaguchi et al., 199; Andl et al., 1998), inactivation of p1 and prb had an inverse correlation which was signi cant by Fisher's exact test (P=0.09). In our series, we found concomitant alteration of p1 and cyclin D 1 in four of 11 primary tumors and two of ve cell lines. These results suggest that these two components are not mutually exclusive and can both be altered in a signi cant proportion of primary tumors. These results 5 Figure Cyclin D 1 FISH on tumor cell lines and a primary tumor. (a) Cyclin D 1 in cell line FaDu interphase appearance showing four chromosome 11 control signals (green; represented as bright gray) and more than 0 cyclin D 1 signals (red; represented as dark gray). (b) Cyclin D 1 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 0 metaphase exhibited cyclin D 1 duplication with ± cyclin D 1 signals on the duplicated long arm of chromosome 11 (arrows) and two normal appearing chromosomes (arrow heads). (d) Primary HNSCC (#1) showing control signals and 1 cyclin D 1 signals (bright gray)

4 5 Table Correlation between the three components of p1/cyclin D1/Rb pathway in HNSCC primary tumors and cell lines p1 p1 prb prb must also be interpreted cautiously, however, because of our small sample size. Using more de nitive methods, our ndings are consistent with previous results in cell lines (Lukas et al., 1995) and primary tumors (Olshan et al., 199). Strikingly, all but two tumors (1/, 91%) contained an alteration in either p1 or prb. This observation con rms the critical importance of the p1/cyclin D 1 /Rb pathway in the progression of HNSCC. At least one in vivo study (Lukas et al., 1995) showed that cyclin D 1 and lack of p1 expression cooperatively provided greater growth advantage to the tumor than either of the two abnormalities alone. Recently, two reports (Zwijen et al., 199; Neuman et al., 199) showed that cyclin D 1 can bind and activate the estrogen receptor indicative of novel role for cyclin D 1 as a CDK-independent activator of the estrogen receptor. This observation implies a novel function of cyclin D 1 in a critical transcriptional pathway, perhaps involving differentiation responses (Jacks and Weinberg, 1998). The most frequent target of inactivation, p1, is lost in most HNSCC and appears critical to the initiation of these neoplasms (van der Riet et al., 199; Mao et al., 199). There is evidence to suggest that both prb and cyclin D 1 occur somewhat later in tumor progression (Yoo et al., 199; Jares et al., 199). Our results suggest that cyclin D 1 may confer an additional advantage to tumor cells in addition to p1/rb inactivation. This advantage may explain the association of cyclin D 1 with more advanced disease and a poorer outcome (Jares et al., 199; Kyomoto et al., 199; Michalides et al., 1995). Materials and methods P value a p1, Cyclin D 1 and Rb in head and neck cancer not ampli ed P value a Cell line and primary tumor samples Five HNSCC cell lines were derived from patients at the Johns Hopkins Hospital in the Department of Otolaryngology (011, 01, 0, and 09) or obtained from the American Type Culture Collection (FaDu). These cell lines were cultured and harvested as previously described (Liggett et al., 199). The p1 and prb status of these cell lines were de ned previously (Lukas et al., 1995; Liggett et al., 199). Twenty-three randomly selected primary HNSCC tumors were collected following surgical resection with prior consent from the Johns Hopkins Hospital patients. DNA from tumor samples and blood as a normal control were extracted as described (van der Riet et al., 199) () present, wt; () absent, inactivated. a P value by Fischer's exact test for p1 was processed on fresh frozen sections as described (Reed et al., 199) using a monoclonal antibody, DCS-50 (Lab Vision, Fremont, CA, USA). prb was stained immunohistochemically on archival formalin- xed and para n-embedded tissue using a monoclonal antibody, G -5 (PharMingen, San Diego, CA, USA) at 1 : 100 ± 00 (Sakaguchi et al., 199; Geradts et al., 199). Brie y, 5 mm sections were cut from blocks and depara nized. After blocking, slides were treated and microwaved in Citra (BioGenex, San Ramon, CA, USA) for 0 min. Slides were incubated with normal serum followed by the primary antibody for h at room temperature and washed in PBS. Immunohistochemical staining was performed with the use of the Vectastain Elite ABC kit (Burlongame, CA, USA),,'-diaminobenzidine tetrahydrochloride development, and counterstained with methylgreen. prb expression was considered normal if de nite nuclear staining could be identi ed in all areas of the tumor. Aberrant expression was de ned if the tumor area showed no nuclear staining and if de nite nuclear staining was seen in immediately adjacent nonneoplastic cells. Genetic analysis Homozygous deletion () of p1 was detected by detailed microsatellite analysis (Reed et al., 199). The criteria for loss of heterozygosity () and was described previously (Reed et al., 199; Cairns et al., 1995; Okami et al., 199). The methylation status of the 5' CpG island in p1 was determined by Southern blot analysis (Merlo et al., 1995; Cairns et al., 199) and sequence analysis of exons 1 and of p1 was carried out as described (Cairns et al., 199). FISH A genomic DNA probe for FISH analysis was obtained by screening a genomic library, Down to the Well system (GenomeSystems, St. Louis, MO, USA), with primers for the ' untranslated region of the cyclin D 1 gene (sense, 5'- GGAAAGCTTCATTCTCCTTGTTG-', anti-sense, 5'- TCTAGGTAAACCTCTGAGGTCC-'). The BAC (bacterial arti cial chromosome) clone #108 containing both the 5' and ' ends of the gene was labeled with digoxigenein-11-dutp (Boehringer Manheim, Indianapolis, IN, USA) by nick translation. The FISH procedure used on cell line preparations and frozen sections was described previously (Okami et al., 199; Cairns et al., 1995; Lichter and Cremer, 199). A biotinlabeled chromosome 11 a satellite probe (Oncor, Gaithersburg MD, USA) was co-hybridized with the cyclin D 1 probe, and both probes were detected using the in situ hybridization kit (Oncor, Gaithersburg MD, USA) according to the manufacturer's instruction. Signals from at least 100 nuclei were counted under a uorescence microscope equipped with a triple-pass lter. In normal epithelia or lymphocytes, the percentage of nuclei with two signals of cyclin D 1 was 5 ± %. Three or more signals were observed in only 5 ± % of the nuclei in normal areas, suggesting a low incidence of non-speci c hybridization for this probe. On the basis of the evaluation of normal non malignant areas, we set the criteria for determination of cyclin D 1 gene as follows. If a nucleus contained; (a) more than three signals of both cyclin D 1 and chromosome 11; and (b) if the ratio of cyclin D 1 /chromosome 11 was over 1.1, the nucleus was considered to have cyclin D 1. Furthermore, if the percentage of the nuclei with amplification exceeded 0%, the sample was interpreted as having cyclin D 1 (Jenkins et al., 199; Simpson et al., 199). For documentation, images were captured by a CCD camera (Photometrics, Tucson, AZ, USA) and processed using the Oncor Image analyzing system (Oncor). Abbreviations prb, retinoblastoma protein; cdk, cyclin dependent kinase; p1, CDKN/MTS-1/INKA; HNSCC, Head and Neck Squamous Cell Carcinoma;, immunohistochemistry; FISH, uorescence in situ hybridization.

5 References Andl T, Kahn T, Pfuhl A, Nicola T, Erber R, Conradt C, KleinW,HelbigM,DietzA,WeidauerHandBoschFX. (1998). Cancer Res., 58, 5±1. Brodeur GM and Hogarty MD. (1998). The Genetic Basis of Human Cancer. Vogelstein B and Kinzler KW. (eds). McGraw Hill: New York, pp.11 ± 1. Cairns P, Mao L, Merlo A, Lee DJ, Schwab D, Eby Y, Tokino K, van der Riet P, Blaugrund JE and Sidransky D. (199). Science (Washington DC), 5, 15 ± 1. Cairns P, Polascik TJ, Eby Y, Tokino K, Califano J, Merlo A, Mao L, Herath J, Jenkins R, Westra W, Rutter JL, BucklerA,GabrielsonE,TockmanM,Cho,KR,Hedrick L,BovaGS,IsaacsW,KochW,SchwabDandSidransky D. (1995). Nat. Genet., 11, 10 ± 1. Cairns P, Okami K, Halachmi S, Halachmi N, Esteller M, Herman, JG, Jen J, Isaacs, WB, Bova GS and Sidransky D. (199). Cancer Res., 5, 9 ± Geradts J, Hu SX, Lincoln CE, Benedict WF and Xu HJ. (199) Int. J. Cancer, 58, 11 ± 1. Jacks T and Weinberg RA. (1998). Science (Washington DC), 80, 105 ± 10. Jares P, Fernandez PL, Campo E, Nadal A, Bosch F, Aiza G, Nayach I, Traserra J and Cardesa A. (199). Cancer Res., 5, 81 ± 81 Jenkins RB, Qian J, Lieber MM and Bostwick DG. (199). Cancer Res., 5, 5 ± 51. Kyomoto R, Kumazawa H, Toda Y, Sakaida N, Okamura A, Iwanaga M, Shintaku M, YamashitaT, Hiai H and Fukumoto M. (199) Int. J. Cancer,, 5 ± 581. Lichter P. and Cremer T. (199). Human Cytogenetics vol. 1, a practical approach (nd ed.). Rooney DE and Czepulkowski BW. (eds). Oxford IRL, pp.15 ± 19. Liggett Jr WH, Sewell DA, Rocco J, Ahrendt SA, Koch W and Sidransky D. (199) Cancer Res., 5, 119 ± 1. Lukas J, Aagaard L, Strauss M and Bartek J. (1995). Cancer Res., 55, 818 ± 8. MaoL,LeeJS,FanYH,RoJY,BatsakisJG,LippmanS, Hittelman W and Hong WK. (199) Nat. Med.,, 8 ± 85. MerloA,HermanJG,MaoL,LeeDJ,GabrielsonE,Burger PC, Baylin SB and Sidransky D. (1995) Nat. Med., 1, 8 ± 9. p1, and Rb in head and neck cancer Michalides R, van Veelen N, Hart A, Loftus B, Wientjens E and Balm A. (1995) Cancer Res., 55, 95 ± 98. Neuman E, Ladha MH, Lin N, Upton TM, Miller SJ, DiRenzoJ,PestellRG,HindsPW,DowdySF,BrownM and Ewen ME. (199) Moll. Cell. Biol., 1, 58 ± 5. Okami K, Cairns P, Westra WH, Linn JF, Ahrendt SA, Wu L, Sidransky D and Jen J. (199) Int. J. Cancer,, 588 ± 59. Olshan AF, Weissler MC, Pei H, Conway K, Anderson S, Fried DB and Yarbrough WG. (199) Oncogene, 1, 811 ± 818. Parry D, Bates S, Mann DJ and Peters G. (1995) EMBO J., 1, 50 ± 511. Reed AL, Califano J, Cairns P, Westra WH, Jones RM, Koch M, Ahrendt S, Eby Y, Sewell D, Nawroz H, Bartek J and Sidransky D. (199) Cancer Res., 5, 0 ±. SakaguchiM,FujiiY,HirabayashiH,YoonHE,KomotoY, Oue T, Kusafuka T, Okada A and Matsuda H. (199) Int. J. Cancer, 5, ± 5. Serrano M, Lee HW, Chin L, Cordon-Cardo C, Beach D and DePinho RA. (199) Cell, 85, ±. Shapiro GI, Edwards CD, Kobzik L, Godleski J, Richards W, Sugarbaker DJ and Rollins BJ. (1995) Cancer Res., 55, 505 ± 509. Sherr CJ. (199). Science (Washington DC),, 1 ± 1. Simpson JF, Quan DE, O'Malley F, Odom Maryon T and Clarke PE. (199) Am.J.Pathol.,151, 11 ± 18. van der Riet P, Nawroz H, Hruben RH, Corio R, Tokino K, Koch W and Sidransky D. (199) Cancer Res., 5, 115 ± Weinberg RA. (1995) Cell, 81, ± 0. YooGH,XuHJ,BrennanJA,WestraW,HrubanRH,Koch W, Benedict WF and Sidransky D. (199) Cancer Res., 5, 0 ± 0. Zwijen RML, Wientjens E, Klompmaker R, van der Sman J, Bernards R and Michalides RJ. (199) Cell, 88, 05 ±

Level of retinoblastoma protein expression correlates with p16 (MTS-1/INK4A/CDKN2) status in bladder cancer

Oncogene (1999) 18, 1197 ± 1203 ã 1999 Stockton Press All rights reserved 0950 ± 9232/99 $12.00 http://www.stockton-press.co.uk/onc Level of retinoblastoma protein expression correlates with p16 (MTS-1/INK4A/CDKN2)