Pancreatic ductal adenocarcinoma (PDAC) exists in a complex desmoplastic microenvironment, which includes cancer-associated fibroblasts [also known as pancreatic stellate cells (PSC)] and immune cells that provide a fibrotic niche that impedes successful cancer therapy. We have found that mast cells are essential for PDAC tumorigenesis. Whether mast cells contribute to the growth of PDAC and/or PSCs is unknown. Here, we tested the hypothesis that mast cells contribute to the growth of PSCs and tumor cells, thus contributing to PDAC development. Tumor cells promoted mast cell migration. Both tumor cells and PSCs stimulated mast cell activation. Conversely, mast cell–derived interleukin (IL)-13 and tryptase stimulated PSC proliferation. Treating tumor-bearing mice with agents that block mast cell migration and function depressed PDAC growth. Our findings suggest that mast cells exacerbate the cellular and extracellular dynamics of the tumor microenvironment found in PDAC. Therefore, targeting mast cells may inhibit stromal formation and improve therapy. Cancer Res; 73(13); 3927–37. ©2013 AACR.
Pancreatic ductal adenocarcinoma (PDAC) is the fourth leading cause of cancer-related death in the United States (1), and from the perspective of cancer mortality it is one of the deadliest diseases (2). With a 5-year relative survival rate of 6% (3), PDAC has an extremely poor prognosis, which is related to its inherent resistance to virtually all therapeutic modalities, including conventional chemotherapy, treatment with targeted agents, radiotherapy, and immunotherapy (3–6).
Tumor cells do not exist in isolation during disease progression. The presence of an intense fibro-inflammatory reaction, composed of immune cells (7) and cancer-associated fibroblasts [also known as the pancreatic stellate cell (PSC); refs. 8, 9], is a prominent pathologic feature of PDAC (10). These components of the microenvironment provide a niche in which pancreatic cancer cells can resist treatment. Essentially, the robust desmoplastic response accompanies the progression from normal histology to a malignant state in human pancreatic cancer (7) and promotes resistance by limiting the delivery of chemotherapeutic agents (5, 9, 11, 12). The cells responsible for the production of the desmoplastic reaction around pancreatic tumors are the PSCs (13).
Several investigators have reported that PSCs promote tumor progression as well as chemoresistance (9, 13–15). PSCs not only increase tumor cell viability in vitro, but they also increase tumor invasion and metastasis in vivo (9, 14, 15). In addition, PSCs produce secreted factors that render PDAC cells more resistant to therapy (9). Despite these advancements, many unanswered questions remain, particularly those relating to maintenance of the tumor fibrotic microenvironment during disease development.
Ehrlich (as reviewed by Ribatti and Crivellato, ref. 16) was the first to observe mast cell infiltration in chronic inflammation, accompanied by a fibrotic tissue response. Mast cells regulate connective tissue turnover in wound healing, liver cirrhosis, and pulmonary fibrosis (17). Mast cells also regulate adaptive immunity to tumors. Three recent reports indicate that increased numbers of mast cells infiltrating into the PDAC tumor microenvironment represent a poor prognostic indicator (18–20). Mast cell infiltration is zone specific in PDAC, and a high mast cell count in the intratumoral border zone and the distance to the tumor border is an independent risk factor for decreased survival of patients with pancreatic cancer (18). However, it is still unknown whether hindering the migration of mast cells to the tumor site will dampen the aggressiveness of PDAC and/or favor increased survival of tumor-bearing hosts. The dynamic cellular interactions and complex interplay between cell components in the human pancreatic cancer microenvironment have not been well defined. The precise contribution of mast cells remains unknown.
In this study, we investigated the interactions among mast cells, cancer cells, and PSCs. We found that pancreatic cancer cells induce mast cell migration. Treating mast cells with conditioned media from tumor cells and PSCs induced mast cell activation. Conversely, mast cell–derived cytokines interleukin (IL)-13 and tryptase induced increased proliferation of PSCs. IL-13 activated proliferation through the TGF-β2 pathway in a STAT6-independent manner. Blocking mast cell migration and function in vivo suppressed PDAC growth and improved survival of PDAC-bearing hosts. Our data suggest that the dynamic cellular and extracellular interactions among mast cells, tumor cells, and stellate cells contribute to the desmoplasia found in pancreatic cancer.
Materials and Methods
Human PDAC, Panc1, BxPC3, and AsPC-1 were obtained from American Type Culture Collection (ATCC). L3.6pl cells were kindly provided by Dr. I.J. Fidler (The University of Texas MD Anderson Cancer Center, Houston, TX). Human PSCs were isolated in primary culture from a surgical specimen of pancreatic adenocarcinoma, and a stable cell line was generated by immortalization (9). The PDAC and PSC cells were cultured in RPMI-1640 (Invitrogen) containing 10% fetal calf serum (FCS) with penicillin and streptomycin (both from Invitrogen) at 37°C in a humidified atmosphere of 5% CO2. The immortal human pancreatic ductal epithelial cell line (HPDE) was provided as a generous gift from Dr. Ming-Sound Tsao (Ontario Cancer Institute, Toronto, ON, Canada; refs. 21, 22) and was cultured in keratinocyte-serum-free medium (SFM; Invitrogen). The human mast cell lines HMC-1 (kindly provided by Dr. J.H. Butterfield, Mayo Clinic, Rochester, MN), and LAD2 (kindly provided by Dr. A.S. Kirshenbaum, NIH, Bethesda, MD) were cultured as described previously (23, 24). Cell lines were validated by short tandem repeat (STR) DNA fingerprinting by the MD Anderson Cancer Center Characterized Cell Line Core using the AmpFℓSTR Identifiler Kit according to the manufacturer's instructions (Applied Biosystems, cat. no. 4322288). The STR profiles were compared with known ATCC fingerprints (ATCC.org), to the Cell Line Integrated Molecular Authentication database (CLIMA) version 0.1.200808 (http://bioinformatics.istge.it/clima/; Nucleic Acids Research 37:D925-D932 PMCID: PMC2686526), and to the MD Anderson fingerprint database. The STR profiles matched known DNA fingerprints or were unique.
Conditioned media preparation
Human PDAC, PSC, and HPDE cells were grown to 70% to 80% confluence in complete culture media. The medium was changed to serum-free RPMI-1640, and cells were cultured for an additional 48 hours. The nonadherent HMC-1 and LAD2 cells (5 × 105/mL) were resuspended in serum-free RPMI-1640 and cultured for 48 hours. Medium was collected and filtered with 0.22-μm filters (Millipore).
The 8-μm pore Transwell polycarbonate membrane chambers (Corning) were used. A total of 5 × 104 HMC-1 or LAD2 cells in 500 μL SFM was added to the upper chamber. Conditioned media from human PDAC, PSC, or HPDE cells was added into the lower chamber. The cells were allowed to migrate through the membrane for 24 hours at 37°C in 5% CO2 atmosphere. The nonmigrating cells on the upper surface of the membrane were removed with a cotton swab. The migrating cells in suspension in the lower chamber were fixed in 4% paraformaldehyde. No mast cells were found adherent to the lower surface of the membrane after fixation. Migrating cells were collected, stained with hematoxylin, and counted.
Cell proliferation assay
Proliferation of mast cells, pancreatic cancer cells, and PSCs was assessed using Alamar Blue. Ten microliters of sterile Alamar Blue was added per 5,000 cells. During the incubation period, the conversion of the nonfluorescent indicator dye (resazurin) to a bright red color (resorufin) by metabolically active cells was monitored by reading absorbance at 570 and 600 nm. Proliferation was defined by calculating the percentage conversion of Alamar Blue, as described previously (25).
Mast cell activation and ELISA
Supernatants from cultures of mast cells, mast cells cocultured with PDAC, or mast cells cocultured with PSCs (50:50 mix of cells) were collected, and tryptase secretion was measured by ELISA. Plates were coated with mouse anti-human tryptase antibody (AbD Serotic) and samples were added. Alkaline phosphatase–conjugated mouse anti-human tryptase antibody (Millipore) was used for detection. After adding the substrate, the absorbance was measured at 405 nm. TNF-α release was measured with a Human TNF-α Immunoassay kit (R&D Systems).
PSCs were cultured in mast cell conditioned medium with the following neutralizing antibodies: (i) anti-human tryptase β-2/TPSB2 antibody [monoclonal mouse immunoglobulin G1 (IgG1) clone #349414; R&D Systems; 4.2 μg/mL equals 3 times the ND50]; and (ii) anti-human IL-13 antibody (monoclonal mouse IgG1 clone #32116; R&D Systems; 2.2 μg/mL equals 3 times the ND50; ref. 26). Mouse IgG1 isotype control (monoclonal mouse IgG1 clone #11711; R&D Systems) was used at the same concentration of the neutralizing antibody. SFM was used as the positive control. Percentage suppression of proliferation = (average proliferation index of antibody blocking group − average proliferation index of negative control group)/(average proliferation index of positive control group − average proliferation index of negative control group) × 100%.
Dye eFluor 670 labeling and flow cytometry analysis
PSCs were labeled with 5 μmol/L eFluor 670 (eBioscience; 10 min at 37°C) and 2 × 104 PSCs were added into each well of a 24-well plate. The cells were then cultured with medium only, mast cell conditioned media, or 2 × 104 HMC-1. Upon 95% confluency (≈8 days), the cells were washed, resuspended in PBS, and supplemented with 2% FBS. Upon cell proliferation, the dye was equally diluted into daughter cells. Fluorescence was measured with a LSRII flow cytometer (BD Biosciences) and analyzed using FlowJo software (Tree Star).
Lysates were prepared from PSCs. Samples containing 40 μg protein per well were loaded in 10% SDS-PAGE and transferred onto polyvinylidene difluoride (PVDF) membranes. Membranes were incubating in blocking buffer (5% nonfat dry milk/PBS) for 1 hour at room temperature, followed by an overnight incubation at 4°C in the presence of antibodies directed against TGF-β2 (R&D Systems), STAT6 (Cell Signaling Technology), pSTAT (Millipore), pSmad2 (Millipore), β-actin (Thermo Scientific), and β-tubulin (Epitomics). Membranes were washed with PBS–0.1% Tween 20 and incubated in the presence of horseradish peroxidase–conjugated secondary antibody. Band detection was carried out by enhanced chemiluminescent substrate (SuperSignal West Dura; Thermo-Fisher Scientific) and captured by X-ray films. Blot quantification was conducted with NIH ImageJ software (http://rsb.info.nih.gov/nih-image/).
Orthotopic implantation of PDAC and AMD3100 or cromolyn treatment
C57BL/6 mice were obtained from The Jackson Laboratory. The mice were maintained in facilities approved by the Association for Assessment and Accreditation of Laboratory Animal Care International in accordance with current regulations and standards of the U.S. Department of Agriculture, Department of Health and Human Services, and NIH. All animal procedures were reviewed and approved by the MD Anderson Cancer Center Animal Care and Use Committee. Orthotopic implantation of PDAC was conducted as described previously (27). Briefly, mice were anesthetized with isoflurane (via inhalation, 4% isoflurane in 96% O2), and a 1-cm incision in the left subcostal region was made. Murine PDAC cells (Panc-02, or K-rasG12Dp53−/−; refs. 5, 28) were injected into the caudal pancreas. The peritoneum and skin were closed with the EZ Clip Wound Closing Kit (Stoelting Co.). Analgesia (buprenorphine; 0.5 mg/kg) was administered postoperatively to minimize pain. Seven days after PDAC implantation, a group of the mice received a CXCR4 chemokine receptor antagonist, AMD3100 (Sigma-Aldrich; 10 μg/mouse, intraperitoneal injection) every Monday, Wednesday, and Friday for 3 weeks, as described previously (27). Another group of the mice received daily injection of cromolyn (10 mg/kg body weight; ref. 29) into the peritoneal cavity. PBS was used as a control. Twenty-eight days after implantation, the mice were sacrificed, and tumor volumes were estimated using the formula (π × long axis × short axis × short axis)/6. Mast cell infiltration into the pancreas was measured using Toluidine blue and analyzed using ImageJ software (NIH) as described previously (27). To determine the effect of AMD3100 on survival, drug treatment continued and mice were sacrificed when they became moribund. In those experiments, we implanted C57BL/6 albino mice with K-rasG12Dp53−/− cells transfected with enhanced firefly luciferase (30). At various times after tumor implantation, the mice were injected with d-luciferin (150 μg/mouse; Caliper Life Sciences) and bioluminescence was measured with the IVIS imaging system (Caliper Life Sciences). Tumor volume was measured every week by imaging luciferase signals until the endpoint was reached.
Statistical differences between control and experimental groups was determined using the unpaired, two-tailed, Student t test; a one-way ANOVA, followed by the Fisher least significant difference (LSD) multiple comparison test; or two-way ANOVA (GraphPad Prism V4 and SPSS for Mac, Inc.). Survival curves were constructed using the Kaplan–Meier method, and statistical significance was determined using the log-rank test. P values less than 0.05 were considered statistically significant. Representative experiments are shown; each experiment was repeated independently at least 3 times.
PDAC cells promote mast cell migration
Few mast cells can be found in the normal pancreas; however, it has been reported that increased mast cell numbers in tumor stroma represent a poor prognosis for patients with pancreatic cancer (18–20). Therefore, we first wanted to determine which cell in the tumor microenvironment promoted mast cell migration. We cultured HMC-1 or LAD2 mast cells in conditioned media derived from different pancreatic cancer cells or nontumorigenic PSC or HPDE cells and measured mast cell migration in vitro. The 4 pancreatic cancer cells that we used represent different stages of human pancreatic cancer: high (L3.6pl, AsPC-1) or low (Panc-1, BxPC-3) metastatic potential; resistant (Panc-1, AsPC-1) or sensitive (BxPC-3, L3.6pl) to gemcitabine chemotherapy. Compared with the negative control (SFM), conditioned medium from all 4 different pancreatic cancer cell lines significantly upregulated mast cell migration (Fig. 1; P < 0.0001). On the other hand, conditioned media from human nontumorigenic pancreatic cell lines did not induce mast cell migration (PSCs vs. SFM, HPDE vs. SFM; P > 0.05).
PDAC cells and PSCs stimulate mast cell activation
Next, we determined which elements within the tumor microenvironment affect mast cell function. We asked if coculturing tumor cells or stellate cells with mast cells would result in mast cell activation. Two different human mast cell lines were used, HMC-1 and LAD2. HMC-1 cells represent immature mast cells, do not express the FcεRI receptor, and grow independent of stem cell factor (23). Cytokine (TNF-α) release is a standard method to measure HMC-1 activation (31). LAD2 cells represent mature mast cells with functional FcεRI receptors and are dependent upon stem cell factor for growth (24). More than 98% of LAD2 cells are tryptase-positive, and tryptase released to the culture supernatant is used as a measure of LAD2 activation and degranulation (32). Using these parameters, we tested the ability of human PDAC and PSC cells to activate human mast cells. Compound 48/80 was used as a positive control (33). A significant increase in tryptase release was observed when LAD2 cells were cocultured with all the PDAC cell lines tested (P < 0.01 vs. mast cells cultured alone). Similarly, culturing LAD2 with PSCs resulted in a small, but significant tryptase release [Fig. 2A; mast cells (MC) + PSC vs. MC only; P = 0.001]. The activation of HMC-1 cells is shown in Fig. 2B. Here again, coculturing the mast cells with the PDAC cell lines or PSCs resulted in a significant increase in TNF-α release (P < 0.001). Our results indicate that both PDAC and PSC cells trigger mast cell activation.
Conditioned media from mast cells stimulate PDAC proliferation
To investigate the potential role of mast cells in driving tumor growth, we cultured PDAC in mast cell conditioned medium (Fig. 3). Compared with the serum-free controls, mast cell conditioned medium increased proliferation of all 4 of the pancreatic cancer cell lines (Fig. 3A, Panc-1, 13.6% of total increment; Fig. 3B, BxPC-3, 12.4% of total increment; Fig. 3C, L3.6pl, 18.9% of total increment; and Fig 3D, AsPC-1, 16.5% of total increment).
Mast cells promote PSC proliferation
Mast cell influx and fibrosis coexist in chronic inflammation (17), suggesting that there may be a role for mast cells in fibrosis of pancreatic cancer. Therefore, we tested the hypothesis that mast cells affect PSC proliferation. When PSCs were cultured with conditioned medium from mast cells, PSC proliferation increased (Fig. 4A; 14.7% of total increment).
We then measured, by flow cytometry, the doubling times of eFluor 670 dye–labeled PSCs cultured in control medium or mast cell conditioned medium or PSCs cocultured with mast cells. We found that the doubling time of PSCs was decreased when the cells were grown in mast cell conditioned medium (24.15 ± 0.06 hours), compared with the control (25.04 ± 0.22 hours; Fig. 4B and C). Strikingly, the doubling time of PSCs cocultured with mast cells was even shorter (22.68 ± 0.43 hours; Fig. 4B and C). Taken together, these results support our hypothesis that mast cells promote desmoplastic response by inducing PSC proliferation. It seems that both secreted mast cell products and direct cell-to-cell contact are involved; however, direct cell-to-cell contact provides the best stimulus for PSC proliferation.
Mast cell–derived IL-13 promotes PSC proliferation
IL-13, a mast cell–derived T-helper cell 2 (TH2)-type cytokine (34), is a major inducer of organ fibrosis. It has been identified as a dominant fibroproliferative cytokine in liver fibrosis induced by schistosomiasis as well as in pulmonary fibrosis caused by asthma via allergic mechanisms (35–37). To determine whether mast cell IL-13 promotes PSC proliferation, we treated mast cell conditioned medium with neutralizing anti-IL-13 antibody. As shown in Fig. 5, neutralizing IL-13 activity suppressed the PSC proliferation (51.53% of blocking rate; P = 0.0015), compared with the positive control (IgG1).
We next investigated if the increase in proliferation in PSCs is due to activation of the downstream effectors of IL-13. Stat6 phosphorylation is the latent cytoplasmic event that plays a central role in IL-13–regulated signal transduction and gene expression (38). IL-13 regulation of the TGF-β and phosphor-Smad2 (39) pathway was recently reported to play a role in fibrosis (40, 41). We hypothesized that IL-13 promotes PSC proliferation by regulating Stat6 phosphorylation or TGF-β–pSmad2 pathway. To examine this hypothesis, we evaluated Stat6, phosphorylated Stat6, TGF-β, and pSmad2 expression by Western blotting. Stat6 and phosphorylated Stat6 (Supplementary Fig. S1) expression levels in PSCs were not affected by either mast cell conditioned media or IL-13 antibody blocking. However, TGF-β2, as well as pSmad2, in PSCs (Fig. 5C and D) were upregulated in mast cell conditioned medium, and TGF-β2 expression and Smad2 phosphorylation were modulated by anti-IL-13 antibody. Therefore, we suggest that IL-13 increases PSC proliferation by enhancing TGF-β2 production and Smad2 activation, which is independent of Stat6 or its phosphorylation.
Mast cell–derived tryptase promotes PSC proliferation
Tryptases are serine proteases implicated in asthma and other inflammatory disorders and are highly expressed in human mast cells (42, 43). Tryptase, a potent mitogen for fibroblasts, has been recognized as a molecular link between mast cell activation and fibrosis (44). A recent study showed the importance of tryptase for inducing a cardiac myofibroblastic phenotype, ultimately leading to the development of cardiac fibrosis (45). We hypothesized that tryptase influences PSC proliferation, which is essential to the fibrotic tumor microenvironment. Adding anti-tryptase antibody to mast cell conditioned medium suppressed PSC proliferation (38.06% of blocking rate; P = 0.0032; Fig. 5).
Blocking mast cell migration to the tumor microenvironment suppresses PDAC growth and promotes increased survival in vivo
Given the fact that PDAC development was suppressed in genetically mast cell–deficient mice (20), we explored pharmacologic approaches to verify that blocking mast cell function can achieve the same effects on PDAC development. CXCR4 is a chemokine receptor that mediates mast cell migration to its ligand, CXCL12 (46). The CXCR4 antagonist (47), AMD3100, blocks mast cell migration in vivo (27). To determine if blocking mast cell migration to the tumor microenvironment affects tumor growth in vivo, we tested the effect of AMD3100 using an orthotopic model of PDAC. Treating tumor-bearing mice with AMD3100 suppressed tumor volume by 50% [286.6 ± 45.3 mm3 vs. 574.9 ± 66.2 mm3 (Fig. 6A; P < 0.005; AMD3100 vs. placebo control] and increased survival (Fig. 6B; P < 0.0142; AMD3100 vs. placebo control). Decreased tumor size and increased survival were associated with decreased mast cell migration into the tumor site (Fig. 6C–E). Collectively, the data indicate that tumor cells activate mast cell migration to the pancreas, and that blocking mast cell migration in vivo suppressed PDAC growth and increased the survival of the tumor-bearing hosts.
Blocking mast cell degranulation results in therapeutic responses to PDAC in vivo
To determine whether targeting mast cell function has a therapeutic effect in PDAC, we injected tumor-bearing C57BL/6 mice with cromolyn, a mast cell stabilizer. The growth of surgically implanted K-rasG12Dp53−/− tumors was significantly suppressed by cromolyn treatment (Fig. 7A–E; P < 0.001; treated mice vs. saline control at day 42 after treatment). Tumor measurement was conducted at the endtime point (day 50) after treatment. The mean tumor sizes in the saline- and cromolyn-treated groups were 31.3 and 5.2 mm3, respectively. A P value of 0.017 was determined using Student t test. Similarly, mast cell degranulation in the tumor microenvironment was inhibited in the cromolyn-treated group compared with the saline-treated control group (Fig. 7A–E). The growth of Panc-02 tumor cells was also significantly suppressed by cromolyn treatment (Supplementary Fig. S2; P < 0.001; treated mice vs. saline control at day 21 after treatment). These data further confirm the critical role of mast cells in PDAC progression and suggest that therapies that target mast cell degranulation may be a useful adjuvant treatment of PDAC.
PDAC is characterized by a prominent desmoplastic stromal matrix, which promotes tumor growth and blocks the efficient delivery of chemotherapeutic agents (5). In addition, 3 recent articles have shown a critical role for mast cells in PDAC (18–20). In this study, we measured the interactions among mast cells, tumor cells, and PSCs to determine how these 3 different elements of the tumor microenvironment contribute to the growth of pancreatic cancer. We found that tumor cells recruit mast cells to the tumor site and stimulate mast cell activation and that blocking mast cell function in vivo suppresses tumor growth and increases survival. To our knowledge, this is the first effective example that targeting mast cells in the tumor microenvironment in the immunocompetent PDAC tumor-bearing host can block tumor growth and promote increased survival. Conversely, mast cells stimulate the proliferation of PDAC and PSC cells. We suggest that the cross-talk between mast cells, tumor cells, and stellate cells promotes mast cell migration and activation, promotes cancer growth, and may contribute to the desmoplastic tumor microenvironment that promotes PDAC tumorigenesis.
We found that both the PDAC and PSC cells can trigger mast cell activation and cytokine release. Conversely, mast cell activation favors tumor cell proliferation, suggesting a feedback loop in the tumor microenvironment. Strouch and colleagues (19) reported that mast cells induced the proliferation of 2 gemcitabine-resistant PDAC cell lines (Panc-1 and AsPC-1). Our results confirm this observation and expand the concept to the most representative PDAC cells, including the gemcitabine-sensitive PDAC cell lines (BxPC-3 and L3.6pl). Thus, mast cells stimulate PDAC proliferation regardless of the metastatic potential or drug resistance of the cell line in question.
More importantly, we identified a central role for mast cells in promoting PSC proliferation. Although mast cell conditioned medium stimulated a significant increase in PSC proliferation, the effect was minimal. This may reflect the culture conditions used in this experiment, in which SFM was used. Because we used SFM in the migration experiments shown in Figs. 1 and 2, we wanted to keep the conditions in each experiment as similar as possible. On the other hand, the maximal effect was found when culture conditions allowed for direct contact of mast cells and PSCs. Mixed culture of mast cells and PSCs led to a much shorter doubling time of PSCs than observed when PSCs were treated with mast cell conditioned medium. This may suggest that in vivo, direct contact between mast cells and PSCs is required to see the maximal increase in PSC proliferation.
In many different tissues and organs and in a variety of benign and malignant diseases (48), mast cell aggregates are accompanied by fibrosis (49). However, despite the frequent occurrence and severity of the fibrosis, little is known about a cause and effect relationship, and this is particularly true in pancreatic cancer. Our findings indicate that activated mast cells secrete IL-13 and tryptase, which accelerates cancer-associated fibroblast proliferation. This suggests that mast cell infiltration and activation in pancreatic cancer contributes to the dense fibrotic stromal formation that is so prevalent in this disease. With regard to the mechanism of IL-13 in PSC proliferation, we tested a panel of extracellular matrix–related factors, including TGF-β1, TGF-β3, TGF-α, TNF-α, IL-8, stromal cell–derived factor-1 (CXCL12), thymic stromal lymphopoietin (TSLP), fibroblast growth factor 2 (FGF-2), and hyaluronic acid. We found that mast cell–derived IL-13 enhanced TGF-β2 production and Smad2 phosphorylation without regulating Stat6 expression or phosphorylation. This was not totally unexpected, as TGF-β2 has been known to induce extracellular matrix formation (50). This growth-stimulating effect of mast cell–derived IL-13 on stellate cells might lead to the fibrogenesis of pancreatic cancer. None of the other factors mentioned earlier in this report were upregulated by mast cell–derived IL-13 in our study.
Of course, the following must be taken into account in interpreting the data presented here. First, our in vitro findings are limited in that only immortalized human tumor cell lines and nontumorigenic cell lines rather than ex vivo primary cells were used here. This in part is due to the difficulty in isolating primary tumor cells. Second, our antibody neutralization studies indicated a role for IL-13 and tryptase in inducing stellate cell proliferation. Antibodies to a whole panel of secreted mast cell products were included in these studies, including TSLP, TSLP receptor, IL-1β, IL-10, and CXCL12; however, only neutralizing tryptase and IL-13 had any effect on PSC proliferation. Also, although we noted a significant reduction in PSC proliferation in the presence of anti-tryptase and/or anti-IL-13, we were not able to reduce PSC proliferation to background levels. This suggests as yet unidentified component(s) in mast cell conditioned medium may be playing a role. Third, although the in vivo data presented in Figs. 6 and 7 indicate that both AMD3100 and cromolyn affect mast cell migration into and function in the tumor site, it is important to keep in mind that these agents can affect other cells in the pancreatic tumor microenvironment. Cromolyn has been shown to bind to S100P on pancreatic cancer cells, which prevents the activation of RAGE, thus inhibiting tumor growth and invasion (51). Cromolyn has also been shown to block angiogenesis (52). However, in our analysis, angiogenesis was not affected by cromolyn. We found that cromolyn treatment did not affect the density of CD31-labeled vessels in the tumor microenvironment (data not shown). AMD3100, a CXCR4 antagonist, was used here because it interferes with the in vivo migration of mast cells (27). Others have shown, however, that AMD3100 also prevented the noncanonical activation of the sonic hedgehog pathway and the induction of epithelial–mesenchymal transition in vitro (53). It is also important to note that both cromolyn and AMD3100 have been shown by others to downregulate the expression of TGF-β (54, 55). On the basis of our observation that mast cell–derived IL-13 upregulation of TGF-β plays a role in the proliferation of stellate cells (Fig. 5), it is entirely possible that both cromolyn and AMD3100 may be playing a similar role here. Regardless of the exact mechanism by which cromolyn and AMD3100 affect pancreatic cancer development, our findings indicate an additional function for these drugs in the tumor microenvironment: inhibition of mast cell migration and function.
In summary, based on our in vitro and in vivo data, we suggest a new paradigm for how mast cells promote the desmoplastic response that favors PDAC development. Initial mast cell migration was initiated by pancreatic cancer cells, not the stroma or benign cells. Pancreatic cancer cells and stromal cells activated the tumor-infiltrating mast cells as evidenced by degranulation (tryptase release) and cytokine release. Degranulation and secretion of the cytokines from mast cells produced a nourishing microenvironment that triggered a feed-forward loop favoring cancer development. As a result, active mast cells stimulated pancreatic cancer and stromal cell proliferation. The presence of mast cells and their cytokines, including IL-13 and tryptase, in the tumor microenvironment of pancreatic cancer leads to stromal cell proliferation with TGF-β2 production and Smad2 phosphorylation, which eventually culminates in tumor progression. Blocking mast cell migration with the CXCR4-antagonist AMD3100 or inhibiting mast cell degranulation with cromolyn disrupted the feedback loop to a certain extent. This finding may help to explain the significant tumor shrinkage in cromolyn-treated tumor-bearing hosts and longer survival in AMD3100-treated tumor-bearing hosts, which suggests that blocking mast cell function to the tumor microenvironment may be a critical step to interrupt the positive feedback loop of mast cell promoting PDAC progression. The findings reported here provide proof of principal that mast cells contribute to the desmoplastic microenvironment found in PDAC by promoting the proliferation of PSCs. Therefore, treatments targeting mast cell function may be one way to overcome stromal formation and improve PDAC therapy.
Disclosure of Potential Conflicts of Interest
No potential conflicts of interest were disclosed.
Conception and design: Y. Ma, R.F. Hwang, C.D. Logsdon, S.E. Ullrich
Development of methodology: Y. Ma, R.F. Hwang, C.D. Logsdon
Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): Y. Ma, S.E. Ullrich
Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): Y. Ma, R.F. Hwang, C.D. Logsdon, S.E. Ullrich
Writing, review, and/or revision of the manuscript: Y. Ma, R.F. Hwang, C.D. Logsdon, S.E. Ullrich
Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): Y. Ma, R.F. Hwang, S.E. Ullrich
Study supervision: C.D. Logsdon, S.E. Ullrich
This work was supported by NIH Grant CA131207, Cancer Prevention & Research Institute of Texas Grant RP120777, and by a grant from G.S. Hogan Gastrointestinal Cancer Research Fund (S.E. Ullrich) and a Hirshberg Foundation pancreatic cancer research seed grant (Y. Ma). The Animal Facilities and the Characterized Cell Line Core were supported in part by a National Cancer Institute Core Grant, CA16672.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
The authors thank Dr. Grzegorz Chodaczek for help with the image analysis, Drs. Patrick Hwu and Weiyi Peng for help with the luciferase labeling experiments, Drs. Chengming Zhu, Omid Tavana, and Zhiqiang Zhang for help with the Western blot analyses, Dr. Stephanie Watowich for critical suggestions, and Nasser Kazimi and Polina Khaskina for help with the animal experiments.
Note: Supplementary data for this article are available at Cancer Research Online (http://cancerres.aacrjournals.org/).
- Received December 11, 2012.
- Revision received April 11, 2013.
- Accepted April 16, 2013.
- ©2013 American Association for Cancer Research.