High SIRT1 expression is a negative prognosticator in pancreatic ductal adenocarcinoma

Pancreatic ductal adenocarcinoma (PDAC) is the fourth leading cause of cancer related deaths in the United States. While substantial progress has been made in the understanding of pancreatic cancer biology [1], therapeutic concepts still provide only modest benefit [2]. The overall 5-year survival rate is approximately 5% [3]. Surgical resection is the only efficient and potentially curative treatment option with 5-year survival rates of around 20% in patients with resectable tumors, but can only be applied in approximately 15-20% of the cases emphasizing the urgent need for early detection strategies [4].

The main prognosticators for surgically resectable PDACs are location, tumor size, resection margin, nodal status, and histological grade. Although these risk factors have been proven to be clinically useful, their ability to reliably predict outcome is limited and mainly reflects tumor distribution rather than tumor biology [4].

Hence, numerous studies have been conducted to identify novel biomarkers that aid outcome prediction and to unravel molecular mechanisms that drive tumor development [5].

Sirt1 (homolog of yeast silent information regulator, Sir2), an isoform of enzymes of the silent information regulatory family (sirtuins), is an evolutionary conserved NAD dependent histone/protein deacetylase (class III HDAC) that mediates epigenetic silencing by modification of lysine residues of histones and deacetylation of numerous non-histone substrates. One of the first substrates identified was p53, whose deacetylation was reported to repress p53-dependent apoptosis in response to cellular stress and DNA damage [6, 7]. Meanwhile, many other targets have been identified, including Ku70 [8], PTEN [9], p73 [10], RelA/p65 [11], FOX01, FOX03a, and FOX04 [12], NICD [13], hypoxia-inducible factors HIF-1α, -2α [14, 15], β-catenin [16], XPA [17], SMAD7 [18], and cortactin [19]. Deacetylation of these targets regulates cell survival, proliferation, and angiogenesis. Overexpression of sirtuins was initially reported to increase lifespan in budding yeast, Caenorhabditis elegans, and Drosophila melanogaster [20‐22] but for the latter two the findings were challenged by a recent study of Burnett and colleagues [23].

The functional role of Sirt1 in cancer is equivocal and suggested to be context dependent [24]. Although there are convincing studies that argue for a tumour suppressive role of Sirt1, recent data provide functional evidence that Sirt1 has oncogenic properties in neuroblastomas by facilitating n-myc stabilization [25]. Serrano [26] reported that transgenic Sirt mice crossed with PTEN-null mice were observed to develop thyroid and prostate cancer further arguing for a tumor promoting function of Sirt1.

While several studies found deregulation of Sirt1 in various tumor entitites including ovary, prostate, gastric, colon, hepatocellular carcinoma as well as melanoma and glioblastoma [27], comprehensive in vivo data in pancreatic cancer is still missing. Reports that explore Sirt1 function in pancreatic cancer are sparse [28].

Hence, we set out to comprehensively investigate Sirt1 expression in a large series of PDACs, its relationship to survival and to assess the functional relevance in cell culture models.

This is the first study that demonstrates Sirt1 to be an independent prognosticator in PDAC with high Sirt1 expression indicating poor outcome. Moreover, our data argue for a functional role of Sirt 1 during tumorigenesis indicating that Sirt1 is not only a biomarker but a potentially oncogenic protein in the PDAC context, whose overexpression leads to increased cell viability in both cell lines, while pharmacological inhibition leads to a concentration-dependent stepwise decrease of viable cells. Cambinol treatment negatively interferes with cell cycle progression (in MiaPaCa-2 cells) and induces apoptosis (in PANC-1 cells) as well as senescence (both cell lines). These observations are in line with Wauters et al. [33] showing an enhancing effect for cell viability and regulatory function of Sirt1 for acinar-to-ductal metaplasia in pancreatic carcinogenesis. The latter results also match data presented by Zhao et al. [28] who reported that utilizing small hairpin RNA Sirt1 knockdown led to increased apoptosis and senescence in PANC-1 cells. However, we failed to observe a synergistic effect of Sirt1 inhibition with Gemcitabine treatment as reported by Zhao et al. [28]. This divergent result may be attributed to the distinct targeting approach in our study, which uses cambinol, a clinically applicable drug with promising anti-cancer effects in animal models of skin cancer and Burkitt’s lymphoma as well as in several cancer cell lines [34]. Interestingly, we detected an application time- and concentration-dependent loss of Sirt1 protein upon cambinol treatment. The underlying cause for this effect, which abrogates Sirt1-function, remains to be elucidated and may be due to protein degradation.

Consistent with the results by Zhao et al. [28] obtained by immunhistochemistry, qPCR and western blotting, we observed a variable expression of Sirt1 in PDACs but did not see a positive correlation of Sirt1 expression with age, tumor size, and lymphatic spread. The different findings may be explained by distinct cohort characteristics including cohort size, age, and sex. However and in contrast to Zhao et al., we observed a strong correlation with higher tumor grades, i.e. the less differentiated the cancer cells are the more Sirt1 expression they exhibit. This finding is of interest since there are reports that implicate Sirt1 in the regulation of cellular differentiation and dedifferentiation processes [35, 36]. Dedifferentiation and the associated phenomenon of epithelial-to-mesenchymal-transition play an essential role in the development of early local and distant tumor spread. Observations that link high Sirt1 expression to poorly differentiated cancers were also made by other investigators for hepatocellular carcinoma [37], prostate cancer [38] and glioblastoma [39].

The association between high Sirt1 expression and poor histological grade may also explain why in our cohort Sirt1 expression is associated with poor outcome regardless of the tumor stage as shown by its prognostic independency in multivariate survival analysis. A Sirt1 positive and poorly differentiated tumor may have acquired a biological profile that allows for e.g. early systemic spread of –clinically undetectable- micrometastases in lymph nodes and distant organs leading to impaired survival regardless of the tumor size and metastases detected at the point of initial tumor diagnosis. A recent study by Nalls and colleagues [40] showed that SAHA-induced micro-RNA 34a (miR34a) expression in human pancreatic cancer cells putatively directly inhibited Sirt1 expression by binding within the 3’UTR of Sirt1. On cellular level, restoration of miR34a expression led to growth inhibition as well as decreased epithelial to mesenchymal transition (EMT) and invasion. Although miR34a does not exclusively target Sirt1, this recent study further argues for an oncogenic role of Sirt1 in PDAC development. Recent results obtained by Pramanik et al. corroborate this view [41].

Functional studies indicate that the subcellular localization of Sirt1 might have functional implications in carcinogenesis. Wauters et al. [33] recently provided evidence that there is nuclear to cytoplasmic shuttling of Sirt1 in rat and mouse acinar cells with potential tumorigenic implications in the acinar to ductal metaplasia carcinogenesis model of PDAC. They also reported on cytoplasmic localization of Sirt1 in exocrine cells of the human pancreas. However, investigating human tissue samples of fully developed pancreatic ductal adenocarcinoma, we only detected nuclear localized Sirt1. This may have several reasons. One potential explanation might be that endogenous cytoplasmic Sirt1 levels in comparison to nuclear expression levels are too low to be detected by our antibody. Another explanation would be that cytoplasmic Sirt1 plays a major role in the development of carcinogenic precursors and nuclear Sirt1 has its place in the fully developed cancer. However, this has to be investigated in future functional studies.

Interestingly, following up the seminal work by Luo et al. and Vasiri et al. [6, 7], a very recent study by Li and coworkers [42] explored the Sirt1-p53 axis in chronic myeloid leukemia (CML) and found that targeting of Sirt1 by either shRNA or the small molecule inhibitor tenovin-6 resulted in increased levels of acetylated p53 in CML CD34+ cells accompanied by increased transcriptional activity of p53. Abrogation of Sirt1 led to growth inhibition and reduced engraftment of the tumor cells. These effects were even more pronounced when cells were synergistically treated with the tyrosine kinase inhibitor imatinib. These data strengthen the view of a context-dependent tumorigenic impact of Sirt1 as also suggested by our results. Since p53 aberrations are commonly involved in PDAC tumorigenesis [43, 44], it is tempting to speculate whether Sirt1 inhibition may help to restore the remaining functionally intact p53 pool. Indeed, recent data [45] indicate that downregulation of Sirt1 by restoration of HIC1 (hypermethylated in cancer 1) leads to increased levels of acetylated p53 and upregulated p21 in pancreatic cancer. On cellular level, overexpressed HIC1, which in turn led to downregulation Sirt1 resulted in cell cycle arrest and apoptosis. Loss of p53 function has also been implicated in resistance to EGFR-targeting strategies [46], the latter having a limited but significant role in the treatment of PDACs [47]. Interestingly, we observed a synergistic impact of combined Sirt1- and EGFR-inhibition suggesting a functional interdependence in PDACs, whose molecular details remain to be explored. In prostatic cancer cells Byles and colleagues [48] observed Sirt1 to modulate EMT upon EGF signalling via the induction of the transcription factor ZEB1. Although it remains to be investigated whether this mechanism works in PDACs, our data and these results may additionally point to a therapeutic rationale for combined EGFR/Sirt1 inhibition.

While a number of small molecule inhibitors of class I and II HDACs are currently in clinical trials for the treatment of malignancies of various organ origins [49], SIRT1 inhibition is currently only investigated in a phase I trial of patients with Huntington’s disease.

In conclusion, there is accumulating evidence that Sirt1 has an oncogenic role in PDACs and provided that further studies are able to reproduce and extent the data presented herein towards mouse model systems, a clinical trial for patients with PDAC, whose outcome and treatment options are extremely limited for the vast majority of patients, may be worthwhile to consider.