Association studies of the copy-number variable ß-defensin cluster on 8p23.1 in adenocarcinoma and chronic pancreatitis

Pancreatitis, a necroinflammatory condition of the pancreas, has both acute and chronic manifestations. In the recent past, our understanding of the pathogenesis of pancreatic inflammation has improved considerably. Whereas acute pancreatitis is known to be initiated by premature activation of digestive enzymes within the exocrine component of the pancreas, chronic pancreatitis (CP) is characterized by progressive and irreversible damage to both the exocrine and endocrine components of the pancreas. CP is believed to result from repeated overt or silent episodes of acute pancreatitis [1]. The key histopathologic features of CP are pancreatic fibrosis, acinar atrophy, chronic inflammation, and distorted and blocked ducts [2]. The annual incidence of CP in industrialized countries ranges from 3.5 to 10 per 100,000. Alcohol abuse is the major risk factor for CP in Western countries, although other mechanisms such as mutations, pancreatic duct obstruction (caused by strictures), hypertriglyceridemia, hypercalcemia, and autoimmunity also have been implicated [3]. Since patients with CP have an approximately 13-fold higher risk to develop pancreatic cancer than the general population, the identification of disease-related genes is essential for understanding the transformation from benign to malignant disorder, and for developing strategies for early diagnosis [4]. Pancreatic cancer (pancreatic ductal adenocarcinoma, PDAC), is the fourth most common cause of cancer-related death in industrialized countries and characterized by extremely low survival rates [5, 6]. Currently, no imaging procedure can reliably differentiate between benign and malignant tumors in CP patients.

Defensins are small cysteine-rich peptides that can be classified as either α-defensin or β-defensin, depending upon the arrangement of six critical cysteine residues. Defensins are synthesized as inactive preproproteins that become post-translationally activated. They are produced in the respiratory, gastrointestinal and genitourinary tracts, the skin, and by circulating blood cells. Defensins are considered a first line of defence against invading pathogens. Of all defensins, the ß-defensins comprise the largest group, with around 40 members encoded in the human genome. Most of the genes are located in defensin (DEF) clusters on chromosomes 8 and 20. In addition to their antimicrobial activity, β-defensins have multifaceted functions in innate and adaptive immunity [7]. The β-defensins are expressed in most epithelial cells and are found to be impaired in many inflammatory diseases, including Crohn's disease, psoriasis, pulmonary inflammation, and periodontal disease [8–13].

Except for DEFB1, all ß-defensin genes (DEFB4, DEFB103 109) harbour a high degree of copy-number variation (CNV). Copy numbers (CN) range from 2 to 13 copies per diploid genome and show large inter-genic concordance because the respective genes bunch in a ~200 kb CNV region, called ‘DEF cluster b’ (Figure 1) [14–20]. In principle, both CNV and sequence variation within a given copy can contribute to clinical phenotypes through variation in gene expression [21]. So far, this has only been demonstrated experimentally for DEFB4 but most probably applies to all other DEF genes with variable CN as well [17]. Anyhow, the analysis of genes located in CNV regions poses two methodological challenges: First, conventional genotyping of single nucleotide variations (SNV) cannot resolve whether a variation occurs between paralogs affected by CNV [22]. To acknowledge this problem, we will address SNVs within the copy number variable DEFB cluster as ‘multi-site variation’ (MSV). Second, the assessment of CNs is also complicated by methodological difficulties. Although widely used, qPCR has questionable reliability so that tightly controlled paralog ratio tests or multiplex ligation-dependent probe amplification (MLPA) have been recommended instead [23–28].

Disease association studies of the CN-variable ß-defensin genes are scarce. Hitherto reported associations mostly either lack replication or are conflicting [21, 29–32]. No CN association, but a skewed distribution of MSV haplotypes, was identified in two prostate cancer groups [33]. The corresponding haplotypes comprise four MSV (rs17843871, rs2680507, rs17843872 and rs4259430) around exon 1 of DEFB104. While haplotypes GGGC and CAAT were significantly under-represented among patients, GAAT and GAAC were significantly over-represented. Moreover, high CNs of the ß-defensin cluster (≥9 copies per diploid genome) were found to be less frequent among prostate cancer patients than among healthy controls.

The aim of the present study was to search for associations between MSV-based DEFB104 haplotypes and DEF cluster b CNs on the one hand, and pancreatic ductal adenocarcinoma and chronic pancreatitis on the other.

Defensins are expressed in the pancreas although it is not entirely clear which cells actually produce and secrete which of these diverse peptides for which purpose. In pancreatic juice, only HNP-3 (α-defensin 3, encoded by DEFA3) has been detected but mRNA expression of ß-defensins has also been demonstrated in pancreatic tissue [35]. The important role of defensins in the innate immune system due to their antimicrobial, chemotactic and regulatory functions, and their involvement in inflammatory processes vindicates the assumption that defensins are also involved in the pathogenesis of pancreatitis and pancreatic cancer.

Pancreatitis may develop as a chronic disease after long-term alcohol abuse. Chronic pancreatitis is a strong risk factor for pancreatic cancer but alcohol does not appear to be an independent causative agent for the disease [36, 37]. Interestingly, acute and chronic pancreatic inflammation occurs as an extra-intestinal co-morbidity of inflammatory bowel disease for which an involvement of defensins is also discussed [27, 38, 39]. Furthermore, in view of the microbicidal properties of defensins, it appears noteworthy that a link between infectious diseases and pancreatic cancer has been drawn both for viral diseases (mumps, HBV infections) and bacterial infections (Helicobacter pylori) [37, 40].

In the present study, both sequence variants in a ß-defensin gene and CN variants of the cluster containing this gene were investigated for a putative association with PDAC and CP. As haplotyping of the CN-variable DEFB104 gene was performed in pools, haplotypes cannot be assigned to the individual diploid CN for DEF cluster b. Respectively, both features had to be tested for association independently. All data from the patient groups were compared to age- and sex-matched healthy controls (CARLA1 and CARLA2).

Earlier, we have demonstrated association between DEFB104 haplotypes and sporadic prostate cancer as well as under-representation of high diploid DEF cluster CN in patients with this disease [33]. In the present study, no statistical support was found for an association between DEFB104 haplotypes and either PDAC or CP. However, analysis of the diploid CN distributions revealed a statistically significant difference between PDAC and CARLA1 that was due mainly to a paucity of 5- and 6-copy samples and an excess of 3-copy ones in the PDAC cohort.

Recently, under-representation of higher diploid DEFB4 CNs (>4) was reported in patients with acute pancreatitis (AP) and severe acute pancreatitis (SAP) [41]. Since DEFB4 is part of DEF cluster b and concordance for the CN of all genes within cluster b has been shown [16], this result is in agreement with our findings in the PDAC cohort. However, for CP, known to increase the risk for developing pancreatic cancer by 10 to 20-fold and a possible outcome of AP, we and others [42] did not observe significant associations with genetic features of the DEF cluster b, potentially pinpointing different roles of defensins in the etiopathogenesis of pancreatic diseases.

Although the functional consequences of the lower DEF cluster b CN observed for PDAC and AP are not yet resolved, lower CNs are rather associated with lower defensin expression [17]. In the light of inflammation as key feature of AP and the established link between inflammation and cancer [43], a low CN would be consistent with an anti-inflammatory effect of defensins described recently [44]. Assuming instead defensins to exert a pro-inflammatory effect [45] would favor a role of perturbed antimicrobial barrier defense in the etiopathogenesis of PDAC and AP. Further studies are necessary to find these missing functional links and to clarify which genetic variants may serve as reliable and feasible markers in the diagnosis and prognosis of pancreatic diseases.

Different DEF cluster b CN distribution between PDAC patients and healthy controls indicate a potential protective effect of higher CNs against the disease. Replication of the study with larger sample numbers are needed to confirm the result and to draw definitive conclusions thereof.

GG, CP and RG collected the disease cohort’s samples and provided the patient’s data. KHG, KW, AK and OK conceived the CARLA study, did the cohort matching and provided the control samples. They also performed the statistical analyses with the support of MK, AW and MN. ST and MG carried out the molecular genetic studies with support by PR. ST, KH and MP drafted the manuscript. All authors read and approved the final manuscript.