Background

Hormone sensitive lipase (HSL) is an intracellular, neutral lipase that catalyzes the hydrolysis of triacylglycerol, diacylglycerol, monoacylglycerol, cholesteryl esters, and retinyl esters [1]. Its activity against diacylglycerol is several-fold higher than against triacylglycerol and monoacylglycerol. Adipose triglyceride lipase (ATGL) and HSL are therefore the major enzymes contributing to triacylglycerol breakdown in the adipose tissue [2]. Interestingly, an increase in the level and activity of HSL has been implicated in the pathogenesis of cachexia [3, 4], a debilitating syndrome characterized by progressive loss of adipose tissue via increased lipolysis [2,3,4,5]. Consequently, pharmacological inhibitors of HSL have been proposed for the treatment of cancer-associated cachexia [4] and a number of compounds have been synthesized and characterized [6, 7]. However, the impact of HSL suppression on cancer development has not been examined.

Pancreatic ductal adenocarcinoma (PDAC) is an extremely aggressive disease with an overall 5-year survival rate of about 8% [8]. Currently, it is the fourth leading cause of cancer deaths in both men and women [8]. PDAC mortality is projected to increase, and before the year 2030 it is expected to become the second leading cause of cancer-related deaths [9]. Cachexia is a prominent condition in PDAC that severely restricts therapeutic options. Therefore, the development of pharmacological agents that can attenuate or reverse cachexia in the context of PDAC is of clinical importance. Administration of inhibitors of HSL have been proposed in the management of cachexia [4] but the precise effect of chronic HSL suppression on the progression of PDAC has never been examined.

HSL null mice appear phenotypically normal (with the exception of infertility in males due to severe oligo- or azoospermia) and are resistant to diet-induced and genetic obesity [10]. In order to assess the role of HSL in PDAC development, we generated conditional KrasG12D mice with HSL deficiency. We found that KC mice with HSL deficiency, compared to KC mice with functional HSL, had similar weight gain and enhanced adipose tissue (AT) and pancreatic inflammation. Surprisingly, KC mice with HSL deficiency exhibited an increased incidence of PDAC. Our results strongly indicate that HSL deficiency is sufficient to accelerate PDAC development in KC mice and therefore imply that chronic suppression of HSL has an unrecognized tumor promoting effect in the KC model.

Methods

Conditional KrasG12D mouse model with HSL deficiency

The conditional KrasG12D (KC) mouse model from Hingorani and colleagues was used for this study [11]. In the KC (LSL-KrasG12D;p48-Cre) strain, expression of oncogenic KrasG12D is activated by Cre-mediated excision of LoxP-Stop-LoxP (LSL) in pancreatic lineages during early embryonic development when the Ptf1a/p48 promoter is active. HSL deficient mice were kindly provided by Fredric Kraemer at Stanford University [10]. Hsl+/− mice were crossed into p48-Cre+/− and LSL-KrasG12D+/− mice to obtain p48-Cre+/−;Hsl+/− and LSL-KrasG12D+/−;Hsl+/− double mutants. The double mutant mice were crossed to generate the desired triple mutant genotypes: LSL-KrasG12D+/−;p48-Cre+/−;Hsl+/+ (KC;Hsl+/+), LSL-KrasG12D+/−;p48-Cre+/−;Hsl+/− (KC;Hsl+/−), and LSL-KrasG12D+/−;p48-Cre+/−;Hsl−/− (KC;Hsl−/−) (Figs. 1a, b). Mice were fed regular chow beginning at one month of age until 20 weeks of age. We analyzed 23 KC;Hsl+/+, 43 KC;Hsl+/−, and 20 KC;Hsl−/− mice. Animal studies were approved by the Chancellor’s Animal Research Committee of the University of California, Los Angeles (UCLA) in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals. All mice were sacrificed under general anesthesia with isoflurane. None of the animals died without euthanasia.

Fig. 1
figure 1

a Breeding scheme to generate KC;Hsl+/+, KC;Hsl+/−, and KC;Hsl−/− mice. b Exon structure of wildtype and HSL deficient mice. Replacing portions of exon 5 and the entire exon 6 with a neo cassette renders mice functionally HSL deficient [10]. Representative genotyping result showing a single 320 bp band in Hsl−/− mice, a single 269 bp band in Hsl+/+ mice, and a double band (320 bp and 269 bp) in Hsl+/− mice. c Weight curves of female and male KC;Hsl+/+ (n = 9 males+ 14 females), KC;Hsl+/− (n = 20 males+ 23 females), and KC;Hsl−/− (n = 10 males+ 10 females) mice fed standard rodent chow for 20 weeks. d Plasma levels of cholesterol, triglycerides, glucose, insulin, and leptin in KC;Hsl+/+ and KC;Hsl−/− mice at 6 months (at sacrifice). N = 6

Full size image

Genotyping analysis

The LSL-KrasG12D, p48-Cre, and Hsl alleles were genotyped by polymerase chain reaction (PCR) analysis as described elsewhere [10, 12].

Blood metabolic panel

Blood samples were collected from mice by intracardiac puncture at euthanasia. Blood chemistry (plasma cholesterol, glucose, and triglycerides levels) was obtained by the DLAM Pathology & Laboratory Medicine Services at UCLA. Levels of plasma insulin and leptin were determined using the MILLIPLEX MAP Mouse Adipokine Magnetic Bead Panel - Endocrine Multiplex Assay (EMD Millipore, Billerica, MA) based on the manufacturer’s instructions.

Adipose tissue inflammation

Formalin-fixed AT was paraffin-embedded and sectioned. Hematoxylin and eosin (H.E.)-stained sections were analyzed on a Nikon Eclipse 90i microscope (Nikon, Melville, NY) equipped with NIS AR4.2 software (Nikon). Crown-like structures (CLS) representing inflammatory foci [13, 14] were quantified and described as number per high-power field (hpf). A CLS was defined as one adipocyte surrounded by inflammatory cells at least partially. For each tissue sample, ten images were taken, and the number of CLS per 10 randomly selected adipocytes were quantified.

Cytokine array

Mesenteric fat homogenates from KC;Hsl+/+ and KC;Hsl−/− mice were profiled using the Mouse Cytokine Antibody Array, C1000 (RayBiotech, Norcross, GA) following the manufacturer’s instructions. The membrane-based proteomic array detects relative levels of 96 different cytokines and chemokines. The complete list of cytokines and chemokines analyzed can be found here: https://www.raybiotech.com/c-series-mouse-cytokine-array-c1000-2/. Following exposure to horseradish peroxidase (HRP) the membranes were imaged using the ChemiDoc™ Touch Imaging System (Bio-Rad Laboratories, Hercules, CA), and the intensity of signals normalized to the internal positive controls was quantified with Multi Gauge V3.0 software (Fujifilm Life Sciences, Tokyo, Japan).

Pancreas histology

Formalin-fixed, paraffin-embedded pancreatic tissue sections were stained with hematoxylin and eosin and histologically analyzed in a blinded fashion. Murine pancreatic intraepithelial neoplasia (PanINs) and invasive PDAC were classified according to histopathologic criteria as previously described [15,16,17].

Quantification of lipid content

Tissue content of triglycerides (TG), phospholipids (PL), and free fatty acids (FFA) and the fatty acid (FA) profile in each lipid fraction was analyzed by the Vanderbilt Mouse Metabolic Phenotyping Center. Briefly, phospholipids, diglycerides, triglycerides, and cholesteryl esters in the extracted lipids were separated by thin layer chromatography, scraped from the plates and methylated. The methylated FA were then extracted and analyzed on an Agilent 7890A gas chromatograph. FA methyl esters were identified by comparing the retention times to those of known standards. Inclusion of lipid standards with odd chain FAs allowed quantitation of lipids in the sample. Dipentadecanoyl phosphatidylcholine (C15:0), diheptadecanoin (C17:0), trieicosenoin (C20:1), and cholesteryl eicosenoate (C20:1) were the standards used.

Immunohistochemistry

Paraffin was removed with xylene and graded alcohol. Heat-induced antigen retrieval was performed with citrate buffer, and endogenous peroxidase activity was blocked with 3% hydrogen peroxide. Slides were then incubated overnight with polyclonal rabbit primary antibodies against HSL (Novus Biologicals, Littleton, CO) or monoclonal rabbit primary antibodies against Ki67 (Cell Signaling Technologies, Danvers, MA). Control images were prepared using isotype matched rabbit IgG (Cell Signaling Technologies). Images were analyzed on a Nikon Eclipse 90i microscope equipped with NIS AR4.2 software (Nikon). Ki67 staining was quantified by counting Ki67 positive cells (stromal and epithelial) in 10 high-power fields per tissue section.

Immunofluorescence

Paraffin was removed with xylene and graded alcohol. Antigen retrieval was performed by using Antigen Unmasking Solution (Vector Laboratories, Burlingame, CA) plus ethylenediaminetetraacetic acid (EDTA). Endogenous peroxidase activity was blocked with 1% hydrogen peroxide. Slides were then incubated overnight with monoclonal rabbit anti-F4/80 antibody (clone SP115, Novus Biologicals) and monoclonal mouse anti-TNF-α (tumor necrosis factor alpha) antibody (clone SPM543, Novus Biologicals). Anti-rabbit IgG antibodies conjugated with Alexa Fluor 555 (Thermo Fisher Scientific, Canoga Park, CA) and anti-mouse IgG antibodies conjugated with Alexa Fluor 488 (Thermo Fisher Scientific) were added at room temperature for one hour. Images were analyzed on a Nikon Eclipse 90i microscope equipped with NIS AR4.2 software (Nikon). F4/80 and TNF-α positive cells were counted in 10 hpf per tissue section and analyzed as 0 (no staining), + (1–5 cells/hpf), ++ (5–10 cells/hpf), and +++ (> 10 cells/hpf).

Western blot analysis

Mouse tissue samples were homogenized in radio-immunoprecipitation assay (RIPA) buffer containing mixture of protease and phosphatase inhibitors (Roche Applied Science, Basel, Switzerland). Tissue homogenates were resolved by SDS-PAGE, electrophoretically transferred onto nitrocellulose membranes, and then immunoblotted for the proteins of interest using the following primary antibodies: Peroxisome proliferator-activated receptor gamma (PPAR-γ) and p44/42 MAPK from Cell Signaling Technologies, and HSL from Novus Biologicals. After incubation with secondary antibodies, the immune-reactive bands detected with enhanced chemiluminescence reagents were imaged and analyzed by the ChemiDoc™ Touch Imaging System (Bio-Rad Laboratories).

PCR analysis of Hsl

Total RNA from tissue or cell lysates were extracted using RNA purification kits (Biomiga, Inc., San Diego, CA). Reverse transcription was performed with the iScript reverse transcription supermix (Bio-Rad Laboratories). The synthesized cDNA was used as template for the PCR analysis of Hsl gene expression. The iTaq™ Supermix (Bio-Rad Laboratories) was used for amplifications. All reactions were performed on the Bio-Rad iQ™5 system. Primers for mouse Hsl: forward, 5′- GCAGTGGTGTGTAACTAGGATTG-3′, and reverse, 5′- CGCTGAGGCTTTGATCTTGC -3′ (spanning exons 1 and 2).

Statistical analysis

Data are presented as mean ± SD. Statistical significance was determined by one-way (or two-way) ANOVA and two-tailed Student’s t-tests assuming unequal variances. For the comparison of PDAC incidence, a two-sided Fisher’s exact test was performed. Significance (p-value less than 0.05) was indicated with an asterisk (*).

Results

Weight gain and metabolic parameters in KC mice with HSL deficiency

After weaning female and male KC;Hsl+/+, KC;Hsl+/−, and KC;Hsl−/− mice were fed regular rodent chow for 20 weeks. Although females gained less weight than male mice, KC;Hsl+/+, KC;Hsl+/−, and KC;Hsl−/− mice within each gender group gained weight similarly throughout the study period (Fig. 1c), which is in accordance with previous reports [10, 18]. Compared to KC;Hsl+/+ mice, KC;Hsl−/− mice showed reduced plasma TG levels, while cholesterol levels were unchanged (Fig. 1d). Our studies demonstrated reduced plasma insulin levels in KC;Hsl−/− mice with normal glucose concentrations (Fig. 1d), suggesting enhanced insulin sensitivity. Furthermore, KC;Hsl−/− mice had reduced plasma levels of leptin (Fig. 1d), a hormone secreted primarily by adipocytes of white adipose tissue (WAT).

KC mice with HSL deficiency display enhanced AT inflammation

We have previously reported an enhanced depot-specific AT inflammation during diet-induced obesity (DIO) in KC mice [14]. In the present study, histological examination revealed increased WAT (mesenteric depot) inflammation in KC;Hsl−/− mice as assessed by quantification of CLS (Fig. 2a). The increase in CLS was accompanied by an enhanced presence of TNF-α expressing F4/80-positive cells (macrophages) in the mesenteric WAT of KC;Hsl−/− mice as assessed by immunofluorescence (Fig. 2b). Detailed tissue lipid analysis revealed decreased TG content in mesenteric WAT of KC;Hsl−/− mice as well as reduced expression of pro-adipogenic PPAR-γ (Fig. 2c), confirming the impairment of adipogenesis in HSL null mice [18]. The anti-adipogenic phenotype associated with HSL deficiency is consistent with the role of HSL in providing intrinsic ligands for PPARγ through release of FAs [19]. The PL content in mesenteric WAT was unchanged in KC;Hsl−/− mice, reflecting the lack of phospholipase activity of HSL. Despite the physiological role of HSL in FA release, the levels of FFA in the mesenteric WAT of KC;Hsl−/− mice were unaltered (Fig. 2c). Due to limited amount of tissue available for each analysis, we did not measure the intracellular levels of diacylglycerols, which are known to accumulate in tissues of Hsl−/− mice due to the importance and preference of HSL to hydrolyze diacylglycerols [20]. Using a Mouse Cytokine Antibody Array, multiple changes in cytokines and chemokines were detected in the mesenteric WAT of KC;Hsl−/− mice (Fig. 2d). Significant elevations of adhesion molecules (e.g. ICAM-1, L-selectin) and chemokines (e.g. MIP-1 gamma, CCL5, CCL22) were found (Fig. 2d) that may reflect AT inflammation with recruitment and infiltration of immune cells into the mesenteric WAT.

Fig. 2
figure 2

a Representative histology (H.E. staining) of the mesenteric WAT of KC;Hsl+/+ and KC;Hsl−/− mice. Scale bar represents 100 μm. Quantification of crown-like structures (CLS) in KC;Hsl+/+ and KC;Hsl−/− mice (right). *: p < 0.01. b Immunofluorescence staining of mesenteric WAT of KC;Hsl+/+ and KC;Hsl−/− mice (two representative mice in each group). Red and green staining denote F4/80 (macrophage marker) and TNF-α (as a marker of tissue inflammation), respectively. Scale bar represents 100 μm. Semi-quantitative analysis of TNF-α expressing macrophages (orange staining) in KC;Hsl+/+ and KC;Hsl−/− mice (table below). c Quantification of PL, TG, and FFA content in mesenteric WAT of KC;Hsl+/+ and KC;Hsl−/− mice (n = 5). Western blot analysis of HSL and PPAR-γ (total ERK as loading control) in KC;Hsl+/+ and KC;Hsl−/− mice (lower left). d Changes (≥ 50% increase or decrease compared to KC;Hsl+/+) in several cytokines and chemokines in the mesenteric WAT of KC;Hsl−/− mice as detected by a Mouse Cytokine Antibody Array (n = 3 biological replicates). Representative array blots (Cytokine Array 3 and 4) in KC;Hsl+/+ and KC;Hsl−/− mice (right)

Full size image

KC mice with HSL deficiency have an increased incidence of PDAC

Consistent with previous reports [21], HSL was expressed in pancreatic islets of wild type (WT) mice (Fig. 3a). Interestingly, immunohistochemistry showed strong HSL expression in PanIN lesions of KC;Hsl+/+ and KC;Hsl−/− mice (Fig. 3a), which has never been described before. Expression of HSL in PanIN cells was confirmed by PCR analysis of primary PanIN cells isolated from KC mice (with functional HSL) and cultured in vitro (Fig. 3b). Despite the functional HSL deficiency in KC;Hsl−/− mice (by virtue of replacing portions of exon 5 and the entire exon 6 with a neo cassette), the antibodies used in the present study seem to be able to detect the non-functional HSL protein in KC;hsl−/− mice. It is unclear whether the apparent lower expression of HSL in KC;Hsl−/− WAT and pancreatic lysates (Fig. 2c and 3b) is caused by an actual decreased expression or by a lower affinity of the antibody to the non-functional HSL protein. Detailed pancreatic tissue lipid analysis demonstrated markedly reduced TG and FFA content in KC;Hsl−/− mice (Fig. 3c). Again, the PL content was unaltered in the pancreas of KC;Hsl−/− mice (Fig. 3c).

Fig. 3
figure 3

a Representative HSL immunohistochemistry of the pancreas in wildtype, KC;Hsl+/+, and KC;Hsl−/− mice showing HSL positivity in pancreatic islets and PanIN cells. Scale bar represents 100 μm. b HSL western blot of total pancreatic lysates in KC;Hsl+/+ and KC;Hsl−/− mice (GAPDH as loading control). PCR analysis of Hsl transcripts in differentiated 3 T3-L1 (adipocytes), murine PanIN (mPanIN) and murine PDAC cells (both isolated from KC mice), murine pancreas and liver (lower). c Quantification of PL, TG, and FFA content in the pancreas of KC;Hsl+/+ and KC;Hsl−/− mice (n = 5). *: p < 0.01

Full size image

The most salient feature of this study is that KC;Hsl−/− mice displayed a significantly increased PDAC incidence at 20 weeks (Fig. 4a). While none of the KC;Hsl+/+ mice developed invasive PDAC at six months of age, 25% (5/20) of KC;Hsl−/− mice had cancer (p = 0.016). The increased PDAC incidence in KC;Hsl−/− mice was thereby associated with an elevated proliferation of PanIN and stromal cells (Fig. 4b) as assessed by Ki67 immunostaining. Furthermore, KC;Hsl−/− mice displayed enhanced pancreatic inflammation as evident histologically and by an increase in TNF-α expressing F4/80-positive macrophages assessed by immunofluorescence (Fig. 4c).

Fig. 4
figure 4

a Representative histology (H.E. staining) of the pancreas of Hsl−/−, KC;Hsl+/+, KC;Hsl+/−, and KC;Hsl−/− mice. Scale bar represents 100 μm. No pancreatic neoplastic lesions were detected in Hsl−/− mice. Analysis of PDAC incidence at 20 weeks (sacrifice) in KC;Hsl+/+, KC;Hsl+/−, and KC;Hsl−/− mice (table right). b Ki67 immunohistochemistry of the pancreas of KC;Hsl+/+ and KC;Hsl−/− mice. Scale bar represents 100 μm. Quantification of Ki67 immunoreactivity in epithelial (ductal, indicated by solid arrows) and stromal cells (indicated by hollow arrows). N = 6, *: p < 0.01. c Immunofluorescence staining of the pancreas of KC;Hsl+/+ and KC;Hsl−/− mice (two representative mice in each group). Scale bar represents 100 μm. Red and green staining denote F4/80 (macrophage marker) and TNF-α (as a marker of tissue inflammation), respectively. Semi-quantitative analysis of TNF-α expressing macrophages (orange staining) in KC;Hsl+/+ and KC;Hsl−/− mice (table below)

Full size image

Patients with PDAC and low expression of LIPE have unfavorable prognosis

To assess the significance of HSL expression in human PDAC, we determined the importance of the expression of the gene encoding HSL as a prognostic marker of survival in patients with PDAC. We used a recently published interactive open-access database (www.proteinatlas.org/pathology) to perform correlation analyses based on mRNA expression levels of LIPE (the gene encoding HSL) in PDAC tissue and the clinical outcome (survival) of the patients. The data in the Pathology Atlas is based on the analysis of transcriptomics and survival in 176 PDAC patients. As illustrated in the Kaplan-Meier plot in Fig. 5, none of the patients of the population with lower levels of LIPE mRNA expression (n = 69) survived for 5 years while 42% of the population (n = 107) with the higher levels of LIPE mRNA survived for 5 years or more.

Fig. 5
figure 5

Kaplan-Meier plot for gene expression of LIPE in PDAC. Image was reproduced from the Human Protein Atlas (version 17) available from www.proteinatlas.org. The link to LIPE is as follows: http://www.proteinatlas.org/ENSG00000079435-LIPE/pathology/tissue/pancreatic+cancer

Full size image

Discussion

A striking feature of the results presented here is that KC mice lacking HSL (KC;Hsl−/−) mice displayed a significant increase in PDAC incidence. To our knowledge, it is the first time that HSL deficiency has been linked to an increased cancer risk. In humans, carriers of a frameshift deletion of exon 9 in the LIPE gene, encoding for HSL, were characterized by metabolic dysfunction, including dyslipidemia, hepatic steatosis, systemic insulin resistance, and diabetes [22]. In the adipose tissue from carriers with the mutation, impaired lipolysis and inflammation were observed [22]. In our study, the exact mechanism(s) involved in promoting PDAC remains incompletely understood. HSL deficiency in KC mice was accompanied by enhanced inflammation in the AT and pancreas. While CLS is a well-characterized feature of adipose tissue inflammation (macrophages surrounding necrotic adipocytes), additional analysis is warranted to further identify inflammatory cell subpopulations. It has been postulated that an increased heterogeneity of adipocytes with necrotic cell death of hypertrophic adipocytes and subsequent infiltration of macrophages may play an important role in inducing WAT inflammation in the context of HSL deficiency [10, 23]. Lipid analysis of the mesenteric WAT in KC;Hsl−/− mice has revealed a decreased TG content, in agreement with previous reports [24]. The mechanistic link between HSL deficiency and TG reduction in WAT was suggested to be a compensatory downregulation of FA esterification enzymes, leading to reduced cellular TG synthesis [25]. This observation is consistent with the reported impairment of white adipocyte differentiation and decreased WAT mass in HSL null mice [18]. The anti-adipogenic phenotype may also be mediated by a decrease in PPAR-γ expression and activity through the reduced intracellular release of FAs, acting as endogenous PPAR-γ ligands, in HSL null mice [19]. This is corroborated by our finding of reduced PPAR-γ expression in the mesenteric WAT of KC;Hsl−/− mice. FFA levels in the mesenteric WAT were not significantly altered in KC;Hsl−/− mice, which can be explained by the presence of additional lipases in the WAT, i.e. adipose triglyceride lipase, which preferentially catalyzes the conversion of triacylglycerol to diacylglycerol and thereby maintaining a FFA pool in WAT [24].

In addition to AT inflammation, KC;Hsl−/− mice also had enhanced pancreatic inflammation as demonstrated histologically and by an increased number of TNF-α producing macrophages in the pancreas. It is currently unclear what mechanisms elicited the pancreatic inflammation seen in KC;Hsl−/− mice. It is possible that HSL deficiency causes enhanced inflammation primarily in the visceral WAT adjacent to the pancreas with an increased production and secretion of pro-inflammatory cytokines, which could elicit a subsequent inflammatory reaction in the pancreas of KC mice. This is supported by our finding that HSL deficiency failed to cause pancreatic inflammation in WT mice. It is conceivable that the robust visceral WAT inflammation in KC;Hsl−/− mice reinforces and amplifies the oncogenic and inflammatory signaling in KrasG12D harboring pancreatic cells, thereby leading to substantial pancreatic inflammation and accelerated PDAC development.

However, we cannot rule out an important role of pancreatic HSL in mediating pancreatic inflammation and tumorigenesis. Indeed, we detected strong HSL expression in pancreatic islets and PanIN lesions and reduced expression of LIPE (the gene encoding HSL) in pancreatic tissue of patients with PDAC is associated with decreased overall survival [26]. HSL deficiency in pancreatic islets may thereby explain the observed reduction in insulin of KC;Hsl−/− mice in our study, as HSL has been described to be important in glucose-stimulated insulin secretion in pancreatic beta cells [27]. HSL deficiency in the pancreas was accompanied by reduced TG and FFA levels, indicating a prominent role of HSL (and a possible absence of additional neutral lipases) in FFA release from intracellular TG depots in the pancreas. It is plausible that in the presence of oncogenic KRAS these lipid changes locally in the pancreas in KC;Hsl−/− mice are driving promotional factors of pancreatic inflammation and PDAC development. These possibilities are not mutually exclusive and the lack of HSL in both visceral WAT and in the pancreas of KC could cooperate in accelerating PDAC development.

Conclusion

Our data demonstrated that HSL deficiency in KC mice leads to inflammation of mesenteric WAT and the pancreas and importantly accelerated PDAC formation. Collectively, our results revealed an unexpected tumor-suppressive role of HSL and emphasize the need of caution in targeting HSL for tumor cachexia or dyslipidemia, as chronic suppression of this enzyme may lead to increased incidence of PDAC.