Alterations of tumor suppressor gene p16 ^INK4a in pancreatic ductal carcinoma

Pancreatic cancer is a malignant neoplasm in the digestive tract the etiology of which is not known fully as yet. Recent advances in molecular oncology have provided explanations at the DNA level that multiple genetic changes contribute to pancreatic cancer development in which the p16 locus of tumor tissue is nearly always altered [1]. The p16 tumor suppressor gene located on 9p21 enodes a 16 KDa protein that acts as a cyclin dependent kinase (CDK) 4/6 inhibitor [2].

p16 belongs to an important group of proteins that includes the p15^INK4b, p21^waf1 and p27^KIP1 which negatively regulate the G₁ phase of cell cycle [2]. The p16 gene product binds to CDK4 and CDK6 inhibiting their interaction with cyclin D1. The inhibition of cyclin D1-CDK4/6 complex activity prevents retinoblastoma protein phosphorylation and release of E2F, leading to the inhibition of cell cycle in G1/S transition [3]. Genetic abnormalities inactivating the p16 gene thus confer a growth advantage to the cell contributing to tumorigenesis.

Inactivating alterations of the gene have been commonly identified in a number of human malignancies [4, 5]. In cancers, functional loss of p16 ^INK4a occurs as a consequence of somatic mutations, homozygous and heterozygous deletions [6, 7]. A high frequency of homozygous deletion and mutation of this gene have been detected in cell lines derived from different types of tumors (glioma, breast, lung, bladder and melanoma) [6, 8] suggesting that p16 may play an important role in the regulation of cellular growth in the majority of cell types. However, homozygous deletions and somatic mutations are rarely observed in primary tumors with p16 genetic alterations [9, 10]. On the other hand, denovo methylation has also been proposed to be an important alternative mechanism of p16 gene inactivation [11]. DNA hypermethylation can inhibit transcription of tumor suppressor genes and mismatch repair genes (p16, hMLH1 and VHL), providing an epigenetic mechanism of selection during tumorigenesis [11, 12].

Abnormalities of tumor suppressor gene p16 have been reported in a variety of human tumors but less information is available regarding alterations of p16 in primary pancreatic ductal carcinoma than in pancreatic cancer derived cell lines and xenografts. There have been a few reports on the p16 alterations in tissue specimen of primary pancreatic ductal adenocarcinomas till date [13‐17]. To provide further evidence that the cell cycle inhibitor p16 might be relevant for pancreatic tumorigenesis, we performed a comprehensive analysis of the mechanisms involved in p16 inactivation such as mutation, hypermethylation and homozygous deletion, in primary ductal adenocarcinomas.

The Rb / p16 tumor suppressive pathway can be abrogated in tumors by inactivation of any of the several members of the pathway [22, 23]. For pancreatic carcinoma, this disruption is caused exclusively by inactivation of p16 ^INK4a gene and, only rarely, the Rb gene [24, 25]. Inactivation of p16 gene occurs through intragenic mutation, homozygous deletion and methylation associated transcriptional silencing. In the present study, all known mechanisms of p16 inactivation were analysed and we found evidence of p16 inactivation in a high proportion (80%) of the pancreatic tumors examined.

Most of the previous studies regarding pancreatic cancer have reported an elevated frequency of p16 gene alterations in pancreatic cancer derived cell line and xenografts [7, 13, 26, 27]. There have been a few reports on the p16 alterations in tissue specimen of primary pancreatic ductal adenocarcinomas. Our report is the sixth such study and the first from India. A comparison of the result of all the reports with the present study is shown in Table 3. Homozygous deletion could not be detected in two studies by Gerdes et al [14] and Ohtsubo et al [15], but were observed in 35% cases in the study by Zhonghua et al [16]. In our study, we were unable to amplify the p16 exons 1–3 in 12% cases presumably indicating homozygous deletion which is in good agreement with 10% deletion frequency reported by Huang et al [13]. We identified mutation in 4 out of 25 (16%) cases which is consistent with 15% mutation frequency reported by Ohtsubo et al [15]. However, this frequency is lower than 22.5% reported by Gerdes et al [14]. Also, the frequency of mutation was higher in exon 2 than exon 1 (3 cases in exon 2 compared with 1 case for exon 1). This result is consistent with the previous reports [13, 14]. Of the 4 points mutations reported here, 3 were transversions and 1 was a missense mutation. Importantly, one out of these four mutations (glutamine → valine at codon 88) was located in the ankyrin repeat III in a highly conserved region believed to form the CDK binding cleft and may have resulted in loss of p16 function [28]. Out of the remaining three mutations, one was in ankyrin repeat I and two others were in ankyrin repeat IV. None of these mutations involved the conserved amino acid and hence may not be crucial to p16 function [28]. On the other hand, two of these mutations resulted in a loss of p16 protein expression and so, it may be postulated that they could have affected the stability and half life of the mRNA or protein. This postulation needs confirmation by further experimental approaches. In 2 out of 18 tumors negative for p16 protein expression, we were unable to detect any genetic alteration in the form of deletion, mutation or promoter hypermethylation. However, presence of a mutation is not completely excluded, as we used PCR-SSCP as a screening tool and it is known from the published studies that this technique detects only about 70% of mutations [19, 29].

We have observed hypermethylation to be a major mechanism of p16 inactivation. This frequency is higher than the 27% frequency reported by Gerdes et al [14] and Fukushima et al [30]. Other studies report and even lower frequency of 15% and 3.3% [15, 17]. The later study was however performed on microdissected tissue samples using restriction enzyme analysis in contrast to our study on fresh samples using MSPCR. Another explanation for the observed differences in the methylation frequencies could be differences in the ethnicity of the population studied and the role of unknown environmental factors. Such a difference in the methylation profile is reported in the context of hepatocellular carcinomas by Shen et al [31]. In pancreatic cancer cell lines also, the prevalence of p16 promoter methylation has ranged from 18%-38% [27, 32]. p16 promoter methylation has also been detected in pancreatic intraductal neoplasia adjacent to pancreatic cancer [28, 33, 34]. Moreover, the detection of p16 promoter methylation in the pancreatic fluid of patients with pancreatic cancer but not in chronic pancreatitis patients has suggested its potential role as a diagnostic marker in the differentiation of benign and malignant pancreatic disease [35].

Most of the previous studies which have correlated p16 alterations to survival data are limited by the fact that a comprehensive analysis of the p16 genetic alterations is lacking. Bartsch et al [36] found reduced survival in patients with p16 mutations but this study did include the analysis of p16 inactivation by homozygous deletions. Hu et al [37], found longer survival in patients with immunohistochemically p16 positive tumors compared with p16-negative ones, but this difference was not significant. Naka et al [38], identified a significantly longer survival in IHC p16 positive tumors, but the analysis of gene mutations was not examined in these studies. The association of p16 alterations with worse prognosis has also been reported by Gerdes et al [14] and Ohtsubo et al [15]. In the present analysis, although correlation to survival was not possible, p16 alterations did not show any association with the stage or histological grade of the tumor.