INK4a Gene Expression and Methylation in Primary Breast ... · cancers using methylation-specific...

INK4a Gene Expression and Methylation in Primary Breast Cancer:Overexpression of p16INK4a Messenger RNA Is a Marker ofPoor Prognosis1

Rina Hui, R. Douglas Macmillan,Frances S. Kenny, Elizabeth A. Musgrove,Roger W. Blamey, Robert I. Nicholson,John F. R. Robertson, and Robert L. Sutherland2

Cancer Research Program, Garvan Institute of Medical Research, St.Vincent’s Hospital, Darlinghurst, Sydney, New South Wales 2010,Australia [R. H., E. A. M., R. L. S.]; Department of Surgery, CityHospital, Nottingham NG5 1PB, United Kingdom [R. D. M., F. S. K.,R. W. B., J. F. R. R.]; and Tenovus Cancer Research Centre, WelshSchool of Pharmacy, Cardiff University, Cardiff CF10 3XF, UnitedKingdom [R. I. N.]

ABSTRACTFrequent deletions or mutations of the INK4 gene,

which encodes the cyclin-dependent kinase 4 inhibitorp16INK4a, have been documented in various human cancers,but little is known about the role of this tumor suppressorgene in primary breast cancer. We examined p16INK4a

mRNA expression and its relationship with cyclin D1 andestrogen receptor (ER) expression in 314 primary breastcancers using Northern blots probed with a p16 exon 1a-specific cDNA. Tumor samples overexpressing p16INK4a

were predominantly ER negative with low levels of cyclinD1. Cyclin D1 and ER mRNA levels in the high p16INK4a

expressers were significantly lower than those in the remain-der of the population (P 5 0.0001). Furthermore, the meanp16INK4a mRNA level in the ER-negative tumors was signif-icantly higher than that in the ER-positive group (P 50.0001). Because theINK4 gene is frequently inactivated byde novomethylation, we investigated the frequency ofINK4aexon 1a methylation in a subset of 120 primary breastcancers using methylation-specific PCR; 24 of these weremethylated. These findings indicate that high expression ofp16INK4a and reduced expression due tode novo INK4amethylation are frequent events in primary breast cancer. Ina subset of 217 patients for whom detailed clinical data were

available, high p16INK4a mRNA expression was associatedwith high tumor grade (P 5 0.006),> 4 axillary lymph nodeinvolvement (P 5 0.004), ER negativity (P5 0.0001), andincreased risk of relapse (P 5 0.006). The significant nega-tive correlation between p16INK4a and ER gene expressionraises issues regarding their functional interrelationshipsand whether high p16INK4a expression may be associatedwith a lack of hormone responsiveness in breast cancer.

INTRODUCTIONp16INK4a and other members of the INK4 family of Cdk3

inhibitors inhibit the G1 cyclin D-dependent kinases, Cdk4 andCdk6, which phosphorylate pRb and facilitate entry into S phase(1, 2). The fact that p16INK4a can block G1-S-phase progressionand that mutant p16INK4a proteins are nonfunctional in cell cyclearrest or Cdk inhibition suggests that p16INK4a plays an impor-tant role in negative growth control (3, 4). Moreover, the findingthat theINK4a gene is frequently deleted in cancer cell lines (5,6) indicates that it may be a tumor suppressor gene. This wassubsequently confirmed in a p16INK4a knockout mouse model inwhich there is direct evidence that p16INK4a deficiency facili-tates tumor development (7). Furthermore, deletions and muta-tions of theINK4a gene develop in early lesions of Barrett’sesophagus, head and neck cancer, and bladder cancer (8–11),suggesting that molecular alterations leading to p16INK4a inac-tivation occur relatively early during carcinogenesis at sometissue sites.

The absence of p16INK4a expression is seen predominantlyin cells that retain wild-typeRB (12). However, p16INK4a isoverexpressed in cancer cell lines and tumors in which pRb isdysfunctional (13–16), providing evidence for a negative feed-back loop in which the functionally inactive pRb fails to se-quester transcription factors, which, in turn, induceINK4a geneexpression. There is accumulating evidence to suggest thatexpression of p16INK4a and pRb is mutually counterbalanced tomaintain growth-inhibitory activity in the cyclin D1-Cdk4-p16INK4a-pRb pathway of cell cycle control. In cultured cells,pRb represses transcription of theINK4a gene (17), whereasexpression of p16INK4a induces transcriptional down-regulationof the RB gene (18). In the event of inactivating mutations ordeletion of one of these tumor suppressor genes, the other wouldbe overexpressed (17, 18); however, high expression of eitherpRb or p16INK4a alone is incapable of inhibiting cell cycleprogression (19).

p16INK4a and p15INK4b are colocalized to chromosome

Received 11/24/99; revised 4/3/00; accepted 4/10/00.The costs of publication of this article were defrayed in part by thepayment of page charges. This article must therefore be hereby markedadvertisementin accordance with 18 U.S.C. Section 1734 solely toindicate this fact.1 Supported by grants from the National Health and Medical ResearchCouncil of Australia, the New South Wales State Cancer Council, andthe Tenovus Organization. R. H. is a recipient of a National Health andMedical Research Council Medical Postgraduate Research Scholarshipand the Beng Kang Kho Scholarship.2 To whom requests for reprints should be addressed, at Cancer Re-search Program, Garvan Institute of Medical Research, St. Vincent’sHospital, 384 Victoria Street, Darlinghurst, Sydney, New South Wales2010, Australia. Phone: 61-2-92958320; Fax: 61-2-92958321.

3 The abbreviations used are: Cdk, cyclin-dependent kinase; ER, estro-gen receptor; pRB, retinoblastoma protein; LOH, loss of heterozygosity;PBL, peripheral blood lymphocyte; LN, lymph node.

2777Vol. 6, 2777–2787, July 2000 Clinical Cancer Research

Research. on March 26, 2021. © 2000 American Association for Cancerclincancerres.aacrjournals.org Downloaded from

http://clincancerres.aacrjournals.org/

9p21, a locus commonly deleted in several human cancers. TheINK4agene has a complex structure. Transcription of theINK4agene can yield two distinct transcripts (a or b mRNAs) codingfor two functionally distinct proteins, p16INK4a and p19ARF (foralternative reading frame). The two transcripts have the sameexons 2 and 3 but contain a different exon 1, designated exon 1aand exon 1b(20–23). Whereas thea transcript is selectivelyexpressed in some tissues in humans as well as mice, thebtranscript is ubiquitously expressed. Given the dissimilarity instructure with other known Cdk inhibitors and the failure tocoprecipitate p19ARF with Cdc2, Cdk2, Cdk4, Cdk6, cyclin D2,cyclin D3, cyclin E, and cyclin A, p19ARF is not a direct Cdkinhibitor. However, ectopic expression of p19ARF in cells withhomozygous deletion ofINK4a induces cell cycle arrest in boththe G1 phase and G2-M phases of the cell cycle with a concom-itant loss of cells in S phase (22). More recent work hasindicated that p19ARF interacts with MDM2, promotes MDM2degradation, and, in turn, stabilizes p53 (24, 25). Thus bothp16INK4a and p19ARF induce cell cycle arrest at apparentlydifferent points in the cell cycle and via distinct mechanisms;consequently, deletion ofINK4a impairs both the pRb and p53tumor suppressor pathways.

Inactivation ofINK4aoccurs frequently in a wide spectrumof sporadic primary cancers and familial melanoma. Mutation orhomozygous deletion ofINK4aoccurs with a frequency rangingfrom approximately 20% in sporadic melanoma, non-small celllung cancer, head and neck cancer, esophageal cancer, andmalignant mesothelioma to 30% in transitional cell cancer of thebladder, 35% in gliomas, and 50% in pancreatic cancer andsquamous cell carcinoma of the bladder (6, 26–29). However,unlike the situation in cancer cell lines, homozygous deletionand mutation ofINK4a are very rarely observed in primarybreast cancers (26, 30, 31). DNA methylation of the humanINK4a gene is associated with gene silencing and hence inacti-vation of INK4a in some human cancers including head andneck, lung, brain, colon, esophageal, and bladder cancers andalso in a small series of breast cancers (32, 33).

In general, p16INK4a inactivation is associated with a moreaggressive phenotype and worse prognosis in a wide range ofneoplasms including pancreatic carcinoma, malignant mela-noma, glioma, leukemia, non-Hodgkin’s lymphoma, and non-small cell lung cancer (34–45). In breast cancer, one publishedstudy (46) reported an association between LOH at 9p21–22 andworse prognostic features including high S-phase fraction, ane-uploidy, and large tumor size ($2 cm), although no associationwith patient survival was demonstrated in a short period of 3years median follow-up. Conversely, another smaller study (47)failed to demonstrate a relationship between LOH at 9p21–22and other clinicopathological parameters in 68 breast cancers.The only study available in the literature specifically examiningthe prognostic significance of p16INK4a in breast cancer reportedthat poor outcome was associated with high expression ofp16INK4a as assessed by strong immunohistochemical staining(16).

Given the paucity of data on the prognostic significance ofthe tumor suppressor geneINK4a in breast cancer, a study ofp16INK4a mRNA expression, including its relationship with cy-clin D1 and ER status, was initiated in a series of tumors from314 patients. The relationship between p16INK4a and various

clinicopathological features and clinical outcome was studied ina subset of 217 patients. Furthermore, the frequency ofINK4amethylation was also investigated using methylation-specificPCR in a subset of 120 samples.

MATERIALS AND METHODSClinicopathological Features. The clinical details of the

series of patients included in this study have been describedpreviously (48–50). In brief, Northern blots containing totalRNA extracted from breast cancers of 364 patients who under-went surgery at the Nottingham Breast Unit during the periodbetween February 1987 and December 1993 were probed withp16INK4a exon 1acDNA. Of the 314 breast cancers evaluablefor p16INK4a mRNA expression, full clinical follow-up datawere available on 217 patients with stage I or II disease. Allpatients under 70 years of age underwent axillary LN samplingwith either simple mastectomy or wide local excision followedby adjuvant postoperative radiotherapy. Adjuvant chemotherapywith cyclophosphamide, methotrexate, and 5-fluorouracil or ta-moxifen was introduced from 1989, based on the NottinghamPrognostic Index (51), age, and ER status.

Data on cyclin D1, ER, and 36B4 mRNA expression wereavailable from previous studies on this series of breast cancers(49, 50).

Breast Cancer Cell Lines and cDNA Probes. Thesources of the breast cancer cell lines were as described below.BT-20, BT-483, BT-549, DU-4475, Hs-578T, MDA-MB-134,MDA-MB-175, MDA-MB-361, MDA-MB-436, MDA-MB-453, MDA-MB-468, SK-BR-3, and ZR-75-1 were obtainedfrom the American Type Culture Collection (Manassas, VA).HBL-100, MCF-7M, MDA-MB-157, MDA-MB-231, MDA-MB-330, and T-47D were obtained from the EG & G MasonResearch Institute (Worcester, MA). The full-length 960-bpp16INK4a cDNA was provided by Dr. David Beach (Cold SpringHarbor Laboratory, Cold Spring Harbor, NY).

A 340-bp p16INK4a exon 1a fragment was generated byPCR from human PBL DNA using exon 1a primers 59-GAA-GAAAGAGGAGGGGCTG and 59-GCGCTACCTGATTC-CAATTC. PCR reactions were performed in a total volume of50 ml. The reaction contained 5ml of 103 PCR buffer [pH 8.3;100 mM Tris-HCl, 500 mM KCl, 15 mM MgCl2, and 0.01%gelatin], 10 mmol of deoxynucleotide triphosphate, 10 pmol ofeach primer, 0.1mg of DNA template, 3.6% formamide, and 5units of Taq polymerase (Boehringer Mannheim, Mannheim,Germany). PCR reactions were overlayed with mineral oil. Thecycling parameters included an initial denaturing step at 94°Cfor 4 min, followed by 92°C for 1 min, 52°C for 30 s, and 72°Cfor 30 s for 30 cycles, and a final elongation cycle of 72°C for5 min. The amplified PCR product was purified through twoSepharose 6 L CB (Sigma Chemical Co., St. Louis, MO) col-umns equilibrated with 13TE buffer [pH 7.4; 10 mM Tris and1 mM EDTA] and centrifuged for 5 min at 6003 g. The 340-bpp16INK4a exon 1acDNA was validated by sequencing.

Analysis of p16INK4a mRNA Expression. The details oftotal RNA extraction from the primary breast cancer samplesand Northern blotting to determine the expression of cyclin D1,ER mRNA, and the ribosomal protein 36B4 as a control forRNA loading were described previously (49, 50). The filters

2778p16INK4a and Prognosis in Breast Cancer



were strip washed in 0.1% SSC and 0.1% SDS at 100°C for 5min and reprobed with the specific p16INK4a exon 1a PCRproduct. Each filter contained four control cell lines for normal-ization between the 22 filters. The ratio of p16INK4a:36B4mRNA signal intensity for each sample was normalized to thatof HBL-100, which was defined arbitrarily as 10 to yield “rel-ative expression of p16INK4a mRNA.”

DNA Methylation Assay. Methylation of the INK4agene in breast tumors and cell lines was detected by methyla-tion-specific PCR (52) using the CpG WIZ Methylation Assay[Oncor, Gaithersburg, MD]. Onemg of tumor or cell line DNAwas modified by sodium bisulfite according to the protocol andsubsequently purified by ethanol precipitation. All unmethyl-ated cytosines were converted to uracils, whereas 5-methylcy-tosines remained unaltered. The methylation status of the treatedDNA was then determined by PCR amplification using specificprimers within the promoter region. The reaction contained 2.5ml of 103 PCR buffer [20 mM Tris-HCl (pH 7.5), 100 mM KCl,1 mM DTT, 0.1 mM EDTA, 15 mM MgCl2, 0.5% Tween 20,0.5% NP40, and 50% glycerol], 2.5ml of 2.5 mM deoxynucle-otide triphosphate mix, 1ml of each primer (the CpG WIZMethylation Assay Kit), 2ml of modified DNA template, and2.6 units of Taq polymerase from the Expand High Fidelity PCRSystem (Boehringer Mannheim). PCR reactions were overlayedwith mineral oil (Sigma Chemical Co.). The cycling parametersincluded an initial denaturing step at 95°C for 5 min, followedby 30 cycles of 95°C for 45 s, 60°C for 45 s, and 72°C for 1 min.Five ml of PCR reactions were analyzed by electrophoresisthrough a 2% agarose gel, followed by ethidium bromide stain-ing. The size of the methylatedINK4a product was 145 bp,whereas that of the unmethylatedINK4a product was 154 bp.

Statistical Analysis. Follow-up data were taken fromtime of last clinic appointment or date of death. Median fol-low-up was 74 months. Of the 217 patients, 81 patients hadrelapsed, 54 had developed distant metastases, and 44 had diedof breast cancer at the time of analysis. Deaths from unrelatedcauses were censored for purposes of survival analyses. Allstatistical calculations were performed using the SPSS DataAnalysis Program (SPSS for Windows 6.1.3; SPSS UK Ltd.).The association between p16INK4a mRNA expression and otherclinicopathological variables was determined by thex2 testusing Fisher’s exact (two-tailed) test. Survival outcomes wereassessed using univariate Cox regression analysis, multivariateCox proportional hazards model, and life table analysis usingthe Wilcoxon Gehan statistic (53). The relationship between theexpression of p16INK4a mRNA and cyclin D1 or ER mRNA wasdetermined using the nonparametric Mann-WhitneyU test.

RESULTSp16INK4a Exon 1a Expression in Breast Cancer Cell

Lines. Given the complexity of theINK4a gene with twopartially overlapping transcripts produced from separate pro-moters, a specific probe for the first exon of human p16INK4a

(exon 1a) was generated by PCR to avoid cross-reaction withthe first exon of p19ARF (exon 1b). A 1.6-kb transcript wasclearly defined on Northern blots of RNA from a panel of breastcancer cell lines probed with full-length p16INK4a or p16INK4a

exon 1acDNA (Fig. 1A). The cell lines with known homozy-

gous deletion [e.g.,MDA-MB-231 and MCF-7 (31)] did notexpress p16INK4a mRNA when probed with either full-lengthp16INK4a or p16INK4a exon 1acDNA (Fig. 1A). However, threeadditional cell lines (T-47D, MDA-MB-134, and DU-4475)lacked evidence of gene expression when the specific exon 1aPCR product was used as a probe, in contrast to the clearlydefined transcript demonstrated when blotted for full-lengthp16INK4a, indicating that p19ARF but not p16INK4a was ex-pressed in these cell lines. Thus, the exon 1a-specific probe forp16INK4a expression was used throughout this study.

p16INK4a Exon 1a Expression in 314 Primary BreastCancers. Total RNA from 364 primary breast cancers wasNorthern blotted on 22 filters and probed sequentially withcDNAs for cyclin D1, ER, p16INK4a exon 1a, and the ribosomalprotein 36B4, the last of which was used as a control for RNAloading [Fig. 1B(49, 50)]. RNAs from 50 breast cancer sampleswere degraded and unable to be evaluated for p16INK4a expres-sion and were therefore excluded from the analysis. The fre-quency distribution of p16INK4a mRNA was unimodal and pos-itively skewed, with the relative p16 mRNA levels ranging up to28.26 (Fig. 1C). However, most of the samples expressed verylow levels of p16INK4a with a median level of 1.76.

Relationship between p16INK4a and Cyclin D1 Expres-sion. In cell cycle regulation, p16INK4a induces a conforma-tional change in Cdk4 and Cdk6 that reduces their affinity forcyclin D1, thereby preventing cyclin D1-Cdk4/Cdk6 assemblyand pRb phosphorylation. Thus, the relative abundance ofp16INK4a and cyclin D1 may affect Cdk4/Cdk6 activity. We firstexamined the relationship between p16INK4a and cyclin D1expression in 314 primary breast cancers. An inverse relation-ship between p16INK4a and cyclin D1 expression was clearlydemonstrated (Fig. 2A). Cyclin D1 mRNA levels in thep16INK4a high expressers (using the median of p16INK4a as thecutoff) were significantly lower than those in the remainder ofthe population (P5 0.0002). Similarly, when the tumors weredivided into halves according to their cyclin D1 levels, therelative p16INK4a mRNA expression in the half with lowercyclin D1 levels was significantly higher than that in the halfwith higher cyclin D1 mRNA (P5 0.0001).

Relationship between p16INK4a and ER Expression.Given that ER is a known marker of prognosis and a predictorfor therapeutic responsiveness to endocrine treatment in breastcancer, the relationship between p16INK4a and ER expressionwas also examined. An inverse relationship between p16INK4a

and ER mRNA levels was clearly shown, and this relationshipwas even more striking in the high expressers of p16INK4a (Fig.2B). ER mRNA levels in the p16INK4a overexpressers weresignificantly lower than those in the remainder of the population(P 5 0.035). p16INK4a mRNA levels in the ER-negative groupwere significantly higher than those in the ER-positive group(P 5 0.0001). Of the 10% of tumors (i.e., 32 samples) thatexpressed the highest levels of p16INK4a mRNA, only six wereER positive, and their relative ER mRNA levels were all verylow (,0.3 arbitrary units). When p16INK4a mRNA levels weredivided into quartiles, the proportion of ER-negative samplesincreased from 15% in all of the lowest three quartiles to 49%in the highest quartile.

2779Clinical Cancer Research



Fig. 1 A, expression of full-length p16INK4a andp16INK4a exon 1a mRNA in breast cancer celllines. A Northern blot containing 30mg of RNAfrom 11 breast cancer cell lines is shown. Thefilter was probed sequentially with p16INK4a exon1a and full-length p16INK4a cDNAs. The ER statusand pRB status of these cell lines are shown (31).B, expression of p16INK4a mRNA in primarybreast cancers. A representative Northern blot con-taining the control breast cancer cell lines HBL-100, MDA-MB-231, MCF-7M, and MDA-MB-134 (20, 10, 20, and 5mg of RNA, respectively)and 20mg of RNA from each of the 18 primarybreast cancer samples is shown. The filters wereprobed sequentially with [a-32P]dCTP-labeled cy-clin D1, ER, p16INK4a exon 1a, and 36B4 cDNAs.The sizes of the transcripts are indicated.C, fre-quency distribution of p16INK4a mRNA levels in314 primary breast cancers.




INK4a Methylation in Breast Cancer Cell Lines andPrimary Breast Cancers. Given that a substantial proportionof breast tumors expressed very low levels of p16INK4a mRNAand that homozygous deletions and point mutations of theINK4a gene are uncommon in breast cancers (30, 31), weinvestigated whether the gene was inactivated by DNA meth-ylation. Exon 1 and 2 coding sequences of theINK4a genecontain a number of CpG islands, and methylation of CpGislands within or near the promoters of theINK4a gene canresult in a loss of gene expression. Using methylation-specificPCR, we examined the frequency ofINK4a methylation in apanel of 19 breast cancer cell lines with known p16INK4a ex-pression and a subset of 120 primary breast cancers for whichmatched DNA was available. Of the 18 breast cancer cell linestested, three exhibited completeINK4a methylation [DU-4475,MDA-MB-134, and T-47D (Fig. 3A)], and one exhibited partialINK4a methylation (ZR 75-1). As expected, no PCR productwas produced in cell lines with homozygous deletion ofINK4a(MCF-7, MDA-MB-231, Hs-578T, and BT-20) using eithermethylated or unmethylated primer sets (Table 1). Unmethyl-atedINK4awas demonstrated in 10 of 18 breast cancer cell lines

(BT-483, BT-549, HBL-100, MDA-MB-157, MDA-MB-175,MDA-MB-361, MDA-MB-436, MDA-MB-453, MDA-MB-468, and SK-BR3) and in both normal cell types [human PBLsand normal breast epithelial cells (184)]. Of 120 primary breastcancers, 24 (20%) exhibitedINK4amethylation (Fig. 3B). Whenthe relative p16INK4a mRNA levels were divided into tertiles, 57tumors were examined forINK4a methylation status in the firsttertile (i.e., lowest expression), 33 were examined forINK4amethylation status in the second tertile, and 30 were examinedfor INK4a methylation status in the third tertile. Of the 24methylated samples, 18 (32%) were in the lowest tertile ofp16INK4a expression, 5 (15%) were in the second tertileof p16INK4a expression, and 1 (3%) was in the highest tertile ofp16INK4a expression (Fig. 3C).

Relationship between p16INK4a mRNA and Prognosis ina Subset of 217 Patients. Of the 314 patients, 217 had fullclinical follow-up data to allow survival analysis. The p16INK4a

levels were corrected for RNA loading with the non-estrogen-regulated ribosomal protein 36B4. For analyses of the relation-ship between p16INK4a mRNA expression and clinicopatholog-ical parameters or survival outcome, patients were divided intotwo equal groups using the median p16INK4a level as the cutoffpoint.

The relationship between p16INK4a mRNA expression andvarious clinicopathological parameters is summarized in Table2. High p16INK4a mRNA expression was associated with hightumor grade (P5 0.006),$4 axillary LN involvement (P50.021), and ER negativity (P5 0.0001). In contrast, no rela-tionships between high p16INK4a mRNA and patient age, men-opausal status, and tumor size were evident.

Life-table analysis revealed that high p16INK4a mRNAexpression was associated with increased risk of relapse in thewhole population of 217 patients (P 5 0.006; Fig. 4A). Al-though the overall survival statistics failed to reach significanceusing the median expression as the cutoff, there is an apparentdivergence of the cumulative proportion overall survival curveswith a trend for worse prognosis in patients with high p16INK4a

mRNA levels (P5 0.082; Fig. 4B). The association with in-creased risk of death became statistically significant when theanalysis was performed using p16 expression as a continuousvariable (P5 0.0037). The Cox multivariate analyses of severalpathological features using relapse-free survival and overallsurvival as end points showed that axillary LN involvement.3(P 5 0.0330) and tumor size.2 cm (P 5 0.0289) wereassociated with increased risk of relapse, whereas axillary LN.3 (P 5 0.0433) and high tumor grade (P5 0.0096) wereassociated with increased risk of death from breast cancer. Incontrast, p16 overexpression was not an independent predictorof early relapse or death (P5 0.6558 and 0.0766, respectively).This is probably not surprising given the relationship betweenp16 overexpression and LN status, which is the single mostimportant independent predictor of outcome.

Cyclin D1 mRNA overexpression is associated with worseprognosis in patients with ER-positive but not ER-negativebreast cancers (48). Moreover, patients with ER-negative dis-ease generally have a less favorable outcome, and given the tightinverse relationship between p16INK4a mRNA expression andER, the reduced disease-free survival in patients with highp16INK4a mRNA expression may be accounted for by its asso-

Fig. 2 Relationship between p16INK4a mRNA expression and (A) cy-clin D1 mRNA expression and (B) ER mRNA expression in primarybreast cancers. p16INK4a, cyclin D1, and ER mRNA levels in RNAsamples extracted from 314 primary breast cancers are plotted.




ciation with ER negativity. Thus, survival analyses within theER subgroups were performed. The association between highp16INK4a mRNA expression and early relapse was upheld withinthe ER-positive subgroup (P5 0.04; Fig. 5A). Unfortunately,the small sample size within the ER-negative group (n 5 54) didnot provide enough events to allow meaningful statistical anal-ysis. There was no association between p16INK4a mRNA ex-pression and disease-free survival within the ER-negative sub-group.

Given that axillary LN status is the single most importantprognostic indicator in breast cancer and that high p16INK4a

mRNA expression is associated with increased nodal involve-ment ($4 LNs), survival analyses within node-negative andnode-positive subgroups were performed. In common with thewhole population survival analyses (n5 217), the associationbetween high p16INK4a expression and early relapse was main-tained in the larger node-positive subgroup (n 5 154; P 50.0382; Fig. 5B). However this association was lost in the much

Fig. 3 Methylation status ofINK4a in (A) breastcancer cell lines and (B) primary breast cancers.Representative gel analyses of methylation-specific PCR reactions on DNA from six breastcancer cell lines and PBLs or six breast tumorsamples are shown. A PCR fragment of 145 or 154bp in size was evident ifINK4a was originallymethylated or unmethylated in the cell line orbreast cancer sample, respectively.C, relationshipbetween INK4a methylation status andINK4agene expression. Data are presented as the percent-age of primary breast cancer samples with meth-ylated and unmethylatedINK4a in each tertile ofrelative p16INK4a mRNA expression.




smaller node-negative subgroup (n 5 33). High p16INK4a

mRNA expression had no impact on overall survival in eithernodal subgroup.

Because high cyclin D1 (48) and high p16INK4a mRNAexpression were each a marker of worse prognosis in this seriesof patients, we examined the potential prognostic significance ofthe combination of these two parameters. Although there wasclearly an inverse relationship between p16INK4a and cyclin D1,there was a significant proportion of tumors exhibiting moderateoverexpression of both parameters when the median was used asthe cutoff. Patients with tumors expressing high cyclin D1/highp16INK4a mRNA levels had increased risk of early relapse ascompared with the group with low cyclin D1/low p16INK4a

mRNA expression (Fig. 6A;P 5 0.0164). Similarly, the groupwith low cyclin D1 levels but high p16INK4a mRNA levels hadan increased risk of early relapse as compared with the groupwith concurrently low cyclin D1 and p16INK4a mRNA expres-sion (P5 0.0340). The association between high expression ofboth cyclin D1/p16INK4a mRNA and early relapse as comparedwith low expression of both cyclin D1/p16INK4a mRNA reachedgreater statistical significance within the ER-positive subgroup(Fig. 6B; P 5 0.0054). There was no survival difference be-tween patients with high cyclin D1/low p16INK4a and low cyclinD1/high p16INK4a mRNA levels in the whole population orwithin ER subgroups.

These data demonstrate that independent and concurrentoverexpression of cyclin D1 and p16INK4a mRNAs were mark-ers of poor prognosis in this series of breast cancers, but little isknown of the prognostic value of p16INK4a inactivation. Con-sequently, survival analyses were carried out in relation toINK4a methylation status. Of the subset of 120 breast cancersfor which methylation data were available, 97 patients hadadequate follow-up data to allow survival analyses. No associ-ation was demonstrated betweenINK4a methylation status andthe overall or disease-free survival in this group of patients.

DISCUSSIONCyclin D1 and Cdk4 accelerate, whereas p16INK4a and pRb

inhibit, cell cycle progression at the late G1 phase of the cellcycle. The loss of functional p16INK4a or pRb has been identi-fied in a variety of human cancers but has not been well studiedin breast cancer. This study provides extensive clinical data onp16INK4a expression in breast cancer. For the first time, a tightinverse relationship between p16INK4a and ER mRNA expres-sion was demonstrated. Moreover, the findings also indicatedthat both p16INK4a inactivation by hypermethylation andp16INK4a overexpression occur frequently in primary breastcancer, and overexpression of p16INK4a mRNA is a marker ofpoor prognosis.

Growth-suppressive effects of p16INK4a generally requirefunctional pRb (54, 55). The observation of high levels ofp16INK4a expression in pRb-negative cells, which is likely dueto loss of a feedback loop regulated by pRb (13–16, 56), and thefact that pRb expression is transcriptionally repressed by ectopicexpression of p16INK4a (18) suggest that high p16INK4a levelsmay be a marker of pRb inactivation or low pRb expression. Onthe other hand, the loss of p16INK4a would likely lead to in-creased cyclin D1/Cdk4 activity. One study indicated that cyclinD1 and p16INK4a alterations can cooperate to deregulate G1

control, resulting in multistep tumorigenesis (3). However, theinverse relationship between p16INK4a and cyclin D1 in thisseries of breast cancers supports the hypothesis that there is noselective advantage for aberration of more than one of thesegenes. Given that overexpression of cyclin D1, p16INK4a inac-tivation, and pRb inactivation are frequently mutually exclusive,

Table 1 INK4amethylation status in normal and breast cancercell lines

Cell linesUnmethylated

reactionMethylated

reactionStatus ofINK4a

gene locus

184 1 2 UnmethylatedPBLs 1 2 UnmethylatedBT-20 2 2 Homozygous deletionBT-483 1 2 UnmethylatedBT-549 1 2 UnmethylatedDU-4475 2 1 MethylatedHs-578T 2 2 Homozygous deletionHBL-100 1 2 UnmethylatedMCF-7M 2 2 Homozygous deletionMDA-MB-134 2 1 MethylatedMDA-MB-157 1 2 UnmethylatedMDA-MB-175 1 2 UnmethylatedMDA-MB-231 2 2 Homozygous deletionMDA-MB-361 1 2 UnmethylatedMDA-MB-436 1 2 UnmethylatedMDA-MB-453 1 2 UnmethylatedMDA-MB-468 1 2 UnmethylatedSK-BR-3 1 2 UnmethylatedT-47D 2 1 MethylatedZR-75-1 1 1 Partially methylated

Table 2 The relationship between p16INK4a mRNA expression andclinicopathological features in 217 breast cancer patients

Features

p16INK4a (n 5 217)

PLow

(n 5 109)High

(n 5 108)

Age (yrs),50 (n 5 78) 36 42$50 (n 5 139) 73 66 0.368

Menopausal statusNot known (n5 7) 3 4Premenopausal (n5 79) 37 42Postmenopausal (n5 131) 69 62 0.412

Tumor size (cm)Not known (n5 10) 7 3,2 (n 5 88) 48 40$2 (n 5 119) 54 65 0.192

Tumor gradeNot known (n5 12) 11 11 (n 5 80) 47 332 (n 5 97) 44 533 (n 5 28) 7 21 0.006a

LN status:Not known (n5 29) 16 13Negative (n5 33) 20 13Positive 1–3 nodes (n5 53) 32 21Positive$4 nodes (n5 102) 41 61 0.021a

ER status:Negative (n5 54) 12 42Positive (n5 163) 97 66 0.0001a

a Significant (P, 0.05).




this study, together with the previous observation of cyclin D1mRNA overexpression in 45% of breast tumors from a similarseries (57), indicated an overall rate of perturbation of the Rbpathway of at least 80% in primary breast cancer (i.e., ;45%cyclin D1 overexpression,;20% INK4a hypermethylation, and;16% p16INK4a overexpression.

The evidence for cyclin D1 induction by estrogen (58–60)and the demonstration of a tight correlation between cyclin D1and ER gene expression in breast cancers (49, 61) may account,at least in part, for the inverse relationship between p16INK4a

and ER mRNA expression reported in this study. ER functioncan up-regulate cyclin D1 expression, which, in turn, increasesphosphorylation of functional pRb, and this may then negativelymodulateINK4a transcription. However, this inverse relation-ship appeared to be even tighter in patients with high p16INK4a

expression, considering that only six tumors in the top 10% ofp16INK4a levels were ER positive, and all had very low levels ofER. This is unlikely to be fully explained by the associationbetween low cyclin D1 expression and ER negativity. One study(62) indicated that estrogen decreases the expression of pRb atthe level of protein and mRNA by a posttranscriptional mech-anism. However, the relationship between pRb and ER in breastcancer has been controversial (16, 63). The inverse relationshipbetween p16INK4a and ER status in this study may suggest thathigh p16INK4a levels could reduce the requirement for estrogenfor proliferation of breast cancer cells. Thus, further investiga-

tion will be required to define the precise mechanisms respon-sible for the relationship between ER and p16INK4a or pRb geneexpression. Furthermore, this tight inverse relationship betweenp16INK4a and ER may indicate that high expression of p16INK4a

may be associated with a lack of hormone responsiveness inbreast cancer.

De novomethylation of CpG islands within the gene pro-moter of tumor suppressor genes is an alternative pathway oftranscriptional inactivation providing a selective growth advan-tage to tumor cells (64). The methylation status ofINK4a in afew breast cancer cell lines has been determined previously (65)using Southern analysis to detect differential restriction enzymecleavage from non-methylation-sensitive and methylation-sensitive restriction enzymes. Methylation-specific PCR, how-ever, eliminates the false positive results inherent in Southernanalysis. In this study, the MCF-7, MDA-MB-231, Hs-578T,and BT-20 cell lines were negative in PCR reactions using bothmethylated- and unmethylated-specific primers, indicating thatINK4a was homozygously deleted, a result consistent with pre-vious findings (31). As expected, the three breast cancer celllines T-47D (65), MDA-MB-134 (66), and DU-4475 with meth-ylated INK4a had undetectable p16INK4a mRNA expression(Fig. 1). Similarly, all breast cancer cell lines with high p16INK4a

expression (MDA-MB-157, MDA-MB-436, BT-549, MDA-MB-468, and HBL-100; Fig. 1) had unmethylatedINK4a(Table 1).

Fig. 4 Life-table analysis with cumulative proportion of (A) disease-free survival and (B) overall survival for the total population of patients(n 5 217) in relation to low (E) or high (F) p16INK4a mRNA levels.

Fig. 5 Relationship between p16INK4a mRNA expression and disease-free survival in patients with (A) ER-positive and (B) axillary LN-positive breast cancers. Low p16INK4a mRNA levels,E; high p16INK4a

mRNA levels,F.




Unlike breast cancer cell lines, most of the breast tumorsamples, as shown in Fig. 3B, displayed products from bothmethylated and unmethylated PCR reactions. This is most likelydue to the inevitable admixture of DNA extracted from bothcancer cells and the surrounding normal stromal cells. Thus, ifa PCR fragment was evident in the methylated reaction, irre-spective of the unmethylated reaction,INK4a was regarded asmethylated in this tumor sample. The marked reduction in thenumber of tumor samples exhibitingINK4a methylation withascending tertiles of p16INK4a mRNA expression was expected.However, the discovery of fiveINK4amethylated tumors withinthe middle tertile of p16INK4a expression and oneINK4a meth-ylated tumor in the highest tertile may indicate that Northernanalysis is limited in detecting complete loss of tumor p16INK4a

expression, given the admixture of normal cells. Moreover,methylation of p16 exon 1 was assessed in this study, rather thanmethylation of its upstream promoter region, which might cor-relate better with expression. In the current series of breastcancers,;20% demonstratedINK4a methylation, which is con-sistent with a previously published smaller study (65). Thus,unlike the low frequency of gene deletion or mutation,INK4a

hypermethylation occurs reasonably frequently, and to date, it isby far the most commonly documented mechanism of inactiva-tion of the INK4a gene in primary breast cancer. However, thesurvival data failed to show even a trend or a relationshipbetween INK4a methylation status and outcome. AdditionalINK4a methylation assays in a larger series of patients will benecessary to fully investigate any prognostic significance ofINK4a methylation.

A recently published study (16) indicated that strong im-munohistochemical staining of p16INK4a was associated withincreased risk of death in 191 breast cancer patients.INK4a is atumor suppressor gene, and tumorigenesis is expected as aconsequence of inactivation of the gene, but not from geneoverexpression. However, high p16INK4a expression may beindicative of inactivation of pRb (13, 16, 56). The relationshipbetween pRb inactivation and clinical outcome in breast cancerhas been controversial. One study indicated an association be-tween abnormal pRb expression and an aggressive phenotype(67), whereas another reported an association betweenRBgenealterations and favorable prognostic factors (63) in breast can-cer. A number of groups failed to demonstrate a relationshipbetween pRb aberration and patient outcome (16, 63, 67). Onthe other hand, overexpression of p16INK4a may be independentof pRb mutation, as indicated in recent studies in ovarian (68)and prostate cancer.4 Nevertheless, although Northern blot anal-ysis was not sensitive enough to detect all p16INK4a inactivationand may have underestimated the number of samples with realoverexpression of p16INK4a, the current study supports earlierevidence (16) that p16INK4a overexpression is an indicator ofpoor prognosis in primary breast cancer.

In conclusion, loss of p16INK4a expression, overexpressionof cyclin D1, and loss of pRb function may have similar effectson G1 progression and may represent a common pathway intumorigenesis. Our findings suggest that both overexpression ofp16INK4a and de novo INK4amethylation occur frequently inprimary breast cancers. Furthermore, high p16INK4a mRNAexpression is associated with aggressive clinicopathological fea-tures in primary breast cancer. The demonstration of a signifi-cant negative correlation between the expression of theINK4aand ER genes raises issues regarding their functional interrela-tionships and, consequently, their roles as potential therapeuticresponse parameters.

ACKNOWLEDGMENTSWe thank Matthew Mitchell for invaluable assistance with the

statistical analysis. We are also indebted to Ann L. Cornish and RichardA. McClelland for help with preparation of the tumor RNA samples.

REFERENCES1. Serrano, M., Hannon, G. J., and Beach, D. A new regulatory motif incell-cycle control causing specific inhibition of cyclin D/CDK4. Nature(Lond.), 366: 704–707, 1993.

4 S. M. Henshall, D. I. Quinn, C. S. Lee, D. R. Head, D. Golovsky, P. C.Brenner, W. Delprado, J. F. Finlayson, P. D. Stricker, J. J. Grygiel, andR. L. Sutherland. Overexpression of the cell cycle inhibitor p16INK4a inhigh grade prostatic intraepithelial neoplasia predicts early relapse inprostate cancer patients, submitted for publication.

Fig. 6 A, life-table analysis of cumulative proportion of disease-freesurvival in patients with (�) low cyclin D1 and low p16INK4a (n 5 59),(L) low cyclin D1 and high p16INK4a (n 5 43), (E) high cyclin D1 andlow p16INK4a (n 5 49), and (Œ) high cyclin D1 and high p16INK4a (n 563). B, life-table analysis of cumulative proportion of disease-freesurvival in patients with (�) low cyclin D1 and low p16INK4a (n 5 52),(L) low cyclin D1 and high p16INK4a (n 5 18), (E) high cyclin D1 andlow p16INK4a (n 5 45), and (Œ) high cyclin D1 and high p16INK4a (n 546) in the ER-positive subgroup (n5 161).




2. Hunter, T., and Pines, J. Cyclins and cancer. II. Cyclin D and CDKinhibitors come of age. Cell,79: 573–582, 1994.

3. Lukas, J., Aagaard, L., Strauss, M., and Bartek, J. Oncogenic aber-rations of p16INK4/CDKN2 and cyclin D1 cooperate to deregulate G1

control. Cancer Res.,55: 4818–4823, 1995.

4. Koh, J., Enders, G. H., Dynlacht, B. D., and Harlow, E. Tumour-derived p16 alleles encoding proteins defective in cell-cycle inhibition.Nature (Lond.),375: 506–510, 1995.

5. Kamb, A., Gruis, N. A., Weaver, F. J., Liu, Q., Harshman, K.,Tavtigian, S. V., Stockert, E., Day, R. S., III, Johnson, B. E., andSkolnick, M. H. A cell cycle regulator potentially involved in genesis ofmany tumor types. Science (Washington DC),264: 436–440, 1994.

6. Nobori, T., Miura, K., Wu, D. J., Lois, A., Takabayashi, K., andCarson, D. A. Deletions of the cyclin-dependent kinase-4 inhibitor genein multiple human cancers. Nature (Lond.),368: 753–756, 1994.

7. Serrano, M., Lee, H., Chin, L., Cordon-Cardo, C., Beach, D., andDePinho, R. A. Role of the INK4a locus in tumor suppression and cellmortality. Cell,85: 27–37, 1996.

8. Barrett, M. T., Sanchez, C. A., Gaalipeau, P. C., Neshat, K., Emond,M., and Reid, B. J. Allelic loss of 9p21 and mutation of theCDKN2/p16gene develop as early lesions during neoplastic progression in Barrett’sesophagus. Oncogene,13: 1867–1873, 1996.

9. Cairns, P., Shaw, M. F., and Knowles, M. A. Initiation of bladdercancer may involve deletion of a tumor suppressor gene on chromosome9. Oncogene,8: 1083–1085, 1993.

10. van der Riet, P. Frequent loss of chromosome 9p21–22 early inhead and neck cancer progression. Cancer Res.,54: 1156–1158, 1994.

11. Papadimitrakopoulou, V., Izzo, J., Lippman, S. M., Lee, J. S., Fan,Y. H., Clayman, G., Ro, J. Y., Hittelman, W. N., Lotan, R., Hong,W. K., and Mao, L. Frequent inactivation of p16INk4a in oral premalig-nant lesions. Oncogene,14: 1799–1803, 1997.

12. Otterson, G. A., Kratzke, R. A., Coxon, A., Kim, Y. W., and Kaye,F. J. Absence of p16INK4a protein is restricted to the subset of lungcancer lines that retains wild type RB. Oncogene,9: 3375–3378, 1994.

13. Parry, D., Bates, S., Mann, D. J., and Peters, G. Lack of cyclinD-Cdk complexes in Rb-negative cells correlates with high levels ofp16INK4/MTS1 tumour suppressor gene product. EMBO J.,14: 503–511, 1995.

14. Kinoshita, I., Dosaka-Akita, H., Mishina, T., Akie, K., Nishi, M.,Hiroumi, H., Hommura, F., and Kawakami, Y. Altered p16INK4 andretinoblastoma protein status in non-small cell lung cancer: potentialsynergistic effect with altered p53 protein on proliferative activity.Cancer Res.,56: 5557–5562, 1996.

15. Khleif, S. N., DeGregori, J., Yee, C. L., Otterson, G. A., Kaye, F. J.,Nevins, J. R., and Howley, P. M. Inhibition of cyclin D-CDK4/CDK6activity is associated with an E2F-mediated induction of cyclin kinaseinhibitor activity. Proc. Natl. Acad. Sci. USA,93: 4350–4354, 1996.

16. Dublin, E. A., Patel, N. K., Gillett, C. E., Smith, P., Peters, G., andBarnes, D. M. Retinoblastoma and p16 proteins in mammary carcinoma:their relationship to cyclin D1 and histopathological parameters. Int. J.Cancer,79: 71–75, 1998.

17. Li, Y., Nichols, M. A., Shay, J. W., and Xiong, Y. Transcriptionalrepression of the D-type cyclin-dependent kinase inhibitor p16 by theretinoblastoma susceptibility gene product pRb. Cancer Res.,54: 6078–6082, 1994.

18. Fang, X. J., Jin, X. M., Xu, H. J., Liu, L., Peng, H. Q., Hogg, D.,Roth, J. A., Yu, Y. H., Xu, F. J., Bast, R. C., and Mills, G. B. Expressionof p16 induces transcriptional downregulation of theRB gene. Onco-gene,16: 1–8, 1998.

19. Lukas, J., Parry, D., Aagaard, L., Mann, D. J., Bartkova, J., Strauss,M., Peters, G., and Bartek, J. Retinoblastoma-protein-dependent cell-cycle inhibition by the tumour suppressor p16. Nature (Lond.),375:503–506, 1995.

20. Mao, L., Merlo, A., Bedi, G., Sharpiro, G. I., Edwards, C. D.,Rollins, B. J., and Sidransky, D. A novel p16INK4a transcript. CancerRes.,55: 2995–2997, 1995.

21. Stone, S., Jiang, P., Dayananth, P., Tavtigian, S. V., Katcher, H.,Parry, D., Peters, G., and Kamb, A. Complex structure and regulation ofthe p16 (MTS1) locus. Cancer Res.,55: 2988–2994, 1995.

22. Quelle, D. E., Zindy, F., Ashmun, R. A., and Sherr, C. J. Alternativereading frames of the INK4a tumor suppressor gene encode two unre-lated proteins capable of inducing cell cycle arrest. Cell,83: 993–1000,1995.

23. Larsen, C-J.p16INK4a: a gene with a dual capacity to encodeunrelated proteins that inhibit cell cycle progression. Oncogene,12:2041–2044, 1996.

24. Pomerantz, J., Schreiber-Angus, N., Liegeois, N. J., Silverman, A.,Alland, L., Chin, L., Potes, K., Chen, K., Orlow, I., Lee, H-W., Cordon-Cardo, C., and DePinho, A. The Ink4a tumor suppressor gene product,p19Arf, interacts with MDM2 and neutralizes MDM2’s inhibition ofp53. Cell,92: 713–723, 1998.

25. Zhang, Y., Xiong, Y., and Yarbrough, W. G. ARF promotes MDM2degradation and stabilizes p53: ARF-INK4a locus deletion impairs boththe Rb and p53 tumor suppression pathways. Cell,92: 725–734, 1998.

26. Ruas, M., and Peters, G. p16INK4a/CDKN2A tumor suppressor andits relatives. Biochim. Biophys. Acta,1378: F115–F177, 1998.

27. Reed, A. L., Califano, J., Cairns, P., Westra, W. H., Jones, R. M.,Koch, W., Ahrendt, S., Eby, Y., Sewell, D., Nawroz, H., Bartek, J., andSidransky, D. High frequency of p16 (CDKN2/MTS-1/INK4A) inacti-vation in head and neck squamous cell carcinoma. Cancer Res.,56:3630–3633, 1996.

28. Caldas, C., Hahn, S. A., da Costa, L. T., Redston, M. S., Schutte,M., Seymour, A. B., Weinstein, C. L., Hruban, R. H., Yeo, C. J., andKern, S. E. Frequent somatic mutations and homozygous deletions ofthe p16 (MTS1) gene in pancreatic adenocarcinoma. Nat. Genet.,8:27–32, 1994.

29. Schutte, M., Hruban, R. H., Geradts, J., Maynard, R., Hilgers, W.,Rabindran, S. K., Moskaluk, C. A., Hahn, S. A., Schwartewaldhoff, I.,Schmiegel, W., Baylin, S. B., Kern, S. E., and Herman, J. G. Abrogationof the Rb/p16 tumor-suppressive pathway in virtually all pancreaticcarcinomas. Cancer Res.,57: 3126–3130, 1997.

30. Berns, E. M., Klijn, J. G., Smid, M., van Staveren, I. L., Gruis,N. A., and Foekens, J. A. Infrequent CDKN2 (MTS1/p16) gene alter-ations in human primary breast cancer. Br. J. Cancer,72: 964–967,1995.

31. Musgrove, E. A., Lilischkis, R., Cornish, A. L., Lee, C. S. L., Setlur,V., Seshadri, R., and Sutherland, R. L. Expression of the cyclin-depend-ent kinase inhibitors p16INK4a, p15INK4B, and p21WAF1/CIP1 in humanbreast cancer. Int. J. Cancer,63: 584–591, 1995.

32. Merlo, A., Herman, J. G., Mao, L., Lee, D. J., Gabrielson, E.,Burger, P. C., Baylin, S. B., and Sidransky, D. 59CpG island methyl-ation is associated with transcriptional silencing of the tumour suppres-sor p16/CDKN2/MTS1 in human cancers. Nat. Med.,1: 686–692,1995.

33. Herman, J. G., Jen, J., Merlo, A., and Baylin, S. B. Hypermeth-ylation-associated inactivation indicates a tumor suppressor role forp15INK4B. Cancer Res.,56: 722–727, 1996.34. Fizzotti, M., Cimino, G., Pisegna, S., Alimena, G., Quartarone, C.,Mandelli, F., Pelicci, P. G., and Coco, F. L. Detection of homozygousdeletions of the cyclin-dependent kinase 4 inhibitor (p16) gene in acutelymphoblastic leukaemia and association with adverse prognostic fea-tures. Blood,85: 2685–2690, 1995.35. Nishikawa, R., Furnari, F. B., Lin, H., Arap, W., Berger, M. S.,Cavenee, W. S., and Su Huang, H-J. Loss of p16INK4 expression isfrequent in high-grade gliomas. Cancer Res.,55: 1941–1945, 1995.36. Reed, J. A., Loganzo, F. J., Shea, C. R., Walker, G. J., Flores, J. F.,Glendening, J. M., Bogdany, J. K., Shiel, M. J., Haluska, F. G., Foun-tain, J. W., and Albino, A. P. Loss of expression of the p16/cyclin-dependent kinase inhibitor 2 tumor suppressor gene in melanocyticlesions correlates with invasive stage of tumor progression. Cancer Res.,55: 2713–2718, 1995.37. Kratzke, R. A., Greatens, T. M., Rubins, J. B., Maddaus, M. A.,Niewoehner, D. E., Niehans, G. A., and Geradts, J. Rb and p16INK4a




expression in resected non-small cell lung tumors. Cancer Res.,56:3415–3420, 1996.38. Zhou, M., Gu, L., Yeager, A. M., and Findley, H. W. Incidence andclinical significance ofCDKN2/MTS1/p16ink4a andMTS2/p15ink4b genedeletions in childhood acute lymphoblastic leukaemia. Pediatr. Haema-tol. Oncol.,14: 141–150, 1997.39. Kees, U. R., Burton, P. R., Lu, C., and Baker, D. L. Homozygousdeletion of thep16/MTS1gene in pediatric acute lymphoblastic leukae-mia is associated with unfavorable clinical outcome. Blood,89: 4161–4166, 1997.40. Taga, S., Osaki, T., Ohgami, A., Imoto, H., Yoshimatsu, T.,Yoshino, I., Yano, K., Nakanishi, R., Ichiyoshi, Y., and Yasumoto, K.Prognostic value of the immunohistochemical detection of p16INK4

expression in non-small cell lung carcinoma. Cancer (Phila.),80: 389–395, 1997.41. Hu, Y-X., Watanabe, H., Ohtsubo, K., Yamaguchi, Y., Ha, A.,Okai, T., and Sawabu, N. Frequent loss of p16 expression and itscorrelation with clinicopathological parameters in pancreatic carcinoma.Clin. Cancer Res.,3: 1473–1477, 1997.42. Straume, O., and Akslen, L. A. Alterations and prognostic signifi-cance of p16 and p53 protein expression in subgroups of cutaneousmelanoma. Int. J. Cancer,74: 535–539, 1997.43. Garcia-Sanz, R., Gonzalez, M., Vargas, M., Chillon, M. C., Balan-zategui, A., Barbon, M., Flores, M. T., and San Miguel, J. F. Deletionsand rearrangements of cyclin-dependent kinase 4 inhibitor genep16areassociated with poor prognosis in B cell non-Hodgkin’s lymphoma.Leukemia (Baltimore),11: 1915–1920, 1997.44. Takeuchi, H., Ozawa, S., Ando, N., Shih, C. H., Koyanagi, K.,Ueda, M., and Kitajima, M. Alteredp16/MTS1/CDKN2and cyclinD1/PRAD-1gene expression is associated with the prognosis of squa-mous cell carcinoma of the esophagus. Clin. Cancer Res.,3: 2229–2236, 1997.45. Naka, T., Kobayashi, M., Ashida, K., Toyota, N., Kaneko, T., andKaibara, N. Aberrant p16INK4 expression related to clinical stage andprognosis in patients with pancreatic cancer. Int. J. Oncol.,12: 1111–1116, 1998.46. Eiriksdottir, G., Sigurdsson, A., Jonasson, J. G., Agnarsson, B. A.,Sigurdsson, H., Gudmundsson, J., Bergthorsson, J. T., Barkardottir,R. B., Egilsson, V., and Ingvarsson, S. Loss of heterozygosity onchromosome 9 in human breast cancer: association with clinical vari-ables and genetic changes at other chromosome regions. Int. J. Cancer,64: 378–382, 1995.47. An, H. X., Niederacher, D., Picard, F., van Roeyen, C., Bender,H. G., and Beckmann, M. W. Frequent allele loss on 9p21–22 defines asmallest common region in the vicinity of theCDKN2gene in sporadicbreast cancer. Genes Chromosomes Cancer,17: 14–20, 1996.48. Kenny, F. S., Hui, R., Musgrove, E. A., Gee, J. M., Blamey, R. W.,Nicholson, R. I., Sutherland, R. L., and Robertson, J. F. R. Overexpres-sion of cyclin D1 mRNA predicts for poor prognosis in estrogenreceptor-positive breast cancer. Clin. Cancer Res.,5: 2069–2075, 1999.49. Hui, R., Cornish, A. L., McClelland, R. A., Robertson, J. F. R.,Blamey, R. W., Musgrove, E. A., Nicholson, R. I., and Sutherland, R. L.Cyclin D1 and estrogen receptor mRNA expression are positively cor-related in primary breast cancer. Clin. Cancer Res.,2: 923–928, 1996.50. Hui, R., Ball, J. R., MacMillan, R. D., Prall, O. W. J., Campbell,D. H., Cornish, A. L., McClelland, R. A., Daly, R. J., Forbes, J. F.,Blamey, R. W., Musgrove, E. A., Robertson, J. F. R., Nicholson, R. I.,and Sutherland, R. L.EMS1gene expression in primary breast cancer:relationship to cyclin D1 and oestrogen receptor expression and patientsurvival. Oncogene,16: 1053–1059, 1998.51. Galea, M. H., Blamey, R. W., Elston, C. E., and Ellis, I. O. TheNottingham Prognostic Index in primary breast cancer. Breast CancerRes. Treat.,22: 207–219, 1992.52. Herman, J. G., Graff, J. R., Myohanen, S., Nelkin, B. D., andBaylin, S. B. Methylation-specific PCR: a novel PCR assay for meth-ylation status of CpG islands. Proc. Natl. Acad. Sci. USA,93: 9821–9826, 1996.

53. Lee, E. T. Statistical Methods for Survival Data Analysis. NewYork: John Wiley and Sons, 1992.

54. Guan, K. L., Jenkins, C. W., Li, Y., Nichols, M. A., Wu, X.,O’Keefe, C. L., Matera, A. G., and Xiong, Y. Growth suppression byp18, a p16INK4/MTS1- and p14INK4B/MTS2-related CDK6 inhibitor,correlates with wild-type pRb function. Genes Dev.,8: 2939–2952,1994.

55. Medema, R. H., Herrera, R. E., Lam, F., and Weinberg, R. A.Growth suppression by p16ink4 requires functional retinoblastoma pro-tein. Proc. Natl. Acad. Sci. USA,92: 6289–6293, 1995.

56. Nielsen, N. H., Emdin, S. O., Cajander, J., and Landberg, G.Deregulation of cyclin E and D1 in breast cancer is associated withinactivation of the retinoblastoma protein. Oncogene,14: 295–304,1997.

57. Buckley, M. F., Sweeney, K. J., Hamilton, J. A., Sini, R. L.,Manning, D. L., Nicholson, R. I., deFazio, A., Watts, C. K., Musgrove,E. A., and Sutherland, R. L. Expression and amplification of cyclingenes in human breast cancer. Oncogene,8: 2127–2133, 1993.

58. Altucci, L., Addeo, R., Cicatiello, L., Dauvois, S., Parker, M. G.,Truss, M., Beato, M., Sica, V., Bresciani, F., and Weisz, A.17b-Estradiol induces cyclin D1 gene transcription, p36D1–p34cdk4

complex activation and p105Rb phosphorylation during mitogenic stim-ulation of G1-arrested human breast cancer cells. Oncogene,12: 2315–2324, 1996.

59. Foster, J. S., and Wimalasena, J. Estrogen regulates activity ofcyclin-dependent kinases and retinoblastoma protein phosphorylation inbreast cancer cells. Mol. Endocrinol.,10: 488–498, 1996.

60. Prall, O. W. J., Sarcevic, B., Musgrove, E. A., Watts, C. K. W., andSutherland, R. L. Estrogen-induced activation of cdk4 and cdk2 duringG1-S phase progression is accompanied by increased cyclin D1 expres-sion and decreased cyclin-dependent kinase inhibitor association withcyclin E-cdk2. J. Biol. Chem.,272: 10882–10894, 1997.

61. Michalides, R., Hageman, P., Vantinteren, H., Houben, L.,Wientjens, E., Klompmaker, R., and Peterse, J. A clinicopathologicalstudy on overexpression of cyclin D1 and of p53 in a series of 248patients with operable breast cancer. Br. J. Cancer,73: 728–734, 1996.

62. Gottardis, M. M., Saceda, M., Garcia-Morales, P., Fung, Y-K.,Solomon, H., Sholler, P. F., Lippman, M. E., and Martin, M. B.Regulation of retinoblastoma gene expression in hormone-dependentbreast cancer. Endocrinology,136: 5659–5665, 1995.

63. Berns, E. M., de Klein, A., van Putten, W. L., van Staveren, I. L.,Bootsma, A., Klijn, J. G., and Foekens, J. A. Association betweenRB-1gene alterations and factors of favourable prognosis in human breastcancer, without effect on survival. Int. J. Cancer,64: 140–145, 1995.

64. Jones, P. A. DNA methylation errors and cancer. Cancer Res.,56:2463–2467, 1996.

65. Herman, J. G., Merlo, A., Mao, L., Lapidus, R. G., Issa, J. P.,Davidson, N. E., Sidransky, D., and Baylin, S. B. Inactivation of theCDKN2/p16/MTS1gene is frequently associated with aberrant DNAmethylation in all common human cancers. Cancer Res.,55: 4525–4530, 1995.

66. Tam, S. W., Shay, J. W., and Pagano, M. Differential expressionand cell cycle regulation of the cyclin-dependent kinase 4 inhibitorp16Ink4. Cancer Res.,54: 5816–5820, 1994.

67. Pietilainen, T., Lipponen, P., Aaltomaa, S., Eskelinen, M., Kosma,V. M., and Syrjanen, K. Expression of retinoblastoma gene protein (Rb)in breast cancer as related to established prognostic factors and survival.Eur. J. Cancer,3: 329–333, 1995.

68. Dong, Y., Walsh, M. D., McGuckin, M. A., Gabrielli, B. G.,Cummings, M. C., Wright, R. G., Hurst, T., Khoo, S. K., and Parsons,P. G. Increased expression of cyclin-dependent kinase inhibitor 2(CDKN2A) gene product p16INK4A in ovarian cancer is associated withprogression and unfavourable prognosis. Int. J. Cancer,74: 57–63,1997.




2000;6:2777-2787. Clin Cancer Res Rina Hui, R. Douglas Macmillan, Frances S. Kenny, et al. Marker of Poor Prognosis

Messenger RNA Is aINK4aCancer: Overexpression of p16 Gene Expression and Methylation in Primary BreastINK4a

Updated version

http://clincancerres.aacrjournals.org/content/6/7/2777

Access the most recent version of this article at:

Cited articles

http://clincancerres.aacrjournals.org/content/6/7/2777.full#ref-list-1

This article cites 65 articles, 27 of which you can access for free at:

Citing articles

http://clincancerres.aacrjournals.org/content/6/7/2777.full#related-urls

This article has been cited by 9 HighWire-hosted articles. Access the articles at:

E-mail alerts related to this article or journal.Sign up to receive free email-alerts

Subscriptions

Reprints and

[email protected] at

To order reprints of this article or to subscribe to the journal, contact the AACR Publications

Permissions

Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://clincancerres.aacrjournals.org/content/6/7/2777To request permission to re-use all or part of this article, use this link



http://clincancerres.aacrjournals.org/content/6/7/2777.full#ref-list-1

http://clincancerres.aacrjournals.org/content/6/7/2777.full#related-urls

http://clincancerres.aacrjournals.org/cgi/alerts

mailto:[email protected]



INK4a Gene Expression and Methylation in Primary Breast ... · cancers using methylation-specific...

Documents

Transcript of INK4a Gene Expression and Methylation in Primary Breast ... · cancers using methylation-specific...