Tumor-Derived CCL2 Mediates Resistance to Radiotherapy in Pancreatic Ductal Adenocarcinoma

Clinical outcomes in pancreatic ductal adenocarcinoma (PDAC) are poor with a 5-year overall survival of only 7% (1, 2). Although metastasis is the most common cause of mortality, 30% of patients will succumb to disease due to complications of local tumor growth (3). In addition to its effect on mortality, local tumor growth can also cause significant morbidity including obstructive biliary sepsis, duodenal obstruction, and intractable pain related to celiac plexus involvement (4–6).

Approximately 25% of patients with PDAC will present with localized but unresectable disease (7); for these patients, radiotherapy is a major treatment modality for local control. However, local progression despite radiotherapy is common (8–10). Recently, stereotactic ablative body radiotherapy (SBRT or SABR) has garnered significant attention as a means to deliver a higher biologically effective radiation dose (11–16). Ablative radiotherapy has also been found to alter the immune response to cancer (17). In particular, for immunologically sensitive models of cancer, ablative radiotherapy can increase T-cell infiltration leading to durable remissions, as shown in a subcutaneous murine model of colon cancer (18). However, the impact of ablative radiotherapy on the immune microenvironment of immunologically resistant tumors, such as PDAC, remains ill-defined.

In PDAC, the tumor microenvironment is characterized by poor infiltration of effector T cells and a predominance of innate inflammatory cells, including macrophages (19, 20). This microenvironment is profoundly immunosuppressive and establishes a site of T-cell immune privilege (21). In addition, the microenvironment that surrounds PDAC is supportive of cancer growth, invasion, and metastasis (22, 23). However, the immune reaction to PDAC is inherently pliable such that myeloid cells responding to PDAC can also be induced with antitumor properties capable of enhancing the efficacy of standard cytotoxic therapies (19, 24).

In this study, we examined the effect of ablative radiotherapy on the immune response to PDAC. Unlike immunologically sensitive tumors, radiotherapy did not induce T-cell infiltration into tumor tissue. In contrast, PDAC cells responded to radiotherapy by releasing inflammatory cytokines and chemokines including CCL2. Tumor-derived CCL2 induced an increase in inflammatory monocytes and macrophages within the tumor microenvironment and inhibited antitumor activity mediated by radiotherapy. In contrast, neutralization of CCL2 blocked inflammatory monocyte recruitment to tumors and resulted in improved tumor control when administered in combination with radiotherapy.

C57BL/6 mice were purchased from The Jackson Laboratories. Animal protocols were reviewed and approved by the Institute of Animal Care and Use Committee of the University of Pennsylvania (Philadelphia, PA).

Mouse pancreatic cancer cell lines were derived from spontaneously arising tumors in Kras^LSL-G12D/+, Trp53^LSL-R172H/+, Pdx1-Cre (KPC) mice as described previously (24, 25). Cell lines were authenticated on the basis of histologic analysis of the implanted cell line in comparison with the primary tumor from which the cell line was derived as described previously (24). Cell lines were tested for mycoplasma contamination, cultured at 37°C in DMEM supplemented with 10% FCS, 83 μg/mL gentamicin, and 1% l-glutamine, and used in experiments between 6 and 8 passages.

PDAC cell lines were implanted subcutaneously at 4–5 × 10⁵ cells into syngeneic C57BL/6 mice. For orthotopic implantation of tumor cells, syngeneic C57BL/6 mice were first anesthetized and the abdomen prepared in a sterile fashion. A small (5–10 mm) incision was made over the left upper quadrant of the abdomen, and the peritoneal cavity was exposed. The pancreas was then located and exteriorized onto a sterile field. PDAC cell lines (5 × 10⁵ cells) were implanted into the tail of the pancreas. The pancreas was then placed back into the peritoneal cavity, and the peritoneum and skin were closed with suture and wound clips, respectively. Tumors were allowed to develop over 14–17 days to approximately 5 mm in diameter. Established tumors were irradiated in a single fraction (14–20 Gy) using the Small Animal Radiation Research Platform (SARRP). Anti-CCL2 (clone 2H5) neutralizing antibody, anti-Ly6C (clone Monts1) depleting antibody, hamster isotype control (hamster IgG), and rat isotype control (clone 2A3) were administered via intraperitoneal injection on days −1, 0, +1, and +3 of radiotherapy. Anti-CD4 (clone GK1.5) and anti-CD8 (clone 2.43) depleting antibodies were administered on day −1. All neutralizing and depleting antibodies were purchased from BioXcell and were endotoxin free. Every 3 to 4 days, the longest tumor dimension (L) and its perpendicular diameter (W) were measured using calipers; volume was calculated as (L × W²)/2. For survival studies, mice were euthanized when tumors reached 1,000 mm³.

Histopathology and IHC were performed on frozen tissue sections. Frozen sections were air dried and fixed with 3% formaldehyde. Hydrogen peroxide was used to quench endogenous peroxidases. Tissue was blocked with 10% normal goat serum in PBS with 0.1% Tween-20. Primary antibodies against mouse antigens included rat anti-F4/80 (eBioscience), rat anti-Ly6C (Abcam), rat anti-Ki67 clone TEC3 (Dako), rat anti-CD31 (BD Pharmingen), rat anti-CD45 (BD Pharmingen), and rabbit anti-cleaved caspase-3 (R&D Systems). Tissues were incubated in blocking buffer containing primary antibodies overnight at 4°C. After washing with PBS, tissues were incubated with PBS containing a secondary biotinylated goat anti-rat antibody (BD Pharmingen). Tissues were developed using Vectastain ABC Kit and counterstained with hematoxylin. Senescence-associated β-galactosidase staining was performed using the Senescence β-Galactosidase Staining Kit (Cell Signaling Technology). Fluorescent imaging was performed on an IX83 inverted microscope (Olympus) and bright-field images were acquired using a BX43 upright (Olympus) microscope.

Mice were euthanized, and subcutaneous tumors were harvested, measured, and weighed. Peritumoral lymph nodes were excluded from analysis. Tumors were minced at 4°C in digestion media containing collagenase (1 mg/mL, Sigma-Aldrich), DNase (150 U/mL, Roche), and Dispase (1 U/mL, Worthington), and then incubated at 37°C for 30 minutes with intermittent agitation. After 70-μm filtration and washing with PBS, single-cell suspensions were stained with antibodies in PBS containing 0.2 mmol/L EDTA with 2% FCS at 4°C for 15 minutes. Anti-mouse antibodies were purchased from BD Biosciences, unless stated otherwise: CD45 (30-F11, PE-Cy7), CD11b (BioLegend, M1/70, APC), F4/80 (eBiosciences; BM8, FITC), Ly6C (AL-21, APC-Cy7), Ly6G (1A8, Percp-Cy5.5), CCR2 (R&D Systems; clone #475301, PE). Antibody-stained single-cell suspensions were examined on a FACSCanto II (BD Biosciences) and analyzed using FlowJo version 10 (FlowJo, LLC). For gene expression analysis of leukocyte subsets obtained from implanted tumors, antibody-stained single-cell suspensions were sorted on a FACSAria II (BD Biosciences) and RNA isolation was performed within 5 to 6 hours of the initial tumor digest. This protocol required 5 × 10⁴ sorted cells per sample. Gating strategy for inflammatory monocytes (IM), granulocytes, and tumor-associated macrophages (TAM) is illustrated in Supplementary Fig. S1.

For in vivo experiments, tumors were harvested, placed at 4°C in serum-free DMEM at 1 mg of tissue per 10 μL of media, and then minced. Tumor suspensions were centrifuged at 12,470 × g for 5 minutes, and supernatant was collected and stored at −20°C. For in vitro experiments, when tumor cell lines reached 70% to 80% confluence in 10-mm plates, cells were washed and incubated in fresh serum-free DMEM at 37°C; supernatant was then collected after 24 hours and stored at −20°C. Cytokines from in vitro and in vivo tumor supernatants were quantified using cytometric bead analysis (CBA; BD Biosciences), using references to recombinant murine standards.

Bone marrow–derived cells (2 × 10⁶/mL) from C57BL/6 mice were placed above a transwell membrane in DMEM containing 1% FCS, which was incubated in tumor supernatant collected in vitro as described above, in the presence or absence of a CCL2-neutralizing antibody (2H5, 10 ng/mL). After incubation at 37°C for 5 hours, transwell membranes were collected, fixed with formaldehyde, stained with crystal violet, and dried. Transmigrated cells were counted at 40× magnification using an upright bright-field microscope (Olympus BX43).

PDAC cell lines at 70% to 80% confluence were cultured in DMEM containing 5% FCS at 0.5-cm depth and irradiated at a dose rate of 2.8 Gy/minute using the X-RAD 320ix (Precision X-ray, Inc.). Sham irradiation involved placing cell culture plates at a similar temperature for the length of irradiation.

Tumor tissue was processed and stored in TRIzol at −80°C. Tumor lysates were thawed on ice and allowed to equilibrate to room temperature before RNA was isolated using a Qiagen RNeasy Mini Kit, according to manufacturer's protocol. For in vitro experiments, tumor cells were washed and harvested using TRIzol. Flow-sorted samples were collected in TRIzol LS and RNA extraction was performed immediately. RNA was collected in RNase-free water and quantified using a NanoDrop Spectrophotometer. cDNA was synthesized from 1 μg of RNA per sample using MultiScribe Reverse Transcriptase and random hexamers (Applied Biosystems). Primers for qRT-PCR were designed using the Primer 3 online program (26, 27) and synthesized by Integrated DNA Technologies. SYBR Green chemistry (Applied Biosystems) was used to determine relative quantifications of all products; expression was normalized to β-actin and calculated using the ΔC_t formula. Fold change was calculated relative to the average of the control group. Primer sequences were as follows:

•  Murine Ccl2: forward 5′-AGCAGCAGGTGTCCCAAAGA-3′

•  Murine Ccl2: reverse 5′-GATCTCATTTGGTTCCGATCCA-3′

•  Human Ccl2: forward 5′-GAAGAATCACCAGCAGCAAGT-3′

•  Human Ccl2: reverse 5′-TCCTGAACCCACTTCTGCTT-3′

Microarray services were provided by the UPENN Molecular Profiling Facility, including quality control tests of the total RNA samples by Agilent Bioanalyzer and Nanodrop spectrophotometry. All protocols were conducted as described in the NuGEN Ovation Pico WTA system v2 user guide and the Affymetrix GeneChip Expression Analysis Technical Manual. Briefly, 25 ng of total RNA was converted to first-strand cDNA using reverse transcriptase primed by poly(T) and random oligomers that incorporated an RNA priming region. Second-strand cDNA synthesis was followed by ribo-SPIA linear amplification of each transcript using an isothermal reaction with RNase, RNA primer, and DNA polymerase, and the resulting ssDNA was assessed by Bioanalyzer, fragmented, and biotinylated by terminal transferase end labeling. Labeled cDNA (5.5 μg) were added to Affymetrix hybridization cocktails, heated at 99°C for 5 minutes and hybridized for 16 hours at 45°C to Mouse Gene 2.0 ST GeneChips (Affymetrix Inc.) using the GeneChip Hybridization oven 645. The microarrays were then washed at low (6× SSPE) and high (100 mmol/L MES, 0.1 mol/L NaCl) stringency and stained with streptavidin–phycoerythrin. Fluorescence was amplified by adding biotinylated anti-streptavidin and an additional aliquot of streptavidin–phycoerythrin stain. A GeneChip 3000 7G scanner was used to collect fluorescence signal. Affymetrix Command Console and Expression Console were used to quantitate expression levels for targeted genes; default values provided by Affymetrix were applied to all analysis parameters. Heatmaps were generated using GENE-E matrix visualization and analysis platform. Gene expression data is available through the NCBI GEO database (accession number GSE82276).

Oligonucleotides (GCAAGATGATCCCAATGAGT and GGCCTGCTGTTCACAGTTGC, Integrated DNA Technologies) encoding guide RNAs that target murine CCL2 exons were cloned into TOPO-expression vectors (Life Technology). A clonal tumor line, derived from a parental KPC-derived PDAC cell line, was then cotransfected using 5 μg LentiCRISPR V2 plasmid (Addgene) with 5 μg of TOPO-CCL2 gRNA plasmids or a control vector targeting GFP. One day after transfection, transfected cells were selected for 48 hours with 10 ng/mL puromycin. Isogenic clones of transfected cells were isolated to produce clonal lines, in which CCL2 knockout was confirmed by cytometric bead array (CBA; BD Biosciences) of tumor cell supernatant.

Statistical significance was determined by Student t test performed with or without Welch correction for unequal variance using GraphPad Prism Version 6.05 (GraphPad Software, Inc.). Significance testing for tumor growth over time was performed using repeated measures two-way ANOVA with Tukey multiple comparisons of means as a post hoc test to evaluate differences between two groups. In some cases, tumor volumes were compared at each time point using Student t test with Holm–Šídák method to correct for multiple comparisons. Significance of overall survival was determined using log-rank analysis.

To investigate the impact of ablative radiotherapy in an immunologically resistant tumor model (21), we derived PDAC cell lines from tumors arising spontaneously in KPC mice and injected them subcutaneously into syngeneic C57BL/6 mice. Approximately 2 weeks later, established tumors were treated with 20 Gy ablative radiotherapy, a dose that is comparable with that used clinically for human PDAC (13). Similar to results reported in patients with PDAC, we found that a single dose of radiotherapy only partially delayed tumor growth (Fig. 1A). However, in contrast to findings in immunogenic murine melanoma and colon cancer models (18, 28), depletion of CD4⁺ and CD8⁺ T cells did not have an impact on the delay in tumor growth induced by radiotherapy (Fig. 1A). Furthermore, we observed no significant difference in T-cell infiltration into tumors after radiotherapy (Supplementary Fig. S2). These findings revealed that ablative radiotherapy alone is insufficient to invoke a productive antitumor immune response against an immunologically resistant tumor such as PDAC.

To investigate potential immune mechanisms regulating the efficacy of ablative radiotherapy, we next examined the intratumoral cytokine profile induced by radiotherapy. At baseline, tumors demonstrated high levels of CCL2, MIG, and G-CSF among a panel of cytokines (Fig. 1B), which is consistent with the robust inflammatory response seen in PDAC (29). However, after radiotherapy, there were small but significant increases in the levels of CXCL1, IL1β, and IL6. Most notably, although, there was a marked increase in CCL2 (Δ = 1,178 ± 294 pg/mL; Fig. 1B). This increase in CCL2 levels in response to radiotherapy was also seen at the RNA level (Fig. 1C) and was reproducible using a second KPC-derived PDAC cell line in vivo (Supplementary Fig. S3).

We next examined the capacity of tumor cells to secrete CCL2 in response to radiotherapy. We found that Ccl2 expression increased in vitro in response to radiotherapy (Fig. 1C), suggesting that tumor cells, not just stromal or immune cells, likely contribute to the increase in CCL2 within the tumor microenvironment. This same radiotherapy-induced CCL2 release was also seen with two human PDAC cell lines. Here, radiotherapy augmented CCL2 expression in both MIA PaCa-2 and AsPC-1 human PDAC cell lines by 2.8 ± 0.5-fold and 5.1 ± 0.8-fold, respectively (Fig. 1D). Thus, PDAC responds to radiotherapy by releasing CCL2.

Consistent with clinical data suggesting that CCL2 expression is increased in human PDAC (30), we detected high levels of CCL2 produced in vitro by an array of KPC-derived PDAC cell lines. In contrast, other chemokines important for the recruitment of myeloid cells, such as CCL5, CXCL1, CCL3, and CCL4, were not uniformly detected (Fig. 2A).

We next used an in vitro chemotaxis assay to understand the significance of tumor-derived factors in recruiting myeloid cells to the tumor microenvironment. We found that soluble factors produced by PDAC cell lines induced the migration of bone marrow–derived cells across a transwell membrane (Fig. 2B). The magnitude of cell migration induced by PDAC was comparable with myeloid cell migration toward 10 ng/mL CCL2 (Fig. 2B). Furthermore, bone marrow–derived cells that migrated across the transwell membrane were enriched for CCR2 expression (Supplementary Fig. S4). To determine the role of tumor-derived CCL2 in directing myeloid cell chemotaxis, we neutralized CCL2 in vitro using a CCL2-specific antibody and found a significant reduction in the migration of bone marrow–derived cells (Fig. 2C) demonstrating a major role for PDAC-derived CCL2 for myeloid cell chemotaxis.

Given the effect of ablative radiotherapy on intratumoral chemokine levels, we examined the dynamics of myeloid cell composition within the tumor microenvironment after radiotherapy. By H&E staining, we detected an increase in cellular content 3 days after radiotherapy, suggesting an acute inflammatory response (Fig. 3A). On more detailed examination of the immune microenvironment, we detected an increase in the frequency of F4/80⁺ monocytes/macrophages within tumors (Fig. 3B). Using flow cytometry, we also found an increase in leukocytes as a percentage of live cells within tumors after radiotherapy (Fig. 3C). Specifically, we observed an increase in TAMs (CD45⁺CD11b⁺Ly6C^loF4/80⁺) and inflammatory monocytes/macrophages (CD45⁺CD11b⁺Ly6C^hi). In contrast, we detected no change in granulocyte (CD45⁺CD11b⁺Ly6G⁺) recruitment. Similar findings were observed in an orthotopic model of PDAC, where ablative radiotherapy induced an increase in inflammatory monocytes/macrophages without a change in granulocyte infiltration (Supplementary Fig. S5).

The increase in inflammatory monocyte/macrophage infiltration seen after radiotherapy was consistent with the increase in intratumoral CCL2 levels induced by radiotherapy (Fig. 1). Inflammatory monocytes/macrophages express CCR2, the receptor for CCL2, which can direct both the egress of immature inflammatory monocytes from the bone marrow into the peripheral blood and their subsequent infiltration into tissues (24, 31). Therefore, we hypothesized that inflammatory monocytes/macrophages may respond to radiation-induced injury by acquiring an altered phenotype that regulates the efficacy of ablative radiotherapy. To test this hypothesis, we examined the gene expression profile of flow-sorted inflammatory monocytes/macrophages isolated from tumors 1 day after ablative radiotherapy compared with control tumors. This analysis revealed only eight of 34,365 transcripts that reached a predetermined threshold of twofold change in expression with a <25% FDR (Supplementary Fig. S6). Four of these, including Phlda3, Ccna1, Ddias, and Trp53inp1 are recognized as cell-cycle or stress response genes. However, none of these genes is known as major determinants of cellular phenotype to suggest a role in redirecting the biology of tumor-infiltrating monocytes/macrophages. Therefore, the minimal impact of ablative radiotherapy on the gene expression profile of inflammatory monocytes/macrophages supports a role for radiotherapy primarily in altering the quantity rather than the quality of inflammatory monocytes/macrophages recruited to the tumor microenvironment.

CCL2 expression and inflammatory monocyte recruitment have been implicated as poor prognostic factors in a variety of malignancies (30, 32, 33). We hypothesized that radiotherapy-induced expression of CCL2 within the tumor microenvironment may recruit inflammatory monocytes/macrophages and TAMs to promote resistance to ablative radiotherapy. To test this hypothesis, mice bearing established PDAC tumors injected subcutaneously were treated with a neutralizing antibody to CCL2 (or control) on days −1, 0, +1, and +3 relative to radiotherapy versus sham treatment. Compared to radiotherapy alone, the combination of radiotherapy and a CCL2-neutralizing antibody significantly impacted tumor growth and prolonged survival (Fig. 4A). This effect was reproduced using a second KPC-derived PDAC cell line (Supplementary Fig. S7).

Although some tumor cells have been reported to express CCR2, we found that CCR2 expression within PDAC tumors was limited to a subset of CD45⁺ leukocytes (Supplementary Fig. S8). In addition, CCL2 neutralization alone did not impact tumor growth (Fig. 4A). Therefore, to understand the impact of CCL2 neutralization on the efficacy of radiotherapy, we examined the myeloid compartment of tumors 3 days after radiotherapy. We found that the acute increase in inflammatory monocytes induced by radiotherapy was abrogated by CCL2 neutralization in vivo (Fig. 4B). In contrast, the presence of granulocytes within tumors was not affected by radiotherapy or the combination of radiotherapy and CCL2 neutralization (Fig. 4B).

Our findings with CCL2 neutralization suggested a key role for inflammatory monocytes/macrophages as an immune resistance mechanism to ablative radiotherapy in PDAC. To further address this possibility, we depleted Ly6C^hi inflammatory monocytes/macrophages using a Ly6C-depleting antibody (24). Using this approach, we found that depletion of inflammatory monocytes significantly delayed tumor growth after radiotherapy, and mirrored the impact of CCL2 neutralization (Fig. 4C). To understand the mechanism by which inflammatory monocytes promote tumor growth after radiotherapy, we examined tumor specimens by IHC. After radiotherapy alone there was no change in vascularity or proliferation, but increased apoptosis (cleaved caspase-3) and a trend toward increased senescence (Fig. 4D). Compared with radiotherapy alone, we found that the combination of radiotherapy and CCL2 neutralization significantly reduced tumor vascularity (decrease in CD31⁺ cells per high-power field) and proliferation (decreased Ki67⁺ cells per high-power field) without further having an impact on apoptosis (cleaved caspase-3) or senescence (β-galactosidase; Fig. 4D). Furthermore, in contrast to immunogenic cancer models where CCL2 neutralization can induce T-cell infiltration, we found that CCL2 neutralization with or without radiotherapy failed to lead to recruitment of CD3⁺ T cells to the tumor microenvironment (Supplementary Fig. S9). Thus, PDAC responds to ablative radiotherapy by releasing CCL2, which then recruits inflammatory monocytes/macrophages to support tumor proliferation and vascularization.

Both malignant and nonmalignant cells can secrete CCL2 within the tumor microenvironment. In a breast cancer metastasis model, CCL2 produced by both malignant and nonmalignant cells was found to be important for recruiting inflammatory monocytes/macrophages to promote metastasis (32). However, our findings in models of PDAC demonstrate that malignant cells respond to ablative radiotherapy by significantly increasing CCL2 production. This raised the possibility that CCL2 produced by malignant cells may be the main mediator of immune resistance to radiotherapy. Therefore, we next examined the role of tumor-derived CCL2 in promoting resistance to radiotherapy. To do this, we genetically eliminated CCL2 from a clonal KPC-derived PDAC cell line using CRISPR technology. As expected, compared with the wild-type clonal cell line (CCL2^wt), the CCL2^null cell line did not produce CCL2 (Fig. 5A). We next examined the biology of CCL2^null versus CCL2^wt tumors implanted into syngeneic mice. We found that levels of CCL2 were significantly decreased in CCL2^null PDAC tumors, although not completely eliminated (Fig. 5A), which likely represents nonmalignant sources of CCL2. In contrast, we did not observe any difference in the engraftment or outgrowth of CCL2^null compared with CCL2^wt tumors (Fig. 5B). Finally, we determined the capacity of radiotherapy to have an impact on the in vivo growth of CCL2^null versus CCL2^wt tumors in vivo. We found that the responsiveness of CCL2^null tumors to radiotherapy was significantly increased compared with CCL2^wt tumors as seen by deeper tumor regressions (Fig. 5C). In addition, consistent with the importance of tumor-derived CCL2 in regulating the efficacy of ablative radiotherapy, CCL2 neutralization did not improve radiotherapy efficacy in CCL2^null tumors (Fig. 5D), as had been seen in CCL2-expressing tumors (Fig. 4A). Together, our findings support a role for tumor-derived CCL2 as a key mechanism of resistance to ablative radiotherapy in PDAC.

Local disease progression can be a major complication of unresectable PDAC with patients developing (i) sepsis secondary to biliary obstruction, (ii) hemorrhage and bowel obstruction related to duodenal invasion, and (iii) intractable pain from celiac plexus involvement. Overall, nearly one-third of patients with PDAC will die from uncontrolled local disease (3). Although radiotherapy remains a mainstay for disease control in the locally advanced setting, human PDAC has demonstrated significant resistance to radiotherapy. This resistance has been attributed to an inherent radioresistance of tumor cells, increased hypoxia, poor vascularity, and dense fibrosis (34–37). Our findings now identify a novel mechanism of radiotherapy resistance wherein PDAC cells appear inherently programmed to respond to radiotherapy-induced stress by releasing high levels of CCL2 that recruit inflammatory monocytes to promote tumor proliferation and vascularity (Fig. 6). Thus, our data support a role for CCL2-dependent inflammation as a major checkpoint to the efficacy of radiotherapy in PDAC.

In this study, we found that ablative radiotherapy, although limited in its antitumor efficacy, induced the accumulation of chemokines and cytokines within the tumor microenvironment, in particular, CCL2. Although chemokine production within the tumor microenvironment can be driven by both malignant and nonmalignant cells, in vitro studies of murine and human PDAC cell lines showed that tumor cells respond to the stress of ablative radiotherapy by augmenting CCL2 expression. Tumor-derived CCL2 was found to be critical for inducing chemotaxis of inflammatory monocytes and directed the recruitment of inflammatory monocytes/macrophages and TAMs to the tumor microenvironment. We found that selective blockade of inflammatory monocyte recruitment, although ineffective alone, improved the antitumor effect of ablative radiotherapy and prolonged survival in a murine model of PDAC. This approach was contingent on tumor-derived CCL2 expression. Using cell lines in which CCL2 was genetically deleted, we observed that tumor-derived CCL2 was necessary for the antitumor effect of CCL2 neutralization in combination with radiotherapy, despite continued expression of CCL2 by nonmalignant cells within the tumor microenvironment. This finding suggests that the proximity of inflammatory monocytes/macrophages to tumor cells which may be directed by tumor-derived CCL2 is critical for monocyte/macrophage protumor activity in the setting of radiotherapy (38).

Using a poorly immunogenic model of PDAC, we found that T cells were not required for CCL2 neutralization to improve the efficacy of ablative radiotherapy. Rather, our findings demonstrate that CCL2 neutralization inhibits inflammatory monocyte recruitment to tumors and their subsequent role in supporting tumor recovery from the cytotoxic effects of radiotherapy. This finding contrasts the use of CCL2–CCR2 blockade in other more immunogenic tumors, where it has been shown to induce T-cell–dependent antitumor immunity when administered in combination with cytotoxic chemotherapy (30, 33). In addition, our findings contrast recent reports in immunogenic cancer models describing a role for T cells in the antitumor effects of ablative radiotherapy (18, 28). These distinct outcomes observed with ablative radiotherapy across immunogenic and poorly immunogenic models suggest a role for baseline tumor immunogenicity in defining rational therapeutic combinations with radiotherapy. As the PDAC tumor microenvironment commonly excludes effector T cells, therapeutic strategies that combine ablative radiotherapy with T-cell–based therapies such as immune checkpoint blockade may be unlikely to be effective without first restoring productive T-cell immunosurveillance, such as with vaccines (39). Our findings in a poorly immunogenic model of PDAC demonstrate that ablative radiotherapy alone is insufficient to overcome T-cell exclusion and, thus, is unlikely to potentiate the activity of immunotherapies that rely on a preexisting T-cell response.

Inflammatory monocytes and the CCL2–CCR2 axis are commonly associated with a poor prognosis in several malignancies, including PDAC, hepatocellular carcinoma, and breast cancer (30, 32, 33). Our findings suggest that radiotherapy does not alter the phenotype of tumor-infiltrating monocytes/macrophages but rather enhances their recruitment to the tumor microenvironment. We performed this analysis on tumor-infiltrating monocytes 1 day after radiotherapy reasoning that an early time point would be necessary to capture changes in gene expression prior to their differentiation within the tumor microenvironment. Under basal conditions, inflammatory monocytes/macrophages appear to comprise a small subset of leukocytes in the PDAC microenvironment, but unlike granulocytes and TAMs, this subset expresses the chemokine receptor CCR2. Inflammatory monocytes recruited to the tumor microenvironment in a CCL2–CCR2- dependent manner can undergo rapid differentiation into TAMs and in doing so, lose Ly6C expression. This is consistent with inflammatory monocytes/macrophages as precursors of TAMs and implies that their recruitment to the tumor microenvironment in response to ablative radiotherapy may help reconstitute the myeloid compartment. However, the transient nature of Ly6C expression may lead to an underestimation of the magnitude of this recruitment when examined on a quantitative basis.

We have shown a therapeutic role for inhibiting monocyte infiltration into tumors after ablative radiotherapy. However, the robust innate immune response that is induced by radiotherapy within tumors also raises the possibility that this immune reaction may be redirected with antitumor activity. For example, IFNγ released systemically in response to a CD40 agonist can redirect tumor-infiltrating monocytes with antifibrotic and antitumor properties (19, 24). Thus, we propose that combining radiotherapy with immune adjuvants may be one way to harness the inflammatory response invoked by ablative radiotherapy.

The myeloid response to cancer is inherently pliable with its phenotype dependent on local signals received within the tumor microenvironment. Our findings show that this myeloid response is a key determinant of tumor cell survival and resilience to radiotherapy. In our study, we demonstrate that modulating ablative radiotherapy-induced inflammation using CCL2-neutralizing antibodies can enhance the efficacy of radiotherapy. These findings support targeting the CCL2–CCR2 axis in combination with ablative radiotherapy to improve local disease control in patients with locally advanced PDAC. This translational approach may have particular promise in patients with genetic signatures that suggest a more locally invasive disease phenotype (40).

Our study was mainly conducted in a subcutaneous model of PDAC using KPC-derived cell lines. Although there may be differences in the efficacy of ablative radiotherapy in PDAC depending on the anatomical location of the treated tumor, we have found that ablative radiotherapy can induce inflammatory monocyte recruitment into PDAC tumors irrespective of whether they were implanted subcutaneously or orthotopically. Nonetheless, it is possible that PDAC tumors arising spontaneously may establish a microenvironment that is distinct from tumors implanted subcutaneously or orthotopically. Our study does not address this possibility given current technical limitations in delivering focal image-guided radiotherapy to abdominal tumors in mice. However, this is not a limitation to treating patients with locally invasive pancreatic tumors where conformal techniques for delivering ablative doses of radiotherapy are well established.

Multiple CCL2 and CCR2 antagonists have been developed and their clinical investigation is ongoing (33, 41–43). However, efforts to combine these agents with radiotherapy have not yet been explored. In contrast, agents targeting the CCL2–CCR2 axis in PDAC are actively under investigation in combination with cytotoxic chemotherapy and early results appear promising (44). The efficacy of chemotherapy in PDAC, although, is limited by the dense fibrotic reaction that surrounds PDAC, whereas ablative doses of radiotherapy can be effectively delivered to produce potent cell injury that is associated with a robust inflammatory response. Thus, combining CCL2/CCR2 antagonists with ablative radiotherapy may be particularly effective at enhancing the efficacy of radiotherapy and warrants clinical investigation based on our findings.

In summary, CCL2 plays an important role in the myeloid response to ablative radiotherapy in PDAC, a tumor that is inherently poorly immunogenic. Our data demonstrate that ablative radiotherapy induces tumor-derived CCL2 expression, which drives recruitment of inflammatory monocytes and macrophages into the tumor microenvironment to promote resistance to radiotherapy (Fig. 6). We propose that inhibition of inflammatory monocyte recruitment, by antagonizing the CCL2–CCR2 chemokine axis in combination with ablative radiotherapy, is a novel therapeutic approach for improving disease control in PDAC, where local tumor growth is responsible for significant morbidity and mortality.

No potential conflicts of interest were disclosed.

Conception and design: A. Kalbasi, G.M. Tooker, E. Ben-Josef, G.L. Beatty

Development of methodology: A. Kalbasi, G.M. Tooker, M. Liu, G.L. Beatty

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): A. Kalbasi, C. Komar, G.M. Tooker, M. Liu, J.W. Lee, W.L. Gladney, G.L. Beatty

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): A. Kalbasi, C. Komar, G.M. Tooker, M. Liu, E. Ben-Josef, G.L. Beatty

Writing, review, and/or revision of the manuscript: A. Kalbasi, C. Komar, G.M. Tooker, M. Liu, J.W. Lee, W.L. Gladney, E. Ben-Josef, G.L. Beatty

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): A. Kalbasi, G.L. Beatty

Study supervision: G.L. Beatty

We thank Kristen B. Long and Ioannis I. Verginadis for technical assistance and helpful discussions; W. Tim Jenkins and Cameron J. Koch for assistance with the Small Animal Radiation Research Platform; and Adam Bedenbaugh for assistance with frozen sectioning of tumors.

This work was supported by a Conquer Cancer Foundation of ASCO Young Investigator Award (to A. Kalbasi), NIH grants K08 CA138907 (to G. L. Beatty), F30 CA196106 (to J.W. Lee), and F30 CA196124 (to M. Liu), Abramson Cancer Center-Radiation Oncology Joint Pilot Award (to A. Kalbasi, E. Ben-Josef, and G. L. Beatty), and Damon Runyon Cancer Research Foundation for which G. L. Beatty is Nadia's Gift Foundation Innovator of the Damon Runyon-Rachleff Innovation Award (DRR-15-12).

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.