Analysis of genetic and epigenetic processes that influence p14ARF expression in breast cancer


The INK4a/ARF locus encodes two unrelated cell cycle-regulatory proteins that both function in tumor suppression, p16INK4a and p14ARF. In human tumors including breast cancer, alterations affecting selectively p14ARF have been poorly analysed. We have performed a comprehensive analysis of the inactivation mechanisms (mutation, homozygous and hemizygous deletion, and promoter hypermethylation) in a large series of 100 primary breast carcinomas. RT–PCR showed expression variable of the p14ARF transcript, with 17% demonstrating overexpression and 26% demonstrating decreased expression. No detectable alterations were observed in the majority of cases with overexpressed p14ARF mRNA, but 77% of tumors with decreased expression presented at least one of these genetic/epigenetic alterations. Nevertheless, a statistically significant correlation was observed between decreased p14ARF expression and several poor prognostic parameters.


The INK4a/ARF locus, located at chromosomal region 9p21, encodes two unrelated cell cycle-regulatory proteins that both function in tumor suppression. p16INK4a (CDKN2, MTS1) is a tumor suppressor that encodes a cyclin-dependent kinase inhibitor (p16INK4a) which restrains cell growth through preventing phosphorylation of the retinoblastoma protein (Serrano et al., 1993). The second product of the locus, p14ARF (p19ARF murine) is a protein of 15 kDa generated through an alternative reading frame that replaces the first exon, 1α, of p16INK4a with exon 1β, located >15 kb upstream of exon 1α (Quelle et al., 1995). The p14ARF protein has been shown to act as a cell inhibitor arresting cell growth in G1-S and also G2-M (Quelle et al., 1995). p14ARF interacts with the Mdm2 oncoprotein and inhibits the nuclear export of Mdm2 by tethering it in the nucleolus (Tao and Levine, 1999). This fact avoids the Mdm2-p53 association and blocks the Mdm2-induced p53 degradation in the proteosome, thereby stabilizing p53 (Zhang et al., 1998). Genes such as Myc, adenovirus E1A and E2F-1 when overexpressed in MEFs rapidly activate the ARF-p53 pathway and trigger replicative crisis by inducing apoptosis (Zindy et al., 1998). p16INK4a and p14ARF expressions are controlled by separate promoters responding to different stimuli and they therefore show independent transcriptional regulations. The unusual structure of this locus has lead to ambiguity regarding the biological role of each gene.

The p16INK4a gene was implicated as a tumor suppressor gene by its frequent mutation, deletion or promoter hypermethylation in a wide variety of human tumors. Thus, inactivation by mutations is frequently observed in esophagus, and pancreas carcinomas. Homozygous deletions are a common event in a large variety of tumors, including bladder tumors, mesotheliomas, high grade gliomas, head and neck carcinomas, and hematological cancers. Furthermore, aberrant methylation of the alpha promoter has been frequently shown to silence the gene in colon carcinomas and breast neoplasias (reviewed in Liggett and Sidransky, 1998).

Whereas inactivation of p16INK4a has been documented as a frequent event in human cancer, the presence and consequence of p14ARF alterations are still poorly understood and remain to be clearly elucidated. Homozygous deletion of the locus causes a loss of both the p16INK4a gene and the p14ARF gene, spawning complications in determining the individual contributions of the two genes in tumor suppression. Mice lacking p14ARF but retaining intact p16INK4a develop lymphomas and sarcomas at an early age (Kamijo et al., 1997). Mutations at the specific exon 1β of p14ARF, necessary and sufficient for inhibits Mdm2 activity, stabilization of the p53 tumor suppressor and cell cycle inhibition, have not been reported in primary tumors (Fitzgerald et al., 1996) and those found in exon 2 part common to both proteins do not appear to be deleterious for p14ARF activity (Quelle et al., 1997). Exon 1β-specific deletions only have been described in metastatic melanoma cell lines (Kumar et al., 1998). Deletion inactivation of p14ARF (and often p16INK4a) has been reported in several human cancers (Markl and Jones, 1998; Sanchez-Céspedes et al., 1999; Iwato et al., 2000). Epigenetic alterations as a hypermethylation of the CpG islands at p14ARF promoter have been reported in primary colorectal cancer (Esteller et al., 2000) and in primary gastric carcinomas (Iida et al., 2000). Also, alterations in p14ARF expression levels have been observed in lung cancer (Gazzeri et al., 1998; Vonlanthen et al., 1998), in breast carcinomas (Brenner et al., 1996), and in colorectal tumors (Zheng et al., 2000). Interestingly, the majority of genomic alterations have been found in hematological malignancies. Thus, deletions and genomic instability are frequent events in non-Hodkin's lymphomas (Herranz et al., 1999) and in several lymphomas and adult leukaemia (Takemoto et al., 2000).

However, inactivation of p14ARF in breast cancer has not been sufficiently analysed. To help elucidate the role of aberrations affecting p14ARF in this type of neoplasia we have performed a comprehensive analysis of the inactivation mechanisms and the mRNA expression levels in a large series of primary breast carcinomas. Furthermore, we correlate this results with several clinico-pathological parameters to analyse their influence on the prognosis.

All assay conditions and primer sequences are summarized at Table 1. SSCP analysis was performed on the 100 primary breast carcinomas studied. As reported previously by others (Gazzeri et al., 1998), we did not detect any mutation in exon 1β (data not shown). SSCP was also performed in 10 normal breast tissues and no mutation was observed in exon 1β of these DNA samples.

Table 1 Assay conditions and primer sequences

Two polymorphic microsatellite markers were used to map our primary breast carcinomas series at the INK4a/ARF locus. Overall, 81 cases were informatives to at least one of the two microsatellite markers. LOH was detected in 17 of these 81 tumors (21%).

The common exon 2 is transcribed in the formation of both proteins, p16 and p14, so it was chosen to the homozygous deletions assay. This alteration was observed in only four breast carcinomas (4%).

p14ARF is a candidate for hypermethylation-associated inactivation because contains a documented CpG island which can be silenced by this epigenetic alteration (Robertson and Jones, 1998). To establish the methylation status of the 5′ region, tumor DNA was obtained from the 100 primary breast carcinomas and was subjected to p14ARF promoter methylation study by methylation-specific PCR (MSP) as previously described (Herman et al., 1996). Of all tumor samples amplified, 24 (24%) presented p14ARF methylation. Bisulfite-modified DNA from 30 normal tissues was not amplified with primers for p14ARF on PCR. These results confirm that methylation observed in primary breast cancers is a tumor-specific change.

As previously suggested (Stone et al., 1995), the similarity in size and sequence of the α and β transcripts may cause difficulties to determinate each one specifically. Only an assay of transcripts using the unique sequences of exon 1β would be able to assess the true levels of p14ARF expression. Thus, an specific analysis of the p14ARF expression mRNA was carried out by RT–PCR. Among 100 primary breast tumors, 26 (26%) expressed no detectable p14ARF mRNA or very low levels, and 17 cases (17%) expressed high levels of p14ARF mRNA when normalized with β-actin.

Of the 26 cases with decreased p14ARF expression levels, 20 (77%) presented at least one genetic or epigenetic alteration. Thus, homozygous deletion and LOH were found in three (12%) and seven (27%) respectively, and aberrant promoter hypermethylation was observed in 13 (50%) of these tumors (Table 2). One of these tumors (#77) presented homozygous deletion and LOH at hMp16α-l1 marker concomitantly. This fact is possible because markers used to determinate both alterations were located at exon 2 and upstream at exon 1α respectively, showing that homozygous deletion did not affect the complete locus at one of alleles.

Table 2 Genetic/epigenetic alterations detected in tumors with p14ARF mRNA decreased expression

Interestingly, no genetic or epigenetic alterations were observed in 15 (88%) of the 17 tumors with p14ARF mRNA overexpression. The others two cases showed loss of heterozigosity. A statistically significant association was found between tumors with overexpressed mRNA and negative ERBB2 expression (P=0.018).

A correlation study was designed with the objective of investigating the prognostic value of several parameters obtained from the medical records of the 100 patients (100 tumors and corresponding normal tissues). When we analysed and compared tumors with decreased p14ARF mRNA expression and the clinico-pathological parameters we observed a statistically significant correlation with some poor prognostic parameters as peritumoral vessel involvement (P=0.001), p53 mutational status (P=0.02) and negative progesterone receptors (P=0.04) (Table 3).

Table 3 Statistically significant correlation between p14ARF mRNA expression levels and clinico-pathological parameters

As reviewed previously (Chin et al., 1998), the evidence for a direct role of p14ARF alterations in human cancers through genetic analysis has been lacking. In our study p14ARF mRNA expression was decreased in 26% of the tumor analysed, suggesting that p14ARF may be involved in the tumorigenic process in primary breast carcinomas. Our results suggest that mutations of p14ARF and homozygous deletion may be rare, and loss of heterozygosity or methylation rather than these may play a role in p14ARF inactivation in breast cancer. Figure 1 shows several representative cases.

Figure 1

Representative cases showing different inactivation mechanisms causing decreased p14ARF mRNA expression. All amplifications were carried out in a 10 μl reaction volume containing 50–100 ng/μl of sample DNA, 0.3 units of AmpliTaq Gold DNA polymerase, 1 μl of 10×PCR buffer, 200 μM dNTP, 0.6 μM of each primer (except in homozygous deletion assay in which primer concentrations were 0.3 and 0.9 μM of INK4a/ARF and β-globin respectively) and 2 mM of MgCl2. PCR condition were: 94°C for 9 min, 35 cycles of 94°C for 30 s, specific annealing temperature for 30 s and 72°C for 30 s, followed by incubation at 72°C for 11 min. Following amplification, all samples were run on nondenaturing 6% polyacrylamide gels at room temperature for 2–4 h at 400 v, (except in the PCR–SSCP assay where electrophoresis was run on nondenaturing 8–12% polyacrylamide gels at 4°C for 12–15 h at 250 v). The allelic band intensity was detected by a non-radioisotopic method using a commercially available silver staining method (Oto et al., 1993). In homozygous deletion assay β-globin served as an internal control. Shifts intensity were analysed by densitometry. When the intensity of the exon 2 shift corresponding to the tumor was less of 25% of its normal counterpart, normalized to β-globin, we considered the existence of homozygous deletion. In methylation assay placental DNA treated with SssI methyltransferase and DNA of KMH2 and L-540 cell lines were used as positive control for methylated alleles. DNA from normal lymphocytes was used as a negative control for methylated genes. In RT–PCR β-actin served as a positive control. Elevated expression was arbitrarily defined as a fourfold or higher increase over the β-actin ratio, and decreased expression as a fourfold or lower under the β-actin ratio

Nevertheless, our data establish that, in primary breast carcinomas, the inactivation of p14ARF is not restricted to those tumors that presented intact p53. Evidences supporting direct biochemical interactions between p14ARF and p53 have been obtained (Zhang et al., 1998). It was hypothesized that p14ARF inactivation and p53 mutation in human cancers should be mutually exclusive because both act on the same pathway. This hypothesis has been observed in some human cancers (Markl and Jones, 1998; Vonlanthen et al., 1998), however, we did not find this inverse correlation in our primary breast tumors. Moreover, we found that inactivation of p14ARF frequently coexists with p53 mutation. This fact is supported by others reports (Sanchez-Céspedes et al., 1999; Esteller et al., 2000) and make evidences that is possible that p14ARF inactivation is not functionally equivalent to abrogation of the p53 pathway by a p53 mutation. Thus, it is clear from this data that p14ARF alterations and p53 mutation are not mutually exclusives in primary breast carcinomas and both can be inactivated simultaneously in the same tumor.

High p14ARF mRNA levels were detected in 17% of all tumors analysed. As reported previously this fact can be produced by multiple events. Overexpression of Myc, adenovirus E1A or E2F-1 in primary MEFs rapidly induces p14 ARF expression and leads to p53-dependent apoptosis (Zindy et al., 1998), and this mechanism is probably what happens in human cancers. The p14ARF overexpression and the following p53-dependent cell cycle arrest appear as the normal cellular response when hyperproliferative signals exist. Probably, in these tumors another alteration in the p14ARF/Mdm2/p53 pathway were happening and this situation eludes the cell cycle control.

Interestingly, a significant correlation (P=0.018) was observed between p14ARF overexpressed tumors and ERBB2 negative. Mechanisms explaining this association are not elucidated. We could hypothesise that p14ARF downregulates ERBB2 expression but this fact has not been demonstrated.

Nevertheless, our comprehensive analysis of mutation, homozygous and hemizygous deletion and methylation suggests that the tumor suppressor p14ARF is cumulatively affected in approximately 41 (41%) of the primary breast carcinomas analysed. Twenty-six (63%) of these tumors presented p14ARF mRNA decreased expression being the p14ARF promoter hypermethylation the first cause of genetic silencing followed by loss of heterozygosity and homozygous deletion.


  1. Brenner AJ, Paladugu A, Wang H, Olopade OI, Dreyling MH, Aldaz CM . 1996 Clin. Cancer Res. 2: 1993–1998

  2. Chin L, Pomerantz J, DePinho RA . 1998 Trends Biochem. Sci. 23: 291–296

  3. Esteller M, Tortola S, Toyota M, Capella G, Peinado MA, Baylin SB, Herman JG . 2000 Cancer Res. 60: 129–133

  4. Fitzgerald MG, Harkin DP, Silva-Arrieta S, MacDonald DJ, Lucchina LC, Unsal H, O'Neill E, Koh J, Finkelstein DM, Isselbacher KJ, Sober AJ, Haber DA . 1996 Proc. Natl. Acad. Sci. USA 93: 8541–8545

  5. Gazzeri S, Della Valle V, Chaussade L, Brambilla C, Larsen CJ, Brambilla E . 1998 Cancer Res. 58: 3926–3931

  6. Herman JG, Graff JR, Myohanen S, Nelkin BD, Baylin SB . 1996 Proc. Natl. Acad. Sci. USA 93: 9821–9826

  7. Herranz M, Urioste M, Santos J, Rivas C, Martinez B, Benitez J, Fernandez-Piqueras J . 1999 Leukemia 13: 808–810

  8. Iida S, Akiyama Y, Nakajima T, Ichikawa W, Nihei Z, Sugihara K, Yuasa Y . 2000 Int. J. Cancer 87: 654–658

  9. Iwato M, Tachibana O, Tohma Y, Arakawa Y, Nitta H, Hasegawa M, Yamashita J, Hayashi Y . 2000 Cancer Res. 60: 2113–2115

  10. Kamijo T, Zindy F, Roussel MF, Quelle DQ, Downing JR, Ashmun RA, Grosveld G, Sherr CJ . 1997 Cell 91: 649–659

  11. Kumar R, Sauroja I, Punnonen K, Jansen C, Hemminki K . 1998 Gene. Chromosome. Canc. 23: 273–277

  12. Liggett WH, Sidransky D . 1998 J. Clin. Oncol. 16: 1197–1206

  13. Markl IDC, Jones PA . 1998 Cancer Res. 58: 5348–5353

  14. Oto M, Miyake S, Yuasa Y . 1993 Ann. Biochem. 213: 19–22

  15. Quelle DE, Zindy F, Ashmun RA, Sherr CJ . 1995 Cell 83: 993–100

  16. Quelle DE, Cheng M, Ashmun RA, Sherr CJ . 1997 Proc. Natl. Acad. Sci. USA 94: 669–673

  17. Robertson KD, Jones PA . 1998 Mol. Cell. Biol. 18: 6457–6473

  18. Sanchez-Céspedes M, Reed AL, Buta M, Wu L, Westra WH, Herman JG, Yang SC, Jen J, Sidransky D . 1999 Oncogene 18: 5843–5849

  19. Serrano M, Hannon GJ, Beach D . 1993 Nature 366: 704–707

  20. Stone S, Jiang P, Dayananth P, Tavtigian SV, Katcher H, Parry D, Peters G, Kamb A . 1995 Cancer Res. 55: 2988–2994

  21. Takemoto S, Trovato R, Cereseto A, Nicot C, Kislyakova T, Casareto L, Waldmann T, Torelli G, Franchini G . 2000 Blood 95: 3939–3944

  22. Tao W, Levine AJ . 1999 Proc. Natl. Acad. Sci. USA 96: 6937–6941

  23. Vonlanthen S, Heighway J, Tschan MP, Borner MM, Altermatt HJ, Kappeler A, Tobler A, Fey MF, Thatcher N, Yarbrough WG, Betticher DC . 1998 Oncogene 17: 2779–2785

  24. Zhang Y, Xiong Y, Yarbrough WG . 1998 Cell 92: 725–734

  25. Zheng S, Chen P, McMillan A, Lafuente A, Lafuente MJ, Ballesta A, Trias M, Wiencke JK . 2000 Carcinogenesis 21: 2057–2064

  26. Zindy F, Eischen CM, Randle D, Kamijo T, Cleveland JL, Sherr CJ, Roussel MF . 1998 Genes Dev. 12: 2424–2434

Download references


We are grateful to Mr Robin Rycroft for his assistance with the English language, revision and preparation of the manuscript. This work was supported by grants from the Fundación Banco Santander Central Hispano, Aventis Pharma S.A and CAM 08.1/0069/20002.

Author information



Corresponding author

Correspondence to Félix Bonilla.

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Silva, J., Domínguez, G., Silva, J. et al. Analysis of genetic and epigenetic processes that influence p14ARF expression in breast cancer. Oncogene 20, 4586–4590 (2001).

Download citation


  • allelic losses
  • p14ARF
  • breast cancer
  • hypermethylation
  • overexpression
  • mutations

Further reading