Toll Like Receptor 2, 4, and 9 Signaling Promotes Autoregulative Tumor Cell Growth and VEGF/PDGF Expression in Human Pancreatic Cancer

Toll like receptor (TLR) signaling has been suggested to play an important role in the inflammatory microenvironment of solid tumors and through this inflammation-mediated tumor growth. Here, we studied the role of tumor cells in their process of self-maintaining TLR expression independent of inflammatory cells and cytokine milieu for autoregulative tumor growth signaling in pancreatic cancer. We analyzed the expression of TLR2, -4, and -9 in primary human cancers and their impact on tumor growth via induced activation in several established pancreatic cancers. TLR-stimulated pancreatic cancer cells were specifically investigated for activated signaling pathways of VEGF/PDGF and anti-apoptotic Bcl-xL expression as well as tumor cell growth. The primary pancreatic cancers and cell lines expressed TLR2, -4, and -9. TLR-specific stimulation resulted in activated MAP-kinase signaling, most likely via autoregulative stimulation of demonstrated TLR-induced VEGF and PDGF expression. Moreover, TLR activation prompted the expression of Bcl-xL and has been demonstrated for the first time to induce tumor cell proliferation in pancreatic cancer. These findings strongly suggest that pancreatic cancer cells use specific Toll like receptor signaling to promote tumor cell proliferation and emphasize the particular role of TLR2, -4, and -9 in this autoregulative process of tumor cell activation and proliferation in pancreatic cancer.


Introduction
Pancreatic ductal adenocarcinoma (PDAC) is the most lethal type of digestive cancer with a five-year survival rate of about 7%. In contrast to increasing survival rates for patients with various other cancer types, therapeutic improvements over the last four decades in pancreatic cancer raised the chances for recovery only about 3% [1]. Besides advanced age, smoking, obesity and diabetes, chronic pancreatitis is a major risk factor for the development of PDAC [2]. During chronic pancreatitis, the pancreas undergoes damaging events promoted through inflammation processes which result in the release of damage-associated molecular patterns (DAMPs) as well as growth factors like vascular endothelial growth factor (VEGF) during subsequent healing [3,4].
In addition to a panel of pathogen-associated molecular patterns (PAMPs) including cell-wall components like lipopolysaccharide (LPS) as well as microbial DNA and RNA, DAMPs which arise from inflammation and cellular injury can stimulate TLRs and consequently induce TLR signaling that supports an inflammatory microenvironment [5,6]. Recently, enhanced expression of TLRs has been described in a variety of different tumor entities and depending on the cancer type, could be linked to either favorable or poor prognosis [7][8][9][10][11]. TLR3 and -4 were identified as predictors of poor survival in breast cancer, whereas high TLR9 predicted enhanced survival [12,13]. In esophageal cancer TLR3, -4, -7, and -9 overexpression has been demonstrated in 70%, 72%, 67%, and 78% of patient tumors, respectively, and, as in breast cancer, poor prognosis was associated with elevated TLR3 expression on tumor cells, however, patients expressing increased TLR9 associated with fibroblasts exhibited improved survival [12][13][14]. Enhanced colorectal tumor expression of TLR7/8 colocalized with the cancer stem cell marker CD133 and correlated with reduced overall survival [7,13]. In one report, enhanced TLR4 expression was identified in 69% of patients with pancreatic cancer and correlated with increased NF-κB signaling, enhanced hypoxia-inducible factor-1α (HIF-1α) expression, and dramatically reduced patient survival [13,15].
In controversy to studies examining the anti-tumor effects of TLR signaling, several tumor entities, particularly those linked to inflammatory processes, have demonstrated unfavorable effects of TLR activation. TLR7 ligands are currently in use for the treatment of various cancer types such as melanoma, breast cancer and basal cell carcinoma [16]. Moreover, TLR3 and -9 activation has been demonstrated to reduce tumor growth in renal cell carcinoma and metastatic colorectal cancer [16,17] Besides those data that demonstrate favorable effects of TLR activation it is known that TLR ligands can also promote cancer cell survival, migration and tumor progression. For example, TLR agonists have been shown to induce tumor cell viability and metastasis in human lung cancer [18]. Activation of TLR3 resulted in increased proliferation in human myeloma cells [19]. Additionally, TLR ligation is related to adhesion and metastasis in human colorectal cancer cells and migration in human glioblastoma (via TLR4) and human breast cancer cells (via TLR2) [18,20,21]. Previously, we showed that TLR7 or -8 expression is associated with UICC stage in PDAC and stimulation increases tumor cell proliferation and resistance to the cytostatic agent 5-fluorouracil (5-FU) in pancreatic cancer cells [8].
Next to TLRs, it is well-known that the expression of growth factors such as VEGF or platelet-derived growth factor (PDGF) is increased in various tumors. Additionally, growth factor expression is associated with augmented angiogenesis and tumor cell proliferation and therefore linked to poor outcome [22][23][24].
In this study, we analyzed the expression of TLR2, -4, and -9 in pancreatic cancer. Besides their pathogen-associated ligands (TLR2: multiple lipoproteins and lipoteichoic acid; TLR4: LPS; and TLR9: unmethylated CpG oligodeoxynucleotide DNA), these TLRs are also known to be induced by several endogenous ligands such as heat shock proteins, fibrinogen, hyaluronic acid fragments and high-mobility group box 1 (HMGB1) [25]. These ligands arise from damaging events promoted through inflammatory processes [3]. Consequently, TLR2, -4, and -9 are highly suspected to play a role in inflammation linked cancers such as pancreatic cancer.
To examine the impact of TLR2, -4, and -9 expression and activation in pancreatic cancer cells, effects of TLR agonists CpG oligonucleotides 2006 (ODN, TLR9 specific), lipoteichnic acid of Staphylococcus aureus (LTA, TLR2 specific), lipopolysaccharide (LPS, TLR4 specific), and HMGB1 (non-specific) on growth factor expression, tumor cell signaling and cancer proliferation were analyzed to elucidate the potential of TLR signaling as a target for therapeutic strategies in PDAC.

TLR2, -4, and -9 Are Expressed in Human Pancreatic Cancer Tissue
Western blot analysis of pancreatic tissue showed no protein expression of TLR2, -4, and -9 in normal pancreatic tissue (NT) compared to increased expression in tissue of chronic pancreatitis (CP) and in particular in primary pancreatic cancer at all stages (UICC I, IIA, IIB, III, and IV) ( Figure 1A).

TLR2, -4, and -9 Are Expressed in Human Pancreatic Cancer Tissue
Western blot analysis of pancreatic tissue showed no protein expression of TLR2, -4, and -9 in normal pancreatic tissue (NT) compared to increased expression in tissue of chronic pancreatitis (CP) and in particular in primary pancreatic cancer at all stages (UICC I, IIA, IIB, III, and IV) ( Figure 1A).  (UICC I, IIA, IIB, III, IV). β-actin probe was used as a control for protein loading; (B) RT-qPCR of normal pancreatic tissue (NT, n = 4), tissue from chronic pancreatitis (CP, n = 4), and primary pancreatic tumor tissue at UICC stages II and III (n = 14). Values for normal pancreatic tissue were standardized to baseline. The relative gene expression is expressed as 2 −ΔΔCq (fold difference, FD). * p < 0.05, ** p < 0.005.

TLR2, -4, and -9 Are Expressed in Human Pancreatic Cancer Tissue
Western blot analysis of pancreatic tissue showed no protein expression of TLR2, -4, and -9 in normal pancreatic tissue (NT) compared to increased expression in tissue of chronic pancreatitis (CP) and in particular in primary pancreatic cancer at all stages (UICC I, IIA, IIB, III, and IV) ( Figure 1A).  (UICC I, IIA, IIB, III, IV). β-actin probe was used as a control for protein loading; (B) RT-qPCR of normal pancreatic tissue (NT, n = 4), tissue from chronic pancreatitis (CP, n = 4), and primary pancreatic tumor tissue at UICC stages II and III (n = 14). Values for normal pancreatic tissue were standardized to baseline. The relative gene expression is expressed as 2 −ΔΔCq (fold difference, FD). * p < 0.05, ** p < 0.005.
In RT-qPCR, elevated relative gene expression of TLR2 (fold difference, FD = 29.8, p < 0.05), TLR4 (FD = 39.9, p < 0.005), and TLR9 (FD = 10.3, p < 0.005) was observed in pancreatic tumor tissues compared to normal tissues ( Figure 1B). Additionally, TLR2 and -4 gene expression was not significantly increased in tissue of chronic pancreatitis compared to normal tissue (TLR2 FD = 2.0 and TLR4 FD = 2.2, respectively) ( Figure 1B).  To substantiate that elevated TLR expression found in ex vivo pancreatic cancer tissue by Western blot and RT-qPCR is associated with pancreatic cancer cells rather than tumor infiltrating cells of the immune system, immunofluorescence double staining of cryo sections was performed. Co-staining of TLR2, -4, and -9 with the epithelial marker EpCAM clearly indicated TLR expressing tumor cells in primary tumor tissue of all UICC stages (data not shown). In Figure 2 representative specimens for TLR2, -4, and -9 staining in pancreatic tumor tissues at UICC stage II are demonstrated and examples for TLR and EpCAM co-expressing cells are marked with white arrows.

TLR2, -4, and -9 Are Expressed in Human Pancreatic Cancer Cell Lines
Expression of TLR2, -4, and -9 was analyzed by RT-qPCR and Western blot in five established human pancreatic cancer cell lines (Panc1, MIAPaCa-2, BxPC-3, AsPC-1, and SW1990) as well as in three primary human pancreatic cancer cell lines (PaCaDD135, PaCaDD159, and PaCaDD185). TLR mRNA was detected in all investigated cell lines indicating constitutive expression of TLR2, -4, and -9 in unstimulated human pancreatic cancer cells. To allow for comparison of RT-qPCR results, cell lines with the lowest expression were standardized to baseline (fold difference, FD = 1).

TLR2, -4, and -9 Are Expressed in Human Pancreatic Cancer Cell Lines
Expression of TLR2, -4, and -9 was analyzed by RT-qPCR and Western blot in five established human pancreatic cancer cell lines (Panc1, MIAPaCa-2, BxPC-3, AsPC-1, and SW1990) as well as in three primary human pancreatic cancer cell lines (PaCaDD135, PaCaDD159, and PaCaDD185). TLR mRNA was detected in all investigated cell lines indicating constitutive expression of TLR2, -4, and -9 in unstimulated human pancreatic cancer cells. To allow for comparison of RT-qPCR results, cell lines with the lowest expression were standardized to baseline (fold difference, FD = 1).

Discussion
Recently, we reported about the role of the two Toll like receptors TLR7 and -8 in solid tumor growth. Expression of the two intracellularly expressed receptors is up-regulated in primary tumors from patients with colon as well as pancreatic cancer. Interestingly, in both human cancers TLR7 and -8 expression was related rather to cancer cells and rarely detected in tumor-infiltrating immune cells. Moreover, our results indicated that TLR7 and -8 expression is associated with tumor cell growth and tumor progression in both colon and pancreatic cancer [7,8]. Here, we analyzed whether other Toll like receptors such as TLR2, -4, and -9 are additionally expressed in pancreatic cancer and may influence tumor cell signaling and proliferation to elucidate their potential for therapeutic strategies in this devastating tumor.
We demonstrated that all three receptors TLR2, -4, and -9 are constitutively expressed in the analyzed established and primary pancreatic cancer cell lines. Moreover, TLR expression in tumor cells was also detected in ex vivo tissue samples of pancreatic cancer. Consistently, in several solid human cancers, TLR expression has been assigned not only to tumor infiltrating cells of the immune system, but rather to tumor cells themselves. Interestingly, more recent data point to tumor cell located TLR2 and -9 as effective markers to distinguish cancerous from normal epithelial cells in human pancreatic cancer [16]. Our data support these observations. While primary pancreatic cancers at all UICC stages demonstrated substantial protein and gene expression of TLR2 and -9, no protein expression and only minor gene expression was observed in normal pancreatic tissue, compared to the moderate protein expression in controls with chronic pancreatitis, supporting their role in the inflammatory response. However, to this point it needs to be further analyzed whether TLR expression can be assigned not only to tumor infiltrating immune cells but significantly to tumor cells and secondly what is the functional role for this tumor cell-mediated TLR expression.
Additionally, we extended current knowledge about the expression of TLR2, -4, and -9 in the various established to non-commercially available primary human pancreatic cancer cell lines. To dissect the effects of TLR activation, we treated two of the established pancreatic cancer cell lines (BxPC-3 and MIAPaCa-2) and one primary pancreatic cancer cell line (PaCaDD135) with TLR specific ligands (LTA for TLR2, LPS for TLR4, ODN for TLR9) and a non-specific ligand (HMGB1). HMGB1 is a damage-associated inflammatory factor that plays an important role in the pathogenesis of numerous chronic inflammatory diseases [26]. It arises from inflammation and subsequent cellular injury as so called damage-associated molecular pattern (DAMP). This implies, that even in the absence of pathogens, disrupted or injured cells can create TLR ligands such as HMGB1 or nucleic acids that may activate TLR signaling when released outside the cell following tissue injury [6].
Interestingly, stimulation of TLR9 and in part TLR2 resulted in upregulated VEGF gene expression (BxPC-3) and particularly of PDGF in all three analyzed human pancreatic cancer cell lines. Moreover, significantly increased protein expression of both growth factors VEGF and PDGF was observed in all investigated cancer cells. VEGF works as a key regulator of angiogenesis and lymphangiogenesis and additionally mediates tumor cell proliferation, migration, differentiation, and cancer cell survival [27][28][29][30]. Next to VEGF, PDGF has more recently been found to be involved in migration and angiogenesis [31]. Additionally, it supports tumor cell proliferation and metabolic alterations as demonstrated in previous studies [32]. Activation of receptor tyrosine kinases, such as the PDGF receptor (PDGFR) and VEGF receptor (VEGFR), has also been implicated in tumor progression and metastasis in human pancreatic cancer [33]. VEGFR expression has been observed in about 50% of human pancreatic cancer cells [34]. Concerning PDGF receptors, high PDGFRb expression was found to be correlated with poor disease-free survival in pancreatic cancer and additionally colon and ovarian cancer patients [35]. Overall, growth factors and their receptors have been shown to result in progression and metastasis formation of pancreatic cancer and engagement of factors such as VEGF and PDGF and their signaling cascades appear to be crucial for the progression of this devastating disease [36].
After ligation with their specific receptors, both growth factors VEGF and PDGF induce subsequent receptor phosphorylation and activation of downstream signaling cascades including the PI3K/Akt/mTOR and MAPK pathway [28,31]. In our functional analysis, we observed increased phosphorylation of Erk after treatment with all TLR agonists for TLR2, -4, and -9 (except only one cancer cell line for TLR2 activation), strongly indicating activation of the MAPK signaling pathway. Moreover, upregulated phosphorylation of Akt after TLR activation, in particular TLR4, also points to an engagement of PI3K/Akt/mTOR signaling.
Notably, stimulation of all analyzed TLRs resulted in significantly increased tumor cell proliferation in all the herein investigated human pancreatic cancer cell lines. Particularly activation of TLR9 demonstrated pronounced proliferation promoting effects. Whether TLR-mediated VEGF and PDGF expression from tumor cells may lead to an autoregulative self-induction of VEGF and PDGF receptors on pancreatic cancer cells and thus acts as a feedback loop inducing tumor cell proliferation remains so far hypothetic. However, it is supported by our observed effects on both PI3K/mTOR and MAPK signaling, since both pathways are downstream signaling cascades of VEGFR and PDGFR.
In line with the proliferation promoting effects, we also demonstrated TLR-induced increase of anti-apoptotic Bcl-xL. In particular MIAPaCa-2 and PaCaDD135 pancreatic cancer cells showed considerably elevated TLR ligand-mediated Bcl-xL expression, indicating TLR signaling to induce cellular survival pathways in pancreatic cancer [37]. As described by others, a direct link between Erk activation and induction of Bcl-xL has been proposed [38]. These findings support the hypothesis that autoregulative expression of VEGF and PDGF receptors on pancreatic cancer cells may create a feedback loop to induce tumor cell proliferation and survival.
To date, a considerable number of studies suggest that TLR activation may reduce tumor growth and therefore has beneficial effects. One of the main mechanisms that explain the antitumor activity of TLR signaling is the induction of a tumor-specific immune response. TLR activation provokes the infiltration of natural killer (NK) cells, cytotoxic T cells and type I T helper cells into the tumor, where they induce tumor cell lysis via secretion of e.g., perforin and IFN-γ (interferon gamma) which results in the release of type I IFNs (IFN-α, β) [39][40][41]. Consequently, several TLR agonists are currently in use or analyzed in clinical trials for their potential in tumor treatment. In chronic lymphatic leukemia and basal cell carcinoma, TLR7 and -8 agonists of both natural (e.g., single stranded RNA, ssRNA) and synthetic (e.g., Imiquimod) origin have demonstrated promising antitumor activity. Additionally, the TLR9 ligand CpG has been shown to suppress cancer growth in lymphomas and tumors of the brain, kidney and skin. Induction of TLR3 by its ligand poly(IC) has also been described to reduce viability in tumor cells of renal cell carcinoma and metastatic colorectal cancer [16,39,[42][43][44].
Despite the numerous reports that demonstrate favorable effects of TLR stimulation, the association of about 15% of human tumors with inflammatory processes and chronic inflammation remains puzzling [17]. e.g., the association of stomach cancer with chronic inflammation induced by Helicobacter pylori is well described. Additionally, chronic inflammation of the digestive tract such as inflammatory bowel disease or chronic pancreatitis can be linked to an increased risk for colorectal or pancreatic cancer [2,39,45]. TLR signaling is known to play a major role in processes of inflammatory response and a large number of studies points to the implication of TLR activation in tumor formation and development, angiogenesis, and tumor growth as well as chemoresistance and immune evasion by the induction of regulatory T cells (Treg) [39,46]. For example, TLR ligands demonstrated anti-apoptotic and metastatic effects in human lung cancer cells and induced proliferation in human myeloma cells (TLR3). Adhesion and metastasis in human colon cancer has been linked to TLR4 activation as well as migration in glioblastoma. TLR2 signaling has been demonstrated to induce migration in human breast cancer cells) and ligation of TLR7 and -8 has been associated with proliferation and chemoresistance in pancreatic cancer (TLR7 and -8) [8,[18][19][20][21]. In the here presented study, we clearly demonstrated for the first time that ligation of TLR2, -4, and -9 results in increased tumor cell proliferation in pancreatic cancer.
Based on these observations, we conclude that inflammation-mediated cancer cell proliferation, tumor progression, and metastatic potential are closely linked to specific TLR expression and signaling in pancreatic cancer. Therefore, targeting of TLR signaling represents a key mechanism to reduce tumor growth, resistance to apoptosis and growth factor induced tumor cell proliferation. As TLR activation can be advantageous for the proliferation, invasiveness, and/or survival of tumor cells, the effects of TLR agonists on tumor cells besides tumor infiltrating immune cells depend on the tumor type, and need to be carefully taken in account in preclinical studies. Based on the current data, particularly in inflammation-mediated tumor growth and progression as is the case for pancreatic cancer, patients may finally benefit from TLR blockade rather than activation.

Patients and Human Tissues
In a retrospective analysis, 20 patients with histologically confirmed pancreatic cancer (adenocarcinoma) of the exocrine pancreas and available representative tumor tissues as well as completed 5-year follow up (January 2007 until December 2009) and 12 patients with chronic pancreatitis were evaluated. Patients were followed up in our Comprehensive Cancer Center (completeness index 0.96). The study was conducted in accordance with the Declaration of Helsinki. Ethical approval was obtained from the Human Research Ethics Committee of the University of Wuerzburg. All patients providing pancreatic cancer samples signed a consent form to allow for this research. The histological stage of the pancreatic cancer specimen was determined according to the UICC (Union Internationale Contre le Cancer) staging system. After acquisition, Tumor localization, stage, and differentiation grade was evaluated in our Department of Pathology. Tumor samples were snap-frozen in liquid nitrogen instantly upon surgical removal. For cryostat-sections, tissue samples were embedded in Tissue Tek (Sakura, Torrence, CA, USA) using suitable cryo molds. Samples were transferred to −80 • C until analyzed.
In our studies, we compared tissue samples of UICC stage I/II (UICC I n = 2, UICC II n = 8) and UICC stage III/IV (UICC III n = 7, UICC IV n = 3), specimens from individuals operated on histologically confirmed chronic pancreatitis (n = 8) and normal tissue of healthy controls (n = 8).

Immunofluorescence Staining
The TLR2 antibody was purchased from Novus Biologicals (Littelton, CO, USA), TLR4 and TLR9 antibodies were provided by abcam (Cambridge, UK), and the antibody for EpCAM was obtained from Biorbyt (Riverside, UK). Secondary antibodies were indocarbocyanin (Cy3)-conjugated anti-rabbit and AlexaFluor488-conjugated anti-mouse IgG (both Jackson ImmunoResearch, West Grove, PA, USA). The staining was performed on serial cryostat sections of the snap-frozen specimens of pancreatic cancers and cytospin preparations of human pancreatic cancer cell lines. For immune staining procedures samples were fixed in acetone and incubated with the primary antibody in Tris-buffered saline (TBS) plus 0.5% (m/v) bovine serum albumin (BSA) overnight at 4 • C in a humidified chamber. Treatment with secondary fluorochrome conjugated antibody was performed for 30 min at room temperature in a humidified chamber. In case of double staining, slides were incubated with the second primary antibody diluted in TBS plus 0.5% (m/v) BSA overnight at 4 • C in a humidified chamber followed by incubation with secondary fluorochrome conjugated antibody for 30 min at room temperature in a humidified chamber. Subsequently, slides were mounted with 4 ,6-diamidin-2-phenylindol (DAPI) Fluoromount-G (Southern Biotech, Birmingham, AL, USA) and analyzed using an Olympus BX51 microscope and the CellSens Dimension software (Olympus, Hamburg, Germany).

Quantitative Real Time RT-qPCR
Gene expression for TLR2, -4, and -9 as well as VEGF and PDGF was determined using quantitative real-time PCR (RT-qPCR). For analysis of pancreatic tissue samples human pancreatic matched cDNA was purchased from Pharmingen (Heidelberg, Germany) and used as a control. Total cellular RNA was extracted using RNeasy Minikit (Qiagen, Hilden, Germany) on the QIAcube platform (Qiagen) according to the manufacturer's instructions. Reverse transcription was performed using ImProm-II reverse transcriptase system (Promega, Mannheim, Germany) and carried out on a Eppendorf Mastercycler (Eppendorf, Hamburg, Germany). qPCR was performed using Taqman Gene Expression Master Mix (Life Technologies, Carlsbad, CA, USA) and Taqman Gene Expression Assays (Life Technologies) in concordance to the manufacturer's instructions and carried out in duplicates on a Biorad CFX96 Touch Real-Time PCR Detection System. Reproducibility was assured by three independent measurements. Housekeeping genes β-actin, glycerinaldehyd-3-phosphat-dehydrogenase (GAPDH), and β2-microglobulin (B2M) were used for calculation of the relative quantification value, fold difference, which is expressed as 2 −∆∆Cq .

Western Blot
Protein extraction from tissue samples and pancreatic cancer cell lines was performed using RIPA buffer containing dithiothreitol (DTT) and protease/phosphatase inhibitor cocktails (Merck Millipore, Billerica, MA, USA). Tumor and normal tissues were cut in small pieces and homogenized for 10 min in prepared RIPA buffer using TissueLyser (Qiagen) before centrifugation (full speed, 4 • C, 20 min). The supernatant was collected and stored at −80 • C.
For preparation of cell lysates adherent cells were detached using accutase solution, washed with Dulbecco's phosphate buffered saline (DPBS, Thermo Fisher Scientific, Waltham, MA, USA) and pelleted at 300× g for 10 min. After resuspension in prepared RIPA buffer, cells were incubated on a rotator at 4 • C for 10 min, then centrifuged (full speed, 4 • C, 20 min) and supernatant was collected and transferred to −80 • C. To determin the protein concentration of the prepared lysates Bradford assay was performed using Roti-Quant solution (Carl Roth, Karlsruhe, Germany). For gel electrophoresis, NuPAGE SDS Buffer and NuPAGE Novex Mini Gels (Thermo Fisher Scientific) were used and handled according to the manufacturer's instructions. Western blotting on nitrocellulose was