Stromal cues regulate the pancreatic cancer epigenome and metabolome
Contributed by Ronald M. Evans, December 14, 2016 (sent for review October 18, 2016; reviewed by Scott M. Lippman and Richard Posner)
Significance
Stromal fibroblasts of the pancreatic tumor microenvironment (TME) have been shown to play both tumor-supportive and tumor-suppressive roles in enacting a dysregulated wound-healing response. This apparent complexity suggests that an improved understanding of the molecular basis of cell–cell interactions in the TME is required to identify and target stroma-derived, growth-permissive mechanisms. Here we show that stromal cues induce transcriptional and metabolic changes in pancreatic cancer cells implicated in anabolic metabolism, which overlap with those previously demonstrated downstream of oncogenic Kras. Stromal signals broadly induce histone acetylation in the pancreatic cancer epigenome, and we highlight inhibition of acetyl-lysine sensing by the bromodomain and extraterminal (BET) bromodomain family, Bromodomain-containing protein 2 (BRD2) in particular, as a potential therapeutic strategy.
Abstract
A fibroinflammatory stromal reaction cooperates with oncogenic signaling to influence pancreatic ductal adenocarcinoma (PDAC) initiation, progression, and therapeutic outcome, yet the mechanistic underpinning of this crosstalk remains poorly understood. Here we show that stromal cues elicit an adaptive response in the cancer cell including the rapid mobilization of a transcriptional network implicated in accelerated growth, along with anabolic changes of an altered metabolome. The close overlap of stroma-induced changes in vitro with those previously shown to be regulated by oncogenic Kras in vivo suggests that oncogenic Kras signaling—a hallmark and key driver of PDAC—is contingent on stromal inputs. Mechanistically, stroma-activated cancer cells show widespread increases in histone acetylation at transcriptionally enhanced genes, implicating the PDAC epigenome as a presumptive point of convergence between these pathways and a potential therapeutic target. Notably, inhibition of the bromodomain and extraterminal (BET) family of epigenetic readers, and of Bromodomain-containing protein 2 (BRD2) in particular, blocks stroma-inducible transcriptional regulation in vitro and tumor progression in vivo. Our work suggests the existence of a molecular “AND-gate” such that tumor activation is the consequence of mutant Kras and stromal cues, providing insight into the role of the tumor microenvironment in the origin and treatment of Ras-driven tumors.
Sign up for PNAS alerts.
Get alerts for new articles, or get an alert when an article is cited.
The tissue microenvironment plays important roles in directing development and differentiation but is also required for normal tissue homeostasis, working in part through secreted factors that control epithelial cell growth and survival (1). Disruptions in this tissue homeostatic mechanism accompany solid tumor progression, enabling cancer cells to overcome the barriers to tumorigenesis imposed by normal tissue architecture (2). In pancreatic ductal adenocarcinoma (PDAC), the dense fibroinflammatory stroma is largely attributed to soluble and extracellular matrix components secreted by cancer-associated fibroblasts [CAFs; derived predominantly from activated pancreatic stellate cells (PSCs)] (3).
The microenvironment surrounding PDAC undergoes dynamic alterations including aberrant immune responses and the accumulation of a fibroinflammatory stroma (4, 5), alterations co-opted from the wound-healing response (6). This cellular compartment not only cooperates with oncogenic alterations to drive pancreatic tumorigenesis (7, 8) but compromises the efficacies of cytotoxic and immune-targeted therapies (9–12). Stromal alterations during pancreatic tumorigenesis include activation of tissue-resident stellate cells: Although stellate cells contribute to tissue homeostasis and maintenance of the basement membrane under normal conditions (13), these cells transdifferentiate into myofibroblast-like cells in the context of tissue damage or during pancreatic tumor progression (14). PSC activation features robust induction of a transcriptional program that drives a fibroinflammatory response, including extracellular matrix components, cytokines, and growth factors (15). Activated PSCs are a rich repository for secreted factors that may have diverse effects on neighboring epithelial cells, yet the nature of this interaction remains incompletely understood. Indeed, recent work suggests both tumor-supportive (16, 17) and tumor-suppressive or homeostatic (18–20) roles for the PDAC-associated stroma, suggesting that an improved understanding of the molecular basis of tumor–stroma interactions may enable identification and targeting of tumor-supportive mechanisms.
Results
To explore the role of CAF-derived signals in gene regulation in the PDAC epithelial compartment, pancreatic cancer cells were grown in hyaluronic acid-based hydrogels in the presence (stromal) or absence (astromal) of collagens and soluble stromal cues (from patient-derived CAFs or in vitro activated PSCs, which secrete similar levels of cytokines, growth factors, ECM components, and remodeling factors) (15). In response to stromal cues, PDAC cell lines grown in this 3D culture system increased the expression of the secreted factors CSF2 and CXCL1, without any apparent change in morphology or initial cell growth (Fig. 1 A and B and Fig. S1A). Furthermore, RNA-seq analyses showed broad effects of stromal inputs on transcription, with the expression of 332 and 128 genes up- and down-regulated, respectively, in MIAPaCa2 cells and 480 up-regulated and 406 down-regulated genes in p53 2.1.1 cells (>1.5-fold; Fig. 1C). KEGG enrichment analyses of significantly up-regulated genes identified similar pathways activated in p53 2.1.1 and MIAPaCa2 cells, including cell cycle, DNA replication, and metabolic pathways (Fig. 1D and Fig. S1B). Indeed, a stroma-inducible signature, defined as the 86 genes up-regulated by stromal cues in both MIAPaCa2 and p53 2.1.1 cell lines, is enriched for genes in metabolic pathways as well as nucleoporins (Fig. 1E), pathways shown to be up-regulated during pancreatic cancer progression in mice (21). The induction of selected genes was validated by quantitative PCR (qPCR) (Fig. 1F and Fig. S1C). Importantly, similar effects of stromal cues are evident in laser-microdissected human pancreatic cancer (Fig. 1G; low IL6 expression indicates little stromal contamination of tumor samples). Collectively, these results indicate that the stromal secretome is sufficient to regulate the expression of diverse gene classes in pancreatic cancer cells.
Fig. 1.
Fig. S1.
Interestingly, the gene sets induced by stromal cues in vitro overlap significantly with those activated by an inducible allele of KrasG12D in vivo (22), suggesting that permissive signals from stroma drive oncogene-driven transcription in vivo (Fig. 2A). This finding is consistent with reports that a fibroinflammatory stromal reaction cooperates with oncogenic alterations in Kras to drive pancreatic cancer progression, whereas mutant Kras alone is insufficient to drive tumorigenesis (8). Supporting this apparent functional convergence, qPCR analyses confirm that the stromal secretome up-regulates genes across multiple functional categories, including downstream KrasG12D targets involved in immune modulation (e.g., Csf2) (23, 24) and anabolic metabolism (e.g., Rrm2, Sc4mol) (Fig. 2B). Consistent with changes in metabolic genes, stroma-activated PDAC cells consumed more glucose (∼40–90%; Fig. 2C and Fig. S2 A and B) and secreted more lactate (40–60%; Fig. 2D), without significant effects on glutamine consumption (Fig. S2C). Furthermore, intracellular metabolomics revealed that stromal cues dramatically alter cellular metabolism, increasing intermediates in glycolysis, the pentose phosphate pathway, downstream steps in nucleic acid synthesis, and the TCA cycle (Fig. 2E and Fig. S2D). Although PDAC cells grow to a similar extent in the presence or absence of stromal cues under nutrient-replete conditions (Fig. S1A), human pancreatic tumors are nutrient-poor and in particular feature low levels of glucose and glutamine (25). Under nutrient-deprived conditions relevant to the microenvironment of human PDAC, stromal cues significantly increased the viability of PDAC cells, consistent with paracrine activation of an adaptive transcriptional survival program (Fig. 2F). Together, these results suggest that the stromal reaction marshals genomic and metabolic responses in PDAC cells that enforce oncogenic signaling.
Fig. 2.
Fig. S2.
To understand how the stroma authorizes gene expression in PDAC cells, we explored acute changes in histone acetylation marks that differ between stroma-rich and stroma-poor regions of human PDAC (26), although we do not exclude effects of additional histone marks. Accordingly, PDAC cells exposed to soluble stromal cues showed increased levels of acetylation at both H3K9 (a marker of active promoters) and H3K27 (a marker of active enhancers; Fig. 3A). Oncogenic Kras induces histone acetylation in vivo (27), implicating epigenetic alterations as a putative point of convergence between oncogenic and microenvironmental signals. A time course revealed surprisingly rapid changes in histone acetylation, detectable within 1–3 h (Fig. 3B), arguing against a requirement for de novo protein synthesis to induce these acetylation changes. To understand the transcriptional consequences of these epigenetic alterations, we localized sites of differential histone acetylation using targeted (ChIP-qPCR) and global (ChIP-Seq) analyses in PDAC cells grown under astromal or stromal conditions. Notably, stromal cues dramatically increased H3K9 acetylation in promoter regions across the genome (Fig. 3 C and D). The top 5% of differentially acetylated protein-coding genes are enriched for cancer pathways, including pancreatic cancer (Fig. 3E), suggesting that stromal cues achieve their impact in part by coordinated chromatin-dependent remodeling of functionally related gene sets. Consistent with these genome-wide studies, site-specific ChIP-qPCR revealed increased acetylation at promoter and enhancer regions in the context of stromal signals (Fig. 3F). These results demonstrate that microenvironmental cues rapidly regulate histone acetylation at a subset of promoters and enhancers in PDAC cells.
Fig. 3.
Given the extent of the epigenetic changes, we postulated that effectors of this paracrine regulation may be susceptible to therapeutic inhibition. To explore this notion, we targeted the interaction between acetylated histone tails and chromatin readers. As acetylated histone marks are known targets of bromodomain and extraterminal (BET) bromodomain-containing proteins (28–31), we explored the impact of the BET bromodomain inhibitor JQ1 on stroma-inducible gene expression in vitro and on tumor growth in vivo. Interestingly, although treatment of pancreatic tumor cells with JQ1 reduced the expression of hundreds of genes under both astromal and stromal conditions, about 450–475 genes were uniquely down-regulated by JQ1 under stromal conditions (Fig. S3A). Notably, in the PDAC cell lines examined, this subset of genes (Fig. 4 A and B and Fig. S3B) significantly overlaps (35–60%) with those induced by stromal cues, suggesting inverse transcriptional regulation by microenvironmental signals and JQ1. By extension, viability in the context of stromal cues was significantly reduced by JQ1 treatment in multiple PDAC cell lines, albeit to varying degrees (Fig. S3C). Antitumor activity of JQ1 in pancreatic cancer has been described previously by Mazur et al. (32); in this study, the authors suggest that the antitumor effects of BET inhibition are likely due to effects on MYC as well as negative regulation of inflammatory gene expression. In agreement with this work, we observed down-regulation of Myc with JQ1 treatment in the p53 2.1.1 line by RNA-seq but not in the MIAPaCa2 line. However, a number of inflammatory genes were down-regulated by JQ1 in both cell lines, consistent with the findings of Mazur et al., including IFIT2, IFIT3, IRF5, CXCL5, and CXCL2 in MIAPaCa2 cells and Il34, C3, Cxcl1, Ccl2, and Cxcl5 in p53 2.1.1 cells. The expression of BET family members Bromodomain-containing proteins 2, 3, and 4 (BRD2, BRD3, and BRD4) (33, 34) in PDAC lines (Fig. S3D) raises the question as to which might be dominant in stroma-activated gene expression. Unexpectedly, BRD2 knockdown most closely phenocopied JQ1 treatment and down-regulated genes induced under stromal conditions (Fig. 4C). This is confirmed by increased BRD2 binding to promoters of stroma-inducible genes, measured by ChIP assay (Fig. 4D). These findings make an interesting contrast with the recent observation that stellate cell activation during the stromal response is principally mediated by BRD4 (35), suggesting the potential for cell-specific functions of different BET bromodomain proteins in the tumor ecosystem.
Fig. 4.
Fig. S3.
To assess whether disruption of BET family activity could affect pancreatic tumor growth in vivo, we used orthotopic transplantation of luciferase-labeled PDAC cells into the pancreata of syngeneic hosts (Fig. S4A). Impressively, JQ1 treatment significantly reduced pancreatic tumor growth and pancreas wet weight at the study endpoint (Fig. 4 E and F and Fig. S4B), consistent with previous results (32, 36). In addition, significant reductions in proliferation and immune cell infiltration were seen (Fig. S4 C and D), consistent with the negative regulation of transcriptional programs promoting anabolic metabolism and inflammation. To assess the specific role of BRD2, we transduced luciferase-labeled p53 2.1.1 PDAC cells with a doxycycline-inducible shRNA targeting Brd2 (Fig. S5 A and B) before orthotopic transplantation (Fig. S5 C and D). Notably, doxycycline treatment significantly reduced pancreatic tumor growth by over 50%, correlating with the reduced BRD2 protein levels in the tumors (Fig. S5 E–G). When this study was performed alongside a doxycycline-inducible nontargeting control hairpin and an independent inducible shBrd2 clone, tumor growth in the control group was unaffected by doxycycline, whereas tumor growth in the inducible shBrd2 group was significantly reduced by docycycline (Fig. S5H). To confirm these results in a different system, FC1245 PDAC cells derived from KPC mice were transduced with inducible shBrd2 or an inducible control hairpin as described above and used for orthotopic transplantation. Doxycycline significantly reduced tumor growth in the inducible shBrd2 line, whereas the control cell line was unaffected by doxycycline treatment (Fig. 4G). Significant growth differences were noted between the shControl and shBrd2 cell in vivo in the absence of doxycycline, which may reflect leaky expression of shBrd2 in vivo in this system or may indicate a less aggressive clone. These results directly implicate the BET family of proteins as a viable therapeutic target in pancreatic cancer, in part due to effects on gene expression via transduction of the cancer epigenome.
Fig. S4.
Fig. S5.
Discussion
Despite the well-acknowledged role of the TME in modulating tumor growth and progression, an underlying molecular mechanism illuminating its effect is rudimentary at best. We show that a cell-free extract of secreted factors from cultured stromal cells is sufficient to dramatically alter survival of cultured PDAC cell lines under relevant conditions of nutrient challenge. This survival response is linked to a rapid change in transcriptional networks driving core metabolic pathways in the TCA cycle, anabolic metabolism and cell growth. The demonstration that the microenvironmental context broadly influences the epigenetic state of the tumor suggests that either side of this dynamic equation—tumor or tumor microenvironment (TME)—can individually or possibly cooperatively serve as therapeutic targets. Interestingly, recent work has shown that tissue-specific microenvironmental signals can shape chromatin landscapes to support cellular function in distinct populations of resident macrophages (37, 38). Together with our work, these studies raise the possibility that microenvironmental context broadly influences the epigenetic state of tissue-resident cell types under physiological and pathological conditions. In the context of the wound-like TME, TME-induced transcriptional and metabolic changes overlap with those induced by oncogenic Kras, which presumably on its own cannot fully enable proliferation and viability in a nutrient-poor microenvironment. Although we have not identified a specific microenvironment-derived factor that regulates transcription and metabolism in a paracrine manner, our previous work identified numerous secreted factors that are transcriptionally up-regulated during stromal activation in pancreatic cancer and that may serve to regulate transcription and metabolism in the epithelial compartment (15). These include factors that activate Ras-MAPK signaling via cell surface receptors to augment MYC activity, including connective tissue growth factor (39), hepatocyte growth factor (40), and insulin-like growth factor binding proteins 2, 3, and 7 (which signal to EGFR indirectly via sphingosine-1-phosphate) (41). Stromal PSCs also produce Il6, which can activate STAT3 in a paracrine manner to transcriptionally activate MYC (42–44) and induce associated metabolic and transcriptional changes. We propose that microenvironmental regulation of gene expression cooperates with cell-autonomous pathways to fortify cancer cell viability by co-opting an otherwise normal, and usually transient, wound-healing response. Collectively, this work describes a connection between the stromal response and a driver oncogene (Fig. 4H) and may offer new insights into the role of the microenvironment and the treatment of Ras-driven tumors.
Materials and Methods
Cell Culture.
Cell lines MIAPaCa2, Panc1, and AsPC1 were obtained from the American Type Culture Collection. The p53 2.1.1 cell line was provided by Eric Collisson, University of California, San Francisco (45) and was derived from PDAC from a FVB/n KrasLSL-G12D/+; Trp53flox/+; Ptf1a-Cre mouse. The FC1245 cell line was provided by David Tuveson, Cold Spring Harbor Laboratory, Cold Spring Harbor, NY, and was derived from a C57BL/6J KrasLSL-G12D/+; Trp53LSL-R172H/+; Pdx1-Cre mouse. Mouse PSCs were isolated from healthy pancreata of wild-type C57BL/6J mice aged 6–12 wk by density centrifugation, as previously described (15). Human cancer-associated stromal myofibroblasts were grown out of surgically resected PDAC samples as described previously (33) and provided by J. A. Drebin and P. J. O’Dwyer, University of Pennsylvania, in accordance with the Institutional Review Boards of the Salk Institute and the University of Pennsylvania. The stromal population was confirmed to lack KRAS exon 2 mutations and was immortalized with SV40 large T antigen (pLenti-SV40-T, Applied Biological Materials Inc.) at a MOI of 2. This immortalized cell line was used for consistent generation of conditioned media for stromal cultures (described in 3D Astromal and Stromal Culture Generation below.). DMEM and RPMI 1640 were purchased from Life Technologies, and characterized FBS was purchased from HyClone. DMEM containing 10% (vol/vol) FBS was used to culture MIAPaCa2, Panc1, p53 2.1.1, and FC1245 cell lines as well as mouse PSCs and human PDAC-associated stromal myofibroblasts. RPMI 1640 containing 10% (vol/vol) FBS was used to culture AsPC1 cells.
3D Astromal and Stromal Culture Generation.
Astromal and stromal cultures were generated using the ESI BIO HyStem-C kit (BioTime, Inc.). Kit components were equilibrated to room temperature, and the suggested volume of degassed H2O was transferred to vials of Extralink (PEGDA), Gelin-S (thiol-modified denatured collagen), and Glycosil (thiol-modified hyaluronan) using a syringe and 25G needle to avoid opening the vials. Vials were vortexed every 5–10 min to aid reconstitution, and all reagents were in solution within 30 min. To generate astromal cultures, PDAC cells were resuspended in a solution of 1:1:1 degassed H2O:Glycosil:Extralink at a concentration of 5 × 105 cells per mL After careful mixing by pipetting, the suspension was added to a culture dish and placed in a 37 °C incubator for 30 min to polymerize. To generate stromal cultures, PDAC cells were resuspended in a solution of 1:1:1 Gelin-S (collagens):Glycosil:Extralink at a concentration of 5 × 105 cells per mL, mixed by pipetting, and added to a culture dish. After 30 min in a 37 °C incubator, both astromal and stromal cultures polymerized and the appropriate volume of DMEM containing 10% (vol/vol) FBS was layered over the hydrogels. The next day, media was removed, gels washed twice with sterile PBS, and DMEM containing 2% (vol/vol) FBS was added to astromal cultures or conditioned media containing 2% (vol/vol) FBS was added to stromal cultures. Cells appeared in single-cell suspension upon plating and expanded as spheroids over several days in 3D culture. Conditioned media for human PDAC lines was generated from immortalized stromal myofibroblasts grown out of surgically resected human PDAC. Stromal cells were grown to confluency and changed to fresh DMEM containing 2% (vol/vol) FBS. After 48 h, media was collected, spun down at 300 × g for 5 min to remove dead cells and debris, and added to stromal cultures. Conditioned media for murine PDAC lines was generated from activated mouse PSCs. PSCs were isolated from wild-type C57BL/6J mice 6–12 wk of age; yield from five mice was pooled into one T25 flask and cultured for 7 d in DMEM containing 10% (vol/vol) FBS to generate confluent activated PSCs or myofibroblast-like cells. On day 7, media was changed to fresh DMEM containing 2% (vol/vol) FBS, and after 48 h, media was collected, spun down at 300 × g for 5 min to remove dead cells and debris, and added to stromal cultures. For applications requiring recovery of a single-cell suspension (i.e., ChIP or cell counting for metabolic assays), hydrogels were constructed as described above but using PEGSSDA (sold separately by BioTime, Inc.) as a cross-linker for both astromal and stromal conditions. Volumetric ratios were as listed above, but the PEGSSDA reagent was reconstituted at a 2× concentration to ensure efficient polymerization of the hydrogels.
Gene Expression Analysis from 3D Cultures.
For RNA isolation from 3D cultures, TRIzol (Life Technologies) was added to each well (the same volume as the hydrogel volume per well). TRIzol and hydrogel were resuspended by pipetting up and down through a cut pipet tip, transferred to 2.8-mm ceramic bead tubes, and homogenized at 3,000 rpm for 20 s in a PowerLyzer 24 (tubes and homogenizer; MO BIO Laboratories, Inc.). RNA isolation was then performed per the manufacturer’s protocol. Reverse transcription was performed using the iScript cDNA Synthesis Kit per the manufacturer’s protocol (Bio-Rad Laboratories, Inc.), and qPCR was carried out using SYBR Green master mix on a CFX384 Real-Time PCR Detection System (Bio-Rad Laboratories, Inc.). Primer sequences are provided in Dataset S1.
RNA-Seq.
RNA-Seq from astromal and stromal conditions was performed after 24 h of conditioned media addition or 48 h of growth in 3D culture. JQ1 treatments (500 nM, provided by J. Bradner, Dana-Farber/Harvard Cancer Center) were administered for the last 16 h of culture. Total RNA was isolated using TRIzol (Life Technologies) and the Rneasy mini kit with on-column Dnase digestion (QIAGEN). Library preparation, sequencing, and analysis procedures are described in SI Materials and Methods.
Metabolite Measurements.
Glucose, lactate, and glutamine consumption were measured using a YSI 2950 analyzer. For this, PDAC cells were seeded in astromal and stromal conditions with a PEGSSDA cross-linker. After 48 h in 3D culture in DMEM or CM, hydrogels were washed three times with sterile PBS and switched to fresh DMEM containing 2% (vol/vol) FBS. After 24 h, spent media samples were collected and compared with fresh media in biological triplicate using the YSI analyzer. PEGSSDA-based hydrogels were dissociated using 40 mM N-acetyl-cysteine (NAC) for 60–120 min per the manufacturer’s protocol, and cells were counted. Metabolite consumption was calculated by subtracting spent media measurements from fresh media measurements and normalizing to cell number.
Metabolomics.
For steady-state intracellular metabolite profiling, PDAC cells were seeded in 3D cultures (5 × 105 cells per mL hydrogel, 1 mL hydrogel per well in six-well plates) on day 1, and changed to DMEM containing 2% (vol/vol) FBS (astromal) or CM containing 2% (vol/vol) FBS (stromal) on day 2. On day 3, cell media was replaced with fresh media 2 h before extraction. Media was aspirated, and cold (–80 °C) methanol was added to each well while plates were on dry ice. After 30 min of incubation at –80 °C, hydrogels and methanol were scraped and transferred to prechilled tubes, and samples were centrifuged at 3,000 × g for 10 min at 4 °C. Supernatants were recovered, centrifugation repeated, and supernatants dried down in a refrigerated speedvac and stored at –80 °C until analysis. Cells seeded in parallel were analyzed by CellTiter-Glo, and results were used for normalization. Metabolites were analyzed in biological triplicate by liquid chromatography–MS/MS using selected reaction monitoring (SRM) in a 5500 QTRAP hybrid triple quadrupole mass spectrometer (AB/SCIEX) coupled to a Prominence UFLC HPLC system (Shimadzu). Data analysis was performed in MetaboAnalyst 3.0.
Western Blot Analysis.
Western blots were performed as previously described (15). Antibody information and additional details can be found in SI Materials and Methods.
Chromatin Immunoprecipitation and ChIP-Seq.
For ChIP and ChIP-Seq experiments under astromal and stromal conditions, 3D cultures were constructed using the PEGSSDA cross-linker to facilitate rapid dissociation. ChIP, deep sequencing, and analysis were performed as described previously (35). For additional details, SI Materials and Methods.
Orthotopic Transplant Studies.
The p53 2.1.1 luciferase-expressing PDAC cell line was used for orthotopic transplantation into immune-competent FVB/n hosts, as previously described (45). All mouse experiments were performed with the approval of the institutional animal care and use committee (IACUC) of the Salk Institute. For further details, SI Materials and Methods.
SI Materials and Methods
Gene Expression Analysis from 3D Cultures.
For RNA isolation from 3D cultures, TRIzol (Life Technologies) was added to each well (the same volume as the hydrogel volume per well). TRIzol and hydrogel were resuspended by pipetting up and down through a cut pipet tip, transferred to 2.8-mm ceramic bead tubes, and homogenized at 3,000 rpm for 20 s in a PowerLyzer 24 (tubes and homogenizer, MO BIO Laboratories, Inc.). RNA isolation was then performed per the manufacturer’s protocol. Reverse transcription was performed using the iScript cDNA Synthesis Kit per the manufacturer’s protocol (Bio-Rad Laboratories, Inc.), and qPCR was carried out using SYBR Green master mix on a CFX384 Real-Time PCR Detection System (Bio-Rad Laboratories, Inc.). Primer sequences are provided in Dataset S1.
Metabolite Measurements.
Glucose, lactate, and glutamine consumption were measured using a YSI 2950 analyzer. For this, PDAC cells were seeded in astromal and stromal conditions with a PEGSSDA cross-linker. After 48 h in 3D culture in DMEM or CM, hydrogels were washed three times with sterile PBS and switched to fresh DMEM containing 2% FBS. After 24 h, spent media samples were collected and compared with fresh media in biological triplicate using the YSI analyzer. PEGSSDA-based hydrogels were dissociated using 40 mM NAC for 60–120 min per the manufacturer’s protocol, and cells were counted. Metabolite consumption was calculated by subtracting spent media measurements from fresh media measurements and normalizing to cell number.
Metabolomics.
For steady-state intracellular metabolite profiling, PDAC cells were seeded in 3D cultures (5 × 105 cells per mL hydrogel, 1 mL hydrogel per well in six-well plates) on day 1 and changed to DMEM containing 2% FBS (astromal) or CM containing 2% FBS (stromal) on day 2. On day 3, cell media was replaced with fresh media 2 h before extraction. Media was aspirated, and cold (–80 °C) methanol was added to each well while plates were on dry ice. After 30 min of incubation at –80 °C, hydrogels and methanol were scraped and transferred to prechilled tubes, and samples were centrifuged at 3,000 × g for 10 min at 4 °C. Supernatants were recovered, centrifugation repeated, and supernatants dried down in a refrigerated speedvac and stored at –80 °C until analysis. Cells seeded in parallel were analyzed by CellTiter-Glo, and results were used for normalization. Metabolites were analyzed in biological triplicate by liquid chromatography–MS/MS using SRM in a 5500 QTRAP hybrid triple quadrupole mass spectrometer (AB/SCIEX) coupled to a Prominence UFLC HPLC system (Shimadzu). Data analysis was performed in MetaboAnalyst 3.0.
RNA-Seq.
RNA-Seq from astromal and stromal conditions was performed after 24 h of conditioned media addition or 48 h of growth in 3D culture. JQ1 treatments (500 nM, provided by J. Bradner, Dana-Farber/Harvard Cancer Center) were administered for the last 16 h of culture. Total RNA was isolated using TRIzol (Life Technologies) and the Rneasy mini kit with on-column Dnase digestion (QIAGEN). Sequencing libraries were prepared from 100 to 500 ng total RNA using the TruSeq RNA Sample Preparation Kit v2 (Illumina). For this, mRNA was purified, fragmented, and used for first-, then second-strand cDNA synthesis followed by adenylation of 3′ ends. Samples were ligated to unique adapters and subject to PCR amplification. Libraries were then validated using the 2100 BioAnalyzer (Agilent), normalized, and pooled for sequencing. Sequencing was carried out on the Illumina HiSEq. 2500 using bar-coded multiplexing and a 100-bp read length. Image analysis and base calling were performed with Illumina CASAVA-1.8.2. This yielded a median of 29.9 M usable reads per sample. Short read sequences were mapped to a UCSC hg19 reference sequence using the RNA-seq aligner STAR (46). Known splice junctions from hg19 were supplied to the aligner, and de novo junction discovery was also permitted. Differential gene expression analysis, statistical testing, and annotation were performed using Cuffdiff 2 (47). Transcript expression was calculated as gene-level relative abundance in fragments per kilobase of exon model per million mapped fragments and used correction for transcript abundance bias (48). RNA-Seq results for genes of interest were also explored visually using the UCSC Genome Browser. Heat maps were generated using the open source clustering software (Cluster 3.0) and Java TreeView (49). Pathway analysis for significantly regulated genes was performed using the KEGG enrichment analysis function of WebGestalt (50, 51). Previously published microarray data (22) were analyzed using the VAMPIRE microarray analysis suite (52) and compared with our RNA-Seq data using the Gene Set Enrichment Analysis software (53, 54).
Gene Expression Analysis in Human PDAC.
Fresh-frozen, OCT-embedded human PDAC samples were provided by Hubert Stoppler and the UCSF Helen Diller Family Comprehensive Cancer Center Tissue Core. Samples were sectioned onto nuclease-free membrane slides (PET 1.0) and stored at –80 °C until use. Before microdissection, samples were stained with hematoxylin and eosin, and stroma-rich and stroma-poor regions were qualitatively identified by M.L., as previously described (26). On the day of laser capture microdissection, serial sections from the reviewed samples were briefly thawed, fixed in ice-cold 70% ethanol, and stained with hematoxylin and eosin. Samples were dehydrated through an ethanol series, and stroma-rich and stroma-poor regions were isolated by laser capture microdissection using a Zeiss PALM Microbeam V4 Laser Capture Microdissection System and adhesive cap tubes (Carl Zeiss Microscopy, LLC). Depending on region size, between 3 and 10 regions of each description were captured from each of four tumors. RNA was extracted from microdissected regions using the RecoverAll Total Nucleic Acid Isolation Kit (Ambion/Life Technologies). Reverse transcription was carried out using the iScript cDNA Synthesis Kit, and a preamplification reaction was carried out with cDNA and pooled primers for genes of interest using SsoAdvanced PreAmp Supermix (Bio-Rad Laboratories, Inc.) for 12 cycles, per manufacturer’s protocol. Samples were then diluted 20-fold and used for qPCR.
Chromatin Immunoprecipitation and ChIP-Seq.
MIAPaCa2 cells were seeded at 5 × 105 cells per mL in 15-cm plates (15 mL per plate). Two 15-cm plates per condition were prepared for each round of ChIP or ChIP-Seq. After 24 h in 3D culture, media was changed to DMEM containing 2% FBS (astromal) or CM containing 2% FBS (Stromal) and cells were incubated another 24 h for histone ChIP or 6 h for BRD2 and MYC ChIP. For dissociation, a solution of 40 mM NAC was prepared in DMEM, pH 7.4. Media was aspirated from the hydrogels, and 20 mL 40 mM NAC in DMEM was layered over each plate; plates were returned to the 37 °C incubator for 30 min. At this time, the hydrogel and NAC solution were mixed by pipeting up and down, and plates were returned to the incubator for 30-min intervals until the hydrogels were adequately dissolved, typically a total of 90 min of NAC treatment. Liquid was removed from the plates and transferred to conical tubes. An equal volume of PBS was added to each tube of dissociated hydrogel solution, and tubes were centrifuged at 300 × g for 5 min. Media were aspirated and cell pellets were washed with PBS and pelleted at 300 × g for 5 min. Cells were fixed in 1% formaldehyde in PBS at room temperature for 10 min, followed by quenching of the cross-linking reaction by addition of 125 mM glycine (final concentration) and incubation at room temperature for 5 min. Fixed cells were centrifuged at 300 × g for 5 min at 4 °C and washed twice with cold PBS. To isolate nuclei, cell pellets were resuspended in ice-cold cell lysis buffer [150 mM NaCl, 50 mM Tris, pH 7.5, 5 mM EDTA, 0.5% Nonidet P-40, 1% Triton X-100, 1× Complete EDTA-free (Roche)], and nuclei were collected by centrifugation for 1 min at 12,000 × g. The nuclei were washed in the cell lysis buffer twice and resuspended in ChIP Shearing Buffer [50 mM Tris·HCl, pH 8, 10 mM EDTA, pH 8, 1% SDS, 1× complete EDTA-free (Roche)] before proceeding to sonication for ChIP, deep sequencing, and analysis as described previously (55). Protein A Dynabeads (Life Technologies) were used for immunoprecipitation, with 100 μg chromatin and 5 μg H3K9Ac (Abcam, ab4441) or H3K27Ac (Active Motif, 39133) antibody per IP for histone ChIP or 300 μg chromatin and 5 μg BRD2 (Bethyl Laboratories, A302-583A) or MYC (Santa Cruz Biotechnology, sc-764) per IP. Sequencing libraries were constructed and sequenced as previously described (55). Briefly, short DNA reads were aligned against the human hg18 reference genome (NCBI Build 36.1) using the Illumina Pipeline Suite v1.7. Reads were aligned using the Bowtie aligner, allowing up to two mismatches in the read. Only tags that map uniquely to the genome were considered for further analysis. Subsequent peak calling and motif analysis were conducted using HOMER, a software suite for ChIP-seq analysis (56). One tag from each unique position was considered to eliminate peaks resulting from clonal amplification of fragments during the ChIP-seq protocol. Peaks were identified by searching for clusters of tags within a sliding 200-bp window, requiring adjacent clusters to be at least 1 kb away from each other. The threshold for the number of tags that determine a valid peak was selected for an FDR < 0.0001, as empirically determined by repeating the peak finding procedure using randomized tag positions. Peaks are required to have at least fourfold more tags (normalized to total count) than input or IgG control samples and fourfold more tags relative to the local background region (10 kb) to avoid identifying regions with genomic duplications or nonlocalized binding. Peaks are annotated to gene products by identifying the nearest RefSeq transcriptional start site. Visualization of ChIP-seq results was achieved by uploading custom tracks onto the UCSC genome browser.
Viability Assays.
Cellular viability was measured using the CellTiter-Glo Luminescent Cell Viability Assay (Promega) per the manufacturer’s protocol but with a 45-min lysis step for 3D cultures, and luminescence was measured on a Tecan Infinite M1000 PRO. To generate nutrient-poor conditions for viability assays, conditioned media was generated as described above but using serum-free DMEM. After 48 h, conditioned media was harvested and spun at 300 × g for 5 min to pellet debris. Supernatant was concentrated 30-fold in an Amicon Ultra Centrifugal filter with a 10-kDa cutoff. Concentrate was brought to the original media volume with DMEM containing 1 mM glucose and 0.65 mM glutamine; this mixture was filtered and concentrated 30-fold again, and the concentrate again brought up to the original media volume with DMEM containing 1 mM glucose and 0.65 mM glutamine. Glucose in the CM was measured alongside DMEM (1 mM glucose) to ensure equal glucose concentrations (QuantiChrom Glucose Assay Kit, BioAssay Systems). FBS (2%) was added to the low-glucose DMEM and low-glucose CM before use in the viability assays. High-glucose (25 mM glucose, 4 mM glutamine) DMEM and CM were used for the nutrient-rich condition, as in other experiments in the paper.
Histone Acetylation Measurements.
PDAC cells were seeded into six-well plates at a density of 2 × 105 cells per well. The next day, media was aspirated and switched to DMEM containing 2% FBS or CM containing 2% FBS. At the indicated time point, media was aspirated, and cells were washed with ice-cold PBS. Cells were scraped and pelleted in a microcentrifuge at 1,000 × g for 3 min at 4 °C. PBS was aspirated and the cell pellet was resuspended in 50 μL Triton extraction buffer (PBS containing 0.5% Triton X-100) containing protease inhibitors (complete EDTA-free, Roche) and 5 mM sodium butyrate (Sigma). Cells were lysed on ice for 10 min and pelleted at 425 × g for 10 min at 4 °C. Supernatant was discarded, and the pellet washed with 50 µL Triton extraction buffer, followed by another centrifugation step. Supernatant was aspirated and pellet resuspended in 30 μL 0.2N HCl. Histones were acid-extracted at 4 °C overnight. The next day, samples were centrifuged at 425 × g for 10 min at 4 °C. Supernatants were collected and concentrations determined by Bradford assay. Samples were stored at –20 °C before use. To measure histone acetylation levels, 5 μg of acid-extracted histones per sample were analyzed by Western blot.
Radiolabeled Glucose Uptake Assay.
Cellular glucose uptake using radiolabeled 2-deoxy-glucose was performed as described previously. Briefly, cells cultured in 12-well plates were incubated at 37 °C in 200 μL KRBH buffer (120 mM NaCl, 4 mM KH2PO4, 1 mM MgSO4, 0.75 mM CaCl2, 30 mM Hepes, and 10 mM NaHCO3) for 20 min followed by the addition of 0.5 uCi of 3H-deoxy-d-glucose (Perkin-Elmer) in 50 μL loading buffer (KRBH + 0.5 mM 2-deoxy-glucose). A nonspecific binding control well was set up with simultaneous addition of labeled 2-deoxy-glucose in loading buffer and 10 μL 1.5 mM cytochalasin B in DMSO. Cells were incubated for an additional 20 min at 37 °C. Glucose uptake assay was terminated upon media removal and extensive wash in PBS. Cells were lysed in 0.1 N NaOH, and the radioactivity was determined in a scintillation counter. Protein concentration was measured by Bradford method. Cellular glucose uptake was calculated as counts per million (cpm)/µg protein subtracting the normalized radioactivity in nonspecific binding control.
Orthotopic Transplant Studies.
For the JQ1 study, wild-type FVB/n mice aged 8–10 wk were injected with 5,000 p53 2.1.1 cells in 50% DMEM, 50% Matrigel into the tail of the pancreas. After 14 d, bioluminescence imaging was performed and mice were randomized into two treatment groups. Mice received i.p. injections of 75 mg/kg JQ1 or an appropriate volume of vehicle (10% hydroxypropyl-β-cyclodextrin) daily for 14 d, at which time they were again subject to bioluminescence imaging. On day 28 of the study, animals were killed and pancreata collected, weighed, and fixed in formalin or embedded in OCT and frozen for histology. Fresh-frozen, OCT-embedded tumor tissues were sectioned, fixed, stained for CD45 (BD Biosciences, 550539), and counterstained with DAPI-containing VECTASHIELD mounting medium (Vector Laboratories) for fluorescence microscopy. Formalin-fixed, paraffin-embedded samples were sectioned, stained for phosho-histone H3 (CST, 9701), counterstained with hematoxylin, developed with the Vectastain Elite ABC staining kit (Vector Laboratories), and imaged using a Zeiss Axio Imager.M2 microscope. Images were acquired using Nuance 3.0.1.2 multispectral imaging software, and positive cells were identified and scored using inForm 1.4.0 Advanced Image Analysis software (PerkinElmer). For the Brd2 knockdown studies, p53 2.1.1 or FC1245 cells were transduced with an inducible lentiviral shRNA targeting Brd2 (mature antisense, TATCCGCTTTCCGTTTAAC; MOI, 2) in which the PGK promoter drives expression of the Tet-On 3G transactivator protein, and doxycycline induces expression of Brd2 shRNA and turboGFP (VSM6404-221925803, Dharmacon/GE Healthcare); the same lines were transduced with a lentiviral shRNA nontargeting control in parallel. Stable cell lines were generated following puromycin selection, and the inducible shBrd2 or shControl cell lines were used for orthotopic transplantation as described above. For the FVB/n-derived lines, 50,000 cells were transplanted into immune-competent FVB/n hosts; after 14 d, mice were imaged and randomized into two treatment groups. One group was administered drinking water containing 2 g/L doxycycline (D43020, Research Products International Corp.) containing 20 g/L sucrose (Sigma); the other group received regular drinking water. On day 28, mice were imaged, and on day 31, mice were killed and pancreata harvested, weighed, and fixed for histology. For the C57BL/6J-derived lines, 5,000 cells were transplanted into immune-competent C57BL/6J ROSA-rtTA*M2/+ hosts; after 11 d, mice were randomized and administered dox water or regular water for another 18 d and then collected.
BET Family siRNA Studies.
PDAC cells were seeded in six-well plates at a density of 3 × 105 cells per well. The next day, two wells per cell line were transfected with individual siRNAs targeting BRD2 (J-004935-06-0005), BRD3 (J-004936-05-0005), or BRD4 (J-004937-06-0005) or with a nontargeting control (D-001810–01-05; all ON-TARGETplus individual siRNAs, Dharmacon/GE Healthcare). The next day, 5 × 105 cells transfected with each siRNA were embedded into astromal or stromal hydrogels. The following day, media was replaced with DMEM containing 2% FBS (astromal) or CM containing 2% FBS (stromal), and the next day, RNA was collected for gene expression analysis. In parallel, cells transfected with the same siRNAs were maintained on plastic until 72 h posttransfection and were subject to nuclear extraction using the NE-PER nuclear and cytoplasmic extraction kit (ThermoFisher Scientific). Nuclear extracts were used for Western blot analysis of BET family proteins, as described above.
Western Blot Analysis.
Proteins were separated by SDS/PAGE and transferred to a PVDF membrane (Bio-Rad Laboratories, Inc.). Membranes were blocked in Tris-buffered saline containing 5% nonfat dry milk and 0.1% Tween 20 (TBS-T), before incubation with the primary antibody overnight at 4 °C. The membranes were then washed with TBS-T followed by exposure to the appropriate horseradish peroxidase-conjugated secondary antibody for 1 h and visualized on a ChemiDoc XRS+ (Bio-Rad Laboratories, Inc.) using the enhanced chemiluminescence (ECL Prime) detection system (Amersham/GE Healthcare). The following antibodies were used: Histone H3 (Abcam, ab1791), H3K9Ac (Abcam, ab4441), H3K27Ac (Active Motif, 39133), BRD2 (CST, 5848), BRD3 (SCBT, sc-81202), BRD4 (CST, 13440), HDAC1 (SCBT, sc-7872), and β-actin (Sigma, A2066).
Data Availability
Data deposition: RNA-Seq data can be accessed in the NCBI Sequence Read Archive (accession nos. SRP059641 and SRP059642), and ChIP-Seq data have been deposited in Gene Expression Omnibus (GEO) (accession no. GSE70971).
Acknowledgments
We thank P. Phojanakong, D. Wang, J. Alvarez, and H. Juguilon for technical assistance; J. Bradner for providing JQ1; and H. Stoppler and the UCSF Helen Diller Family Comprehensive Cancer Center Tissue Core for providing fresh-frozen human PDAC samples. This work was funded by NIH Grants DK057978, HL105278, DK090962, HL088093, ES010337, and CA014195 (to R.M.E.), K99CA188259 (to M.H.S.), and R01GM095567 (to A.C.K.); National Cancer Institute Grants R01CA15749, R01CA188048, and P01CA117969; ACS Research Scholar Grant RSG-13-298-01-TBG; the Glenn Foundation for Medical Research; Leona M. and Harry B. Helmsley Charitable Trust Grant 2012-PG-MED002; and Ipsen/Biomeasure. R.M.E. and M.D. were supported in part by a Stand Up to Cancer Dream Team Translational Cancer Research Grant, a Program of the Entertainment Industry Foundation (SU2C-AACR-DT0509). R.M.E. is an investigator of the Howard Hughes Medical Institute and March of Dimes Chair in Molecular and Developmental Biology at the Salk Institute and is supported by a grant from the Lustgarten Foundation and a gift from the Freeberg Foundation.
Supporting Information
Supporting Information (PDF)
Supporting Information
- Download
- 1.27 MB
pnas.1620164114.sd01.xlsx
- Download
- 11.84 KB
References
1
JD Potter, Morphogens, morphostats, microarchitecture and malignancy. Nat Rev Cancer 7, 464–474 (2007).
2
MJ Bissell, WC Hines, Why don’t we get more cancer? A proposed role of the microenvironment in restraining cancer progression. Nat Med 17, 320–329 (2011).
3
M Apte, RC Pirola, JS Wilson, Pancreatic stellate cell: Physiologic role, role in fibrosis and cancer. Curr Opin Gastroenterol 31, 416–423 (2015).
4
D Hanahan, LM Coussens, Accessories to the crime: Functions of cells recruited to the tumor microenvironment. Cancer Cell 21, 309–322 (2012).
5
H Ying, et al., Genetics and biology of pancreatic ductal adenocarcinoma. Genes Dev 30, 355–385 (2016).
6
J Riss, et al., Cancers as wounds that do not heal: Differences and similarities between renal regeneration/repair and renal cell carcinoma. Cancer Res 66, 7216–7224 (2006).
7
DF Quail, JA Joyce, Microenvironmental regulation of tumor progression and metastasis. Nat Med 19, 1423–1437 (2013).
8
C Guerra, et al., Chronic pancreatitis is essential for induction of pancreatic ductal adenocarcinoma by K-Ras oncogenes in adult mice. Cancer Cell 11, 291–302 (2007).
9
PP Provenzano, et al., Enzymatic targeting of the stroma ablates physical barriers to treatment of pancreatic ductal adenocarcinoma. Cancer Cell 21, 418–429 (2012).
10
MA Jacobetz, et al., Hyaluronan impairs vascular function and drug delivery in a mouse model of pancreatic cancer. Gut 62, 112–120 (2013).
11
C Feig, et al., Targeting CXCL12 from FAP-expressing carcinoma-associated fibroblasts synergizes with anti-PD-L1 immunotherapy in pancreatic cancer. Proc Natl Acad Sci USA 110, 20212–20217 (2013).
12
GL Beatty, et al., Exclusion of T cells from pancreatic carcinomas in mice is regulated by Ly6C F4/80 extra-tumor macrophages. Gastroenterology 149, 201–210 (2015).
13
MM Riopel, J Li, S Liu, A Leask, R Wang, β1 integrin-extracellular matrix interactions are essential for maintaining exocrine pancreas architecture and function. Lab Invest 93, 31–40 (2013).
14
JS Wilson, RC Pirola, MV Apte, Stars and stripes in pancreatic cancer: Role of stellate cells and stroma in cancer progression. Front Physiol 5, 52 (2014).
15
MH Sherman, et al., Vitamin D receptor-mediated stromal reprogramming suppresses pancreatitis and enhances pancreatic cancer therapy. Cell 159, 80–93 (2014).
16
M Waghray, et al., GM-CSF mediates mesenchymal-epithelial cross-talk in pancreatic cancer. Cancer Discov 6, 886–899 (2016).
17
CM Sousa, et al., Pancreatic stellate cells support tumour metabolism through autophagic alanine secretion. Nature 536, 479–483 (2016).
18
AD Rhim, et al., Stromal elements act to restrain, rather than support, pancreatic ductal adenocarcinoma. Cancer Cell 25, 735–747 (2014).
19
BC Özdemir, et al., Depletion of carcinoma-associated fibroblasts and fibrosis induces immunosuppression and accelerates pancreas cancer with reduced survival. Cancer Cell 25, 719–734 (2014).
20
JJ Lee, et al., Stromal response to Hedgehog signaling restrains pancreatic cancer progression. Proc Natl Acad Sci USA 111, E3091–E3100 (2014).
21
SF Boj, et al., Organoid models of human and mouse ductal pancreatic cancer. Cell 160, 324–338 (2015).
22
H Ying, et al., Oncogenic Kras maintains pancreatic tumors through regulation of anabolic glucose metabolism. Cell 149, 656–670 (2012).
23
LJ Bayne, et al., Tumor-derived granulocyte-macrophage colony-stimulating factor regulates myeloid inflammation and T cell immunity in pancreatic cancer. Cancer Cell 21, 822–835 (2012).
24
Y Pylayeva-Gupta, KE Lee, CH Hajdu, G Miller, D Bar-Sagi, Oncogenic Kras-induced GM-CSF production promotes the development of pancreatic neoplasia. Cancer Cell 21, 836–847 (2012).
25
JJ Kamphorst, et al., Human pancreatic cancer tumors are nutrient poor and tumor cells actively scavenge extracellular protein. Cancer Res 75, 544–553 (2015).
26
S Dangi-Garimella, V Sahai, K Ebine, K Kumar, HG Munshi, Three-dimensional collagen I promotes gemcitabine resistance in vitro in pancreatic cancer cells through HMGA2-dependent histone acetyltransferase expression. PLoS One 8, e64566 (2013).
27
JV Lee, et al., Akt-dependent metabolic reprogramming regulates tumor cell histone acetylation. Cell Metab 20, 306–319 (2014).
28
JE Delmore, et al., BET bromodomain inhibition as a therapeutic strategy to target c-Myc. Cell 146, 904–917 (2011).
29
F Winston, CD Allis, The bromodomain: A chromatin-targeting module? Nat Struct Biol 6, 601–604 (1999).
30
C Dhalluin, et al., Structure and ligand of a histone acetyltransferase bromodomain. Nature 399, 491–496 (1999).
31
MA Dawson, et al., Inhibition of BET recruitment to chromatin as an effective treatment for MLL-fusion leukaemia. Nature 478, 529–533 (2011).
32
PK Mazur, et al., Combined inhibition of BET family proteins and histone deacetylases as a potential epigenetics-based therapy for pancreatic ductal adenocarcinoma. Nat Med 21, 1163–1171 (2015).
33
L Zeng, MM Zhou, Bromodomain: An acetyl-lysine binding domain. FEBS Lett 513, 124–128 (2002).
34
B Florence, DV Faller, You bet-cha: A novel family of transcriptional regulators. Front Biosci 6, D1008–D1018 (2001).
35
N Ding, et al., A vitamin D receptor/SMAD genomic circuit gates hepatic fibrotic response. Cell 153, 601–613 (2013).
36
N Roy, et al., Brg1 promotes both tumor-suppressive and oncogenic activities at distinct stages of pancreatic cancer formation. Genes Dev 29, 658–671 (2015).
37
D Gosselin, et al., Environment drives selection and function of enhancers controlling tissue-specific macrophage identities. Cell 159, 1327–1340 (2014).
38
Y Lavin, et al., Tissue-resident macrophage enhancer landscapes are shaped by the local microenvironment. Cell 159, 1312–1326 (2014).
39
G Yosimichi, et al., CTGF/Hcs24 induces chondrocyte differentiation through a p38 mitogen-activated protein kinase (p38MAPK), and proliferation through a p44/42 MAPK/extracellular-signal regulated kinase (ERK). Eur J Biochem 268, 6058–6065 (2001).
40
RM Day, V Cioce, D Breckenridge, P Castagnino, DP Bottaro, Differential signaling by alternative HGF isoforms through c-Met: Activation of both MAP kinase and PI 3-kinase pathways is insufficient for mitogenesis. Oncogene 18, 3399–3406 (1999).
41
RC Baxter, IGF binding proteins in cancer: Mechanistic and clinical insights. Nat Rev Cancer 14, 329–341 (2014).
42
N Kiuchi, et al., STAT3 is required for the gp130-mediated full activation of the c-myc gene. J Exp Med 189, 63–73 (1999).
43
T Hirano, K Ishihara, M Hibi, Roles of STAT3 in mediating the cell growth, differentiation and survival signals relayed through the IL-6 family of cytokine receptors. Oncogene 19, 2548–2556 (2000).
44
DG Nowak, et al., MYC drives Pten/Trp53-deficient proliferation and metastasis due to IL6 secretion and AKT suppression via PHLPP2. Cancer Discov 5, 636–651 (2015).
45
EA Collisson, et al., A central role for RAF→MEK→ERK signaling in the genesis of pancreatic ductal adenocarcinoma. Cancer Discov 2, 685–693 (2012).
46
A Dobin, et al., STAR: Ultrafast universal RNA-seq aligner. Bioinformatics 29, 15–21 (2013).
47
C Trapnell, et al., Differential analysis of gene regulation at transcript resolution with RNA-seq. Nat Biotechnol 31, 46–53 (2013).
48
A Roberts, C Trapnell, J Donaghey, JL Rinn, L Pachter, Improving RNA-Seq expression estimates by correcting for fragment bias. Genome Biol 12, R22 (2011).
49
AJ Saldanha, Java Treeview--Extensible visualization of microarray data. Bioinformatics 20, 3246–3248 (2004).
50
J Wang, D Duncan, Z Shi, B Zhang, WEB-based GEne SeT AnaLysis Toolkit (WebGestalt): Update 2013. Nucleic Acids Res 41, W77–W83 (2013).
51
B Zhang, S Kirov, J Snoddy, WebGestalt: An integrated system for exploring gene sets in various biological contexts. Nucleic Acids Res 33, W741–W748 (2005).
52
A Hsiao, DS Worrall, JM Olefsky, S Subramaniam, Variance-modeled posterior inference of microarray data: Detecting gene-expression changes in 3T3-L1 adipocytes. Bioinformatics 20, 3108–3127 (2004).
53
A Subramanian, et al., Gene set enrichment analysis: A knowledge-based approach for interpreting genome-wide expression profiles. Proc Natl Acad Sci USA 102, 15545–15550 (2005).
54
VK Mootha, et al., PGC-1alpha-responsive genes involved in oxidative phosphorylation are coordinately downregulated in human diabetes. Nat Genet 34, 267–273 (2003).
55
GD Barish, et al., Bcl-6 and NF-kappaB cistromes mediate opposing regulation of the innate immune response. Genes Dev 24, 2760–2765 (2010).
56
S Heinz, et al., Simple combinations of lineage-determining transcription factors prime cis-regulatory elements required for macrophage and B cell identities. Mol Cell 38, 576–589 (2010).
Information & Authors
Information
Published in
Classifications
Copyright
Freely available online through the PNAS open access option.
Data Availability
Data deposition: RNA-Seq data can be accessed in the NCBI Sequence Read Archive (accession nos. SRP059641 and SRP059642), and ChIP-Seq data have been deposited in Gene Expression Omnibus (GEO) (accession no. GSE70971).
Submission history
Published online: January 17, 2017
Published in issue: January 31, 2017
Keywords
Acknowledgments
We thank P. Phojanakong, D. Wang, J. Alvarez, and H. Juguilon for technical assistance; J. Bradner for providing JQ1; and H. Stoppler and the UCSF Helen Diller Family Comprehensive Cancer Center Tissue Core for providing fresh-frozen human PDAC samples. This work was funded by NIH Grants DK057978, HL105278, DK090962, HL088093, ES010337, and CA014195 (to R.M.E.), K99CA188259 (to M.H.S.), and R01GM095567 (to A.C.K.); National Cancer Institute Grants R01CA15749, R01CA188048, and P01CA117969; ACS Research Scholar Grant RSG-13-298-01-TBG; the Glenn Foundation for Medical Research; Leona M. and Harry B. Helmsley Charitable Trust Grant 2012-PG-MED002; and Ipsen/Biomeasure. R.M.E. and M.D. were supported in part by a Stand Up to Cancer Dream Team Translational Cancer Research Grant, a Program of the Entertainment Industry Foundation (SU2C-AACR-DT0509). R.M.E. is an investigator of the Howard Hughes Medical Institute and March of Dimes Chair in Molecular and Developmental Biology at the Salk Institute and is supported by a grant from the Lustgarten Foundation and a gift from the Freeberg Foundation.
Authors
Competing Interests
Conflict of interest statement: M.H.S., R.T.Y., N.D., A.R.A., M.D., and R.M.E. are coinventors of technologies using BET inhibitors and may be entitled to royalties. A.C.K. is an inventor on patents pertaining to Kras regulated metabolic pathways; redox control pathways in pancreatic cancer, targeting GOT1 as a therapeutic approach; and the autophagic control of iron metabolism. A.C.K. is on the Scientific Advisory Board of Cornerstone Pharmaceuticals and is a founder of Vescor Therapeutics.
Metrics & Citations
Metrics
Citation statements
Altmetrics
Citations
If you have the appropriate software installed, you can download article citation data to the citation manager of your choice. Simply select your manager software from the list below and click Download.
Cited by
Loading...
View Options
View options
PDF format
Download this article as a PDF file
DOWNLOAD PDFGet Access
Login options
Check if you have access through your login credentials or your institution to get full access on this article.
Personal login Institutional LoginRecommend to a librarian
Recommend PNAS to a LibrarianPurchase options
Purchase this article to get full access to it.