The Immune-microenvironment Confers Chemoresistance of Colorectal Cancer through Macrophage-Derived IL6

Cancer development and progression are complex processes that are not only caused by accumulated genetic modifications in cancer cells but are also influenced by the surrounding microenvironment. Cancer cells recruit vasculature and stroma (including immune cells, fibroblasts, cytokines, and the extracellular matrix that surrounds them) to the tumor microenvironment (TME), and the activated TME in turn modifies the malignant behaviors of cancer cells (1). Numerous studies have indicated that infiltrating immune cells in TME fail to execute antitumor functions but interact intimately with the tumor cells to promote oncogenesis and progression. The innate immunity cells and cells of the adaptive immune are both involved in the cross-talk (2). Tumor-associated macrophages (TAMs) are one of the most abundant types of cells in TME, directly affecting tumor progression in many cases (3).

Colorectal cancer is the third most common cancer worldwide (4). Systemic chemotherapy is one of the standard treatments for colorectal cancer. However, many patients with colorectal cancer do not respond to conventional chemotherapy due to drug resistance. Chemoresistance, including inherent and acquired drug resistance, is considered to be mediated by multiple factors, such as drug inactivation, accelerated drug efflux, and alterations in the target cells (5). Recent studies have suggested that immune cells in TME play important roles in mediating acquired drug resistance and confer resistance to physiologic mediators of cell death (6, 7). For example, Shree and colleagues showed that macrophages protect breast cancer cells from cell death induced by taxol, etoposide, and doxorubicin (8). In advanced non–small cell lung cancer, a high Fox3p⁺/CD8⁺ T-cell ratio is associated with a poor response to platinum-based chemotherapy (9). Ding and colleagues revealed that targeting CD4⁺ T cells constitutes a new type of immunotherapy to increase therapeutic efficacy (10). In colorectal cancer, the localization and density of immune cells in TME are also associated with response to chemotherapy (11). The presence of high numbers of TAMs was also reported to be associated with poor clinical outcome in breast, pancreatic, bladder, ovarian, and gastric cancers (12–16). These reports suggest that immune cells, especially macrophages in TME, may play key roles in modulating biological phenotypes of cancer cells. In colorectal cancer, contradictory conclusions about the effects of TAMs on patient prognosis have been reported (17–22), and the detailed roles of TAMs in colorectal cancer chemoresistance are largely unclear.

In this study, we observed that the macrophage infiltration in human colorectal cancer tissues is associated with chemoresistance and poor prognosis. Further functional and mechanistic studies revealed that maladjusted miR-155-5p/C/EBPβ/IL6 signaling in TAMs induce the chemoresistance in colorectal cancer cells by regulating the IL6R/STAT3/miR-204-5p axis, revealing a new cross-talk between immune cells and tumor cells in colorectal cancer TME. These findings might contribute to insight concerning TME and the poor response of colorectal cancer cells to conventional chemotherapy.

Four human colorectal cancer cell lines (DLD1, HCT-8, HT-29, and LoVo), human embryo intestinal mucosa cells CCC-HIE-2, human monocyte cell line THP-1, mouse colorectal cancer cell line CT26.WT, and macrophage cell line RAW264.7 were purchased from ATCC. All of the cell lines were obtained during 2008 to 2014 and cultured following the instructions recommended by ATCC as we described previously (23). Cells were confirmed to be mycoplasma free and passaged no more than 18 to 25 times after thawing. Cell lines were characterized by Genewiz Inc. using short tandem repeat markers (last tested in 2017). THP-1 cells were treated with 200 nmol/L phorbol 12-myristate 13-acetate (Sigma) for 24 hours to differentiate into adhered macrophages. In some experiments, colorectal cancer cells were treated with BSA or 25 ng/mL IL6 (Novoprotein) for 3 days; the media were replaced with serum-free media and the cells were harvested 24 hours later.

For coculture experiments, macrophages (5 × 10⁵) were placed into the lower chamber in 12-well plate and colorectal cancer cells (5 × 10⁵) were added into the upper chamber of a transwell insert with a 0.4-μm pore size (Millipore). In some experiments, cocultures were treated with neutralizing antibodies to IL6 (anti-IL6; Abcam) at 1:400 following the manufacturer's instructions. The colorectal cancer–conditioned macrophage cells were washed with PBS and added with fresh serum-free media. Then, 24 hours later, conditioned media (CM) were collected and filtered with a 0.22-μm filter.

Tumor tissues from 81 patients with colorectal cancer were obtained from the Affiliated Hospital of Jiangnan University (Wuxi, Jiangsu, China; Supplementary Table S1). These cases were followed up for more than 5 years. The samples were gathered with informed consent according to the Institutional Review Board of Ethical Committee–approved protocol.

IHC staining was performed on 4-mm sections of paraffin-embedded tissue samples to detect the protein expression levels of CD68, P-gp, Bcl2, IL6, and RAB22A. In brief, the slides were incubated in anti-CD68 (1:200, Santa Cruz Biotechnology), anti-P-gp (1:200, Santa Cruz Biotechnology) antibodies, anti-Bcl2 (1:100, Santa Cruz Biotechnology), C/EBPβ (1:200, Abcam), anti-IL6 (1:600, Santa Cruz Biotechnology), STAT3 (1:500, Cell Signaling Technology), pSTAT3 (1:200, Cell Signaling Technology), and anti-RAB22A (1:200, Proteintech) at 4°C overnight. The subsequent steps were performed using the EnVision FLEX High pH 9.0 Visualization System (DAKO). All slides were independently evaluated by two pathologists without the knowledge of the patients' clinical information. The staining intensity was visually scored and stratified (score 0-3) as described previously: negatively stained 0 (−), weakly stained 1 (+), moderately stained 2 (++), and strongly stained 3 (+++; refs. 23–26). Additional methods are described in Supplementary Materials and Methods.

Total RNA was extracted from tissues or cells using RNAiso Plus (TaKaRa). The concentrations of RNA were determined using a NanoDrop 2000 (Thermo Fisher Scientific). cDNA was synthesized using HiFiScript cDNA Synthesis Kit (CWBIO). qRT-PCR analyses were conducted to quantitate the relative mRNA expression using UltraSYBR Mixture (CWBIO), with β-actin as an internal control. Stem-loop qRT-PCR assays using TaqMan miRNA probes (Applied Biosystems) were performed to quantify the levels of mature miRNAs. The primers used are listed in Supplementary Table S2.

The human pri-miR-155 sequence was amplified from normal human genomic DNA and cloned into the lentivirus expression vector pWPXL to generate pWPXL-miR-155. The promoter region of IL6 containing the potential C/EBPβ-binding sites was amplified from human genomic DNA using PrimerSTAR Premix (TaKaRa) and was then cloned into the region directly upstream of modified firefly luciferase cassette in a pGL3 Luciferase Reporter Vector (p-Luc). Duplex siRNAs were purchased from GenePharma.

The pWPXL-GFP or pWPXL-miR-155 plasmid was cotransfected into HEK-293T cells along with the packaging plasmid psPAX2 and the envelope plasmid pMD2G using Lipofectamine 2000 (Invitrogen) as we described previously (27). Virus particles were harvested 48 hours after cotransfection and were individually used to infect THP-1 cells. The cells were then harvested at 3 days after infection for Western blotting and qRT-PCR validation.

CT26.WT cells were subcutaneously injected into the left armpits of male BALB/c mouse at 5 weeks of age with or without RAW264.7 cells. Daily bolus doses of 5-fluorouracil (5-FU) at 25 mg/kg were given intraperitoneally for two 4-day cycles 10 days after the injection. The mice were sacrificed after a period of 21 days and examined for the growth of subcutaneous tumors. All animal care and handling procedures were performed in accordance with the NIH's Guide for the Care and Use of Laboratory Animals and were approved by the Ethical Committee of Affiliated Hospital of Jiangnan University.

Colorectal cancer cells were treated with 5-FU or oxaliplatin (LOHP; range, 0–100 mg/mL), and cell inhibition was then assessed by a CCK-8 assay (Dojindo). The IC₅₀ was calculated. For the apoptosis analysis, colorectal cancer cells were treated with 5 mg/mL 5-FU or LOHP for 48 hours. These cells were then harvested and subjected to apoptosis analysis using an Annexin V-FITC and Propidium Iodide Labeling Kit (CWBIO).

A Human Cytokine Antibody Arrays Kit (RayBiotech) was used to detect cytokines in the culture medium according to the manufacturer's instructions. Briefly, the arrays were incubated with 100 μL of medium, biotin-conjugated antibodies, and a Streptavidin-Fluor–linked secondary antibody one by one for 2 hours. The glass chip was scanned with a laser scanner (Axon, Genepix) using a Cy3-compatible (green; 532 nm) laser. Quantitative array analysis was performed using GenePixPro6.0 (Axon).

ELISA development reagents (duo-set kit) for human IL6 and mouse IL6 were purchased from ExCell Bio, and the assay was performed according to the manufacturer's instructions. Absorbance was measured using a microplate reader (HIDEX).

Protein extracts were probed with antibodies against human phospho-STAT3 (Tyr705) (1:1,500, CST), STAT3 (1:1,000, Proteintech), IL6R (1:500, Santa Cruz Biotechnology), RAB22A (1:1,000, Proteintech), P-gp (1:1,000, Santa Cruz Biotechnology) antibodies, Bcl2 (1:1,000, Santa Cruz Biotechnology), C/EBPβ (1:1,000, Abcam), or β-actin (1:2,000, Abcam). Peroxidase-conjugated anti-mouse or rabbit antibody (The Jackson Laboratory) was used as a secondary antibody and the antigen–antibody reaction was visualized by an enhanced chemiluminescence assay (CWBIO).

THP-1–differentiated macrophages were cultured in 96-well plates and cotransfected with 50 nmol/L of miR-155-5p mimic, CEBPB siRNA (siCEBPB), 50 ng of luciferase reporter, and 10 ng of pRL-CMV Renilla luciferase reporter using Lipofectamine 2000 (Invitrogen). Forty-eight hours after transfection, the luciferase activities were assayed using a Luciferase Assay Kit (Promega).

miRCURY LNA Detection probe for hsa-miR-155-5p obtained from TSINGKE labeled with FAM (488 nm, green) at the 5′ and 3′ terminus. In brief, preliminary hybrid liquid was added to cover slides at 37°C for 1 hour. Then, a total of 100 μL diluted probe was added to each slide, and slides were kept in the hybridization chamber and incubated in oven at 37°C overnight. Posthybridization washes were performed and counterstained with DAPI. For immunofluorescence staining, serial sections of the same specimens were incubated with CD68 antibodies labeled with Cy3 (red, eBioscience). All images were captured on an Olympus fluorescence microscope equipped with vision software.

The data were expressed as the mean ± SEM and were subjected to the Student t test or the Mann–Whitney U test unless otherwise specified (χ² test or a Spearman correlation). Cox proportional hazards regression analysis was used to calculate HRs and the 95% confidence intervals. A value of P < 0.05 was considered statistically significant. The SPSS 16.0 package was used for the statistical analyses.

In our previous colorectal cancer cohort studies, different overall survival (OS) times have been observed in stage III patients who received surgery and standardized chemotherapy (23, 24). Patients with poor survival seemed to be resistant to the chemotherapy. We supposed that the TME, especially the infiltration of macrophages, may be associated with the chemoresistance and prognosis of colorectal cancer. To assess potential effects of macrophage infiltration on the chemoresistance of colorectal cancer, we assessed the status of TAMs in colorectal cancer tissues. First, we selected 65 patients with stage III colorectal cancer who received radical operations and were treated with 5-FU, L-OHP, and leucovorin from the colorectal cancer cohort reported previously (23, 24). Thirty-seven living patients with an OS longer than 60 months were labeled as chemotherapy sensitive (sensitive group), and 28 deceased patients with OS shorter than 24 months were regarded as chemotherapy resistant (resistant group). IHC staining of CD68 (the marker of human macrophage) was performed to measure macrophage infiltration (CD68⁺) in these tumor tissues. Stronger macrophage infiltration (CD68 scores 2 or 3) was frequently observed in the resistant group compared with the sensitive group and was associated with poor survival (P < 0.05, Fig. 1A; Supplementary Table S3; Supplementary Fig. S1A). Second, to confirm the correlation of macrophage infiltration to drug resistance, we measured the expression of MDR1 (P-gp), a key marker of multidrug resistance (MDR), and the apoptosis maker BCL2 (Bcl2) in these tumor tissues. The results of qRT-PCR revealed significant upregulation of MDR1 and BCL2 in the resistant group compared with the sensitive group (Supplementary Fig. S1B). IHC staining of P-gp and Bcl2 proteins showed the same trend (Fig. 1B and C; Supplementary Table S4 and S5). In addition, as chemotherapy usually induces acquired chemoresistance of cancer cells, the surviving cancer cells after chemotherapy often display MDR phenotypes (28). Therefore, we further assessed the macrophage infiltration and P-gp/Bcl2 expression in tumors from 16 patients with colorectal cancer receiving neoadjuvant chemotherapy (NCT, neoadjuvant group). As expected, these tumors did show stronger macrophage infiltration and the expression of P-gp and Bcl2 than the Sensitive tumors (Fig. 1A–C). Moreover, the expression levels of CD68, P-gp, and Bcl2 are relatively higher in the postresection specimens compared with their corresponding untreated biopsies (Supplementary Fig. S1C). Taken together, these results suggest that TAMs are associated with chemoresistance and poor survival in colorectal cancer.

To further confirm the relationship between TAMs and drug resistance in colorectal cancer, a tumor formation assay in a BalB/C mouse model was performed. The mouse macrophage cells RAW264.7 preincubated with mouse colorectal cancer cells CT26.WT for 3 days were used to mimic TAMs in human colorectal cancer tissues. As shown in Fig. 1D, CT26.WT cells (control group) or CT26.WT mixed with RAW264.7 (pre-cocultured; 1:1; coinjection group) were injected into flanks of BalB/C mice. Ten days after injection, daily bolus doses of 5-FU at 25 mg/kg were administered intraperitoneally to each mouse for two 4-day cycles. The tumors of the coinjection mice were more resistant to 5-FU therapy than those of the control mice (Fig. 1D), indicating that TAMs promote in vivo chemoresistance in colorectal cancer.

To mimic the effects of TAMs on drug resistance in vitro, the human monocyte cell line THP-1 was induced into macrophages and then cocultured with different colorectal cancer cell lines (HT-29, LoVo, DLD1, or HCT-8) for 3 days. Subsequently, these colorectal cancer cells were subjected to drug sensitivity and apoptosis analysis. The results showed that the IC₅₀s of cocultured colorectal cancer cells were significantly higher than those of their corresponding control cells (P < 0.01; Fig. 1E; Supplementary Fig. S2A), and coculture with macrophages significantly increased the resistance to chemotherapeutic agents (5-FU and LOHP) and reduced drug-induced apoptosis in colorectal cancer cells (Fig. 1F; Supplementary Fig. S2B). Taken together, these results suggest that TAMs strongly increase resistance of colorectal cancer cells to chemotherapeutic drugs.

To explore how TAMs increase the in vitro chemoresistance of colorectal cancer cells, we treated colorectal cancer cells with CM from the colorectal cancer–conditioned macrophages. After CM treatment, the colorectal cancer cells exhibited increased chemoresistance compared with the control cells (Supplementary Fig. S3A), suggesting that some secreted factors in TAMs could affect colorectal cancer cells. Given the key signaling transduction role of cytokines between different types of cells in TME, we speculated that TAMs promote chemoresistance by secreting certain cytokines. Therefore, we measured the cytokine profiles of CM from macrophages cocultured with colorectal cancer cells (HT-29) or embryo intestinal mucosa cells (CCC-HIE-2), and three cytokines (IL6, IL1β, and MCP2) were significantly altered in the CM from the macrophages cocultured with HT-29 compared with the macrophage control (Fig. 2A). Of the three cytokines, IL6 emerged as the most prominently upregulated and abundant cytokine. Therefore, IL6 was selected for subsequent analyses. An ELISA assay further confirmed the increase of IL6 in the CMs from macrophages cocultured with four different colorectal cancer cells (Fig. 2B), and increased IL6 staining was also observed in the drug-resistant tumors and the tumor tissues receiving NCT (Fig. 2C; Supplementary Table S6).

To evaluate whether IL6 is critical for chemoresistance in colorectal cancer, an exogenous recombinant IL6 was added in the culture medium of several colorectal cancer cell lines. The results showed that IL6 significantly increased the resistance of these colorectal cancer cells to chemotherapeutic drugs (5-FU and LOHP) and reduced drug-induced apoptosis (Fig. 2D and E; Supplementary Fig. S3B and S3C). In addition, the mRNA levels of IL6R, MDR1, and BCL2 were also upregulated in IL6-treated colorectal cancer cells (Fig. 2F).

To further investigate whether colorectal cancer–conditioned macrophages affected chemoresistance in colorectal cancer through IL6, the drug sensitivity and apoptosis were assessed in colorectal cancer cells cocultured with macrophages and an anti-IL6 neutralizing antibody. The results showed that the addition of neutralizing anti-IL6 antibody to the coculture system suppressed the macrophage-induced chemoresistance in colorectal cancer cells (Fig. 3A and B; Supplementary Fig. S4A and S4B). In addition, the macrophage-induced upregulation of IL6R, MDR1, and BCL2 in colorectal cancer cells was also inhibited by the anti-IL6 antibody (Fig. 3C). Collectively, these data indicate that colorectal cancer–conditioned macrophages induce chemoresistance in colorectal cancer cells by secreting IL6.

To decipher the molecular mechanism that mediates the effects of macrophage-secreted IL6 on the chemoresistance of colorectal cancer, IL6R and STAT3, key mediators of IL6 signaling, were analyzed in colorectal cancer cells. The results showed that IL6R expression and STAT3 activation (Tyr705 phosphorylation) were considerably enhanced in colorectal cancer cells cocultured with macrophages, which could be blocked by a neutralizing anti-IL6 antibody (Fig. 4A). We further detected the levels of STAT3 and pSTAT3 in colorectal cancer tissues and revealed that the pSTAT3 staining is stronger in the resistant and neoadjuvant groups than in the sensitive group (Fig. 4B). Activated Stat3 can bind to three putative STAT-binding sites located in the TRPM3/miR-204 promoter region, leading to the transcriptional repression of miR-204 (29, 30). Moreover, our previous study revealed that miR-204-5p increases the sensitivity of colorectal cancer cells to 5-FU and LOHP (23, 24). We speculated that miR-204-5p may be the key functional target of macrophage-secreted IL6 in colorectal cancer cells. The qRT-PCR results did reveal that miR-204-5p was significantly downregulated in the Resistant tumors and the tumors receiving NCT (Fig. 4C). The downregulation of miR-204-5p was also observed in colorectal cancer cells cocultured with macrophages as well as in IL6-treated colorectal cancer cells compared with the control cells (Fig. 4D).

We previously reported that miR-204-5p enhances the chemotherapeutic sensitivity in colorectal cancer cells by inhibiting RAB22A and Bcl2 (23, 24). As expected, in contrast to the decreased expression of miR-204-5p in the drug-resistant tumors and IL6-treated colorectal cancer cells, colorectal cancer–conditioned macrophages or IL6 could induce RAB22A and Bcl2 upregulation in colorectal cancer cells, which could be reversed by IL6 blocking or ectopic miR-204-5p expression (Fig. 4E and F). IHC staining of RAB22A performed on clinical colorectal cancer tissues further confirmed these in vitro results (Fig. 4B; Supplementary Table S7). We further observed that ectopic miR-204-5p expression could reverse macrophage-induced chemoresistance in colorectal cancer cells (Fig. 4G; Supplementary Fig. S5), suggesting that macrophage/IL6 could induce MDR via miR-204-5p. In addition, ectopic miR-204-5p expression could not block the macrophage-induced upregulation of P-gp in colorectal cancer cells, indicating a miR-204-5p–independent manner (Fig. 4F). Taken together, these results imply that TAM-secreted IL6 modulates colorectal cancer chemoresistance partly by inhibiting miR-204-5p signaling.

We previously demonstrated that miR-155, an miRNA downregulated in TAMs and M2 macrophages, is a critical molecule controlling the macrophage phenotypic switch, and miR-155–modified TAMs can be reprogrammed and regain antitumor capacity (31). Interestingly, we observed that miR-155-5p levels were markedly lower in the macrophages cocultured with colorectal cancer cells than in their corresponding control macrophages (Fig. 5A). Functional analysis revealed that ectopic miR-155 expression (Fig. 5B) significantly reverses colorectal cancer chemoresistance induced by these colorectal cancer–conditioned macrophages, suggesting a key role of miR-155 in the TAM-induced chemoresistance in colorectal cancer (Fig. 5C; Supplementary Fig. S6A).

We have previously reported that miR-155-5p promotes the M1 polarization of macrophages by suppressing the C/EBPβ (CAAT/enhancer-binding protein β) signaling cascade (31), and C/EBP-binding motifs have been identified in the promoter of IL6 (32). We speculated that decreased miR-155-5p signaling result in increased IL6 production in TAMs. As expected, colorectal cancer–conditioned macrophages increased C/EBPβ expression, which could be blocked by miR-155-5p overexpression (Fig. 5D). A negative association was also observed between miR-155-5p and C/EBPβ expression in TAMs (Fig. 5E). Moreover, restoring miR-155-5p expression significantly inhibited IL6 expression and secretion in colorectal cancer–conditioned macrophages (Fig. 5F). To further validate the regulation of the miR-155-5p/C/EBPβ signaling on the IL6 expression in macrophages, the promoter region of IL6 containing the C/EBP-binding motifs was cloned into a luciferase reporter plasmid, and luciferase assays showed that transfection of miR-155-5p or siCEBPB could inhibit the IL6 promoter-derived luciferase expression in macrophages (Fig. 5G; Supplementary Fig. S6B), indicating that miR-155-5p inhibits IL6 expression by targeting C/EBPβ in macrophages. Consequently, the miR-204-5p expression was significantly increased, whereas IL6R, MDR1, and BCL2 were downregulated in colorectal cancer cells cocultured with miR-155–overexpressing macrophages (Fig. 5H). Moreover, we used peripheral blood monocytes (PBM) from healthy donors and TAMs isolated from human colorectal cancer tissues to confirm the miR-155-5p/C/EBPβ/IL6 axis. LPS and IFNγ were used to induce inflammatory macrophages from PBMs (IPBMs). The results were in accordance with those in colorectal cancer–cocultured macrophages/TAMs (C/EBPβ and IL6 are upregulated and miR-155-5p is downregulated in TAMs). Interestingly, miR-155-5p is also upregulated in LPS-induced IPBM, suggesting that LPS may regulate macrophage activation via a mechanism different from that in tumor microenvironment (Supplementary Fig. S7).

Collectively, these data indicate that decreased miR-155-5p expression in TAMs results in C/EBPβ overexpression, which activated IL6 transcription and led to increased expression and secretion of IL6.

To further test the regulation of miR-155 on the TAM-induced colorectal cancer chemoresistance in vivo, we stably overexpressed miR-155 in mouse macrophages (RAW264.7-miR-155) using lentivirus transduction (Fig. 6A). Consistent with the results in human cell lines, both C/EBPβ and IL6 levels were significantly decreased in RAW264.7-miR-155 cells compared with the control cells (Fig. 6A). Consequently, the expression of IL6R, BCL2, and MDR1 were downregulated, and miR-204-5p expression was increased in colorectal cancer cells (CT26.WT) coincubated with RAW264.7-miR-155 macrophages compared with the control cells (Fig. 6B). Then, CT26.WT cells mixed with RAW264.7-GFP or RAW264.7-miR-155 (1:1) were injected into the flanks of BalB/C mice and treated with 5-FU. As shown in Fig. 6C and D, the xenograft tumors of CT26.WT/RAW264.7-miR-155 were more sensitive to the 5-FU therapy than the CT26.WT/RAW264.7 control. IHC staining in these tumors also confirmed the inhibitory effects of miR-155 overexpression in macrophages on the expression of P-gp, Bcl2, and RAB22A in colorectal cancer cells (Fig. 6E). Taken together, these data demonstrate that stably overexpressing miR-155 in TAMs could reverse the TAM-induced colorectal cancer chemoresistance.

Recent data revealed that the malignant behaviors of cancer cells are regulated by other types of cells in TME, especially immune cells. MDR is a major reason for chemotherapy failure, resulting in tumor metastasis and relapse. In this study, we found that increased macrophage infiltration in colorectal cancer tissues was significantly associated with increased chemoresistance and poor prognosis. Further studies revealed that TAM-derived IL6 induces chemoresistance in colorectal cancer cells by partly regulating the IL6R/STAT3/miR-204-5p pathway, uncovering a new cross-talk between TAMs and colorectal cancer cells.

Clinical observations have shown that high macrophage infiltration is generally associated with poor prognosis in human cancer, including breast, gastric, prostatic, pancreatic, ovarian, and cervical carcinomas (12–16). Interestingly, contradictory results have also been reported in several types of human cancers, including colorectal cancer (17–22). In colorectal cancer, an increased infiltration of TAMs was associated with lymph node metastases and tumor progression (19, 21), and M2 macrophage infiltration was associated with poor prognosis in colorectal cancer (20, 22). However, increased macrophage infiltration was also reported to correlate with improved survival in colorectal cancer (17, 18). Gut macrophages are highly plastic cells, and their phenotype is dependent on the local microenvironment in colorectal cancer (33). It is difficult to establish a standard assay to measure TAMs in different tumors, which may partly explain these inconsistencies in colorectal cancer. Despite the inconsistent results concerning the prognostic role of TAMs in colorectal cancer, recent data showed that TAMs exert tumor-promoting functions by regulating cell growth, metastasis, and drug resistance in colorectal cancer (34–38). Because the phenotypes of TAMs are highly plastic and associated with tumor stage and their location in TME (33, 39, 40), we measured the total macrophage infiltration other than M1 or M2 macrophage in representative patients with colorectal cancer to investigate its clinical significance. We observed that increased macrophage infiltration is associated with poor survival and increased expression of P-gp and Bcl2 in colorectal cancer cells, suggesting the key role of TAMs on chemoresistance.

Chemotherapy is a standard treatment for most patients with colorectal cancer; however, many patients did not respond to chemotherapeutic agents due to chemoresistance. Chemoresistance is usually due to two types of factors: intrinsic factors within the colorectal cancer cells themselves and extrinsic factors in the TME (5). TAMs could mediate resistance to antitumor drugs in various ways. It has been reported that TAMs suppress CD8⁺ T lymphocyte activity and then promote chemoresistance to paclitaxel and metastasis in mammary tumors (41). Angst and colleagues showed that mononuclear cell–derived IL1β confers chemoresistance in pancreatic cancer cells by upregulating cyclooxygenase-2 (42). These studies highlight the key roles of macrophages in chemoresistance. Macrophages constitute one of the major populations of cells that reside in the colorectal cancer TME (43). As increased macrophage infiltration is associated with poor survival in our study, in vitro and in vivo experiments further confirmed the phenomenon observed in clinical colorectal cancer tissues, demonstrating the key regulatory role of TAMs on the MDR of colorectal cancer. Consistent with our results, a recent study showed that macrophages could induce 5-FU resistance in colorectal cancer by releasing putrescine (35).

To explore the interaction between TAMs and colorectal cancer cells, we treated colorectal cancer cells with CM from TAMs. Consistent with the changes in the expression of MDR or apoptosis-related genes, the TAM-derived CM-stimulated colorectal cancer cells exhibited drug-resistant behaviors, suggesting that TAMs secrete a factor that affects the chemoresistance of colorectal cancer cells. Given the key role of cytokines in cell–cell interactions, we screened the changes of the secretory cytokine profile in the CM from colorectal cancer–conditioned macrophages and identified IL6 as the most significantly upregulated cytokine. Subsequent functional assays confirmed that IL6 is accountable for the TAM-induced chemoresistance in colorectal cancer.

As a key proinflammatory factor, IL6 is also implicated in the regulation of tumorigenesis, progression, and MDR (44–47). IL6 can induce MDR in human cancer cells by complicated mechanisms, such as increasing expression of MDR-related genes (MDR1 and GSTP1) and apoptosis-inhibitory proteins (Bcl-2, Bcl-xL, and XIAP), and activation of AKT, ERK, MAPK, and STAT3 signaling (47, 48). Here, we revealed that by binding with its specific ligand IL6R, IL6 subsequently phosphorylates STAT3 and leads to chemoresistance in colorectal cancer. We had previously reported that miR-204-5p could significantly inhibit cell growth, metastasis, and chemoresistance in colorectal cancer cells (23, 24). In this study, we further validated that miR-204-5p expression in TAM-conditioned colorectal cancer cells was significantly inhibited by macrophage-secreted IL6, whereas RAB22A and BCL2, confirmed target genes of miR-204-5p, were highly expressed in these colorectal cancer cells. Mechanistic studies revealed that activated STAT3 by macrophage-derived IL6 transcriptionally inhibited the expression of miR-204-5p. These data uncovered a new mechanism mediating decreased expression of miR-204-5p in colorectal cancer and provide a new mechanism by which TAMs induce chemoresistance in colorectal cancer. Interestingly, a recent study reported that miR-204 targets IL6R and sensitizes epithelial ovarian cancer cells to cisplatin, which partly explained the upregulation of IL6R in colorectal cancer cells cocultured with macrophages or treated with IL6 in our study (49).

MDR1 is often upregulated in chemoresistant tumor cells, and colorectal cancer–conditioned macrophages could induce MDR1 expression by secreting IL6 in a miR-204-5p–independent manner, suggesting the complicated multidimensional regulatory mechanisms of MDR by TAM-derived IL6. Consistent with our results, Wang and colleagues reported that autocrine production of IL6 confers chemoresistance in ovarian cancer cells by activating several pathways and genes, including MDR1 and BCL2 (48). Autocrine production of IL6 could also confer MDR in breast cancer cells by inducing C/EBP and MDR1 expression (50). On the basis of our data and previous studies, we propose a working model for TAM-induced chemoresistance in colorectal cancer (Fig. 6F). The detailed mechanism by which TAMs induce MDR1 expression in colorectal cancer should be investigated in future work.

TAMs are an important area of research for the effective treatment of tumors. Our previous study demonstrated that miR-155 is a critical molecule controlling the macrophage phenotypic switch and that miR-155–modified TAMs can be reprogrammed and regain tumor-killing capacity. In this study, we revealed that miR-155-5p was frequently downregulated in colorectal cancer–associated macrophages, which resulted in increased C/EBPβ expression. C/EBPβ transcriptionally activated IL6 in TAMs and finally induced chemoresistance by activating the IL6R/STAT3/miR-204-5p pathway in colorectal cancer cells. Restoration of the expression of miR-155-5p in TAMs reversed the TAM-induced chemoresistance in colorectal cancer cells. Thus, restoring miR-155-5p expression in TAMs may provide a new therapeutic strategy to reverse MDR in colorectal cancer at the TME level.

In conclusion, we revealed a new cross-talk between immune cells and tumor cells in the colorectal cancer microenvironment. The maladjusted miR-155-5p/C/EBPβ/IL6 signaling in TAMs could induce chemoresistance in colorectal cancer cells by regulating the IL6R/STAT3/miR-204-5p axis. TAM-targeted regulation may represent a promising therapeutic strategy to improve chemotherapeutic efficacy in colorectal cancer.

No potential conflicts of interest were disclosed.

Conception and design: Y. Yin, Z. Huang

Development of methodology: Y. Yin, Z. Huang

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): Y. Yin, S. Yao, Y. Hu, Y. Feng, M. Li, Y. Qin, X. Qi, L. Zhou, B. Fei, Z. Huang

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): Y. Yin, Z. Bian, L. Zhou, Z. Huang

Writing, review, and/or revision of the manuscript: Y. Yin, J. Zou, D. Hua, Z. Huang

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): J. Zou, Z. Huang

Study supervision: Z. Huang

Other (i.e., evaluating scores of the IHC staining): J. Zhang

This study was partially supported by grants from the National Natural Science Foundation of China (81672328, 81772636, 81301784, and 81272299), Natural Science Foundation of Jiangsu Province (BK20151108 and BK20150004), Fundamental Research Funds for the Central Universities (NOJUSRP51619B and JUSRP51710A), Medical Key Professionals Program of Jiangsu Province, Medical Innovation Team Program of Wuxi, and Hospital Management Centre of Wuxi (YGZXZ1401).

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.