The INK4a/ARF locus encodes two cell cycle-regulatory proteins, p16INK4a and p14ARF,which share an exon using different reading frames. p14ARF antagonizes MDM2-dependent p53 degradation. However, no point mutations in p14ARF not altering p16INK4a have been described in primary tumors. We report that p14ARF is epigenetically inactivated in several colorectal cell lines, and its expression is restored by treatment with demethylating agents. In primary colorectal carcinomas, p14ARF promoter hypermethylation was found in 31 of 110 (28%) of the tumors and observed in 13 of 41 (32%)colorectal adenomas but was not present in any normal tissues. p14ARF methylation appears in the context of an adjacent unmethylated p16INK4a promoter in 16 of 31 (52%) of the carcinomas methylated at p14ARF. Although p14ARFhypermethylation was slightly overrepresented in tumors with wild-type p53 compared to tumors harboring p53mutations [19 of 55 (34%) versus 12 of 55 (22%)],this difference did not reach statistical significance. p14ARF aberrant methylation was not related to the presence of K-ras mutations. Our results demonstrate that p14ARF promoter hypermethylation is frequent in colorectal cancer and occurs independently of the p16INK4a methylation status and only marginally in relation to the p53 mutational status.
Disruption of the p53 and Rb3tumor suppressor pathways is a fundamental trend of most human cancer cells (1). In tumorigenesis, loss of Rbfunction can occur by direct inactivation of the Rb gene itself through mutation, sequestration of the Rb protein by viral oncoproteins, or promoter hypermethylation or by deregulation of the genes controlling Rb phosphorylation status (2). These last alterations include cyclin D1 gene amplification, CDK4activating mutations, and also gene amplification and inactivation of the inhibitors of CDK4, the INK4 family (composed of p15INK4b, p16INK4a, p18INK4c, and p19INK4d; Refs.1 and 3). The p16INK4agene was implicated as a tumor suppressor gene by its frequent mutation, deletion, or promoter hypermethylation in a variety of human tumors (1, 4, 5). In addition, p16INK4agerm-line mutations have been associated with familial melanoma(6). Recently, it was demonstrated that a portion of the p16INK4a gene has the capacity to encode a second product, murine p19ARF (7). p19ARF, or human p14ARF, has a unique first exon (denominated exon 1β), located approximately 20 kb centromeric to the first exon of p16INK4a (denoted as exon 1α). Under the control of its own promoter, exon 1β splices into exon 2 of INK4a in an alternative reading frame, producing a different protein than p16INK4a (1, 3, 7, 8).
p19ARF overexpression induces G1- and G2-phase arrest through a p16INK4a-independent mechanism(9). Furthermore, the p19ARF-mediated cell cycle arrest seems to be abolished in mouse embryo fibroblasts lacking functional p53 (8). Recent work suggests that p19ARF and p14ARFinteract in vivo with the MDM2 protein, neutralizing MDM2-mediated degradation of p53(10, 11, 12, 13). Thus, theoretically, inactivation of p14ARF would then be predicted to decrease the frequency for concomitant p53 mutations. p19ARF binds to MDM2 through its NH2 terminus end, which encodes exon 1β(13). p19ARF-specific null mice carrying a disrupted exon 1β are cancer prone at an early age(8), but no human point mutations in exon 1β have been reported. Interestingly, exon 1β-specific deletions have been described in melanoma cell lines (14), and p14ARF genomic alterations are found in a majority of T-cell acute lymphocytic leukemias (15). However, to date, most of the functional studies of p14ARF have involved murine systems, and little is known about the putative p14ARF function as a tumor suppressor gene in humans. Recently, the human p14ARF promoter has been cloned and contains a CpG island that is aberrantly methylated in colorectal cancer cell lines (16). Hypermethylation of normally unmethylated CpG islands in the promoter regions of tumor suppressor and DNA repair genes, including p16INK4a, E-cadherin, hMLH1, GSTP1, and MGMT (4, 17, 18, 19, 20), correlates with loss of transcription. Thus, promoter hypermethylation could be a mechanism for p14ARF inactivation in human tumors. In addition,methylation of p16INK4a and p15INK4b can occur with more tumor-specific patterns than found for homozygous deletions involving the entire 9p region encompassing these genes (21).
In the present study, we studied CpG island promoter methylation of the human p14ARF gene in cancer cell lines and more than 100 primary colorectal carcinomas. Colorectal tumors were chosen because homozygous deletions of the INK4a/ARF locus are not present in this tumor type (22), and colorectal carcinoma has well-characterized mutational inactivation of p53, thus allowing us to examine the relationship to p53 mutations. Our results demonstrate that p14ARF can be silenced by promoter hypermethylation in colorectal cancer cell lines. In the primary colorectal tumors, p14ARF is aberrantly methylated in approximately a quarter of the neoplasms studied, and this methylation represents an event independent of the p16INK4a methylation status. Furthermore, p14ARF hypermethylation, although more frequent in tumors with wild-type p53, is also observed in those cancers with p53 mutations.
MATERIALS AND METHODS
DNA methylation patterns in the CpG islands of the p14ARF gene were determined by MSP(23). MSP distinguishes unmethylated from methylated alleles in a given gene based on sequence changes produced after bisulfite treatment of DNA, which converts unmethylated (but not methylated) cytosines to uracil, and subsequent PCR using primers designed for either methylated or unmethylated DNA (23). The primer sequences designed for p14ARF spanned six CpGs within the 5′ region of the gene. Primer sequences of p14ARF for the unmethylated reaction were 5′-TTTTTGGTGTTAAAGGGTGGTGTAGT-3′ (sense) and 5′-CACAAAAACCCTCACTCACAACAA-3′ (antisense), which amplify a 132-bp product, and primer sequences of p14ARFfor the methylated reaction were 5′-GTGTTAAAGGGCGGCGTAGC-3′ (sense) and 5′-AAAACCCTCACTCGCGACGA-3′ (antisense), which amplify a 122-bp product. The 5′ position of the sense unmethylated and methylated primers corresponds to bp 195 and 201 of GenBank sequence number L41934. Both antisense primers originate from bp 303 of this sequence. The annealing temperature for both the unmethylated and methylated reactions was 60°C. Placental DNA treated in vitro with SssI methyltransferase was used as a positive control for methylated alleles. DNA from normal lymphocytes was used as a negative control for methylated genes. Ten μl of each PCR reaction were loaded directly onto nondenaturing 6% polyacrylamide gels, stained with ethidium bromide, and visualized under UV illumination. DNA methylation patterns in the 5′-CpG island of p16INK4a were determined by MSP as described previously (23).
RT-PCR was performed as described previously (21), using 3μg of total cellular RNA to generate cDNA. This cDNA (100 ng) was amplified by PCR with primers for exon 1β(5′-GGTTTTCGTGGTTCACATCCCGCG-3′) and exon 2(5′-CAGGAAGCCCTCCCGGGCAGC-3′) of p14ARF, which amplify a 254-bp product spanning sequence 204–437 from GenBank accession number S78535. RT-PCR for GAPDH served as a positive control. Ten μl of each PCR reaction were loaded directly onto nondenaturing 6% polyacrylamide gels, stained with ethidium bromide, and visualized under UV illumination.
Detection of K-ras and p53Mutations.
Mutations at codons 12 and 13 of the K-ras gene were detected and characterized by a RFLP/PCR approach (24). p53 mutations in exons 4–9 were analyzed by single-strand conformational polymorphism analysis. Briefly, a first PCR was performed using primers 12979U (GCTGCCGTGTTCCAGTTGCT) and 14875D(AGGCATCACTGCCCCCTGAT). The resulting 1897-bp fragment was then used as a template to separately amplify a fragment of 410 bp including exons 5 and 6 [with primers 13054U (TACTCCCCTGCCCTCAACAAG) and 13463D(CTCCTCCCAGAGACCCCAGT)] and a fragment of 622 bp including exons 7 and 8 [with primers 13966U (CTGGCCTCATCTTGGGCCTG) and 14587D(CTCGCTTAGTGCTCCCTGGG)]. These two fragments were then digested with restriction enzyme HpaII, and the resulting fragments were run on a 6% polyacrylamide gel without glycerol (0.2 h at 30 W and 5–6 h at 6 W) and with 10% glycerol (0.2 h at 30 W and 13–14 h at 6 W) to detect mobility shifts. Mutations were confirmed by direct cycle sequencing of the PCR products using the AmpliCycle Sequencing Kit(Perkin-Elmer, Branchburg, NJ). Exons 4 and 9 were only analyzed on those samples negative for mutations in exons 5–8. Exon 4 was amplified directly from DNA using primers 12019U (GTCCCCCTTGCCGTCCCAAG)and 12349D (TACGGCCAGGCATTGAAGTC). The resulting 331-bp fragment was run without previous digestion on a 6%polyacrylamide/10% glycerol gel for 0.2 h at 30 W and 19 h at 6 W. To analyze exon 9, a fragment of 788 bp including exons 7–9 was amplified with primers 13966U (CTGGCCTCATCTTGGGCCTG) and 14753D(CTGAAGGGTGAAATATTCTCC) and digested with HhaI to produce two fragments of 548 and 240 bp, the latter of which contained exon 9.
Promoter Hypermethylation and Expression of p14ARF in Cancer Cell Lines.
Structurally, p14ARF is a candidate for hypermethylation-associated inactivation because a 5′-CpG island is located around the transcription start site (16). The region chosen for the MSP analysis spans the area of greatest CpG density studied recently for methylation changes in several cell lines(16). Normal lymphocytes, colon, breast, endometrium, and lung were found completely unmethylated at the p14ARF promoter (Fig. 1,A). Normal kidney and liver were also unmethylated at p14ARF. We also examined 16 cancer cell lines derived from these tissues. p14ARF was fully methylated in four (25%) colorectal carcinoma cell lines (SW48, RKO,DLD-1, and LoVo; Fig. 1,B). Among these, expression of the p14ARF transcript by RT-PCR was assessed in LoVo and DLD-1, both of which lack p14ARF expression (Fig. 1,C). Treatment of these cell lines with the demethylating agent 5-aza-2′-deoxycytidine restored the expression of the transcript in both cases (Fig. 1,C). These results agree with those recently reported by Robertson et al.(16), although in that case, the colorectal cancer cell line HCT-15, which is isogenic to the DLD-1 cell line(25), was used. The colorectal cancer cell line RKO methylated at p14ARF showed a low level of expression but demonstrated an increase in RNA after the treatment with 5-aza-2′-deoxycytidine. A similar phenomenon has been described previously in the case of the colorectal cancer cell line SW48(16), which was also methylated at p14ARF in our study. The colorectal cancer cell lines SW1417, SK-CO1, LS180T, and HT-29 were partially methylated, and p14ARF expression was demonstrated in the last one(Fig. 1,C). Normal lymphocytes and the unmethylated colorectal cell line SW480 and the unmethylated leukemia cell line HL-60 demonstrated expression of the p14ARFtranscript (Fig. 1 C).
The proposed role of p14ARF in modulating p53-MDM2 function would predict the diminishing need of p53 mutations in those tumors or cell lines with p14ARF promoter hypermethylation, assuming that only one “hit” in the same pathway would be necessary in the transforming process. In this series of cancer cell lines, cell lines fully methylated at p14ARF (LoVo, DLD-1, SW48, and RKO) were more frequently found to have an intact p53[three of four cell lines (75%), with the exception DLD-1]. Among cell lines with an unmethylated promoter, including colorectal cells,non-small cell lung cancer, and leukemia (SW480, HCT116, SW837,COLO205, H157, H1618, U1752, and HL-60), only one (HCT116) had a wild-type p53. Thus, most cancer cell lines methylated at p14ARF are wild-type p53, and those unmethylated have a mutant p53, but this difference did not reach statistical significance (Fisher’s exact test, P = 0.07).
Promoter Hypermethylation of p14ARF in Primary Colorectal Carcinomas.
DNA obtained from 110 primary colorectal carcinomas was subjected to p14ARF promoter methylation study using MSP. Among all of the colorectal tumors studied, p14ARFpromoter hypermethylation was present in 31 of 110 (28%) samples (Fig. 1,D). In 32 patients from whom normal adjacent mucosa DNA was available, no methylation of the p14ARF promoter was observed in any case (an example is shown in Fig. 1,A). Thus,the methylation observed in the colorectal cancers and cell lines is a tumor-specific change. Abnormal methylation of the p14ARF promoter region in the colorectal carcinomas was not associated with significant differences of gender, age of onset, clinical status, Duke’s stage, DNA ploidy, or the presence of residual disease. When the samples were decoded for their p53 status, p14ARF was hypermethylated in 19 of 55 (34%) colorectal tumors with wild-type p53 and in 12 of 55 (22%) tumors with a mutant p53 (Fig. 2,A). Thus, although the tumors with functional p53more often harbor p14ARF epigenetic inactivation than tumors with p53 mutations, this trend does not reach statistical significance (Fisher’s exact test, P = 0.20)and was certainly not limited to p53 wild-type tumors (Fig. 2 A).
We also wondered whether p14ARF promoter hypermethylation would be related to promoter hypermethylation of the adjacent gene, p16INK4a. Aberrant methylation of p16INK4a was assessed in the same samples analyzed for p14ARF methylation and demonstrated in 41 of 110(37%) of the primary colorectal carcinomas studied, a frequency similar to that described previously (Ref. 26; Fig. 1,D). p14ARF was hypermethylated in the presence of an unmethylated p16INK4a promoter in 16 of 110 (14%) of the cases, whereas p14ARF promoter hypermethylation was coincident with p16INK4apromoter hypermethylation in 15 of 110 (14%) of the cases (Fig. 2,B). In 26 of 110 (24%) of the cases, p16INK4a was methylated without p14ARF methylation (Fig. 2 B). Thus,methylation at the p14ARF and p16INK4a promoters does not seem to be directly related (Fisher’s exact test, P = 0.28). It has been hypothesized that mutations in exons 1–4 of p53 may be concomitant with alterations in both p16INK4a and p14ARF(27). Although we did not address this aspect specifically, in the three cases in which exon 4 mutations were found, p16INK4a was unmethylated in all three, and p14ARF was methylated in only one case.
Because it has been recently reported that p19ARF(the mouse homologue of p14ARF) is essential for the activation of p53 in response to oncogenic Ras(28), we also examined the p14ARFpromoter hypermethylation in relation to the K-ras status of the tumor. A total of 43 of 110 (39%) primary colorectal carcinomas had a K-ras point mutation. The frequency of p14ARF methylation was slightly higher in tumors with K-ras mutations than in those without K-ras mutations [15 of 43(35%) versus 16 of 67 (24%)], but this difference was not statistically significant (Fisher’s exact test, P = 0.28). Thus, no evident linkage between these epigenetic and genetic alterations was observed.
Promoter Hypermethylation and Expression of p14ARF in Colorectal Adenomas.
DNA obtained from 41 primary colorectal adenomas was subjected to p14ARF promoter methylation study using MSP. Among all of the adenomas studied, p14ARF promoter hypermethylation was present in 13 of 41 (32%) samples (Fig. 1,E). We also examined the expression of p14ARF using RT-PCR in these colorectal adenomas. Among 20 early lesions with available cDNA, 12 colorectal adenomas unmethylated at p14ARF expressed high levels of p14ARF mRNA, whereas 8 adenomas with p14ARF methylation expressed no detectable p14ARF mRNA (n = 7) or very little p14ARF mRNA (n = 1), demonstrating an exact correlation of transcriptional loss with p14ARF hypermethylation (Fig. 3). The similar rate of p14ARF methylation in adenomas and carcinomas suggests that like p16INK4amethylation, inactivation of p14ARF appears to be an early event in colorectal tumorigenesis.
Our study represents the first finding of a specific lesion in p14ARF, promoter hypermethylation, in primary human tumors. In many cases, this does not affect the neighboring gene, p16INK4a. Furthermore, our data establish that, at least in colorectal tumors, the inactivation on p14ARF is not restricted to those neoplasms with intact p53. It has been postulated that due to the fact that p53 is directly targeted in approximately 50% of human malignancies (29), other mechanisms are required to shut down the p53 pathway in tumors with wild-type p53. For example, lesions in the MDM2 gene such as amplification can lead to MDM2 overexpression and abrogation of p53 function. MDM2 represses p53 transcriptional activity and mediates the degradation of p53 (30). However, MDM2 gene amplification is restricted to a few tumor types (30). p53 degradation mediated by the viral protein E6 is also limited to a minority of neoplasms (31). Another event contributing to abrogate the p53 pathway in the cancer cell may be p14ARF loss of function. Somatic mutations affecting exons 2 and 3 p14ARF, shared in an alternative reading frame with p16INK4a, have been demonstrated in tumors (1, 5), but no mutations in the specific exon 1β of p14ARF have been reported. p14ARF can also be lost by homozygous deletion in several tumor types (22), but this loss also targets p16INK4a in the vast majority of cases. However, p14ARF has been proposed as the crucial target in T-cell acute lymphocytic leukemias and glioblastomas that exhibit 9p21 lesions (15). Recently, aberrant methylation of the p14ARF promoter has been demonstrated in cancer cell lines (16). Our data corroborate the findings of loss of p14ARF expression associated with CpG island methylation and its restoration by the use of demethylating agents. We now demonstrate that such hypermethylation of p14ARFis not limited to cell lines and is not a rare event. Colorectal carcinomas were selected because these cancers do not show homozygous deletion of the 9p21 locus (22) and thus allow the separate study of p14ARF and p16INK4a inactivation. In addition, the considerable rate of p53 mutations in colorectal tumors helps us to address the existence or nonexistence of a relation between p14ARF and p53 status.
Our results demonstrate that p14ARF promoter hypermethylation occurs in approximately one-fourth of primary colorectal carcinomas. p14ARF promoter hypermethylation also seems to be an early event in colorectal tumorigenesis because, like p16INK4a methylation, it can be found in colon adenomas. When the p14ARFpromoter methylation analysis is compared with the p53status, we observe only a slightly decreased occurrence of p53 mutations in those primary tumors with p14ARF inactivation. Thus, p14ARFand p53 can be inactivated simultaneously in the same tumor. Such a scenario is also observed in a mouse model, where tumors arising in p19ARF−/− mice can harbor p53mutations (8).
Finally, it is relevant to note that p14ARF promoter hypermethylation is not dependent on p16INK4apromoter methylation status. In our set of colorectal tumors, the most common situation is the simultaneous absence of methylation in both promoters (48%), but after this, the three possible scenarios(p16INK4a methylated alone, p14ARF methylated alone, and both methylated) are similarly represented. In primary colorectal carcinoma,hypermethylation of p14ARF and p16INK4a are independent events. In colorectal cancer cell lines, because the vast majority of cell lines are methylated at p16INK4a, no cell line has p14ARF methylation with an unmethylated p16INK4a promoter. It is also interesting that the CpG island of p15INK4b, which is located only 14 kb upstream of the p14ARF promoter, remains unmethylated in colorectal tumors and cell lines, although it is methylated in leukemia (21). Thus, the p14ARF promoter demonstrates selective epigenetic silencing in a subset of colorectal tumors, with a hypermethylated promoter (p14ARF) between two unmethylated promoters(p16INK4a and p15INK4b) that are frequently methylated in other tumors.
In summary, our data suggest that promoter hypermethylation of p14ARF is a relatively common and early event in colorectal tumorigenesis and is not necessarily related to the methylation status of neighbor gene p16INK4a or restricted to tumors with intact p53.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Supported in part by NIH Grant CA54396 and Grants from Fondo de Investigacion Sanitaria and Comision Interministerial de Ciencia y Tecnología. M. E. is a recipient of a Spanish Ministerio de Educacion y Cultura Award. J. G. H. is a Valvano Foundation Scholar. S. B. B. and J. G. H. receive research funding and are entitled to sales royalties from ONCOR, which is developing products related to the research described in this article. The terms of this arrangement have been reviewed and approved by The Johns Hopkins University in accordance with its conflict of interest policies.
The abbreviations used are: Rb, retinoblastoma;MSP, methylation-specific PCR; RT-PCR, reverse transcription-PCR;GAPDH, glyceraldehyde-3-phosphate dehydrogenase.
We thank Andrew B. Sparks and Kenneth W. Kinzler for providing cell lines for this study.