miR-34a is a microRNA safeguard for Citrobacter-induced inflammatory colon oncogenesis

Inflammation often induces regeneration to repair the tissue damage. However, chronic inflammation can transform temporary hyperplasia into a fertile ground for tumorigenesis. Here, we demonstrate that the microRNA miR-34a acts as a central safeguard to protect the inflammatory stem cell niche and reparative regeneration. Although playing little role in regular homeostasis, miR-34a deficiency leads to colon tumorigenesis after Citrobacter rodentium infection. miR-34a targets both immune and epithelial cells to restrain inflammation-induced stem cell proliferation. miR-34a targets Interleukin six receptor (IL-6R) and Interleukin 23 receptor (IL-23R) to suppress T helper 17 (Th17) cell differentiation and expansion, targets chemokine CCL22 to hinder Th17 cell recruitment to the colon epithelium, and targets an orphan receptor Interleukin 17 receptor D (IL-17RD) to inhibit IL-17-induced stem cell proliferation. Our study highlights the importance of microRNAs in protecting the stem cell niche during inflammation despite their lack of function in regular tissue homeostasis.


Introduction
The colon epithelium is constantly regenerated by stem cells residing at the bottoms of the intestinal crypts (Humphries and Wright, 2008). Infection of pathogenic bacteria in the colon can disrupt the normal gut microbiome and cause chronic inflammation, which has been linked to diseases including as inflammatory bowel disease (IBD) and recognized as a significant risk factor for colorectal cancer (CRC) development (Gagnière et al., 2016;Wang and Karin, 2015;Collins et al., 2011). It has been estimated that chronic inflammation and persistent infections contribute to a significant portion of human cancers, especially CRC Zur Hausen, 2009).
Inflammation plays a dual role in tissue homeostasis. On one hand, inflammation is associated with damage to the tissue; on the other hand, it triggers stem cell proliferation and reparative regeneration (Karin and Clevers, 2016). Events of damage and inflammation have been associated with regenerative signaling pathways such as Wnt to increase the number of stem cells and cause regeneration and hyperplasia in intestinal and colonic epithelia (Ashton et al., 2010;Miyoshi et al., 2012).
On the other hand, chronic inflammation causes excessive regeneration, and the resulting hyperplasia could eventually lead to cancer. TNF-a is associated with CRC progression (Al Obeed et al., 2014;Zins et al., 2007), and blocking TNF-a reduces the likelihood of colorectal carcinogenesis associated with chronic colitis (Popivanova et al., 2008). IL-17 have also been shown to promote colitis-associated early colorectal carcinogenesis (Grivennikov et al., 2009;Wang et al., 2014), and IL-22 stimulates stem cell growth after injury and promotes CRC stemness (Lindemans et al., 2015;Kryczek et al., 2014). Infiltration of T helper 1 (Th1) cells in CRC tumor specimens is associated with prolonged disease-free survival. However, infiltration of T helper 17 (Th17) cells, which secrete IL-17 and IL-22, is predictive of poor prognosis for CRC patients (Tosolini et al., 2011).
In this study, we demonstrate that miR-34a acts as safeguard to protect the stem cell niche during inflammation-induced reparative regeneration. miR-34a deficiency led to colon tumorigenesis after C. rodentium infection, where Th17 cell infiltration and epithelial stem cell proliferation were observed. During the pro-inflammatory response, miR-34a suppressed Th17 cell differentiation and expansion by targeting IL-23R, Th17 cell recruitment to the colon epithelium by targeting CCL22, and IL-17 induced stem cell proliferation by targeting IL-17RD. Loss of miR-34a results in a reparative regeneration process that goes awry.
After C. rodentium (2 Â 10 9 CFU) infection, both wild-type and miR-34a-/-mice developed diarrhea and weight loss within 2 weeks. Elongation of crypts and loss of goblet cells were observed ( Figure 1A). Histopathological changes of pre-neoplasia and neoplasia were limited to the miR-34a-/-genotype and were first noted at the four-month time point. Microscopic sections from wild type control mice were free of dysplastic and neoplastic changes at four-month and six-month time  The arrows indicate the visible colon tumors. 2 Â 10 9 CFU C. rodentium were used to infect the mice orally. Six months after the infection, the mice were euthanized and the colons were imaged. Scale bar, 5 mm. (C) Frequencies of colonic tumor formation in Figure 1 continued on next page points following infection (Boivin et al., 2003). In miR-34a-/-mice, no dysplasia or early neoplasia was present at a two-month time point (0/4), whereas at four months half the animals (2/4) had dysplastic change microscopically. At the six-month time point, 11 out of 20 miR 34a-/-mice had microscopic changes ranging from dysplasia (2/20), to adenoma (7/20), to adenocarcinoma (2/20) ( Figure 1A-C, Figure 1-figure supplement 1A). All tumors in this model were relatively well differentiated. One animal with a colonic adenocarcinoma in the section of distal colon also had a squamous cell carcinoma of the rectum. Dysplastic and neoplastic changes were characterized by strong intracytoplasmic b-catenin staining and occasional cells with nuclear staining ( Figure 1A). The earliest dysplastic changes are noted in deep reaches of crypts that are within inflamed ulcerated colonic mucosa in several of the sections where the diffuse inflammation of the C. rodentium has subsided and focal long-standing inflammation has set up due to ulceration of the surface. ( The colon stem cells, marked by the Wnt signaling enhancers Lgr5 and Ascl2 (Schuijers et al., 2015), are usually confined at the base of the crypt in wild-type and miR-34a-/-mice but became enriched in C. rodentium-induced colon tumors in miR-34a-/-mice ( Figure 1D). Enrichment of Lgr5 and Ascl2 expression in the colon tumors of infected miR-34a-/-mice was further confirmed by western blot ( Figure 1E). Colonic epithelial cells from infected miR-34a-/-mice (2 months post infection) had significantly higher organoid-forming efficiency and growth rates that the other groups (Sato et al., 2011a;Sato et al., 2011b) ( Figure 1F-H).
Th17 cells are enriched in number and in close proximity to stem cells in miR-34a-/-colon tumors CD4 +T helper (Th) cells are known to infiltrate and accumulate in the inflammatory environment, which can either promote or suppress tissue malignancy (Terzic´et al., 2010;Kim and Bae, 2016). We isolated CD4 +Th cells from the colon epithelium of C. rodentium-infected wild-type and miR-34a-/-mice and analyzed the relative abundance of Th1, Th2, Th17, and Treg subpopulations according to their associated expression of IFN-g, IL-4, IL-17, and FoxP3, respectively. IFN-g, IL-4, and FoxP3 levels were similar between wild-type and miR-34a-/-tissue, but IL-17 was significantly upregulated in miR-34a-/-tissue ( Figure 2A). miR-34a deletion increased the number of CD4 +IL17 +Th17 in the colon 2 and 6 months post C. rodentium infection ( Figure 2B,C). Immunofluorescence suggested that many of the enriched CD4 +IL-17 + Th17 cells were in proximity to Ascl2 +colon stem cells ( Figure 2D,E).
In the infected colon, miR-34a deletion did not significantly increase the number or IL-17 expression level of lineage(CD3e/Ly-6G/Ly-6C/CD11b/CD45R/B220/TER-119)-/CD117+/CD45 +cells, which contain a subset of ILC3 cells that may express IL-17 (Dong, 2008)   miR-34a suppresses Th17 differentiation and expansion by targeting IL6R and IL-23R We then aimed to understand how miR-34a deletion led to accumulation of Th17 cells in the C. rodentium-induced colon tumors. IL-6 is critical for initiating the differentiation of native CD4 +T cells into Th17 cells, and IL-23 is essential for the final step of Th17 cell differentiation, its proliferation, and IL-17 expression (Dong, 2008;Acosta-Rodriguez et al., 2007;Toussirot, 2012). We found that protein levels of IL-6R and IL-23R, the receptors for IL-6 and IL-23, were upregulated in CD4 +T cells isolated from the miR-34a-/-colon compared to the wild-type control ( Figure 3A, Figure 3- 24.3% figure supplement 1A,B). The RNA22 algorithm identified putative miR-34a binding sites in the IL-6r and IL-23r 3'UTRs ( Figure 3B,C), which were then confirmed by the luciferase reporter assay ( Figure 3D,E).
Conditioned medium collected from miR-34a-/-colon tumor organoids enhanced the migration of in vito-differentiated Th17 cells in comparison to medium from the wild-type colon organoids ( Figure 3J). The addition of anti-CCL22 neutralizing antibody in the medium or knockdown of CCL22 in miR-34a-/-colon tumor organoids reduced Th17 migration back to the wild-type level ( Figure 3J, Figure 3-figure supplement 2). Therefore, miR-34a suppresses recruitment of Th17 cells by targeting CCL22 production in colon epithelial cells.

Th17 cells promote colon organoid growth via IL-17
We then tested whether Th17 cells, which were enriched by loss of miR-34a and were located in proximity to Ascl2 +colon stem cells ( Figure 2B-D), regulated colon epithelial cell proliferation. Mouse CD4 +T cells were induced to differentiate into Th17 cells and co-cultured with colon organoids. The presence of Th17 cells significantly increased the organoid sizes. Accordingly, the addition of anti-IL-17 neutralizing antibody suppressed this growth, suggesting that the growth effect was caused by Th17-secreted IL-17 ( Figure 4A). In the absence of Th17 cells, recombinant IL-17 in the medium increased organoid growth ( Figure 4B,C) and also upregulated Lgr5 and Ascl2 (stem cell marker) expression ( Figure 4D).
Consistent with its safeguard role, miR-34a levels were higher in Lgr5-GFP + colon stem cells than in Lgr5-GFP-cells ( Figure 5-figure supplement 2A). On the other hand, miR-34b and miR- 34c, the other two miR-34 family members, were barely detectable in colon epithelial cells (Figure 5-figure supplement 2B). Furthermore, miR-34a expression levels in both CD4 +T cells and colon epithelial cells were upregulated during C. rodentium-induced inflammation ( Figure 5-figure  supplement 2C,D). This is consistent with the dual role miR-34a plays in both CD4 +T cells and colon epithelial cells.
We then performed bone marrow transplantation by replacing the bone marrow in irradiated wildtype C57BL/6J mice with the bone marrow from miR-34a-/-mice, which resulted in mice with wildtype epithelial cells but miR-34a-/-immune cells ( Figure 6D). The transplantation efficiency was validated to be above 90% by flow analysis (Figure 6-figure supplement 3). 6 months after C. rodentium infection, 3 out of the 12 mice with transplanted miR-34a-/-bone marrow developed colon tumors, whereas none of the wildtype mice developed tumors ( Figure 6E). CD4 +IL17+Th17 cells were enriched in the infected colons of the mice with miR-34a-/-bone marrow transplants ( Figure 6F). Caveats of this transplantation experiment include the potential confounding effects of radiation on the intestinal cells (e.g., LGR5 +cells are sensitive to radiation) and radiation-resistant cells, which can only be addressed by additional control groups with bone marrow transplantation from wild-type to wild-type and from miR-34a-/-to miR-34a-/-. Taken together, the combination of miR-34a deficiency in both epithelial and immune cells seems to elicit a stronger tumorigenic effect than miR-34a deficiency in epithelial or immune cells alone, consistent with the interaction between Th17 cells and colon epithelial cells.

IL-17 neutralizing antibody abrogates C. rodentium-induced colon stem cell proliferation and tumorigenesis
To validate the role of IL-17 in C. rodentium-infection-induced colon stem cell proliferation, we infected the mice with C. rodentium on day 0, then intraperitoneally injected the mice with IL-17 neutralizing antibody on days 1, 5, and 10, and sacrificed them on either day 15 or day 30 to harvest the colon ( Figure 6G). Injection of the IL-17 antibody suppressed Ascl2 +stem cells proliferation in the infected colonic crypts according to immunofluorescence ( Figure 6H) and decreased Ascl2 and Lgr5 expression according to western blot ( Figure 6I).

Uninfected infected
WT mice with miR-34a-/bone marrow 0(11) 3 (12) Colon tumorigenesis in C. rodentium infected mice with miR-34a-/-bone marrow (  IL-17 and miR-34a levels are associated with human CRC To investigate whether IL-17 regulates human colon epithelial growth, we grew organoid cultures from human colon tissue using an established protocol (Sato et al., 2011a) and added recombinant human IL-17 into the culture medium. Consistent with mouse organoids, addition of human IL-17 into the medium increased the sizes of human colon organoids ( Figure 7A,B). RT-qPCR  (Figure 7source data 1) suggested that the expression levels of the two Th17 cell markers IL-17 and RORC as well as miR-34a target IL-17rd were higher in tumor tissues than that in normal colon tissues, while miR-34a expression levels were downregulated ( Figure 7C-F).
We then validated whether miR-34a suppresses IL-6R, IL-23R, CCL22, and IL-17RD in human cells. We ectopically expressed miR-34a in Jurkat cells, a human T lymphocyte cell line, and SW480 cells, a human colon epithelial cancer cell line. Western blots indicated that miR-34a suppressed IL-6R expression in Jurkat cells as well as CCL22 and IL-17RD expression in SW480 cells. However, miR-34a did not seem to suppress IL-23R expression (Figure 7-figure supplement 1).
We then checked the miR-34a binding sequences in the 3'UTRs of the human genes using miRanda and RNA22. Luciferase reporter assays containing the 3'UTRs with wild-type or mutated binding sequences confirmed that miR-34a directly binds to these putative binding sites in IL-17rd, CCL22 and IL-6r 3'UTRs and regulates their expression (Figure 7-figure supplement 1). On the other hand, miR-34a did not seem to regulate IL-23R.
Discussion miR-34a is a known tumor suppressor that targets cell proliferation and apoptosis genes. In fact, RNA-seq performed on splenic CD4 +T cells and colon epithelial cells isolated from miR-34a-/-and wildtype mice revealed various changes in gene expression (Figure 3-figure supplement 4), including well-known miR-34a target genes such as Notch1, Snai2, BCL2, and c-Met (Figure 8source data 1 and 2). However, miR-34a deficiency alone does not lead to tumorigenesis, suggesting that mere upregulation of these genes by miR-34a loss is not sufficient to cause cancer. Our study indicates that miR-34a acts as a safeguard for the inflammatory stem cell niche and reparative regeneration by modulating both the immune and epithelial responses to infection and inflammation (Figure 8). First, miR-34a suppresses Th17 cell differentiation and expansion by targeting IL-6R and IL-23R. Second, miR-34a limits Th17 cell recruitment to the epithelium by targeting CCL22. Lastly, miR-34a hinders IL-17-induced stem cell proliferation by targeting IL-17RD. These miR-34a targeting mechanisms are largely conserved between murine and human, although miR-34a does not seem to target IL-23R in human T cells.
Colon stem cells reside at the base of the crypt, relying on the niche to provide necessary signaling cues for self-renewal. The mesenchyme beneath the niche provides many essential factors (Degirmenci et al., 2018). In addition, cKit+/Reg4 +colonic crypt base secretory cells interdigitate with Lgr5 +stem cells, providing the latter with Notch ligands DLL1 and DLL4, and epidermal growth factor (Rothenberg et al., 2012;Sasaki et al., 2016). Normally, stem cells are constrained to this spatial niche and are forced to differentiate when they leave the niche. However, in human colon adenoma and carcinoma samples, Lgr5 +stem like cells are highly upregulated and are not confined to the spatial niche as in normal crypts (Baker et al., 2015). Similarly, we have observed this enrichment and expanded distribution of stem cells in C. rodentium-induced colon tumors in miR-34a-/mice. Inflammatory cytokines such as IL-17 potentially provide an enlarged 'inflammatory niche' by stimulating receptors such as IL-17RD on the stem cells, enabling them to ignore the constraints of the crypt base and proliferate away from the crypt base mesenchyme and secretory cells. Interestingly, IL-17RD amplifies IL-17RA signaling in a way analogous to Lgr5 receptor amplification of Wnt signaling for self-renewal.
Non-coding RNAs occupy the majority of the mammalian genome (Kung et al., 2013;Mattick and Rinn, 2015). Evolutionarily, the percentage of genome devoted to the non-coding region is consistently associated with the complexity of the organism, rising from less than 25% in prokaryotes, 25-50% in simple eukaryotes, more than 50% in fungi, plants and animals, to approximately 98.5% in humans-which have a genome size that is three orders of magnitude larger than prokaryotes (Mattick, 2004). Compared to microRNA, the role of long non-coding RNA (lncRNA) in regulating tumors has just started to be appreciated (Huarte, 2015;Prensner and Chinnaiyan, 2011;Schmitt and Chang, 2016). In fact, lncRNA has been shown to regulate miR-34a in human CRC, especially in cancer stem cells . Similar to miR-34a, many of the lncRNAs with strong functions in tumors are largely dispensable for normal development and tissue homeostasis Nakagawa et al., 2014). It is possible that the abundance of non-coding RNAs in mammals may provide extra surveillance to protect tissue integrity during stress conditions such as inflammation, which are often not captured by laboratory animal models raised in well-controlled circumstances.
The miR-34a mimic was the first microRNA mimic to reach clinical trial for cancer therapy (Bouchie, 2013;Bader, 2012). Previous studies largely focused on the role of miR-34a to induce cell cycle arrest, senescence, and apoptosis. This study suggests that enhancing miR-34a levels may have additional benefits of suppressing Th17 cells and IL-17 stimulation of cancer stem cells in the tumor microenvironment. Furthermore, inflammatory cytokines such as IL-6 can suppress miR-34a to increase epithelial-mesenchymal transition (EMT) (Rokavec et al., 2014), so boosting miR-34a may mitigate inflammation-induced CRC invasion and migration. It might be worth paying extra attention to CRC with Th17 cell enrichment and evaluating therapeutics effects based on CRC classification, especially on the inflammatory and stem cell-like subtypes (Sadanandam et al., 2013).

Materials and methods
Transgenic mice and bacterial infection C57BL/6J and B6(Cg)-Mir34atm1Lhe/J mice (Choi et al., 2011) were ordered from the Jackson Laboratory. Lgr5-EGFP-creER T2 /miR-34a flox/flox mice were generated as described as previously . Cre recombinase was induced by intraperitoneal injection of tamoxifen (Sigma) dissolved in sterile corn oil at a dose of 75 mg/kg before infection with C. rodentium. Mouse maintenance and procedures were approved by Duke University DLAR and followed the protocol (A286-15-10). C. rodentium strain DBS100 was purchased from ATCC and cultured according to previously described methods (Shui et al., 2012). 2 Â 10 9 C.F.U C. rodentium were infected into 8 weeks old mice by oral gavage.

Bone marrow transplantation
The bone marrow transplantation procedures were performed as previously described (Alpdogan et al., 2003). Male C57BL/6J.SJL mice (Ly5.1) with the Ptprc b leukocyte marker CD45.1/ Ly5.1 obtained from the Jackson Laboratory were used as recipients for transplantation at the age of 8-10 weeks. The recipient mice received 1000 Rad (1Gy, filter 4) whole body lethal irradiation on XRad320. After 6 hr, the irradiated mice received bone marrow cells from male C57BL/6J miR34a -/donor mice with Ptprc b leukocyte marker CD45.2/Ly5.2. The donor femurs were collected aseptically, and the bone marrow canals were washed out with sterile media. 6 million cells per mouse were transplanted into lethally irradiated recipients via tail vein injection. Mice were housed and received sulfamethoxazole trimethoprim medicated, acidified water for 4 weeks. 6 weeks after reconstitution, blood was collected from the recipient mice, and the reconstituted blood cell lineages were analyzed by flow cytometry using CD45.1-PE (Biolegend) and CD45.2-FITC (Biolegend).

Clinical specimen and colon organoid culture
Frozen CRC specimens and paired controls were acquired from Weill Cornell Medical College (WCMC) Colon Cancer Biobank for evaluation of Th17 cell-related gene expression. Surgically resected fresh normal human colon tissues were obtained from Duke University hospital. The study was approved by the ethical committee of Duke University hospital, Duke University, and WCMC. All samples were obtained with informed consent.
Mouse and human colon crypt isolation and organoid culturing were performed as described previously (Sato et al., 2011a). To investigate CCL22 regulation on Th17 cell migration and IL-17RA and IL-17RD regulation on organoids growth, lentiviral vector carrying shRNA against CCL22, IL-17ra or IL-17rd were purchased from Sigma and infected into organoids according to previously described protocols (Koo et al., 2011).

CD4 +T cell isolation and Th17 cell differentiation
To investigate Th17 cell enrichment in C. rodentium-infected colons, CD4 +T cells were first isolated from mouse colon as described previously (Weigmann et al., 2007). Briefly, after washing with cold PBS, the mouse colon was cut into 0.5-1 cm pieces and incubated in Ca 2+ and Mg 2+ free PBS containing 0.37 mg/ml EDTA and 0.145 mg/ml DTT in a shaking incubator at 37˚C for 15 min. The supernatant was decanted, and the remaining tissue was further incubated in RPMI-1640 containing 5% fetal calf serum, 20 mM HEPES, 100 U/ml each of penicillin and streptomycin, and 0.1 mg/ml collagenase dispase (Sigma) while shaking at 37˚C for 90 min. After filtering through 70 mm cell strainers, the cells were collected by centrifugation, and the pellet was suspended in 35% percoll solution (Sigma). The cells were then collected by centrifugation at 2000 rpm for 20 min. CD4 +T cells were isolated using a mouse CD4 +T cell isolation kit (StemCell Technology). Then 10,000 CD4 +T cells were counted for flow cytometry. After staining for CD4 and IL-17, Th17 cells were analyzed by flow cytometry.
To evaluate the effect of IL-23R on Th17 cell differentiation, CD4 +T cells were isolated from mouse spleen as described previously (Weigmann et al., 2007). Briefly, the spleen was minced and squeezed through 70 mm cell strainers to obtain single cells. After collection by centrifugation, the cells were suspended into 35% percoll solution (Sigma) with heparin, followed by incubation in red cell lysis buffer (Abcam) to get rid of red blood cells. The cells were then washed, and CD4 +T cells were isolated using a mouse CD4 +T cell isolation kit (StemCell Technology). Isolated CD4 +T cells were cultured in 24-well plate coated with anti-CD3e and anti-CD28 antibodies in 1640 RPMI medium with 10% FCS and recombinant mouse IL-2 (rmIL2, 20 ng/mL) at 1 Â 10 6 / mL according to the previous protocol (Zhong et al., 2010). Lentiviral vectors carrying shRNA against IL-23r were purchased from Sigma and infected into CD4 +T cells following previously described protocols (Bao et al., 2006). After selection by antibiotics, the cells were induced to differentiate into Th17 cells using the FlowCellect Mouse Th17 Differentiation Kit according to the protocol (EMD Millipore). Th17 cell differentiation efficiency was measured by flow cytometry by CD4 and IL-17 staining.

Histochemical staining
Selected colon tissues from wildtype and miR-34a-/-mice euthanized at 2 days, 2, 4, and 6 months after C. rodentium infection were fixed in 10% neutral buffered formalin, processed routinely and embedded in paraffin, sectioned at five microns, and stained with hematoxylin and eosin. Glass slides were scanned via high resolution virtual slide imaging at 40x (Aperio, Leica Biosystems) and then reviewed by a board-certified veterinary pathologist with experience in mouse tumor biology (JE) without knowledge of genotype. Lesions were scored according to established murine pathology (Boivin et al., 2003). Representative proliferative colonic lesions were selected for recuts and bcatenin immunohistochemistry (IHC) was performed. IHC was conducted with a rabbit monoclonal antibody against b-catenin (1:400, Abcam) after epitope retrieval. The secondary detection system was a labelled polymer-HRP anti-rabbit (Dako).

Chemotaxis assays
The chemotaxis assay was performed as described previously . Briefly, 1 Â 10 5 Th17 cells were applied to the upper well of the ChemoTex chambers (96-well, 5 mm pore size; Neu-roProbe). Conditional medium from miR-34a-/-colon organoids or control organoids was added in the lower chamber. To evaluate the effect of CCL22 on Th17 migration, a neutralizing monoclonal antibody against CCL22 (R and D) was included in the conditional medium. After a 2 hr incubation, the cells in the upper wells were removed, and the migrated cells were collected by centrifugation. Migrated cells were counted using a hemocytometer.

Immunofluorescence
Immunofluorescence was performed on paraffin-embedded colon sections. After rehydration and antigen retrieval, the sections were blocked by 2% horse serum in PBS for 2 hr at RT and incubated with anti-Ascl2 (1:200, Santa Cruz), anti-IL17 (1:200, Abcam), anti-CD4 (1:50, R and D Systems) or anti-GFP (1:500, Abcam) primary antibodies in antibody diluent buffer (DAKO) overnight at 4˚C. After washing, the sections were then incubated with Rhodamine Red or Alexa fluor 488 labeled secondary antibodies (Invitrogen) for 1 hr at room temperature. After counterstaining with DAPI (Invitrogen), the slides were observed on an Axio Imager upright microscope (Zeiss).

Flow cytometry analysis
Th17 cells were analyzed by CD4 and IL-17 staining. Briefly, single cells were fixed with 4% formaldehyde and further permeabilized by methanol. The cells were then incubated with anti-IL-17 (1:200, Abcam) and anti-CD4 (1:100, R and D Systems) antibodies, followed by incubation with APC or FITC labeled secondary antibody (Invitrogen). The samples were analyzed using a Beckman Coulter flow cytometer. The raw FACS data were analyzed with the FlowJo software.

Quantitative real-time PCR
Total RNA was extracted from the tissue using the RNeasy mini kit (Qiagen). cDNA was synthesized from 500 ng of total RNA in 20 ml of reaction volume using the High Capacity cDNA Archive Kit (Applied Biosystems). Quantitative PCR was carried out using TaqMan assays (Applied Biosystems) to detect miR-34a and hIL-17, and the SYBR Green System (Applied Biosystems) for all other gene expression measurements. miR-34a primers were purchased from Applied Biosystems, and hIL-17 primers were purchased from Thermo Fisher. Other qPCR primers were synthesized by IDT, and the sequences are listed in Figure 7-source data 2. All relative gene expression measurements utilized at least three biological replicates for both the wild-type and miR34a deficient experimental groups with three technical replicates per biological replicate. The expression of each gene was defined using the threshold cycle (Ct), and the relative expression levels were calculated using the 2-44Ct method after normalization to the ß-actin expression level.
Sequence mutation and gene knockdown gRNA:CAGAATGATGGCGGTGGCAG-TGG was designed for mutation of sequence complementary to miR-34a binding site in the mouse IL-17rd 3UTR. The gRNA was then vector pLentiCRISPR v2. and transfected into mouse single colon stem cells. DNA sequencing of single colonies confirmed successful deletion of miR-34a binding in mouse IL-17rd 3UTR. Mutagenesis for luciferase reporter assay was performed using QuickChange Site-Directed Mutagenesis Kit (Stratagene). shRNAs for knockdown of IL-6R, IL-23R, IL-17RA, IL-17RD and CCL22 were purchased from Sigma. PCR were performed using primers to amply three most APC mutation regions in mouse colon cancer. Primers: 'GCCATCCCTTCACGTTAG' and 'TTCCAC TTTGGCATAAGGC' for DNA sequence contains mutation 1; Primers: 'TGACAGCACAGAATCCAG TG' and 'TACCAAGCATTGAGAG' for DNA sequence contains mutation 2 (B); Primers: 'TAGGCAC TGGACATAAGGGC' and 'GTAACTGTCAAGAATCAATGG' for DNA sequence contains mutation 3.

Statistical analysis
Data were expressed as mean ± standard deviation (s.d.) of three biological replicates. Student T-tests were used for comparisons, with p<0.05 considered significant. Patient data were expressed as mean ± standard error of the mean (s.e.m.). Paired T-tests were used for comparison of the 17 matched patient normal colon and CRC samples.