Expression of p 16, Rb, and Cyclin D1 Gene Products in Oral and Laryngeal Squamous Carcinoma: Biological and Clinical Implications

ADEL K. EDNAGGAR, MD, SYELING LAI, MD, GARY L. CLAYMAN, MD, JAIN-HUA ZHOU, MD, PHD, SUSAN A. TUCKER, PHD, JEFFREY MYERS, MD, MARIO A. LUNA, MD, AND WILLIAM F. BENEDICT, MD

Cyclin DI, pl6, and Rb genes play a critical role in the regulation of the GI-S transition of the cell cycle and are frequently altered in several neoplastic entities. Analysis of the protein products of these genes by molecular and immtmohistochemical methods provides information on their functional status and allows for the phenotypic evaluation of tumor cells, We performed Western blotting and immunohistochemical analysis on tissues from 35 primary oral and laryngeal squamous carcinoma specimens with previous molecular analysis of the pl6 gene and correlated the results with relevant clinicopathologic factors. Our study shows significant concordance between Western blotting and immunostaining results for cyclin D1 (P = .01), pl6 proteins (P = .01), and Rb (P = .04). Heterogeneous staining of tumor cells and the positivity of non-neoplastic host elements for Rb by immunohistochemistry contributed to the discrep-

Regulation of the G1-S transition of the cell cycle is tightly control led by the Rb pathway proteins, which include the p16, Rb, and cyclin D1 gene products. 1-3 Recent evidence suggests that cyclin D1 complexes with CDK4,6 to hypophosphoryla te Rb to active forms be- fore fur ther phosphoryla t ion to the inactive state. 4 This latter hyperphosphoryla t ion leads to its dissociation f rom the E2F transcription factor and the initiation of DNA replication. Under normal regulatory conditions, Rb hyperphosphoryla t ion in turn can be inhibited by pl6 , aiding in the control of cell entry to S-phase. 5-~1 In addition, a feedback autoregulatory mechanism involv- ing activated and inactivated Rb has been implicated in the modula t ion of cyclin D1 and p16 expression. 5J2-14 Studies of these genes in several neoplastic entities have shown f requent alterations suggesting an impor tan t role in tumorigenesis. ~5-26

Analyses of cell cycle genes, however, have been conducted by variable methodologies, materials, and pat ient populations, resulting in a lack of consensus on their clinical application. ~6-24,27 Most of these studies

From the Departments of Pathology, Biomathematics, Head and Neck Surgery, and Genitourinary Medical Ontology, The University of Texas M.D. Anderson Cancer Center, Houston, TX.

Accepted for publication April 20, 1999. Supported by grant 1P50DE11906-01 from the National Institute

of DentaI Health to AEN and CA54672 from the NCI to W.B. Address correspondence and reprint requests to Adel K. E1-

Naggar, MD, PhD, Department of Pathology, The University of Texas, M.D. Anderson Cancer Center, 1515 Holcombe Blvd, Box 85, Hous- ton, TX 77030.

Copyright © 1999 by W.B. Saunders Company 0046-8177/99/3009-0003510.00/0

ancy noted in some tumors by Western blotting. Significant reciprocal relationship between pl6 and Rb proteins was observed (P < .001); in most tumors, absence of pl6 (89%) and detectable Rb (94%) proteins were found. Two tumors had negative cyclin D1 expression, and one third overexpressed this protein. There was a lack of correlation between cyclin D1 overexpression and the clinicopatho- logic factors studied. Our results indicate that the absence of p16 in most of these tumors may constitute an early tumorigenic event and that the loss of the Rb function plays a minor role in HNSC. HUM PATROL 30:1013-1018. Copyright © 1999 by W.B. Saunders Company

Key words: head and neck squamons carcinoma, cell cycle, tumor suppressor gene, immunohistochemistry, Western blotting.

Abbreviations: HNSC, head and neck squamous carcinoma; IHC, immunohistochemistry; WB~ Western blotting.

were p e r f o r m e d by molecular techniques using fresh or frozen tissues, but they are cumbersome, require highly skilled personnel, and do not distinguish between normal , benign, and tumor cells. 24-,26 Immunohis to- chemical analysis of the prote in products of these genes, a l though somewhat subjective and semiquantita- five, allows for their phenotypic localization at the cellular level on both frozen and archival t i ssues . 16,17,27

In this study, we pe r fo rmed concur ren t immunohis- tochemical and Western blott ing analyses of the p l6 , Rb, and cyclin-D1 gene products in a cohor t of p r imary head and neck squamous carcinoma (HNSC) to deter- mine their differential expression and functional roles in these tumors.

MATERIALS AND METHODS

Tumor

Matched pairs of fresh normal mucosa and tumor speci- mens from 35 resected primary HNSC received at the frozen section suite of the Department of Pathology, The University of Texas, M.D. Anderson Cancer Center, from November 1992 to December 1995 formed the materials for this study. Normal mucosa was obtained from the farthest margin of resection after frozen section verification. Invasive carcinoma speci- mens were carefully dissected from grossly viable tumor and were subjected to froze~ section for evaluation of tumor content. Specimens were obtained by 1 head and neck pathologist (A.E.-N.) and promptly snap frozen and kept at -80°C until used. Selected paraffin-embedded, formalin- fixed tumor blocks of these tumors were used for immunostain- ing.

1013

HUMAN PATHOLOGY Volume 30, No, 9 (September 1999)

Protein Extraction

Frozen tissues were homogenized and sonicated for 15 seconds twice in 500 tlL lyse buffer (1 × phosphate-buffered saline, 1% Nonidet-P40, 0.5% sodium deoxycholate, 0.1% sodium-dodecyl sulfate, and 0.1 mg /mL phenylmethylsulfo- nyl fluoride) and incubated on ice for 30 minutes. The lysates were centrifuged and the supernatant was collected (450 tlL). Protein concentration was determined spectrophotometri- cally using a DC protein assay kit (Bio-Rad Laboratory, Hercules, CA.).

RB

For Rb, a tumor was considered to have a normal heterogeneous Rb pattern if various intensities of Rb nuclear staining was seen in malignant cells, (Fig 1, case 17). Such a normal pattern was expected because of the differences in Rb protein expression usually observed throughout the cell cycle. 2s Tumors were scored as Rb negative only if no tumor cell had Rb nuclear staining, but contiguous Rb positive normal cells were present as an internal positive control (Fig 1, Case 18). sl

Western Blot Analysis

One hundred micrograms total protein was size fraction- ated on 8% (for Rb), 10% (for cyclin DI), and 12% (for p16) acrylamide gels. The proteins were blotted onto a nitrocellu- lose membrane. (Hybond-N, Amersham International, United Kingdom) with a transblot cell (Bio-Rad Laboratory). The membranes were blocked in a phosphate-buffered saline containing 5% dried milk. Antibodies used were mouse monoclonal antibody against Rb (Santa Cruz Biotechnology, Inc., Santa Cruz, CA) and rabbit polyclonal antibody for cyclin D1 (gift from Dr Lawrence Zuckerberg, Boston, MA) and p16 i*~k4 (PharMingen, San Diego, CA). Proteins were detected by an enhanced chemiluminescence system (Amersham, Ar- lington Heights, IL) according to the manufacturer's instruc- tion.

To determine whether cyclin D1 is overexpressed, the protein level in Western blotting was measured by a Personal Densitometer SI, Dimension 466V (Molecular Dynamics, Sunnyvale, CA). Density in different blots was standardized by subtracting the background reading. Overexpression was determined if the ratio of tumor to normal band density was higher than 2 SD of the normal tissue.

IHC Analysis and Scoring

Immunohistochemical staining for cyclin D1, Rb, and p16 was performed on formalin-fixed, paraffimembedded tissue sections. The protocol for staining Rb nuclear protein has been previously described. 2s Briefly, after hydrogen perox- ide treatment (0.3% in absolute methanol for 30 minutes) to block the endogenous peroxide activity, the slides were washed in glass distilled water and placed in 0.01 mol /L sodium citrate buffer (pH 6.0) for 15 minutes at 95°C, which was followed by rinsing the distilled water and phosphate- buffered saline. The slides were then processed for staining of Rb protein using the polyclonal anti-Rb antibody, Rb-WL-129 at a dilution of 1:628. The staining protocol for p16 was similar, except that monoclonal anti-pl6 antibody, NCL-p16, clone DCS-50 (Vector Laboratories, Burlingame, CA) at a 1:25 dilution was used. s° For cyclin D1, (gift from Dr. Lawrence Zuckerberg) at 1:200 dilution was used. The primary antibod- ies were visualized, using their respective ABC Elite Kits (Vector Laboratories, Burlingame, CA) with 0.05% 3,3'- diaminobenzidine in Tris-HCL buffer containing 0.01% tolu- idine blue. In addition, all cut sections were kept at 4°C before staining. Staining was performed for p16 within 14 days after sectioning, because unlike staining for Rb, that can be done at greatly extended periods after sectioning, and p16 protein staining can diminish rapidly if staining is delayed after preparation of slides. 3°

Irnmunohistochemical Scoring

Scoring was performed in a blinded fashion by 2 indepen- dent observers for each marker.

p16

Tumors were categorized as p16 negative if no malignant cells had positive staining and at least several contiguous pl6-positive normal cells were present as internal control. Some cytoplasmic p16 staining was seen in most pl6-positive cases but was not considered in the evaluation. Tumor were scored positive if rumor cells showed heterozygous or homoge- neous nuclear staining with and without cytoplasmic reactivity (Fig 1, Cases 18, 2, and 30). Although many tumors had strong positive staining in cellular nuclei, as recently found in bladder cancers s° weaker heterogeneous versus strong p16 homogeneous staining was not evaluated in this study.

Cyclin D1

Nuclear staining for cyclin D1 was non-uniform in all tumors, reflecting the difference in expression with peak in G1 of the cell cycle (Fig 2, case 24). Tumors were scored cyclin D1 negative if fewer than 5% of tumor cells showed nuclear staining and positive if 5% to 20% (1+), 20 to 50 (2+), and more than 50% (3+) of the cells, respectively, showed nuclear staining in tumor cell as that previously reported. 19,s2m

Detection of p 16 Genetic Alteration

Polymerase chain reaction (PCR)-based LOH and muta- tional analysis as well as the methylation status of exons 1 and 2 were carried out in a previous study. 25

RESULTS

Table 1 shows the c l in icopathologic features o f each pa t ien t and status o f p16, Rb, and cyclin D1 in HNSC. Patients compr i sed 23 m e n and 12 w o m e n rang ing in age f rom 35 to 80 years, with a m e a n age o f 56.3 years. The re were 14 larynx and 21 oral cavity tumors. Histologically, 3 were well differentiated, 22 were modera te ly differentiated, and 10 were poor ly differentiated. Patients t u m o r stages compr i sed 6 stage I, 5 stage II, 5 stage III, and 19 stage IV tumors . Twenty tumors were DNA diploid, and 15 were DNA aneuplo id , with S-phase ranges f rom 7% to 28% (mean, 13.0%). All patients were t rea ted primari ly with surgery, and 3 h a d addi t ional postoperat ive chemotherapy . Four pat ients were lost to follow-up, and for the r emain ing 31 patients follow-up r anged f r o m 3 to 173 m o n t h s (mean, 40.3 months ) . At the last t ime o f contact , 27 patients are alive (4 with disease and 21 free o f disease), 1 d ied o f disease, and 3 d ied o f o the r causes. Only 2 tumors each were negative for Rb and cyclin D1. Eleven cases showed h igh expression (+ +) o f cyclin D1.

1014

p16, Rb, AND CYCLIN D1 GENE PRODUCTS (EI-Naggar et al)

Rb

I p16 D1 Rb p16 D1

17 18

FIGURE 1. Composite Western bloffing and immunohisto: chemical staining of cases 17 and 18, Concordant results between WB and immunostaining for RB, p16, and cyclin D1 is shown in case 17. In case 18, concordance between cyclin D1 and p16 expression but a negative immunostaining in tumor cells and a detectable Rb band by WB and negative staining in tumor cells are noted.

Table 2 presents the correlation between p16 protein as measured by immunohistochemicai (IHC) compared with Western blotting (WB) analysis. Concor- dant results between IHC and WB analysis was found in 30 (85.7%) of 35 cases (P = .002); 26 of these 30 tumors were p16 negative, and the 4 tumors were positive. In the 5 tumors with discrepancy between these tech- niques, negative staining by immunohistochemistry and apparently false-positive WB was observed. This is illus- trated in case 17 of Figure 1, where faint p16 band on WB but negative tumor cells staining and positively staining stromal cells are observed. The presence of scattered positive nuclear and cytoplasmic staining in host normal elements in these cases most likely contrib- uted to the WB results. There were no instance of tumors scored as positive by IHC and negative by WB.

The correlation between immunostaining and the molecular analysis z5 of the p16 gene is shown in Table 3. Of the 29 in which molecular analysis was available, 3 tumors had point or frame shift mutation, 6 showed homozygous deletion, and 9 were methylated. Eleven (37.9%) had no molecular or methylation alterations. Twenty-one (72%) tumors manifested concordant re- sults between p16 protein expression by immunohisto- chemistry and genomic analyses. In the 18 tumors with me thylation or DNA alteration, no p 16 nuclear staining was noted. Eight tumors showed negative immunostain- ing and lacked any genomic alterations, and 3 tumors stained positively for p16 and had no genomic alter- ations. The correlation between molecular analysis and IHC for p16 was statistically significant (P = .04). No incidence of molecular alteration and positive p16 protein immunostaining was found.

Table 4 presents the correlation between pRb status as measured by IHC and WB results. Twenty-six (74.3%) of the 35 tumors manifested concordant re- sults; 25 were positive, and 1 tumor was negative by both techniques (Figs 1, 2). Discordant results were noted in 9 tumors (25.7%); 8 tumors were positive by immuno- staining but were negative by WB, and 1 tumor was negative by IHC but positive by WB (Fig 1, case 18). There were positive statistical correlation between the

results of the 2 techniques (P = .04). Eight tumors with positive results by IHC were false-negative WB results. In these cases, Rb staining was heterogeneous, and the ratio of tumor to stromal was generally lower than in cases that were positive by both techniques. The only tumor with negative immunostaining and positive WB showed a strong stromal Rb endothelial cell immunore- activity (Fig 2, case 30).

The correlation between Rb and p16 expression (Table 5) by immunohistochemical staining showed significant reciprocal correlation (P < .001). Thirty-two tumors (91.4%) showed an inverse relationship in protein staining; 31 cases had negative p16 and positive Rb staining, and 1 showed negative Rb and positive p16 staining. Three (8.6%) tumors were positive, and 1 tumor was negative for both proteins.

For cyclin D1, significant correlation between IHC and WB (Table 6) analysis was also noted (P = .04). Twenty-nine (82.8%) tumors were concordant; 27 had positive and 2 tumors had negative results by both techniques (Fig 2, case 30). Protein degradation was ruled out in these 2 tumors because of the presence of Rb or p16 proteins in the same preparation. Only 6 (17.4%) tumors showed positive cyclin D1 immunostain- ing but were negative by WB. All of these cases had weaker cyclin D1 staining than the other 27 cases, which were positive by WB. Eleven tumors (31%) had strong nuclear staining in more than 50% of cells and were considered to represent overexpression for cyclin D1. Because of the insufficient tumor specimens left for Southern blotting, molecular correlation studies were not performed in these specimens.

DISCUSSION

Although considerable information is available on the expression of individual cell cycle-related genes in different tumor types, a consensus on their clinical relevance for many malignancies remains uncertain, mainly because of the variable methodologies, materi- als, and patient populations used in these studies. ~5-~ In

Rb p16 D1 Rb p16 D1

24 30

FIGURE 2. Depicts two examples of immunohistochemical and Western bloffing results in two different squamous carci- noma (cases 24 and 30). In case 24, negative p16 by both methods and positive results for Rb and cyclin D1 are shown. In case 30, a rare instance of lack of Rb and positive p l 6 proteins is illustrated. In this the same case, also a ra re instance, of negative cyclin D1 is found,

1015

HUMAN PATHOLOGY Volume 30, No. 9 (September 1999)

TABLE 1. Clinicopathologic Features and p16, Cyclin D1, and Rb Tumor Expression Status in HNSC

Survival No. Sex Age Differentiation Stage DI S-phase Rx Metastasis (mo) Rb p16 D1

2 M 67 MD IV 1.49 19 S No DUC (32) + - + + 6 F 35 MD II 1.00 !2 S,C No Alive (55) + - + + 7 M 64 MD IV 1.00 9 S Skin Alive (61) + - + 8 M 69 WD I 1.00 8 S Neck Alive (173) + - +

11 M 36 MD IV 1.00 9 S No Alive (52) + - + 12 M 39 WD I 1.00 8 S No Alive (63) + - + + 14 M 51 PD IV 1.66 12 S Bone Alive (6) + + + 15 M 37 MD I 1.00 9 S No Alive (66) + - + 16 M 60 PD IV 1.00 13 S UN UN + - + 17 F 66 MD II !.61 11 S No Alive (57) + - + 19 F 56 PD I 1.00 24 S No Alive (8) + - + 20 M 53 MD IV 1.00 7 S No Alive (46) + - + + 21 M 56 MD IV 1.00 8 S,C No DUC (21) + - + 23 F 80 PD IV 2.20 28 S,C No Alive (3) + - + + 24 M 59 PD IV 1.47 20 S No Alive (16) - - + 25 M 61 MD IV 1.69 24 S Skin Alive (35) + - + + 26 F 55 MD IV 1.00 10 S UN UN + - + 27 F 77 MD IV 1.00 9 S No Alive (50) + - + 28 M 73 PD III 1.00 12 S No Alive (4) ÷ - + 29 F 62 MD II l 1.15 10 S No Alive (67) + - + + 30 M 47 MD III 1.65 13 S Lung Alive (42) - + - 31 M 35 MD I 1.00 11 S No Alive (28) + - + 32 M 67 PD IV 1.32 12 S No Alive (20) + + + + 33 M 63 PD 1V 1.83 14 S No Alive (45) + - - 34 M 35 MD I 1.00 10 S No Alive (28) + - + 35 M 67 MD II 1.00 9 S No Alive (37) + + + 36 M 77 MD III 1.00 28 S UN UN + - + 37 M 52 WD IV 1.12 10 S No Alive (28) + - + 38 F 50 MD IV 1.46 14 S UN UN + - + + 40 M 58 PD III 1.36 12 S Lung DOC (26) + - + 42 F 67 PD IV 1.78 12 S No Alive (9) + - + + 43 F 49 MD II 1.00 15 S No Alive (33) + - + + 44 M 60 MD II 1.42 13 S No DOD (25) + - + 45 F 50 MD IV 1.00 8 S No Alive (49) + - + 46 F 39 MD IV 1.00 13 S No Alive (67) + - +

Abbreviations: WD, well differentiated; MD, moderate ly differentiated; PD, Surgery; C, Chemotherapy DUC, died of unknown cause; DOC, Died of other expressed; + +, overexpression.

poorly differentiated; Diff, differentiation; Rx, t reatment; S, cause; DOD, d ied of disease; UN, unknown; - , negative; +,

molecular-based studies of DNA or RNA, the results may be affected by the inclusion of normal cellular elements and the requirement for optimal acquisition and stor- age of starting materials. Immunohistochemistry, how- ever, can be performed on frozen and archived tissues and allows for the discrimination between normal and benign cells from tumor cell expression.l~18,29

Our comparative analysis shows a good correlation between Western blotting and immunocytochemical analysis for cyclin D1 and p16 genes. P16 protein was undetected by both techniques in most of our tumors, and only a few tumors showed negative staining in cancer cells but a detectable band on Western blotting.

TABLE 2. Correlation Between p16 Protein Western Blotting and Immunohistochemistry

IHC

WB + (%) - (%) P-value

+ 4 (11.4) 5 (14.2) =.002 - 0 26 (74.2)

Abbreviations: +, positive; - , negative; WB, Western blotting; IHC, immunohis tochemical .

In these Cases, the positivity detected by IHC in stromal cells is most likely the source for the positive Western blotting results. Genomic information, including meth- ylation and DNA alterations in these tumors from a previous study, 25 correlated significantly with the immu- nostaining results. A small number of tumors, however, had negative nuclear staining but no detectable ge- nomic abnormalities of the p16. In these cases, techni- cal limitations, posttranscriptional, or promoter region changes cannot be ruled out and should be investigated in future studies. ~2

As previously reported in studies of other tumors,

TABLE 3. Correlation Between p16 Immunostaining and Genomic Alterations

IHC

Genomic Alterat ion - (%) + (%) P-value*

Yes 18 (51.4) 0 =.04 No 8 (27.8) 3 (10.3%)

Abbreviation: IHC, immunohis tochemical ; +, positive; - , nega- tive.

*Six cases lacked genetic information.

1016

p16, Rb, AND CYCLIN D1 GENE PRODUCTS (EI-Naggar eta[)

TABLE 4. Correlation Between Rb Protein Expression by Western Blotting and Immunohistochemistry in HNSC

IHC

TABLE 6. Correlation Between Cyclin D1 Expression by Western Bloffing and Immunohistochemistry in HNSC

IHC

WB + (%) - (%) P-value

+ 25 (71.4) l (2.8) =.04 - 8 (22.8) 1 (2.8)

Abbreviations: +, positive; - , negative; WB, Western blotting; IHC, immunohistochemical.

significant statistical correlation between Western blot- ting and immunostaining analysis for Rb protein was observed) 6,21,2S,27 In 9 instances, however, disagreement in the results of both methodologies was found. Eight of these cases were negative by Western blotting, but positive immunostaining in tumor cells was noted, indicating lower sensitivity of the Western technique, especially in tumors with high stromal-to-tumor ra- tio. 2s,z9 In the only negative tumor by IHC that was positive by Western blotting, the stromal cells were strongly reactive for the Rb protein. We also observed, as have others, a significant reciprocal expression of p 16 and Rb in most HNSC, supporting a cooperative nega- tive feedback mechanism between these genes in the maintenance of the G1-S checkpoint. ~1~,33 Our results are also in agreement with those of other investigators regarding the infrequent Rb alterations in HNSC tumori- genesis. 6-11

Our cyclin D1 analysis indicates a good correlation between positive IHC and Western results in most tumors. Only 2 tumors had no detectable cyclin D1 by both techniques. The mechanism and significance of this finding is unknown and could not be related to protein degradation because the samples were suitable for p16 and Rb analysis. Cyclin D1 overexpression was noted in only one third of this cohort based on more than 50% strongly positive nuclear staining. However, the clinicopathologic feature that reached marginal statistical correlation was DNA ploidy index (P = .09). The biological significance of these findings, however, must await larger studies with long-term follow-up of these patients. This together with the establishment of well-defined and reproducible criteria for accurate assessment of cyclin D1 overexpression s4-38 are neces- sary for assessing the role of this gene in this and other solid tumors.

In conclusion, our study shows (1) high frequency o f p l 6 loss in HNSC, (2) Rb plays a minor role in HNSC tumorigenesis, (3) significant reciprocal relationship between p l 6 / R b gene expression, and (4) that immuno-

TABLE 5. Correlation Between p16 and Rb Genes Expressions by IHC in HNSC

Rb

pl6 + (%) - (%) P-value

+ 2 (5.7) 1 (2.8) <.001 - 31 (88.5) 1 (2.8)

Abbreviations: +, positive; - , negative; IHC, immunohistochem- istry:

WB + (%) - (%) P-value

+ 27 (77.1) 0 - 6 * ( 1 7 . 1 ) 2 ( 5 . 7 ) = . 0 4

Abbreviations: +, positive; - , negative; IHC, immunohistochemi- cal; WB, Western blotting.

*Heterogeneous faint staining.

cytochemistry is a sensitive and reliable technique for screening the functional status o f p l 6 and Rb genes.

REFERENCES

1. Weinberg RA: The retinoblastoma protein and cell cycle control. Cell 81:323-330, 1995.

2. Hall M, Peters G: Genetic alterations of cyelins, cyclin- dependent kinases and cdk inhibitors in human cancer. Adv Cancer Res 68:67-108, 1996

3. Sherr CJ: Cancer cell cycles. Science 274:1672-1677, 1996 4. Ezhevsky SA, Nagahara H, Vocero-Akbani AM, et aI: Hypo-

phosphorylation of the retinoblastoma protein (pRb) by Cyelin D: Cdk4/6 complexes results in active pRb. Proc Natl Acad Sci 94:10699- 10704, 1997

5. Lukas J, Parry D, Aagaard L, et al: Retinoblastoma-protein- dependent cell-cycle inhibition by the tumor suppressor pl6. Nature (Lond) 375:503-506, 1995

6. Otterson GA, Kratzke RA, Coxon A, et al: Absence of p l6 INK4 protein in restricted to the subset of lung cancer lines that retains wild type RB. Oneogene 9:3375-3378, 1994

7. Sakaguchi M, Fujii Y, Hirabayashi H, et al: Inversely correlated expression of p l6 and Rb protein in non-small cell lung cancers: An immunohistoehemical study. In tJ Cancer 65:442-445, 1996

8. Shapiro GI, Edwards CD, Kobzik L, et al: Reciprocal Rb inactivation and p l6 INK4 expression in primary lung cancers and cell lines. Cancer Res 55:505-509, 1995

9. Ueki K, Ono Y, Henson JW, et al: CDKN2/pl6 or Rb alterations occur in the majority of glioblastomas and are inversely correlated. Cancer Res 56:150-153, 1996

10. Yeager T, Stadler W, Belair C, et al: Increased p l6 levels correlate with pRb alterations in human urothelial cells. Cancer Res 55:493-497, 1995

11. I~-atzke RA, Greatens TM, Rubins JB, et al: Rb and p l6 INK4 expression in resected non-small-cell lung tumors. Cancer Res 56:3414- 3420, 1996

12. Parry D, Bates S, Mann DJ, et ah Lack of cyclin D-Cdk complex in Rb-negative cells correlated with high level o f p l 6 ink4/mtsl tumor suppressor gene product. EMBOJ 14:503-511, 1995

13. Muller H, LukasJ, Schneider A, et al: Cyclin D1 expression is regulated by the retinoblastoma protein. Proc Natl Acad Sci USA 91:2945-2949, 1994

14. Fang X, Jin Z, Xu HJ, et ai: Expression of p l6 induces transcriptional downregulation of the Rb gene. Oneogene 16:1-8, 1998

15. Tanaka H, Fuji Y, Hirabayashi H, et al: Disruption of the Rb pathway and cell-proliferative activity in non-small cell lung cancers. IntJ Cancer (Pred Oncol) 79:111-115, 1998

16. Dublin EA, Patel NK, Gillet CE, et al: Retinoblastoma and p l6 proteins in mammary carcinoma: Their relationship to cyclin D1 and pathological parameter. IntJ Cancer (Pred Oncol) 79:71-75, 1998

17. March KL, Varley JM: Frequent alterations of cell cycle regulators in early-stage breast lesions as detected by immunohisto- chemistry. BrJ Cancer 77:1460-1468, 1998

18. Michalides R, Hageman PH, van Tinteren H, et al: A clinicopathologic study on overexpression of cyclin D 1 and of p53 in a series of 248 patients with operable breast cancer. Br J Cancer 73:728-734, 1996

1017

HUMAN PATHOLOGY Volume 30, No. 9 (September 1999)

19. van Diest PJ, Michalides RJ, Jannink I, et al: Cyclin D1 expression in invasive breast cancer: Correlations and prognostic value. AmJ Patho1150:705-711, I997

20. VarleyJM, Armour J, SwallowJE, et al: The retinoblastoma gene is frequently altered leading to loss of expression in primary breast tumours. Oncogene 4:725-729, 1989

21. Stauss M, Gusterson B, BartekJ: Abnormal patterns of D-type cyclin expression and G1 regulation in human head and neck cancer. Cancer Res 55:949-956, 1995

22. Maelandsmo GM, Florenes VA, Hovig E, et al: Involvement of the pRb/pl6/cdk4/cycl in D1 pathway in the tumorigenesis of spo- radic malignant melanomas. BrJ Cancer 73:909-916, 1996

23. Nielsen NH, Emdin SO, Cajander J, et al: Deregulation of cyclin E and D1 in breast cancer is associated with inactivation of the retinoblastoma protein. Oncogene 14:295-304, 1997

24. Reed AL, Califano J, Cairns R et al: High frequency of p16 (CDKN2/MTS-1/INK4A) inactivation of head and neck squamous cell carcinoma. Cancer Res 56:3630-3633, 1996

25. E1-Naggar AK, Lai S, Clayman G, et al: Methylation: A dominant mechanism of pl6/CDKN2 gene inactivation in head and neck squamous carcinoma. AmJ Patho1151:1767-1774, 1997

26. Olshan AF, Weissler MC, Pei H, et al: Alterations of the p16 gene in head and neck cancer: Frequency and association with p53, PRAD-I and HPV. Oncogene 14:811-818, 1997

27. GeradtsJ, Kratzke RA, Niehanx GA, et al: Immunohistochemi- cal detection of the cyclin-dependent kinase inhibitor 2/multiple tumor suppressor gene 1 (cdkn2/mtsl) product p16 ink4 in archival human solid tumors: Correlation with refinoblastoma protein expres- sion. Cancer Res 54:6006-6011, 1995

28. Xu HJ, Cagle PT, Hu SX, et al: Altered retinoblastoma and p53 protein status in non-small cell carcinoma of the lung: Potential synergistic effects on prognosis. Clin Cancer Res 2:1169-1176, 1996

29. Xu HJ, Hu SX, Hashimoto T, et al: The retinoblastoma susceptibility gene product: A characteristic pattern in normal cells and abnormal expression in malignant. Oncogene 4:807-812, 1989

30. Benedict WF, Lerner SP, Zhou J, et al: Level of retinoblas- toma protein expression correlates with p16 (MTS-1/INK4A/ CDKN2) stares in bladder cancer. Oncogene 18:1197-1203, 1999

31. Xu HJ, Cairns P, Hu SX, et al: Loss of Rb protein expression in primary bladder cancer correlates with loss of heterozygosity at the Rb locus and rumor progression. IntJ Cancer 53:781-784, 1993

32. Shapiro GI, ParkJE, Edwards CD, et al: Multiple mechanisms of p16 IN~z4a inactivation in non-small cell lung cancer cell lines. Cancer Res 55:6200-6209, 1995

33. Roncalli M, Bosari S, Marchetti A, et al: Cell cycle-related gene abnormalities and product expression in esophageal carcinoma. Lab Invest 78:1049-1057, 1998

34. Gillett C, Smith P, Gregory W, et al: Cyclin D1 and prognosis in human breast cancer. IntJ Cancer 69:92-99, 1996

35. Barbareschi M, Pelosio P, Orazio C, et al: Cyclin-D1 gene amplification and expression in breast carcinoma: Relation with clinicopathologic characteristics and with retinoblastoma gene prod- uct, p53 and p21 wAF1 immunohistochemical expression. In J Cancer 74:171-174, 1997

36. Schuuring E, Verhoeven E, Tinteren H van, et al: Amplifica- tion of genes within the chromosome 11q13 region is indicative of poor prognosis in patients with operable breast cancer. Cancer Res 52:5229-5234, 1991

37. Laramie G, Fanfl V, Smith R, et al: Dl12287, a putative oncogene on chromosome 11q13 is amplified and expressed in squamous cell and mammary carcinomas and linked to bcl-1. Onco- gene 6:439-444, 1991

38. Michalides R, van Veelee N, Hart A, et al: Overexpression of cyclin D1 correlates with recurrence in a group of forty-seven operable squamous cell carcinomas of the head and neck. Cancer Res 55:975-978, 1995

1018