Colorectal cancer is one of the most common cancers and a major cause of mortality. Proinflammatory and antitumor immune responses play critical roles in colitis-associated colon cancer. CCL17, a chemokine of the C-C family and ligand for CCR4, is expressed by intestinal dendritic cells in the steady state and is upregulated during colitis in mouse models and inflammatory bowel disease patients. In this study, we investigated the expression pattern and functional relevance of CCL17 for colitis-associated colon tumor development using CCL17–enhanced GFP-knockin mice. CCL17 was highly expressed by dendritic cells but also upregulated in macrophages and intermediary monocytes in colon tumors induced by exposure to azoxymethane and dextran sodium sulfate. Despite a similar degree of inflammation in the colon, CCL17-deficient mice developed fewer tumors than did CCL17-competent mice. This protective effect was abrogated by cohousing, indicating a dependency on the microbiota. Changes in microbiota diversity and composition were detected in separately housed CCL17-deficient mice, and these mice were more susceptible to azoxymethane-induced early apoptosis in the colon affecting tumor initiation. Immune cell infiltration in colitis-induced colon tumors was not affected by the lack of CCL17. Taken together, our results indicate that CCL17 promotes colitis-associated tumorigenesis by influencing the composition of the intestinal microbiome and reducing apoptosis during tumor initiation.

Colorectal cancer (CRC) is the third most frequent malignancy worldwide, with increasing incidence (1). The pathophysiology of CRC is multifactorial, with genetic, environmental, and lifestyle factors playing major roles (2). There is clear evidence that inflammation is closely linked to colorectal tumorigenesis, either preceding tumor formation as in ulcerative colitis or elicited by the tumor as in sporadic CRC (35). Transcription factors NF-κB and STAT3 have important functions in inflammation-induced colon tumor development as was shown in inducible mouse models of colitis-associated colon cancer (57). NF-κB activation in tumor-associated myeloid cells leads to inflammatory cytokine production promoting tumor growth. In intestinal epithelial cells that have acquired oncogenic mutation, activation of NF-κB and STAT3 by cytokines leads to enhanced survival and proliferation. Thus, initiated intestinal epithelial cells are rescued from apoptosis and can generate tumors in the inflammatory microenvironment (6, 7).

An important factor in intestinal inflammation and carcinogenesis is the gut microbiome (8). Specific commensal bacterial strains that expand during colitis have been linked to increased tumorigenesis (9), whereas others have been shown to produce protective metabolites (10). Innate myeloid cells, including dendritic cells (DCs) and macrophages, maintain tolerance against commensal bacteria and produce factors influencing the intestinal microbiota in steady state and inflammation (11). In crosstalk with the microbiota, DCs and macrophages can contribute to tumor-promoting inflammation in the intestine. In the developing tumors, myeloid cells are skewed toward a protumor or antitumor activity by the tumor microenvironment (TME), inhibiting or supporting T cell responses against the tumor (5, 12, 13). Chemokines play an important role in migration and homing of immune cells to developing and established tumors. In addition, they can influence immune and cancer cell functionality (14). The C-C chemokine CCL17 binds to CCR4 and induces chemotaxis of CCR4-expressing CD4+ regulatory T cells (Tregs) and Th2, Th17 effector, and NK cells (1518). The chemotactic activity of CCL17 is partially redundant with that of CCL22, which also binds to CCR4 (19, 20) to recruit Tregs into tumors (21). CCL17 promotes migration and activation of murine DCs (2224) and plays an important proinflammatory role in models of organ-specific inflammation, such as atopic dermatitis, neuroinflammation, atherosclerosis, arthritis, lupus nephritis, and acute colitis (2227). CCL17 is expressed by DCs, especially the CD11b+ type 2 conventional DC (cDC2) subset, as well as macrophages under inflammatory conditions (2830). Myeloid cell subsets in murine (3133) and human (34) tumors express CCL17. CCL17 expression was shown to be inhibited by IFN-γ and induced by IL-4 and GM-CSF (29, 35).

The cDC population in the mouse consists of type 1 cDCs (XCR1+CD103+), with superior cross-presentation capacity, and cDC2s (CD11b+SIRPα+), critical for Th cell responses (36). An additional CD103+CD11b+ cDC2 subset plays a critical role in the induction of intestinal Th17 and ILC3 responses (37, 38). In the steady-state resident macrophages (CD64+F4/80+MHC class II (MHC II)+Ly6C) with anti-inflammatory and tissue repair function are constantly replaced by circulating Ly6Chi monocytes (39). During colitis, recruited monocytes differentiate into proinflammatory MHC IIhi macrophages (40). F4/80hiMHC IIlo macrophages with immunosuppressive properties were found to be expanded in colon tumors of APCmin/+ mice independent of monocyte recruitment (41).

CCL17 was shown to have proinflammatory functions (22, 25, 28) that may promote colon tumor formation. In contrast, this chemokine also influences the number and function of various immune cells that can infiltrate tumors and either promote or inhibit tumorigenesis. We therefore investigated its role in the colon tumor model induced by azoxymethane (AOM) and repeated exposure to dextran sodium sulfate (DSS), which closely resembles colitis-associated colon cancer in humans (42). We found that CCL17-deficient mice developed fewer tumors than CCL17-competent mice in this model. The phenotype was dependent on the microbiota and coincided with increased apoptosis but not with reduced inflammation in the colon. CCL17 was highly expressed in DCs and macrophages in colon tumors and influenced gene expression in tumor-associated myeloid cells without affecting the number and function of intratumoral effector lymphocytes.

CCL17–enhanced GFP (eGFP) mice on the C57BL/6J background (at least 10 generations backcrossed) provided by Irmgard Förster, University of Bonn (28), were bred in-house and rederived into C57BL/JRccHsd mice (Envigo, Huntingdon, U.K.). All mice were housed under specific pathogen-free conditions in individually ventilated cages. Health monitoring was performed according to the recommendations of the Federation of European Laboratory Animal Science Association. All experimental procedures involving mice were performed in accordance with the regulations of and were approved by the local government (Regierung von Oberbayern, license nos. ROB-55.2-2532.Vet_02-13-36 and ROB-55.2-2532.Vet_02-17-22).

To induce acute colitis, mice received 4% DSS (MP Biomedicals, Santa Ana, CA, no. 02 160110 80) for 5 d in the drinking water and were sacrificed after 7 d. To induce colitis-associated tumor formation, mice were injected once with 10 mg/kg AOM (Sigma-Aldrich, St. Louis, MO, no. A5486) i.p. and after 5 d were treated with three cycles of DSS (2.5% w/w for 5 d) with 2 wk in between the cycles. Mice were sacrificed on day 85 of the experiment or at an earlier time point when they met the euthanization criteria. On the day of the analysis, the colon was cut longitudinally, washed with ice-cold PBS, and the tumor numbers and sizes were recorded. Tumors of similar sizes were excised and used for flow cytometric analyses, histology, and gene expression analysis, respectively. For short-term AOM/DSS experiments, mice received one injection of AOM and after 5 d were treated with DSS (2.5% w/w) for 2 or 5 d, followed by 4 d of recovery. For short-term exposure to AOM, mice were injected once with 10 mg/kg AOM i.p. and sacrificed after 8 h.

After excision of the tumors the remaining normal intestinal tissue was cut into 5-mm-long pieces and incubated with 2 mM DTT (Sigma-Aldrich, no. D0632), 10 mM HEPES, and 10 mM EDTA in HBSS (without Ca2+/Mg2+) for 10 min at 37°C and 125 rpm. The remaining intestinal tissue pieces and the cut tumor tissue were digested with DNAse (0.5 µg/ml, Sigma-Aldrich, no. 10104159001), collagenase D (2.5 µg/ml, Sigma-Aldrich, no. 11088858001), collagenase V (5 µg/ml, Sigma-Aldrich, no. C9263), and collagenase IV (157 Wünsch units/ml, Worthington Biochemical, Lakewood, NJ, no. LS004188) in RPMI 1640 for 30 min at 37°C with gentle shaking and pushed through a 100- and a 40-µm filter followed by washing with ice-cold HBSS containing Mg2+ and Ca2+ (HBSS+). Single-cell suspensions were then incubated for 10 min on ice with anti-CD16/anti-CD32 containing cell culture supernatant and surface stained with the respective Abs for 20 min on ice protected from light. For intracellular cytokine staining, the cells were restimulated with 20 ng/ml PMA and 1 µg/ml ionomycin for 5 h in the presence of GolgiPlug (0.2% v/v) and GolgiStop (0.14% v/v) (BD Biosciences, Franklin Lakes, NJ, nos. 555029, and 554724). After surface Ab staining, cells were fixed and permeabilized using the Foxp3 fixation/permeabilization kit (Thermo Fisher Scientific, Waltham, MA, no. 00-5521-00) and then stained intracellularly according to the manufacturer’s instructions. The following Abs were used: anti-mouse Ly6C AF700, anti-mouse MHC II BV785, anti-mouse MHC II BV421, anti-mouse CD64 PE-Dazzle 594, anti-mouse F4/80 PE-Dazzle 594, anti-mouse CD11b allophycocyanin-Cy7, anti-mouse CD206 AF647, anti-mouse CD45 PerCP, anti-mouse Ly6G BV650, anti-mouse CD11c PE-Cy7, anti-mouse B220 BV605, anti-mouse CD4 BV605, anti-mouse CD8 AF700, anti-mouse NK1.1 BV650 (all from BioLegend), anti-mouse CD103 PE, anti-mouse IFN-γ PE, anti-mouse CD69 V450 (all from BD Biosciences), and anti-mouse CD209a allophycocyanin, anti-mouse CD3 allophycocyanin-Cy7, anti-mouse Foxp3 PE-Cy7, anti-mouse IL-17A allophycocyanin, and anti-mouse IgA PE (all from Thermo Fisher Scientific). Cells were analyzed on a CytoFLEX S (Beckman Coulter, Brea, CA) or BD LSRFortessa (Becton Dickinson, Franklin Lakes, NJ) flow cytometer or sorted using a BD FACSAria Fusion (Becton Dickinson). Data analysis was conducted with FlowJo v10 (Becton Dickinson).

Selected tumors of similar sizes were fixed in 4% paraformaldehyde for 1 h at 4°C and subsequently incubated in 20% sucrose overnight at 4°C. Tumors were then embedded in OCT (Leica, Wetzlar, Germany) and stored at −80°C. H&E staining of cryosections (5–8 μm) was performed as described. For immunofluorescence staining, cryosections were blocked with PBS containing 5% goat serum (Vector Laboratories, no. S-1000) and 0.5% Triton X-100 for 1 h at room temperature and then stained with primary Abs (anti-GFP, Abcam, Cambridge, U.K., no. 6556; anti-Ki67, Cell Signaling Technologies, Danvers, MA, no. 12202; anti-mouse E-cadherin, BD Biosciences, no. 610181) at 4°C overnight followed by incubation with secondary Abs (anti-mouse IgG AF555, Thermo Fisher Scientific, no. A32727; anti-rabbit IgG AF488, Thermo Fisher Scientific, no. A11008) for 1 h at room temperature and DAPI staining (Sigma-Aldrich, no. D9542). Imaging was conducted with a Leica SP8X WLL upright confocal microscope or an Olympus BX41 fluorescence microscope. Quantification of Ki67+ cells was performed using ImageJ. The number of total cells (DAPI+) and the number of Ki67+ cells (AF488+) were determined and the frequency of Ki67+ cells was calculated.

The colon was opened longitudinally and rolled up. Formalin-fixed, paraffin-embedded sections (5 μm) were deparaffinized/rehydrated, boiled in citrate-based Ag retrieval solution (Dako, Hamburg, Germany, no. S2369) for 20 min, incubated in 3% H2O2/PBS (Carl Roth, no. 9681.1) for 10 min, and blocked with 2.5% normal horse serum (Vector Laboratories). Sections were incubated overnight at 4°C with the primary Ab (anti-cleaved caspase-3 [Asp175], Cell Signaling Technology, no. 9661, or rabbit anti-mouse Muc2, Elabscience, no. E-AB-70212), followed by detection reagent (ImmPRESS HRP horse anti-rabbit IgG polymer detection kit, peroxidase, no. MP-7401) for 1 h at room temperature. Bound Abs were visualized using diaminobenzidine staining (Dako liquid + diaminobenzidine substrate chromogen system, no. K3468). Sections were counterstained with hematoxylin (Vector Laboratories, no. H-3401). Slides (one section per mouse) were scanned with a Vectra Polaris automated quantitative pathology imaging system (PerkinElmer) and evaluated in a blinded fashion by two independent investigators. For the immunohistochemistry score of cleaved caspase-3, at least 38 crypts were evaluated per mouse, and results are given as the number of positive cells per crypt.

Microbiota analysis (shown in Fig. 5) was performed as described in Biehl et al. (43) with feces instead of rectal swabs as sample material. Data analysis was conducted with QIIME v1.9 (44). Open-reference operational taxonomic unit (OTU) clustering and taxonomy assignment of sequences were done with UCLUST (45) against the Silva database release 119 (46) at the 97% similarity level. The α diversity and β diversity metrics were finally calculated after rarefying >2000 sequences per sample. Linear discriminant analysis (LDA) coupled with effect size measurement (LDA effect size [LEfSe] analysis) was conducted (non-parametric Wilcoxon sum-rank test followed by LDA analysis) (47). Bray–Curtis distance matrices were calculated with QIIME for principal component analysis, and statistical testing was done with Adonis (permutational multivariate ANOVA using the Bray–Curtis distance matrices).

Microbiota analysis shown in Supplemental Fig. 3 was performed by amplifying the V3–V4 regions (25 cycles) from 24 ng of template DNA in a two-step process (48) using primer 341F and 785R (49) and a combinatorial dual indexing strategy. PCR products were purified using magnetic beads (Beckman Coulter) and pooled in an equimolar amount of 2 nM. The multiplexed samples were sequenced on an Illumina HiSeq in paired-end mode (2 × 250 bp) using the Rapid v2 chemistry. For amplicon sequence analysis, all data were analyzed using IMNGS (50) and Rhea (51). Sequencing data were preprocessed using the IMNGS pipeline. Seventeen nucleotides on the 5′ end and 21 nt on the 3′ end were trimmed for the R1 and R2 reads, respectively (trim score 20), and an expected error rate of 2. Chimera were removed using UCHIME (52), and the reads of demultiplexed samples were merged and clustered by 97% similarity using UPARSE v11.0 (53). OTUs occurring at a relative abundance <0.25% across all samples were removed to prevent the analysis of spurious OTUs. Taxonomy was assigned using the RDP classifier version 2.11 and confirmed using the SILVA database (46). Downstream analysis was performed using R package Rhea (51). OTUs were normalized and percentage relative abundance was computed. β Diversity analysis was used to assess the diversity between groups based on generalized UniFrac distances (54). α Diversity within species was calculated based on Shannon effective diversity.

Data are accessible at National Center for Biotechnology Information BioProject accession no. PRJNA877569 (https://www.ncbi.nlm.nih.gov/sra/PRJNA877569).

RNA was isolated from snap-frozen tissue or freshly sorted cells using TRIzol reagent (Thermo Fisher Scientific, no. 15596018) and reverse transcribed using SuperScript III (Thermo Fisher Scientific, no. 18080-044) according to the manufacturer’s instructions. Quantitative real-time PCR was performed on a LightCycler 480 (Roche, Basel, Switzerland) using predeveloped primer-probe sets for Ccr4, Ccl17, Csf2, Gzmb, Ifn-g, Il17a, Muc2, Reg3g (Thermo Fisher Scientific), and Hprt (Integrated DNA Technologies). Data were analyzed with the 2−ΔΔCt method (55).

Library preparation for bulk 3′ sequencing of poly(A)-RNA was performed as described by Parekh et al. (56) with minor modifications. Briefly, for each sample, barcoded full-length cDNA was generated with a Maxima RT polymerase (Thermo Fisher Scientific) using oligo(dT) primer–containing barcodes, unique molecular identifiers (UMIs), and an adapter. The addition of a template-switch oligonucleotide resulted in the extension of 5′ ends of the cDNAs, and full-length cDNA was amplified with a primer binding to the template-switch oligonucleotide site and the adapter. cDNA was fragmented with the Nextera XT kit (Illumina), and only the 3′-end fragments were finally amplified using primers with Illumina P5 and P7 overhangs. In comparison with Parekh et al. (56), the P5 and P7 sites were exchanged to allow sequencing of the cDNA in read1 and barcodes and UMIs in read2 to achieve a better cluster recognition. The library was sequenced on the NextSeq 500 platform (Illumina) with 75 cycles for the cDNA and 16 cycles for the barcodes and UMIs. Raw sequencing data were processed with Dropseq-tools version 1.12 using gene annotations from the Ensembl GRCm38.87 database to generate sample- and gene-wise UMI tables. Downstream analysis was conducted with R v3.4.4 and DESeq2 v1.18.1. For gene set enrichment analysis (GSEA) the GSEA v4.0.3 Mac application was used. Data are accessible at National Center for Biotechnology Information Gene Expression Omnibus through accession no. GSE213751 (https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE213751).

Cecal contents were homogenized in 100 µl of 0.5% Tween 20 in PBS per 10 mg of cecal content and centrifuged at 100 × g for 15 min at 4°C to pellet large particles. Supernatants were transferred into fresh tubes and after addition of protease inhibitor, were centrifuged for 10 min at 10,000 × g at 4°C and the supernatants were removed. IgA was quantified in the supernatants and in serum samples using an anti-mouse IgA ELISA kit (Thermo Fisher Scientific, no. 88-50450-22) according to the manufacturer’s instructions. For detection of bacterial IgA coating, feces were homogenized in 100 µl of PBS per 10 mg, centrifuged at 50 × g for 15 min at 4°C, and the pellet was discarded. After two washes in 1% BSA in PBS, the fecal bacteria were stained with anti–IgA-PE Ab (Thermo Fisher Scientific, no. 12-4204-81) in PBS with 20% normal rat serum and 1% BSA and analyzed on a FACSCanto II (Becton Dickinson) flow cytometer.

For standard statistical analyses, Prism v7 (GraphPad Software, San Diego, CA) was used. Unless specified otherwise, individual data points, mean, and SEM are depicted. Unpaired, two-tailed Student t tests (for normally distributed data) or Mann–Whitney tests (for not normally distributed data) were performed for two-group comparisons. For comparisons of three or more groups, one-way ANOVA with a Holm–Sidak multiple comparison test or Kruskal–Wallis test with Dunn’s multiple comparison tests were used. A p value <0.05 and false discovery rate <0.25 were considered statistically significant.

We used heterozygous CCL17-eGFP knockin (ki) (wild-type [wt]/ki) mice (28) to investigate CCL17 expression in myeloid cell subsets in colitis-associated tumors compared with colitis and steady state. Subpopulations of monocytes/macrophages and cDCs were identified as shown in (Fig. 1A, and the percentage of CCL17-eGFP+ cells was measured by flow cytometry using cells isolated from wt mice as a control to set the gate on eGFP+ cells (Supplemental Fig. 1). In the steady state, CCL17-eGFP–expressing cells within CD45+CD11b+Ly6G cells were mainly found in CD11b+ DCs (p6) whereas the frequency of monocytes and macrophages expressing CCL17 was low (Fig. 1B). cDC2s and CD103+CD11b+ DCs showed a higher percentage of CCL17-expressing cells than did type 1 cDCs (Fig. 1B), consistent with previously published data (22). In the inflamed colon during acute DSS-induced colitis, also intermediate monocytes (p2) and MHC IIhi macrophages (p5) in addition to cDCs expressed CCL17 (Fig. 1C). In AOM/DSS-induced colon tumors, intermediate monocytes (p2), MHC IIhi macrophages (p5), and CD11b+ DCs (p6) showed a significantly higher frequency of CCL17-expressing cells compared with the nontumorous colon tissue of the same mice. Intratumoral Ly6CloMHC IIlo macrophages (p4) also contained a small fraction of CCL17-expressing cells (Fig. 1D). A trend toward a higher percentage of CCL17-eGFP+ cells was also observed in DC subpopulations in tumors versus nontumorous colon tissue (not significant). CCL17-eGFP expression was not detected in CD45 cells or CD45+ non-DC/non-macrophages in the tested conditions (Supplemental Fig. 1). The enrichment of CCL17+ cells in the tumors compared with nontumorous tissue was also observed by immunofluorescence (Fig. 1E). Ccl17 mRNA expression was increased 9-fold in the tumors compared with colon tissue of the same mice in accordance with the increased frequency of CCL17-expressing cells (Fig. 1F). Csf2 mRNA encoding GM-CSF, a known inducer of CCL17 expression (30), was concomitantly increased (Fig. 1F). Thus, CCL17 was strongly induced in colitis-associated colon tumors, especially in tumor-associated macrophages.

FIGURE 1.

CCL17 expression in the colon in steady-state, colitis, and colitis-induced tumors. CCL17-eGFP–expressing cells were detected in colon lamina propria cells and cells isolated from colon tumors by flow cytometry using CCL17-eGFP knock-in (wt/ki) mice. (A) Gating strategy for CD11b+Ly6G myeloid cells (left) and for DCs (right). Representative data from colitis samples are shown. (BD) CCL17-eGFP expression in subpopulations of CD45+CD11b+Ly6G myeloid cells (upper panels) and in subpopulations of CD45+CD64MHC IIhiCD11chi DCs (lower panels) (B) in the colon lamina propria fraction of untreated mice (steady state), (C) in the colon lamina propria fraction of mice with acute DSS-induced colitis, and (D) in cells isolated from digested colon tumor (columns with contour) and matched nontumoral colon tissue (columns w/o contour). Mean ± SEM. Each data point represents one mouse (n ≥ 3). Histograms show one representative example. *p < 0.05, unpaired, two-tailed Student t test. Data are representative of n ≥ 3 independent experiments. (E) Immunofluorescence staining of normal colonic (left) and colon tumor (right) tissue stained for GFP (green) and E-cadherin (gray). Scale bars, 10 µm. Data are representative of n = 3 independent experiments. (F) Relative expression of Ccl17 and Csf2 transcripts in AOM/DSS-induced colon tumor tissue (T) and matched nontumorous colon tissue (N) normalized to expression levels in nontumorous colon tissue. Mean ± SEM. Each data point represents one mouse (n ≥ 7). *p < 0.05, unpaired, two-tailed Mann–Whitney test. Data were pooled from three independent experiments.

FIGURE 1.

CCL17 expression in the colon in steady-state, colitis, and colitis-induced tumors. CCL17-eGFP–expressing cells were detected in colon lamina propria cells and cells isolated from colon tumors by flow cytometry using CCL17-eGFP knock-in (wt/ki) mice. (A) Gating strategy for CD11b+Ly6G myeloid cells (left) and for DCs (right). Representative data from colitis samples are shown. (BD) CCL17-eGFP expression in subpopulations of CD45+CD11b+Ly6G myeloid cells (upper panels) and in subpopulations of CD45+CD64MHC IIhiCD11chi DCs (lower panels) (B) in the colon lamina propria fraction of untreated mice (steady state), (C) in the colon lamina propria fraction of mice with acute DSS-induced colitis, and (D) in cells isolated from digested colon tumor (columns with contour) and matched nontumoral colon tissue (columns w/o contour). Mean ± SEM. Each data point represents one mouse (n ≥ 3). Histograms show one representative example. *p < 0.05, unpaired, two-tailed Student t test. Data are representative of n ≥ 3 independent experiments. (E) Immunofluorescence staining of normal colonic (left) and colon tumor (right) tissue stained for GFP (green) and E-cadherin (gray). Scale bars, 10 µm. Data are representative of n = 3 independent experiments. (F) Relative expression of Ccl17 and Csf2 transcripts in AOM/DSS-induced colon tumor tissue (T) and matched nontumorous colon tissue (N) normalized to expression levels in nontumorous colon tissue. Mean ± SEM. Each data point represents one mouse (n ≥ 7). *p < 0.05, unpaired, two-tailed Mann–Whitney test. Data were pooled from three independent experiments.

Close modal

RNA sequencing of myeloid cell subsets isolated from AOM/DSS-induced colon tumors (p1, p2, p3, and p4 as in (Fig. 1A) was performed to verify subset-specific expression of Ccl17 mRNA and characterize these subsets by transcriptional profiling (Supplemental Fig. 2A). Ccl17 expression was mainly detected in Ly6CloMHC IIhi cells (p3) encompassing MHC IIhi macrophages and CD11b+ DCs, consistent with the CCL17-eGFP signal detected by flow cytometry (Supplemental Fig. 2B). These cells also expressed the highest level of related chemokine Ccl22 (Supplemental Fig. 2D). Hierarchical clustering of differentially expressed genes and Gene Ontology term analysis revealed enrichment of genes involved in endocytosis and inflammatory activity in Ly6CloMHC IIhi cells (Supplemental Fig. 2C, 2D). Mrc1 (mannose receptor, CD206), a marker of immunosuppressive M2-like macrophages, was expressed at higher levels in Ly6CloMHC IIhi cells on mRNA and protein levels (Supplemental Fig. 2E), but these cells did not generally express higher levels of M2 marker genes or lower levels of M1 marker genes (Supplemental Fig. 2F). Thus, Ly6CloMHC IIhi cells expressing Ccl17 and Ccl22 in AOM/DSS-induced tumors are marked by higher expression of CD206 and a gene expression profile enriched for inflammatory cytokines and endocytosis.

Given the strong upregulation of CCL17 in colon tumor tissue and specifically in tumor-associated macrophages, we investigated whether CCL17 plays a nonredundant role in the tumorigenesis of AOM/DSS-induced colon tumors. CCL17-deficient (ki/ki) and CCL17-competent (wt/ki and wt/wt) mice received one injection of AOM and were subjected to three cycles of low-dose DSS as depicted in (Fig. 2A. To control for microbiota-mediated effects, which may influence colon tumor formation (8, 9), we bred CCL17-deficient mice from a homozygous ki/ki breeding pair for one generation and compared the offspring (ki single housed [SiHo]) with CCL17-deficient (ki cohoused [CoHo]) and with CCL17-competent (ctrl) mice generated as littermates generated from heterozygous breeding pairs that stem from the same initial breeding pair as the homozygous ki/ki breeding pair from heterozygous breeding pairs (Fig. 2B). There was no difference in the relative bodyweight during AOM/DSS treatment between the experimental groups (Fig. 2C). However, the number of tumors per mouse was significantly decreased in single-housed CCL17-deficient mice (ki SiHo) compared with cohoused CCL17-deficient (ki CoHo) and CCL17-competent (ctrl) mice (Fig. 2D). Although the mean tumor area was slightly larger in the single-housed CCL17-deficient group compared with the other groups (Fig. 2E), the percentage of Ki67+ tumor cells was not significantly different between the groups (Fig. 2F). Exposure of MC38 colon tumor cells to recombinant CCL17 showed no direct effect of this chemokine on tumor cell growth (Supplemental Fig. 4D). Thus, tumor multiplicity but not tumor growth was reduced in single-housed CCL17-deficient mice.

FIGURE 2.

CCL17-deficient single-housed mice show decreased colon tumor formation induced by AOM/DSS. Single-bred/-housed CCL17-deficient (ki/ki) and cohoused CCL17-ki/ki and CCL17-competent (wt/ki and wt/wt) littermates were treated with AOM and three cycles of DSS. (A) Experimental scheme of the AOM/DSS treatment. AOM and DSS were dissolved in drinking water (% wt/wt). (B) Breeding strategy for obtaining the experimental groups ctrl (CCL17-wt/wt and CCL17-wt/ki), ki CoHo (CCL17-ki/ki with CCL17-wt/wt and CCL17-wt/ki cohoused littermates) and ki SiHo (CCL17-ki/ki with CCL17-ki/ki littermates, separately housed from CCL17-wt/wt and CCL17-wt/ki mice). (C) Relative bodyweight during AOM/DSS treatment. The bodyweight was normalized to the starting weight at the beginning of each DSS cycle (indicated by arrows). G, ctrl; @, ki SiHo; X, ki CoHo. Mean ± SEM. Each data point represents one mouse (n ≥ 11, pooled from two experiments). (D) Number of tumors per mouse after AOM/DSS treatment. Each data point represents one mouse (n ≥ 11, pooled from two experiments, @ and C indicate the two experiments). Mean ± SEM. *p < 0.05, two-tailed Mann–Whitney test. (E) Area per tumor. Each data point represents one tumor (n ≥ 98). Line indicates median. *p < 0.05, Tukey’s multiple comparison t test. (F) Left, Percentage of Ki67+ cells, determined by immunofluorescence staining (Ki67+ of DAPI+ cells). Each data point represents one mouse (n ≥ 3). Mean ± SEM. Right, Representative section stained for DAPI (blue), Ki67 (magenta), and E-cadherin (gray). Scale bar, 10 µm.

FIGURE 2.

CCL17-deficient single-housed mice show decreased colon tumor formation induced by AOM/DSS. Single-bred/-housed CCL17-deficient (ki/ki) and cohoused CCL17-ki/ki and CCL17-competent (wt/ki and wt/wt) littermates were treated with AOM and three cycles of DSS. (A) Experimental scheme of the AOM/DSS treatment. AOM and DSS were dissolved in drinking water (% wt/wt). (B) Breeding strategy for obtaining the experimental groups ctrl (CCL17-wt/wt and CCL17-wt/ki), ki CoHo (CCL17-ki/ki with CCL17-wt/wt and CCL17-wt/ki cohoused littermates) and ki SiHo (CCL17-ki/ki with CCL17-ki/ki littermates, separately housed from CCL17-wt/wt and CCL17-wt/ki mice). (C) Relative bodyweight during AOM/DSS treatment. The bodyweight was normalized to the starting weight at the beginning of each DSS cycle (indicated by arrows). G, ctrl; @, ki SiHo; X, ki CoHo. Mean ± SEM. Each data point represents one mouse (n ≥ 11, pooled from two experiments). (D) Number of tumors per mouse after AOM/DSS treatment. Each data point represents one mouse (n ≥ 11, pooled from two experiments, @ and C indicate the two experiments). Mean ± SEM. *p < 0.05, two-tailed Mann–Whitney test. (E) Area per tumor. Each data point represents one tumor (n ≥ 98). Line indicates median. *p < 0.05, Tukey’s multiple comparison t test. (F) Left, Percentage of Ki67+ cells, determined by immunofluorescence staining (Ki67+ of DAPI+ cells). Each data point represents one mouse (n ≥ 3). Mean ± SEM. Right, Representative section stained for DAPI (blue), Ki67 (magenta), and E-cadherin (gray). Scale bar, 10 µm.

Close modal

As chemokine CCL17 was highly expressed in AOM/DSS-induced tumors by macrophages and DCs (Fig. 1), we hypothesized that it might influence the immune cell composition in the tumors, thereby affecting tumor development. Analysis of the immune cell subsets in AOM/DSS-induced tumors by flow cytometry revealed neither statistically significant differences in the frequencies of CD45+ immune cells, nor changes in the percentages of CD11b+ myeloid cells, granulocytes, monocytes, macrophages, and DC subpopulations between single-housed CCL17-deficient mice and the control groups (Fig. 3A, 3B). Global transcriptome analysis of myeloid cells isolated from AOM/DSS-induced tumors showed that monocytes (p1) and Ly6CloMHC IIhi (p3) cells from tumors of single-housed CCL17-deficient mice were enriched for genes involved in proteasomal degradation (e.g., Psmc2 and Psmb1) and oxidative phosphorylation (e.g., Uqcr10 and Atp6v1f). Intratumoral Ly6CloMHC IIhi (p3) cells from tumors of these mice were also enriched for genes involved in lysosomal activity (e.g., Psap and Hexb), suggesting increased endocytosis/phagocytosis levels (Fig. 3C). These results indicate functional differences of the myeloid cell compartment in tumors of the CCL17-deficient single-housed group, which coincided with reduced tumor numbers.

FIGURE 3.

Similar frequencies but altered gene expression signatures in tumor-associated myeloid cells of CCL17-deficient mice. Mice were treated with AOM and three cycles of DSS (as shown in (Fig. 2A). Tumor-infiltrating immune cells in tumors of ctrl, ki SiHo, and ki CoHo were analyzed. (A) Percentages of total CD45+ cells, CD11b+ myeloid cells of CD45+ cells, granulocytes of CD11b+ cells and monocytes, intermediate monocytes, and MHC IIhi and MHC IIlo macrophages within CD11b+Ly6G cells in tumors of ctrl, ki SiHo, and ki CoHo mice. Each data point represents one mouse (n ≥ 4). Mean ± SEM. (B) Percentages of total DCs of CD45+ cells and DC subsets within CD45+CD64MHC IIhiCD11chi DCs in tumors of ctrl, ki SiHo, and ki CoHo mice. Each data point represents one mouse (n ≥ 4). Mean ± SEM. (C) Gene set enrichment analyses of tumor-infiltrating CD11b+Ly6G myeloid cell populations p1–p4 from ki SiHo and ctrl mice. The color intensity of the circles denotes the significance level (false discovery rate [FDR]); the circle diameter reflects the normalized enrichment score (NES); red indicates GSE in ki SiHo mice; blue indicates GSE in ctrl mice; n ≥ 4 biological replicates for each group.

FIGURE 3.

Similar frequencies but altered gene expression signatures in tumor-associated myeloid cells of CCL17-deficient mice. Mice were treated with AOM and three cycles of DSS (as shown in (Fig. 2A). Tumor-infiltrating immune cells in tumors of ctrl, ki SiHo, and ki CoHo were analyzed. (A) Percentages of total CD45+ cells, CD11b+ myeloid cells of CD45+ cells, granulocytes of CD11b+ cells and monocytes, intermediate monocytes, and MHC IIhi and MHC IIlo macrophages within CD11b+Ly6G cells in tumors of ctrl, ki SiHo, and ki CoHo mice. Each data point represents one mouse (n ≥ 4). Mean ± SEM. (B) Percentages of total DCs of CD45+ cells and DC subsets within CD45+CD64MHC IIhiCD11chi DCs in tumors of ctrl, ki SiHo, and ki CoHo mice. Each data point represents one mouse (n ≥ 4). Mean ± SEM. (C) Gene set enrichment analyses of tumor-infiltrating CD11b+Ly6G myeloid cell populations p1–p4 from ki SiHo and ctrl mice. The color intensity of the circles denotes the significance level (false discovery rate [FDR]); the circle diameter reflects the normalized enrichment score (NES); red indicates GSE in ki SiHo mice; blue indicates GSE in ctrl mice; n ≥ 4 biological replicates for each group.

Close modal

Innate and adaptive lymphocytes also influence the TME and may have tumor-promoting as well as antitumor activity (5, 13). Besides a reduction of CD8+ T cells in tumors of single-housed CCL17-deficient mice, we found no statistically significant differences between the experimental groups in the percentages of B cells, CD4+ T cells, Foxp3+ Tregs, NK cells, and NKT cells (Fig. 4A–C). Gzmb, Il17, and Ifng gene expression in tumor tissue (Fig. 4D) and the frequencies of IFN-γ– and IL-17–producing cells within effector lymphocyte subsets were not higher in the ki SiHo group (Fig. 4E, 4F). Thus, CCL17 was not required for recruitment of innate and adaptive immune cells to the tumors, and the changes in gene expression of tumor-associated myeloid cells observed in single-housed CCL17-deficient mice were not associated with increased antitumor effector T cell responses.

FIGURE 4.

No difference in lymphocyte infiltration in AOM/DSS-induced tumors between single-housed CCL17-deficient mice and controls. Mice were treated with AOM and three cycles of DSS (as shown in (Fig. 2A). Tumor-infiltrating immune cells in tumors of ctrl, ki SiHo, and ki CoHo were analyzed. (A) Frequency of B220+ within intratumoral CD45+ cells. Each data point represents one mouse (n ≥ 4). Mean ± SEM. (B) CD4+ T cell, Treg, and CD8+ T cell frequencies (of CD45+ cells). (C) percentages of NK and NKT cells (of CD45+ cells) in ctrl, ki SiHo, and ki CoHo tumors. Each data point represents one mouse (n ≥ 4). Mean ± SEM. (D) Relative gene expression of gzmb, il17, and ifng in tumors of ctrl, ki SiHo, and ki CoHo mice. Each data point represents one mouse (n ≥ 6). Mean ± SEM. (E and F) Percentages of IFN-γ–expressing cells (E) and IL-17–expressing cells (F) in tumors of ctrl, ki SiHo, and ki CoHo mice. Each data point represents one mouse (n ≥ 4). Mean ± SEM.

FIGURE 4.

No difference in lymphocyte infiltration in AOM/DSS-induced tumors between single-housed CCL17-deficient mice and controls. Mice were treated with AOM and three cycles of DSS (as shown in (Fig. 2A). Tumor-infiltrating immune cells in tumors of ctrl, ki SiHo, and ki CoHo were analyzed. (A) Frequency of B220+ within intratumoral CD45+ cells. Each data point represents one mouse (n ≥ 4). Mean ± SEM. (B) CD4+ T cell, Treg, and CD8+ T cell frequencies (of CD45+ cells). (C) percentages of NK and NKT cells (of CD45+ cells) in ctrl, ki SiHo, and ki CoHo tumors. Each data point represents one mouse (n ≥ 4). Mean ± SEM. (D) Relative gene expression of gzmb, il17, and ifng in tumors of ctrl, ki SiHo, and ki CoHo mice. Each data point represents one mouse (n ≥ 6). Mean ± SEM. (E and F) Percentages of IFN-γ–expressing cells (E) and IL-17–expressing cells (F) in tumors of ctrl, ki SiHo, and ki CoHo mice. Each data point represents one mouse (n ≥ 4). Mean ± SEM.

Close modal

The observed reduction in tumor numbers in CCL17-deficient mice appeared to be microbiota-dependent, as CCL17-deficient mice cohoused with CCL17-competent littermates did not show this phenotype. We therefore analyzed the steady-state fecal microbiota of single-housed CCL17-deficient mice in comparison with that of controls and CCL17-deficient mice cohoused with controls. The number of observed OTUs was significantly smaller in feces of single-housed CCL17-deficient mice than in the feces of the cohoused mice (Fig. 5A). The Bray–Curtis principal coordinate analysis revealed a significant alteration of the bacterial composition in the feces of single-housed CCL17-deficient mice versus cohoused control mice (Fig. 5B). Further analysis using the LEfSE method showed an enrichment of the genus Akkermansia (belonging to the Verrucomicrobia family) and the Peptococcaceae family in the feces of single-housed CCL17-deficient mice (Fig. 5C), coinciding with fewer tumors developing in these mice in the AOM/DSS model. Reduced diversity of the fecal microbiota in single-housed CCL17-deficient mice was confirmed in an additional experiment that included also single-housed control mice (wt/wt and wt/ki). Although α diversity was again reduced in single-housed CCL17-deficient mice, no differences in α diversity were detected between single-housed and cohoused controls or between wt/wt and wt/ki controls (Supplemental Fig. 3A–C). Samples from single-housed CCL17-deficient mice grouped together distant from cohoused CCL17-deficient and CCL17-competent control mice in the principal coordinate analysis (Supplemental Fig. 3D).

FIGURE 5.

CCL17 deficiency leads to an altered microbiota. (A) Number of observed OTUs versus number of sequences per sample in the feces of steady-state ki SiHo, ctrl CoHo, and ki CoHo mice. For statistical analysis, ctrl CoHo and ki CoHo samples were analyzed as one group. Mean ± SEM (n ≥ 7). *p < 0.05 unpaired, two-tailed Student t test. (B) Bray–Curtis principal coordinate analysis of the microbiota composition in ki SiHo, ctrl CoHo, and ki CoHo mice. Each data point represents one mouse. *p < 0.05, Adonis (permutational multivariate analysis of variance using the Bray–Curtis distance matrices). (C) Linear discriminant analysis (LDA) effect size (LEfSe) analysis comparing ki SiHo, ctrl CoHo, and ki CoHo groups. Left, taxonomic cladogram generated by LEfSe. Right, LDA score. Red bars represent taxa significantly enriched in ki SiHo mice. (D) Cecal (left) and serum (right) IgA concentrations in ctrl and ki SiHo mice. Each data point represents one mouse (n = 5). Mean ± SEM. *p < 0.05 unpaired, two-tailed Student’s t test. (E) Left, Percentage of IgA-coated bacteria, determined by flow cytometry. Each data point represents one mouse (n = 5). Mean ± SEM. *p < 0.05 unpaired, two-tailed Student t test. Right, Representative contour plots of IgA coating in bacteria from ctrl and ki SiHo fecal samples. Numbers indicate percentages.

FIGURE 5.

CCL17 deficiency leads to an altered microbiota. (A) Number of observed OTUs versus number of sequences per sample in the feces of steady-state ki SiHo, ctrl CoHo, and ki CoHo mice. For statistical analysis, ctrl CoHo and ki CoHo samples were analyzed as one group. Mean ± SEM (n ≥ 7). *p < 0.05 unpaired, two-tailed Student t test. (B) Bray–Curtis principal coordinate analysis of the microbiota composition in ki SiHo, ctrl CoHo, and ki CoHo mice. Each data point represents one mouse. *p < 0.05, Adonis (permutational multivariate analysis of variance using the Bray–Curtis distance matrices). (C) Linear discriminant analysis (LDA) effect size (LEfSe) analysis comparing ki SiHo, ctrl CoHo, and ki CoHo groups. Left, taxonomic cladogram generated by LEfSe. Right, LDA score. Red bars represent taxa significantly enriched in ki SiHo mice. (D) Cecal (left) and serum (right) IgA concentrations in ctrl and ki SiHo mice. Each data point represents one mouse (n = 5). Mean ± SEM. *p < 0.05 unpaired, two-tailed Student’s t test. (E) Left, Percentage of IgA-coated bacteria, determined by flow cytometry. Each data point represents one mouse (n = 5). Mean ± SEM. *p < 0.05 unpaired, two-tailed Student t test. Right, Representative contour plots of IgA coating in bacteria from ctrl and ki SiHo fecal samples. Numbers indicate percentages.

Close modal

Secretory IgA regulates microbiota composition and maintains intestinal homeostasis (57, 58). IgA levels were significantly elevated in the cecal content but not in the serum of single-housed CCL17-deficient mice (Fig. 5D), demonstrating that only intestinal and not systemic IgA production was altered in these mice. Bacterial IgA coating has been suggested to influence microbial composition (58, 59). We found significantly more IgA-coated bacteria in the feces of single-housed CCL17-deficient mice than in control mice (Fig. 5E). Thus, secretory IgA and IgA-coating of bacteria coincided with the establishment of an altered microbiota in CCL17-deficient mice, which was protective against colitis-induced tumor formation.

The degree of inflammation and regeneration after initial exposure to AOM/DSS is important for tumor incidence and early tumor promotion in this model (7). Single-housed CCL17-deficient mice (ki/ki) and control mice (wt/ki) were injected with AOM followed by one cycle of DSS and analyzed until day 14 (Fig. 6A). We did not observe statistically significant differences in weight loss during the inflammation phase (day 8–12) or in weight gain during the regeneration phase (Fig. 6B). The frequencies of CD45+ leukocytes in the colon lamina propria were not altered between the groups, indicating a similar inflammatory response. The percentages of granulocytes, monocytes, and macrophages within CD45+ lamina propria leukocytes were not significantly different between the groups (Fig. 6C). Histological examination did not reveal major differences in epithelial damage, leukocyte infiltration, and regenerative responses of the epithelium (Fig. 6D). It is therefore unlikely that reduced tumor numbers in single-housed CCL17-deficient mice are a consequence of reduced inflammation or regeneration during tumor initiation.

FIGURE 6.

Increased AOM-induced apoptosis of colonic epithelial cells in CCL17-deficient mice. (A) Single-housed CCL17-deficient and control mice were injected with 10 mg/kg AOM followed by one cycle of 2.5% DSS and were analyzed after 14 d during the regeneration phase. (B) Relative bodyweight during AOM/DSS treatment. open squres, ctrl; filled circles, ki SiHo. Mean ± SEM (n ≥ 4). (C) Frequencies of CD45+ cells and myeloid cell subpopulations within CD45+ cells in the colon lamina propria on day 14. Mean ± SEM. Each data point represents one mouse (n = 4). *p < 0.05, unpaired, two-tailed Student t test. (D) Representative images of H&E-stained colon tissue sections from a ctrl and a ki SiHo mouse. Scale bars, 10 µm. (E) Mice were injected once with 10 mg/kg AOM i.p. followed 5 d later by exposure to DSS (2.5%) in drinking water for 2 d. (F and G) Detection of apoptosis in colonic epithelium by cleaved caspase-3 (cC3) immunohistochemistry. Scale bars, 800 µm (top), 50 µm (bottom). (F) The number of positive cells was determined in ≥60 crypts per mouse (untreated mice, n = 2; single-housed CCL17-ki/ki mice, n = 6; CCL17-wt/ki mice, n = 5). (G) Representative images of tissue sections stained for cC3 (brown) and counterstained with hematoxylin (blue). (H) Mice were injected with 10 mg/kg AOM i.p. and sacrificed 8 h later. (I and J) Detection of apoptosis in colonic epithelium by cleaved caspase-3 immunohistochemistry. Scale bars, 20 μm. (I) The number of positive cells was determined in ≥60 crypts per mouse (untreated mice, n = 2; CCL17-ki/ki mice, n = 6; CCL17-wt/ki mice, n = 5). (J) Representative image of colon tissue sections stained for cC3 with detection of cC3+ cells in the base of crypts. In (F) and (I), the mean numbers of cC3+ cells per crypt are shown. (K) Mice were treated as described in (B). Left, Relative Muc2 mRNA expression in colon tissue. Right, Muc2 detected by immunohistochemistry in FFPE sections of the colon. Representative images are shown. Scale bars, 20 µm. In (F), (I), and (K), each data point represents results of one mouse; columns indicate mean values, error bars indicate SEM. *p < 0.05, **p < 0.01 by unpaired Student t test corrected with Holm–Sidak method.

FIGURE 6.

Increased AOM-induced apoptosis of colonic epithelial cells in CCL17-deficient mice. (A) Single-housed CCL17-deficient and control mice were injected with 10 mg/kg AOM followed by one cycle of 2.5% DSS and were analyzed after 14 d during the regeneration phase. (B) Relative bodyweight during AOM/DSS treatment. open squres, ctrl; filled circles, ki SiHo. Mean ± SEM (n ≥ 4). (C) Frequencies of CD45+ cells and myeloid cell subpopulations within CD45+ cells in the colon lamina propria on day 14. Mean ± SEM. Each data point represents one mouse (n = 4). *p < 0.05, unpaired, two-tailed Student t test. (D) Representative images of H&E-stained colon tissue sections from a ctrl and a ki SiHo mouse. Scale bars, 10 µm. (E) Mice were injected once with 10 mg/kg AOM i.p. followed 5 d later by exposure to DSS (2.5%) in drinking water for 2 d. (F and G) Detection of apoptosis in colonic epithelium by cleaved caspase-3 (cC3) immunohistochemistry. Scale bars, 800 µm (top), 50 µm (bottom). (F) The number of positive cells was determined in ≥60 crypts per mouse (untreated mice, n = 2; single-housed CCL17-ki/ki mice, n = 6; CCL17-wt/ki mice, n = 5). (G) Representative images of tissue sections stained for cC3 (brown) and counterstained with hematoxylin (blue). (H) Mice were injected with 10 mg/kg AOM i.p. and sacrificed 8 h later. (I and J) Detection of apoptosis in colonic epithelium by cleaved caspase-3 immunohistochemistry. Scale bars, 20 μm. (I) The number of positive cells was determined in ≥60 crypts per mouse (untreated mice, n = 2; CCL17-ki/ki mice, n = 6; CCL17-wt/ki mice, n = 5). (J) Representative image of colon tissue sections stained for cC3 with detection of cC3+ cells in the base of crypts. In (F) and (I), the mean numbers of cC3+ cells per crypt are shown. (K) Mice were treated as described in (B). Left, Relative Muc2 mRNA expression in colon tissue. Right, Muc2 detected by immunohistochemistry in FFPE sections of the colon. Representative images are shown. Scale bars, 20 µm. In (F), (I), and (K), each data point represents results of one mouse; columns indicate mean values, error bars indicate SEM. *p < 0.05, **p < 0.01 by unpaired Student t test corrected with Holm–Sidak method.

Close modal

Early apoptosis of intestinal epithelial cells in response to AOM/DSS or AOM alone was shown to reduce the number of tumors that ultimately develop in the AOM/DSS model (7, 60, 61). We therefore quantified the apoptotic epithelial cells in colon sections after short-term exposure to AOM/DSS or to AOM alone by staining for cleaved caspase-3. In single-housed CCL17-deficient mice injected with AOM and exposed to DSS for 2 d (Fig. 6E), we observed a significantly higher number of apoptotic cells per colon crypt compared with CCL17 wt/ki mice (Fig. 6F). Cells staining positively for cleaved caspase-3 were detected at the top of the crypts at this time point (Fig. 6G). We therefore investigated the immediate effects of the carcinogen 8 h after injection of AOM (Fig. 6H) and also found a higher number of apoptotic cells per colon crypt in single-housed CCL17-deficient mice than in CCL17 wt/ki mice (Fig. 6I). At this time point, cells positive for cleaved caspase-3 were detected not only at the top of the crypts but also at the base of the crypts (Fig. 6J). These results show that colon epithelial cells in CCL17-deficient mice are more susceptible to early apoptosis induced by AOM or AOM/DSS, which may lead to reduced tumor numbers.

Mucus produced by goblet cells forms a protective layer that reduces contact of the colon epithelium with fecal bacteria and protects against tumor development. We therefore measured the expression of Muc2, the secretory mucin, in the colon of mice after short-term exposure to AOM (as in (Fig. 6H). Muc2 mRNA expression was comparable between single-housed CCL17-deficient and control mice, and Muc2 staining in colon tissue sections did not reveal overt differences between the experimental groups (Fig. 6K). Similarly, in untreated mice the number of mucus-producing goblet cells was not different (Supplemental Fig. 4A, 4B), and no major differences of the mucus layer were found as shown by Muc2 staining in Carnoy’s-fixed colon tissue sections (Supplemental Fig. 4C). Thus, it is unlikely that the increased epithelial apoptosis in response to the carcinogen AOM observed in single-housed CCL17-deficient mice is due to an impaired mucus barrier. Taken together, our results show that separately housed CCL17-deficient mice establish an altered microbiota, which is associated with reduced survival of initiated epithelial cells after exposure to the carcinogen leading to reduced tumor numbers.

In this study we observed a reduced number of colitis-associated colon tumors in CCL17-deficient mice dependent on the microbiota, which was altered in separately housed CCL17-deficient mice. This phenotype was not associated with reduced inflammation but coincided with increased apoptosis in the colon during the early phase of tumor development as well as with functional changes in tumor-infiltrating myeloid cells, whereas recruitment of innate and adaptive immune cells and effector lymphocyte functions were not affected.

We detected CCL17 expression in the colon lamina propria specifically in CD11b+ and CD103+ CD11b+ DC subsets in the steady state. This is in line with previous studies reporting constitutive expression of CCL17 in CD11b+ cDCs in lymph nodes and other organs (23, 28, 29). In the present study, we found that CCL17 was also expressed by intermediate monocytes and macrophages in colitis-induced colon tumors. A similar expression pattern was also seen in spontaneously forming intestinal tumors in AOM-treated Apc1638N/+ mice (32). Induction of CCL17 expression was also found in mesenteric lymph node DCs and MHC IIhi macrophages postinfection with Salmonella typhimurium (62). One of the driving factors for CCL17 expression in tumors may be GM-CSF, which we found to be strongly induced in the tumors. GM-CSF was identified as a key cytokine inducing CCL17 expression in monocytes and macrophages via induction of IFN regulatory factor 4 (IRF4) in models of inflammatory arthritis and peritonitis (35). In addition to GM-CSF, other exogenous factors such as IL-4 (29, 63) or microbial signals may contribute to the induction of CCL17 expression in the TME.

Our finding that effector T cell frequency was not altered in single-housed CCL17-deficient mice indicated that the observed reduction in tumor number was not due to increased recruitment or activation of T cells directed against the tumor. The frequency of Tregs and Th17 cells, which have been shown to migrate in response to CCL17 (18), was not reduced in tumors of CCL17-deficient mice, indicating that CCL17 is redundant for T cell migration to the tumors in this model. The related chemokine CCL22, which was shown to be important for Treg recruitment and functionality, may compensate for the lack of CCL17 in this regard (21, 64). CCL17 was also found to be dispensable for the recruitment of myeloid subpopulations into the AOM/DSS-induced tumors, indicating that the previously described enhancing effect of CCL17 on transendothelial migration and recruitment of DCs in response to CCR7 ligands was not relevant in the present study (23, 24).

Interestingly, the protective effect of CCL17 deficiency was dependent on the altered microbiota that established itself in separately housed CCL17-deficient mice. To our knowledge, this is the first report of a microbiota-dependent phenotype in the CCL17-deficient ki mouse model, which may be relevant for other phenotypes that have been observed in this mouse line (2226). The intestinal microbiota composition was shown to be a crucial determinant of the tumor number in the colitis-associated colon tumor model (6568). Intestinal bacteria and their metabolites modulate tumor initiation and tumor promotion by influencing the genotoxic effect of the carcinogen, by promoting or inhibiting colonic inflammation, and by modifying the TME (5). We found alterations in the abundance of commensal bacteria taxa that may influence colon tumorigenesis. For example, the genus Akkermansia was increased in single-housed CCL17-deficient mice. The role of Akkermansia for intestinal health and disease is controversial. By degrading mucins, Akkermansia muciniphila was shown to adhere to colonic epithelial cells and promote the integrity of the epithelial barrier (69). A higher tumor load in CX3CR1-deficient mice in the AOM/DSS model was associated with reduced A. muciniphila in these mice (70). However, administration of A. muciniphila promoted tumor formation in the AOM/DSS-induced colon tumor model (71). We observed that microbiota diversity and composition were altered in single-housed CCL17-deficient mice, and the reduction in tumor multiplicity was reversed by cohousing, indicating an important role of the intestinal microbiota for the observed phenotype. The role of specific microbiota components for colon tumorigenesis in this context remains to be investigated.

It is unlikely that CCL17 has a direct effect on the intestinal microbiota, because it was shown to have only a low to moderate antimicrobial activity against Escherichia coli and Staphylococcus aureus in vitro at high doses (72). Increased intestinal IgA and IgA-coating of commensal bacteria observed in single-housed CCL17-deficient mice could be involved in the observed changes in the microbiota composition. Intestinal IgA, produced by plasma cells in GALTs, which can bind to intestinal bacteria, is an important regulator of the intestinal microbiota (57, 73). In our study, a higher percentage of IgA-coated fecal bacteria coincided with a lower number of different fecal OTUs in single-housed CCL17-deficient mice. Thus, higher IgA coating may reduce tumorigenic OTUs in the single-housed CCL17-deficient mice. It has been shown that intestinal DCs and macrophages can influence the microbiota (70, 74). These cells also affect intestinal IgA responses for example by releasing cytokines such as TNF-α or BAFF and APRIL, which support survival of IgA-producing plasma cells (7577). Our results suggest that CCL17 is one of the factors produced by DCs that influence microbiota composition and IgA production in the steady state.

Inflammation contributes to the initiation and promotion of colon tumor development (5), and reduced inflammation was observed in CCL17-deficient mice in several models of inflammation, including colitis (2227). Although CCL17-deficient mice showed reduced weight loss in an acute colitis model induced by short high-dose exposure to DSS (4%) (22), decreased inflammation was not seen in CCL17-deficient mice in the AOM/DSS-induced colon tumor model after the first cycle of DSS in our study. This discrepancy may be explained by the lower dose of DSS used, the additional effect of the carcinogen AOM, as well as different housing conditions compared with the previous study. Incongruity or divergence between the degree of colitis and the colon tumor load in the AOM/DSS model has also been observed in other genetically altered mouse strains, demonstrating that additional factors can be decisive for tumor development in this model (7880).

Interestingly, tumor multiplicity, but not tumor growth, was reduced in the single-housed CCL17-deficient mice, suggesting that the tumor initiation is affected. It was shown previously that mice lacking IKKβ or STAT3 in intestinal epithelial cells in which carcinogen-exposed enterocytes were not rescued from apoptosis in the tumor initiation phase had lower tumor numbers in the AOM/DSS model (6, 7). Similarly, we observed an increased epithelial cell apoptosis induced by short-term exposure to AOM alone or AOM/DSS in the colon of single-housed CCL17-deficient mice, which developed fewer tumors. This enhanced susceptibility to apoptosis was not associated with an impaired production of mucus and formation of a mucus layer, which was shown to be protective against colon tumor formation (81).

Our results show that lack of myeloid cell–derived chemokine CCL17 leads to changes in the microbiota, which are associated with increased susceptibility to carcinogen-induced apoptosis in initiated colon epithelial cells as a likely explanation for reduced tumor multiplicity. These effects should be considered when developing drugs that target CCL17 to treat inflammatory diseases or inflammation-induced colon cancer.

We acknowledge the Core Facility Animal Models, the Core Facility Bioimaging, and the Core Facility Flow Cytometry at the Biomedical Center, LMU Munich. We thank Lisa Richter for support in cell sorting. We thank Meryem Gülfem Öner-Ziegler for sharing protocols and giving advice. We thank Irmgard Förster for providing the CCL17-eGFP knockin mouse line.

This work was supported by Deutsche Forschungsgemeinschaft Grants 252441188, 46383290, 210592381 SFB1054 TPA06, and by the Georg and Traud Gravenhorst Stiftung (to A.B.K.). T.B. received funding from Deutsche Forschungsgemeinschaft Grant 210592381 SFB1054 TP B03. R.M. received a Ph.D. scholarship from the Studienstiftung des Deutschen Volkes. V.F. received a Ph.D. scholarship from Quantitative Biosciences Munich. E.W. received a Ph.D. scholarship from the Villigst Foundation. B.S. acknowledges funding from the Center for Gastrointestinal Microbiome Research.

The raw sequence data presented in this article have been submitted to the BioProject under accession number PRJNA877569 and to the Gene Expression Omnibus under accession no. GSE213751.

The online version of this article contains supplemental material.

Abbreviations used in this article:

     
  • AOM

    azoxymethane

  •  
  • cDC2

    type 2 conventional DC

  •  
  • CRC

    colorectal cancer

  •  
  • DC

    dendritic cell

  •  
  • DSS

    dextran sodium sulfate

  •  
  • eGFP

    enhanced GFP

  •  
  • ki

    knockin

  •  
  • ki CoHo

    ki cohoused

  •  
  • ki SiHo

    ki single housed

  •  
  • LDA

    linear discriminant analysis

  •  
  • LEfSe

    LDA effect size

  •  
  • MHC II

    MHC class II

  •  
  • OTU

    operational taxonomic unit

  •  
  • TME

    tumor microenvironment

  •  
  • Treg

    regulatory T cell

  •  
  • UMI

    unique molecular identifier

  •  
  • wt

    wild-type

1.
Bray
F.
,
J.
Ferlay
,
I.
Soerjomataram
,
R. L.
Siegel
,
L. A.
Torre
,
A.
Jemal
.
2018
.
Global cancer statistics 2018: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. [Published erratum appears in 2020 CA Cancer J. Clin. 70: 313.]
CA Cancer J. Clin.
68
:
394
424
.
2.
Keum
N.
,
E.
Giovannucci
.
2019
.
Global burden of colorectal cancer: emerging trends, risk factors and prevention strategies.
Nat. Rev. Gastroenterol. Hepatol.
16
:
713
732
.
3.
Grivennikov
S. I.
,
F.
Cominelli
.
2016
.
Colitis-associated and sporadic colon cancers: different diseases, different mutations?
Gastroenterology
150
:
808
810
.
4.
Rogler
G.
2014
.
Chronic ulcerative colitis and colorectal cancer.
Cancer Lett.
345
:
235
241
.
5.
Schmitt
M.
,
F. R.
Greten
.
2021
.
The inflammatory pathogenesis of colorectal cancer.
Nat. Rev. Immunol.
21
:
653
667
.
6.
Bollrath
J.
,
T. J.
Phesse
,
V. A.
von Burstin
,
T.
Putoczki
,
M.
Bennecke
,
T.
Bateman
,
T.
Nebelsiek
,
T.
Lundgren-May
,
O.
Canli
,
S.
Schwitalla
, et al
2009
.
gp130-mediated Stat3 activation in enterocytes regulates cell survival and cell-cycle progression during colitis-associated tumorigenesis.
Cancer Cell
15
:
91
102
.
7.
Greten
F. R.
,
L.
Eckmann
,
T. F.
Greten
,
J. M.
Park
,
Z.-W.
Li
,
L. J.
Egan
,
M. F.
Kagnoff
,
M.
Karin
.
2004
.
IKKβ links inflammation and tumorigenesis in a mouse model of colitis-associated cancer.
Cell
118
:
285
296
.
8.
Ohtani
N.
2015
.
Microbiome and cancer.
Semin. Immunopathol.
37
:
65
72
.
9.
Arthur
J. C.
,
E.
Perez-Chanona
,
M.
Mühlbauer
,
S.
Tomkovich
,
J. M.
Uronis
,
T. J.
Fan
,
B. J.
Campbell
,
T.
Abujamel
,
B.
Dogan
,
A. B.
Rogers
, et al
2012
.
Intestinal inflammation targets cancer-inducing activity of the microbiota.
Science
338
:
120
123
.
10.
Louis
P.
,
G. L.
Hold
,
H. J.
Flint
.
2014
.
The gut microbiota, bacterial metabolites and colorectal cancer.
Nat. Rev. Microbiol.
12
:
661
672
.
11.
Zheng
D.
,
T.
Liwinski
,
E.
Elinav
.
2020
.
Interaction between microbiota and immunity in health and disease.
Cell Res.
30
:
492
506
.
12.
De Oliveira
T.
,
M.
Ramakrishnan
,
M. A.
Diamanti
,
P. K.
Ziegler
,
F.
Brombacher
,
F. R.
Greten
.
2019
.
Loss of Stat6 affects chromatin condensation in intestinal epithelial cells causing diverse outcome in murine models of inflammation-associated and sporadic colon carcinogenesis.
Oncogene
38
:
1787
1801
.
13.
Grivennikov
S. I.
,
F. R.
Greten
,
M.
Karin
.
2010
.
Immunity, inflammation, and cancer.
Cell
140
:
883
899
.
14.
Halama
N.
,
I.
Zoernig
,
A.
Berthel
,
C.
Kahlert
,
F.
Klupp
,
M.
Suarez-Carmona
,
T.
Suetterlin
,
K.
Brand
,
J.
Krauss
,
F.
Lasitschka
, et al
2016
.
Tumoral immune cell exploitation in colorectal cancer metastases can be targeted effectively by anti-CCR5 therapy in cancer patients.
Cancer Cell
29
:
587
601
.
15.
Iellem
A.
,
M.
Mariani
,
R.
Lang
,
H.
Recalde
,
P.
Panina-Bordignon
,
F.
Sinigaglia
,
D.
D’Ambrosio
.
2001
.
Unique chemotactic response profile and specific expression of chemokine receptors CCR4 and CCR8 by CD4+CD25+ regulatory T cells.
J. Exp. Med.
194
:
847
853
.
16.
Inngjerdingen
M.
,
B.
Damaj
,
A. A.
Maghazachi
.
2000
.
Human NK cells express CC chemokine receptors 4 and 8 and respond to thymus and activation-regulated chemokine, macrophage-derived chemokine, and I-309.
J. Immunol.
164
:
4048
4054
.
17.
Lieberam
I.
,
I.
Förster
.
1999
.
The murine β-chemokine TARC is expressed by subsets of dendritic cells and attracts primed CD4+ T cells.
Eur. J. Immunol.
29
:
2684
2694
.
18.
Scheu
S.
,
S.
Ali
,
C.
Ruland
,
V.
Arolt
,
J.
Alferink
.
2017
.
The C-C chemokines CCL17 and CCL22 and their receptor CCR4 in CNS autoimmunity.
Int. J. Mol. Sci.
18
:
2306
.
19.
Imai
T.
,
D.
Chantry
,
C. J.
Raport
,
C. L.
Wood
,
M.
Nishimura
,
R.
Godiska
,
O.
Yoshie
,
P. W.
Gray
.
1998
.
Macrophage-derived chemokine is a functional ligand for the CC chemokine receptor 4.
J. Biol. Chem.
273
:
1764
1768
.
20.
Imai
T.
,
M.
Baba
,
M.
Nishimura
,
M.
Kakizaki
,
S.
Takagi
,
O.
Yoshie
.
1997
.
The T cell-directed CC chemokine TARC is a highly specific biological ligand for CC chemokine receptor 4.
J. Biol. Chem.
272
:
15036
15042
.
21.
Anz
D.
,
M.
Rapp
,
S.
Eiber
,
V. H.
Koelzer
,
R.
Thaler
,
S.
Haubner
,
M.
Knott
,
S.
Nagel
,
M.
Golic
,
G. M.
Wiedemann
, et al
2015
.
Suppression of intratumoral CCL22 by type I interferon inhibits migration of regulatory T cells and blocks cancer progression.
Cancer Res.
75
:
4483
4493
.
22.
Heiseke
A. F.
,
A. C.
Faul
,
H. A.
Lehr
,
I.
Förster
,
R. M.
Schmid
,
A. B.
Krug
,
W.
Reindl
.
2012
.
CCL17 promotes intestinal inflammation in mice and counteracts regulatory T cell-mediated protection from colitis.
Gastroenterology
142
:
335
345
.
23.
Ruland
C.
,
H.
Renken
,
I.
Kuzmanov
,
A.
Fattahi Mehr
,
K.
Schwarte
,
M.
Cerina
,
A.
Herrmann
,
D.-M.
Otte
,
A.
Zimmer
,
N.
Schwab
, et al
2017
.
Chemokine CCL17 is expressed by dendritic cells in the CNS during experimental autoimmune encephalomyelitis and promotes pathogenesis of disease.
Brain Behav. Immun.
66
:
382
393
.
24.
Stutte
S.
,
T.
Quast
,
N.
Gerbitzki
,
T.
Savinko
,
N.
Novak
,
J.
Reifenberger
,
B.
Homey
,
W.
Kolanus
,
H.
Alenius
,
I.
Förster
.
2010
.
Requirement of CCL17 for CCR7- and CXCR4-dependent migration of cutaneous dendritic cells.
Proc. Natl. Acad. Sci. USA
107
:
8736
8741
.
25.
Weber
C.
,
S.
Meiler
,
Y.
Döring
,
M.
Koch
,
M.
Drechsler
,
R. T.
Megens
,
Z.
Rowinska
,
K.
Bidzhekov
,
C.
Fecher
,
E.
Ribechini
, et al
2011
.
CCL17-expressing dendritic cells drive atherosclerosis by restraining regulatory T cell homeostasis in mice.
J. Clin. Invest.
121
:
2898
2910
.
26.
Lee
M.-C.
,
R.
Saleh
,
A.
Achuthan
,
A. J.
Fleetwood
,
I.
Förster
,
J. A.
Hamilton
,
A. D.
Cook
.
2018
.
CCL17 blockade as a therapy for osteoarthritis pain and disease.
Arthritis Res. Ther.
20
:
62
.
27.
Amo
L.
,
H. K.
Kole
,
B.
Scott
,
C. F.
Qi
,
J.
Wu
,
S.
Bolland
.
2021
.
CCL17-producing cDC2s are essential in end-stage lupus nephritis and averted by a parasitic infection.
J. Clin. Invest.
131
:
e148000
.
28.
Alferink
J.
,
I.
Lieberam
,
W.
Reindl
,
A.
Behrens
,
S.
Weiss
,
N.
Hüser
,
K.
Gerauer
,
R.
Ross
,
A. B.
Reske-Kunz
,
P.
Ahmad-Nejad
, et al
2003
.
Compartmentalized production of CCL17 in vivo: strong inducibility in peripheral dendritic cells contrasts selective absence from the spleen.
J. Exp. Med.
197
:
585
599
.
29.
Globisch
T.
,
N.
Steiner
,
L.
Fülle
,
V.
Lukacs-Kornek
,
D.
Degrandi
,
P.
Dresing
,
J.
Alferink
,
R.
Lang
,
K.
Pfeffer
,
M.
Beyer
, et al
2014
.
Cytokine-dependent regulation of dendritic cell differentiation in the splenic microenvironment.
Eur. J. Immunol.
44
:
500
510
.
30.
Lee
M.-C.
,
D. C.
Lacey
,
A. J.
Fleetwood
,
A.
Achuthan
,
J. A.
Hamilton
,
A. D.
Cook
.
2019
.
GM-CSF- and IRF4-dependent signaling can regulate myeloid cell numbers and the macrophage phenotype during inflammation.
J. Immunol.
202
:
3033
3040
.
31.
Kratochvill
F.
,
G.
Neale
,
J. M.
Haverkamp
,
L. A.
Van de Velde
,
A. M.
Smith
,
D.
Kawauchi
,
J.
McEvoy
,
M. F.
Roussel
,
M. A.
Dyer
,
J. E.
Qualls
,
P. J.
Murray
.
2015
.
TNF counterbalances the emergence of M2 tumor macrophages.
Cell Rep.
12
:
1902
1914
.
32.
Metzger
R.
,
M.
Maruskova
,
S.
Krebs
,
K. P.
Janssen
,
A. B.
Krug
.
2019
.
Increased incidence of colon tumors in AOM-treated Apc1638N/+ mice reveals higher frequency of tumor associated neutrophils in colon than small intestine.
Front. Oncol.
9
:
1001
.
33.
Perry
C. J.
,
A. R.
Muñoz-Rojas
,
K. M.
Meeth
,
L. N.
Kellman
,
R. A.
Amezquita
,
D.
Thakral
,
V. Y.
Du
,
J. X.
Wang
,
W.
Damsky
,
A. L.
Kuhlmann
, et al
2018
.
Myeloid-targeted immunotherapies act in synergy to induce inflammation and antitumor immunity.
J. Exp. Med.
215
:
877
893
.
34.
Zilionis
R.
,
C.
Engblom
,
C.
Pfirschke
,
V.
Savova
,
D.
Zemmour
,
H. D.
Saatcioglu
,
I.
Krishnan
,
G.
Maroni
,
C. V.
Meyerovitz
,
C. M.
Kerwin
, et al
2019
.
Single-cell transcriptomics of human and mouse lung cancers reveals conserved myeloid populations across individuals and species.
Immunity
50
:
1317
1334.e10
.
35.
Achuthan
A.
,
A. D.
Cook
,
M. C.
Lee
,
R.
Saleh
,
H. W.
Khiew
,
M. W.
Chang
,
C.
Louis
,
A. J.
Fleetwood
,
D. C.
Lacey
,
A. D.
Christensen
, et al
2016
.
Granulocyte macrophage colony-stimulating factor induces CCL17 production via IRF4 to mediate inflammation.
J. Clin. Invest.
126
:
3453
3466
.
36.
Cabeza-Cabrerizo
M.
,
A.
Cardoso
,
C. M.
Minutti
,
M.
Pereira da Costa
,
C.
Reis e Sousa
.
2021
.
Dendritic cells revisited.
Annu. Rev. Immunol.
39
:
131
166
.
37.
Persson
E. K.
,
H.
Uronen-Hansson
,
M.
Semmrich
,
A.
Rivollier
,
K.
Hägerbrand
,
J.
Marsal
,
S.
Gudjonsson
,
U.
Håkansson
,
B.
Reizis
,
K.
Kotarsky
,
W. W.
Agace
.
2013
.
IRF4 transcription-factor-dependent CD103+CD11b+ dendritic cells drive mucosal T helper 17 cell differentiation.
Immunity
38
:
958
969
.
38.
Schlitzer
A.
,
N.
McGovern
,
P.
Teo
,
T.
Zelante
,
K.
Atarashi
,
D.
Low
,
A. W.
Ho
,
P.
See
,
A.
Shin
,
P. S.
Wasan
, et al
2013
.
IRF4 transcription factor-dependent CD11b+ dendritic cells in human and mouse control mucosal IL-17 cytokine responses.
Immunity
38
:
970
983
.
39.
Bain
C. C.
,
A.
Bravo-Blas
,
C. L.
Scott
,
E. G.
Perdiguero
,
F.
Geissmann
,
S.
Henri
,
B.
Malissen
,
L. C.
Osborne
,
D.
Artis
,
A. M.
Mowat
.
2014
.
Constant replenishment from circulating monocytes maintains the macrophage pool in the intestine of adult mice. [Published erratum appears in 2014 Nat. Immunol. 15: 1090.]
Nat. Immunol.
15
:
929
937
.
40.
Bain
C. C.
,
C. L.
Scott
,
H.
Uronen-Hansson
,
S.
Gudjonsson
,
O.
Jansson
,
O.
Grip
,
M.
Guilliams
,
B.
Malissen
,
W. W.
Agace
,
A. M.
Mowat
.
2013
.
Resident and pro-inflammatory macrophages in the colon represent alternative context-dependent fates of the same Ly6Chi monocyte precursors.
Mucosal Immunol.
6
:
498
510
.
41.
Soncin
I.
,
J.
Sheng
,
Q.
Chen
,
S.
Foo
,
K.
Duan
,
J.
Lum
,
M.
Poidinger
,
F.
Zolezzi
,
K.
Karjalainen
,
C.
Ruedl
.
2018
.
The tumour microenvironment creates a niche for the self-renewal of tumour-promoting macrophages in colon adenoma.
Nat. Commun.
9
:
582
.
42.
Neufert
C.
,
C.
Becker
,
M. F.
Neurath
.
2007
.
An inducible mouse model of colon carcinogenesis for the analysis of sporadic and inflammation-driven tumor progression.
Nat. Protoc.
2
:
1998
2004
.
43.
Biehl
L. M.
,
D.
Garzetti
,
F.
Farowski
,
D.
Ring
,
M. B.
Koeppel
,
H.
Rohde
,
P.
Schafhausen
,
B.
Stecher
,
M. J. G. T.
Vehreschild
.
2019
.
Usability of rectal swabs for microbiome sampling in a cohort study of hematological and oncological patients.
PLoS One
14
:
e0215428
.
44.
Caporaso
J. G.
,
J.
Kuczynski
,
J.
Stombaugh
,
K.
Bittinger
,
F. D.
Bushman
,
E. K.
Costello
,
N.
Fierer
,
A. G.
Peña
,
J. K.
Goodrich
,
J. I.
Gordon
, et al
2010
.
QIIME allows analysis of high-throughput community sequencing data.
Nat. Methods
7
:
335
336
.
45.
Edgar
R. C.
2010
.
Search and clustering orders of magnitude faster than BLAST.
Bioinformatics
26
:
2460
2461
.
46.
Quast
C.
,
E.
Pruesse
,
P.
Yilmaz
,
J.
Gerken
,
T.
Schweer
,
P.
Yarza
,
J.
Peplies
,
F. O.
Glöckner
.
2013
.
The SILVA ribosomal RNA gene database project: improved data processing and web-based tools.
Nucleic Acids Res.
41
(
D1
):
D590
D596
.
47.
Segata
N.
,
J.
Izard
,
L.
Waldron
,
D.
Gevers
,
L.
Miropolsky
,
W. S.
Garrett
,
C.
Huttenhower
.
2011
.
Metagenomic biomarker discovery and explanation.
Genome Biol.
12
:
R60
.
48.
Berry
D.
,
K.
Ben Mahfoudh
,
M.
Wagner
,
A.
Loy
.
2011
.
Barcoded primers used in multiplex amplicon pyrosequencing bias amplification. [Published erratum appears in 2012 Appl. Environ. Microbiol. 78: 612.]
Appl. Environ. Microbiol.
77
:
7846
7849
.
49.
Klindworth
A.
,
E.
Pruesse
,
T.
Schweer
,
J.
Peplies
,
C.
Quast
,
M.
Horn
,
F. O.
Glöckner
.
2013
.
Evaluation of general 16S ribosomal RNA gene PCR primers for classical and next-generation sequencing-based diversity studies.
Nucleic Acids Res.
41
:
e1
.
50.
Lagkouvardos
I.
,
D.
Joseph
,
M.
Kapfhammer
,
S.
Giritli
,
M.
Horn
,
D.
Haller
,
T.
Clavel
.
2016
.
IMNGS: a comprehensive open resource of processed 16S rRNA microbial profiles for ecology and diversity studies.
Sci. Rep.
6
:
33721
.
51.
Lagkouvardos
I.
,
S.
Fischer
,
N.
Kumar
,
T.
Clavel
.
2017
.
Rhea: a transparent and modular R pipeline for microbial profiling based on 16S rRNA gene amplicons.
PeerJ
5
:
e2836
.
52.
Edgar
R. C.
,
B. J.
Haas
,
J. C.
Clemente
,
C.
Quince
,
R.
Knight
.
2011
.
UCHIME improves sensitivity and speed of chimera detection.
Bioinformatics
27
:
2194
2200
.
53.
Edgar
R. C.
2013
.
UPARSE: highly accurate OTU sequences from microbial amplicon reads.
Nat. Methods
10
:
996
998
.
54.
Chen
J.
,
K.
Bittinger
,
E. S.
Charlson
,
C.
Hoffmann
,
J.
Lewis
,
G. D.
Wu
,
R. G.
Collman
,
F. D.
Bushman
,
H.
Li
.
2012
.
Associating microbiome composition with environmental covariates using generalized UniFrac distances.
Bioinformatics
28
:
2106
2113
.
55.
Pfaffl
M. W.
2001
.
A new mathematical model for relative quantification in real-time RT-PCR.
Nucleic Acids Res.
29
:
e45
.
56.
Parekh
S.
,
C.
Ziegenhain
,
B.
Vieth
,
W.
Enard
,
I.
Hellmann
.
2016
.
The impact of amplification on differential expression analyses by RNA-seq.
Sci. Rep.
6
:
25533
.
57.
Pabst
O.
2012
.
New concepts in the generation and functions of IgA.
Nat. Rev. Immunol.
12
:
821
832
.
58.
Moor
K.
,
M.
Diard
,
M. E.
Sellin
,
B.
Felmy
,
S. Y.
Wotzka
,
A.
Toska
,
E.
Bakkeren
,
M.
Arnoldini
,
F.
Bansept
,
A. D.
Co
, et al
2017
.
High-avidity IgA protects the intestine by enchaining growing bacteria.
Nature
544
:
498
502
.
59.
Chagwedera
D. N.
,
Q. Y.
Ang
,
J. E.
Bisanz
,
Y. A.
Leong
,
K.
Ganeshan
,
J.
Cai
,
A. D.
Patterson
,
P. J.
Turnbaugh
,
A.
Chawla
.
2019
.
Nutrient sensing in CD11c cells alters the gut microbiota to regulate food intake and body mass.
Cell Metab.
30
:
364
373.e7
.
60.
Qiu
W.
,
E. B.
Carson-Walter
,
S. F.
Kuan
,
L.
Zhang
,
J.
Yu
.
2009
.
PUMA suppresses intestinal tumorigenesis in mice.
Cancer Res.
69
:
4999
5006
.
61.
Schwitalla
S.
,
P. K.
Ziegler
,
D.
Horst
,
V.
Becker
,
I.
Kerle
,
Y.
Begus-Nahrmann
,
A.
Lechel
,
K. L.
Rudolph
,
R.
Langer
,
J.
Slotta-Huspenina
, et al
2013
.
Loss of p53 in enterocytes generates an inflammatory microenvironment enabling invasion and lymph node metastasis of carcinogen-induced colorectal tumors.
Cancer Cell
23
:
93
106
.
62.
Erazo
A. B.
,
N.
Wang
,
L.
Standke
,
A. D.
Semeniuk
,
L.
Fülle
,
S. C.
Cengiz
,
M.
Thiem
,
H.
Weighardt
,
R. A.
Strugnell
,
I.
Förster
.
2021
.
CCL17-expressing dendritic cells in the intestine are preferentially infected by Salmonella but CCL17 plays a redundant role in systemic dissemination.
Immun. Inflamm. Dis.
9
:
891
904
.
63.
Valente
M.
,
Y.
Dölen
,
E.
van Dinther
,
L.
Vimeux
,
M.
Fallet
,
V.
Feuillet
,
C. G.
Figdor
.
2019
.
Cross-talk between iNKT cells and CD8 T cells in the spleen requires the IL-4/CCL17 axis for the generation of short-lived effector cells.
Proc. Natl. Acad. Sci. USA
116
:
25816
25827
.
64.
Rapp
M.
,
M. W. M.
Wintergerst
,
W. G.
Kunz
,
V. K.
Vetter
,
M. M. L.
Knott
,
D.
Lisowski
,
S.
Haubner
,
S.
Moder
,
R.
Thaler
,
S.
Eiber
, et al
2019
.
CCL22 controls immunity by promoting regulatory T cell communication with dendritic cells in lymph nodes.
J. Exp. Med.
216
:
1170
1181
.
65.
Tschurtschenthaler
M.
,
J.
Wang
,
C.
Fricke
,
T. M. J.
Fritz
,
L.
Niederreiter
,
T. E.
Adolph
,
E.
Sarcevic
,
S.
Künzel
,
F. A.
Offner
,
U.
Kalinke
, et al
2014
.
Type I interferon signalling in the intestinal epithelium affects Paneth cells, microbial ecology and epithelial regeneration.
Gut
63
:
1921
1931
.
66.
Man
S. M.
,
Q.
Zhu
,
L.
Zhu
,
Z.
Liu
,
R.
Karki
,
A.
Malik
,
D.
Sharma
,
L.
Li
,
R. K.
Malireddi
,
P.
Gurung
, et al
2015
.
Critical role for the DNA sensor AIM2 in stem cell proliferation and cancer.
Cell
162
:
45
58
.
67.
Zhan
Y.
,
P. J.
Chen
,
W. D.
Sadler
,
F.
Wang
,
S.
Poe
,
G.
Núñez
,
K. A.
Eaton
,
G. Y.
Chen
.
2013
.
Gut microbiota protects against gastrointestinal tumorigenesis caused by epithelial injury.
Cancer Res.
73
:
7199
7210
.
68.
Lee
Y. P.
,
C. C.
Chiu
,
T. J.
Lin
,
S. W.
Hung
,
W. C.
Huang
,
C. F.
Chiu
,
Y. T.
Huang
,
Y. H.
Chen
,
T. H.
Chen
,
H. L.
Chuang
.
2019
.
The germ-free mice monocolonization with Bacteroides fragilis improves azoxymethane/dextran sulfate sodium induced colitis-associated colorectal cancer.
Immunopharmacol. Immunotoxicol.
41
:
207
213
.
69.
Reunanen
J.
,
V.
Kainulainen
,
L.
Huuskonen
,
N.
Ottman
,
C.
Belzer
,
H.
Huhtinen
,
W. M.
de Vos
,
R.
Satokari
.
2015
.
Akkermansia muciniphila adheres to enterocytes and strengthens the integrity of the epithelial cell layer.
Appl. Environ. Microbiol.
81
:
3655
3662
.
70.
Marelli
G.
,
M.
Erreni
,
A.
Anselmo
,
V.
Taverniti
,
S.
Guglielmetti
,
A.
Mantovani
,
P.
Allavena
.
2017
.
Heme-oxygenase-1 production by intestinal CX3CR1+ macrophages helps to resolve inflammation and prevents carcinogenesis.
Cancer Res.
77
:
4472
4485
.
71.
Wang
F.
,
K.
Cai
,
Q.
Xiao
,
L.
He
,
L.
Xie
,
Z.
Liu
.
2022
.
Akkermansia muciniphila administration exacerbated the development of colitis-associated colorectal cancer in mice.
J. Cancer
13
:
124
133
.
72.
Yang
D.
,
Q.
Chen
,
D. M.
Hoover
,
P.
Staley
,
K. D.
Tucker
,
J.
Lubkowski
,
J. J.
Oppenheim
.
2003
.
Many chemokines including CCL20/MIP-3α display antimicrobial activity.
J. Leukoc. Biol.
74
:
448
455
.
73.
Palm
N. W.
,
M. R.
de Zoete
,
T. W.
Cullen
,
N. A.
Barry
,
J.
Stefanowski
,
L.
Hao
,
P. H.
Degnan
,
J.
Hu
,
I.
Peter
,
W.
Zhang
, et al
2014
.
Immunoglobulin A coating identifies colitogenic bacteria in inflammatory bowel disease.
Cell
158
:
1000
1010
.
74.
Malik
A.
,
D.
Sharma
,
R. K. S.
Malireddi
,
C. S.
Guy
,
T.-C.
Chang
,
S. R.
Olsen
,
G.
Neale
,
P.
Vogel
,
T.-D.
Kanneganti
.
2018
.
SYK-CARD9 signaling axis promotes gut fungi-mediated inflammasome activation to restrict colitis and colon cancer.
Immunity
49
:
515
530.e5
.
75.
Flores-Langarica
A.
,
K.
Müller Luda
,
E. K.
Persson
,
C. N.
Cook
,
S.
Bobat
,
J. L.
Marshall
,
M. W.
Dahlgren
,
K.
Hägerbrand
,
K. M.
Toellner
,
M. D.
Goodall
, et al
2018
.
CD103+CD11b+ mucosal classical dendritic cells initiate long-term switched antibody responses to flagellin.
Mucosal Immunol.
11
:
681
692
.
76.
Kim
Y.-I.
,
J.-H.
Song
,
H.-J.
Ko
,
M.-N.
Kweon
,
C.-Y.
Kang
,
H.-C.
Reinecker
,
S.-Y.
Chang
.
2018
.
CX3CR1+ macrophages and CD8+ T cells control intestinal IgA production.
J. Immunol.
201
:
1287
1294
.
77.
Reboldi
A.
,
T. I.
Arnon
,
L. B.
Rodda
,
A.
Atakilit
,
D.
Sheppard
,
J. G.
Cyster
.
2016
.
IgA production requires B cell interaction with subepithelial dendritic cells in Peyer’s patches.
Science
352
:
aaf4822
.
78.
Dörsam
B.
,
N.
Seiwert
,
S.
Foersch
,
S.
Stroh
,
G.
Nagel
,
D.
Begaliew
,
E.
Diehl
,
A.
Kraus
,
M.
McKeague
,
V.
Minneker
, et al
2018
.
PARP-1 protects against colorectal tumor induction, but promotes inflammation-driven colorectal tumor progression.
Proc. Natl. Acad. Sci. USA
115
:
E4061
E4070
.
79.
Hu
B.
,
E.
Elinav
,
S.
Huber
,
C. J.
Booth
,
T.
Strowig
,
C.
Jin
,
S. C.
Eisenbarth
,
R. A.
Flavell
.
2010
.
Inflammation-induced tumorigenesis in the colon is regulated by caspase-1 and NLRC4.
Proc. Natl. Acad. Sci. USA
107
:
21635
21640
.
80.
Kesselring
R.
,
J.
Glaesner
,
A.
Hiergeist
,
E.
Naschberger
,
H.
Neumann
,
S. M.
Brunner
,
A. K.
Wege
,
C.
Seebauer
,
G.
Köhl
,
S.
Merkl
, et al
2016
.
IRAK-M expression in tumor cells supports colorectal cancer progression through reduction of antimicrobial defense and stabilization of STAT3.
Cancer Cell
29
:
684
696
.
81.
Velcich
A.
,
W.
Yang
,
J.
Heyer
,
A.
Fragale
,
C.
Nicholas
,
S.
Viani
,
R.
Kucherlapati
,
M.
Lipkin
,
K.
Yang
,
L.
Augenlicht
.
2002
.
Colorectal cancer in mice genetically deficient in the mucin Muc2.
Science
295
:
1726
1729
.

The authors have no financial conflicts of interest.

Supplementary data