Transdifferentiation of tumor infiltrating innate lymphoid cells during progression of colorectal cancer

ABSTRACT Innate lymphoid cells (ILCs) reside in mucosal surfaces to potentiate immune responses, sustain mucosal integrity and maintain tissue homeostasis. However, how tumor infiltrating

ILCs modulate tumor development and progression is unclear. Here we profiled tumor infiltrating ILCs during colorectal cancer (CRC) progression by single-cell RNA sequencing. We identified

six clusters of tumor infiltrating ILCs with unique features. ILC1s expressed inhibitory receptors and underwent inhibitory functional conversion at the late stage of CRC. ILC2s were

classified into three subsets (called ILC2-A, -B, -C), of which ILC2-C subset could facilitate tumor progression. HS3ST1 and PD1 were highly expressed in ILC2s of late stage CRC tumors and

deficiency of HS3ST1 or PD1 in ILC2s suppressed tumor growth. Moreover, ILC3s transdifferentiated into ILCregs during CRC progression and ILCregs promoted tumor growth. Of note, TGF-β

signaling initiated the conversion of ILC3s to ILCregs and blockade of TGF-β signaling could disrupt the ILCreg transdifferentiation and inhibited tumor growth. Thus, intervention of ILC

conversions might be a potential strategy for CRC immunotherapy. You have full access to this article via your institution. Download PDF SIMILAR CONTENT BEING VIEWED BY OTHERS THE C-TYPE

LECTIN DCIR CONTRIBUTES TO THE IMMUNE RESPONSE AND PATHOGENESIS OF COLORECTAL CANCER Article Open access 26 March 2024 IL-33-ACTIVATED ILC2S INDUCE TERTIARY LYMPHOID STRUCTURES IN PANCREATIC

CANCER Article Open access 15 January 2025 SIRPΑ ON TUMOR-ASSOCIATED MYELOID CELLS RESTRAINS ANTITUMOR IMMUNITY IN COLORECTAL CANCER INDEPENDENT OF ITS INTERACTION WITH CD47 Article 10

January 2024 INTRODUCTION The innate immune system provides first-line defence against invading pathogens by triggering inflammatory and antimicrobial responses. Although robust inflammation

contributes to the clearance of pathogens, inflammatory disorders and chronic inflammation are also correlated with tumor development and progression.1,2,3,4 In addition, several cancers

display characteristics of chronic inflammation during their progression. Colorectal cancer (CRC) is considered as a good example of a tumor correlated with chronic inflammation.2,5

Accumulating evidence shows that inflammatory disorder has substantial effect on CRC development and progression.4,6 Tumor infiltrating lymphocytes (TILs) are thought to be a sign of

attacking transformed cells by the immune system at an early stage.7 Tumor cells can be recognized by TILs, which produce large amounts of IFN-γ (such as NK, NKT, and γδT cells) or directly

kill transformed cells (such as NK and cytotoxic T lymphocyte (CTL cells)), leading to elimination of tumor cells.8,9 TILs withstand immune selection pressure against tumor cells and

immunoevasive mutations might be acquired, which lead to tumor escape from immunosurveillance.10 Meanwhile, chronic inflammation caused by immune dysregulation promotes tumor development.

Inflammatory microenvironment alters the fate of lymphocytes by various mediators such as cytokines and chemokines. For instance, tumor cells produce immunosuppressive cytokines such as

vascular endothelial growth factor (VEGF), transforming growth factor-β (TGF-β), which convert effector T cells into regulatory T cells (Tregs). Tregs secrete IL-10 and produce PD-L1 and

CTLA-4 to suppress CTL functions.11 In addition, continuous tumor antigen stimulation maintains the expression of inhibitory receptors on T cells, such as PD-1 and CTLA-4, which disarms

effector T cells leading to tumor development and progression.12 At present, therapeutic applications targeting these immunosuppressive molecules become promising strategies for antitumor

therapy. Innate lymphoid cells (ILCs) are innate lymphocytes that lack other lineage markers and contribute to early pathogen defence.13 ILCs were previously categorized into three

subgroups, namely group 1 ILC (ILC1s), group 2 ILC (ILC2s) and group 3 ILC (ILC3s), according to their featured cytokine profiles and unique transcription factors.14 ILC1s are characterized

by expressing T-bet and producing IFN-γ, which is essential for clearance of intracellular microbial infections.15,16 ILC2s require Gata3 for their differentiation and maintenance.17 After

activation by IL-25 and IL-33, ILC2s produce cytokines, such as IL-5 and IL-13, to promote resolution of helminth infections and to participate in pathogenesis of asthma.18,19 ILC3s are

defined by RORγt expression and generate cytokines such as IL-22 and IL-17, which are extremely critical for resistance against bacterial infections.20,21 We recently defined a new

regulatory subpopulation of ILCs named ILCregs, which contain Id3 and produce IL-10 to modulate inflammatory response.22 However, it is unknown about intratumoral heterogeneity of ILC

populations and their associations with tumor progression. Here we used single-cell RNA sequencing (scRNA-seq) to profile tumor infiltrating ILCs in CRC at early and late stages. We

identified six clusters of ILCs inside colon tumors and revealed conversions of their functions and subsets during CRC progression, and blocking ILC transdifferentiation could suppress tumor

development. RESULTS PROFILING OF TUMOR INFILTRATING ILCS IN EARLY AND LATE STAGES OF COLORECTAL TUMOR In order to explore dynamic regulation of tumor infiltrating ILCs during CRC

progression, we used azoxymethane/dextran sodium sulfate (AOM/DSS)-induced colitis-associated CRC, which mimics tumor progression caused by chronic colitis as seen in inflammatory bowel

disease (IBD).23,24 We isolated tumor infiltrating ILCs and conducted droplet-based scRNA-seq to profile ILCs inside colon tumors at day 60 (early stage) and day 120 (late stage) (Fig. 1a).

We excluded contaminating lineage cells and low quality cells, and finally obtained 1,654 ILC cells (Supplementary information, Fig. S1a–f). Unsupervised graph clustering divided ILCs into

six groups, and we labeled them with known ILC markers (Supplementary information, Fig. S1g).25,26 Six ILC subsets inside CRC tumors included ILC1s, three groups of ILC2s (termed ILC2-A, B,

C), ILC3s and ILCregs (Fig. 1b). Based on scRNA-seq data, cell proportions were dynamically changed during tumor progression (Fig. 1c–e). ILC1 cells slightly reduced at the late stage of CRC

tumors (Fig. 1c–e). ILC-2A cells were dominant ILC2 cells at the early stage of CRC tumors (Fig. 1c–e). ILC2-B cells distributed similarly at both early and late stages. However, ILC2-C

cells mainly existed in colon tumors at the late stage (Fig. 1c–e). Of note, ILC3 cells mainly resided at the early stage of CRC tumors, whereas ILCregs appeared at the late stage of CRC

tumors (Fig. 1c–e). We next assessed signature genes for each cell type and defined many known and newly identified markers (Fig. 1f). Tumor infiltrating ILC2 subsets shared several common

ILC2 markers, and also harbored their unique signature expression patterns (Fig. 1g). All of the identified ILC subsets in CRC tumors expressed common ILC signature genes such as _Il7r_ and

_Il2ra_ (Fig. 1g). Notably, three ILC2 subsets highly expressed _Klrg1_, _Il1rl1_ (encoding ST2), and _Gata3_. Thus, unbiased analyses of tumor infiltrating ILCs reveal dynamic changes of

ILC subsets and different gene expression profiles inside CRC tumors during CRC progression. CHARACTERIZING ILC1 CELLS IN COLORECTAL TUMOR It has been reported that ILC1-like cells in

mammary tumor exert tumor suppressive functions.27 The exact role of ILC1s inside CRC tumor is unknown. According to our scRNA seq data, we identified top 9 signature genes of tumor

infiltrating ILC1s. Of these 9 genes, _Tyrobp, Klrd1, Klrk1, Klrb1c_ and _Klrc2_ were previously identified as ILC1-related genes.25,28 In addition, _Ctsw_ (encoding cathepsin W

preproprotein), _Txk_ (encoding RIk kinase), and _Cd7_ were also highly expressed in ILC1 cells (Fig. 2a). We next compared gene expression profiles of tumor infiltrating ILC1s at early

stage versus late stage. Intriguingly, ILC1s at the early stage expressed high levels of activating receptors (_Klrd1, Ncr1, Klrc2, Klrb1c_), whereas they expressed inhibitory receptors

(_Klre1, Klra7_) at the late stage (Fig. 2b).29 These observations were further validated by flow cytometry (Fig. 2c). However, expression levels of these receptors did not show significant

changes on ILC1s of peri-tumor tissues (Supplementary information, Fig. S2a). We found that ILC1 cells inside CRC tumors slightly declined at the late stage (Fig. 2d), but no overt changes

in peri-tumor tissues between the early stage and the late stage (Supplementary information, Fig. S2b). Furthermore, _Il12rb2_ was decreased in ILC1s in the late stage (Fig. 2b), suggesting

impaired response of ILC1s to IL-12 stimulation. IFN-γ plays a critical role in eradication of early tumors.8,30 We isolated ILC1s from early and late stages of CRC tumors and stimulated

with IL-12 plus IL-18 in vitro. We found that IFN-γ production was remarkably decreased (Fig. 2e). We also assessed tumor infiltrating ILC1s in CRC patients according to the defined markers

on human ILC1s.16 We observed that tumor infiltrating ILC1s in advanced CRC patients showed lower frequencies with high levels of inhibitory receptors (Fig. 2f–h). In contrast, these

receptors did not show apparent changes in peri-tumor tissues (Supplementary information, Fig. S2c). Taken together, ILC1s undergo functional change during CRC progression. HS3ST1 AND PD1

ARE HIGHLY EXPRESSED IN TUMOR FILTRATING ILC2 CELLS OF LATE STAGE CRC TUMORS As shown in Fig. 1g, all three ILC2 subsets expressed ILC2 signature genes. Besides these signature genes, we

also identified several genes in tumor infiltrating ILC2s that were gradually induced during CRC progression, including _Hs3st1, Ctla2a, Ltb4r1, Pdcd1, Tnfrsf18_, and _Hes1_ (Fig. 3a, b).

These genes were indeed highly expressed in the late stage of CRC tumors (Supplementary information, Fig. S3a). Of these induced genes in ILC2s, we focused on _Hs3st1_ and _Pdcd1_. _Hs3st1_,

encoding heparan sulfate 3-O-sulfotransferase 1 (HS3ST1), catalyzes biosynthesis of heparan sulfate.31,32 _Hs3st1_ was restrictedly expressed in ILC2s and substantially induced during CRC

progression (Fig. 3a, b). _Pdcd1_, encoding PD1, is an immune checkpoint receptor,33,34 which was also highly expressed in ILC2 cells over CRC progression, suggesting a critical role of

_Pdcd1_ for ILC2 function. Consistent with the mRNA levels, these two molecules were indeed highly expressed in ILC2s in late stage CRC tumors by flow cytometry (Fig. 3c). These observations

were further validated by immunofluorescence staining (Fig. 3d, e). Intriguingly, HS3ST1 and PD1 were substantially expressed in ILC2s of human advanced CRC tumor tissues (Fig. 3f–h).

Furthermore, these two molecules were restrictedly expressed in tumor tissues but not in peri-tumor tissues (Fig. 3i). Collectively, HS3ST1 and PD1 are highly expressed in tumor filtrating

ILC2 cells of late stage CRC tumors. ILC2-C SUBSET PROMOTES CRC PROGRESSION Tumor infiltrating ILC2s expressed high levels of PD1 in human advanced CRC tumors. ILC2s expressed differential

levels of PD1 over CRC stages (Fig. 3f). We thus defined PD1low ILC2s as ILC2-A in human CRC tumors and PD1high ILC2s as ILC2-C. In order to determine the role of ILC2s in the tumor growth

of human CRC, we isolated human PD1high ILC2-C cells from advanced CRC samples and engrafted them with tumor cells into B-NSG mice. We found that ILC2-C cells apparently promoted tumor

growth, whereas ILC2-A cells had no such effect (Fig. 4a, b; Supplementary information, Fig. S3b). Additionally, tumor cells showed increased proliferation rates with ILC2-C cell engraftment

(Fig. 4c). These data suggest that ILC2-C cells facilitate tumor growth. We next generated HS3ST1- or PD1-deficient ILC2-C cells by CRISRP-Cas9 technology and transferred them together with

tumor cells into B-NSG mice (Fig. 4d; Supplementary information, Fig. S3c). Deletion of HS3ST1 or PD1 in ILC2-C cells did not apparently changed cell death (Supplementary information, Fig.

S3d). Deletion of HS3ST1 or PD1 in human ILC2-C cells markedly suppressed tumor growth (Fig. 4d, f; Supplementary information, Fig. S3e), and inhibited tumor cell proliferation rates as well

(Fig. 4e, g). In order to verify the role of HS3ST1 of ILC2s in CRC progression, we conditionally knocked out _Hs3st1_ in ILC2s (Supplementary information, Fig. S3f). Given that _Hs3st1_

was extremely highly expressed in ILC2s compared with other immune cells (Supplementary information, Fig. S3g), we thus crossed _Hs3st1__flox/flox_ mice with Id2-CreERT2 mice to delete

_Hs3st1_ in ILCs. With administration of tamoxifen, HS3ST1 was efficiently depleted in ILC2s (Supplementary information, Fig. S3h). Of note, deficiency of HS3ST1 suppressed AOM/DSS-induced

colon tumor development (Fig. 4h–i), and proliferation rates of tumor cells (Fig. 4j). Whereas tamoxifen did not affect tumor growth (Supplementary information, Fig. S3i). In addition, we

used anti-PD1 antibody to inject into CRC patient-derived tumor cell (PDC) tumors once a week. Of note, anti-PD1 antibody treatment remarkably suppressed tumor growth (Supplementary

information, Fig. S3j), suggesting that blocking PD1 on ILC2s might be a potential target against CRC patients. Collectively, ILC2-Cs promote tumor growth and HS3ST1 and PD1 in ILC2s are

involved in the CRC tumor progression. TRANSDIFFERENTIATION OF ILC3 TO ILCREG DURING CRC PROGRESSION Based on our scRNA seq data, the proportion of ILC3s was decreased with increased ILCregs

over CRC tumor progression. We used diffusion map to analyze distributions of ILC3 and ILCreg. We observed that conversion of ILC3s to ILCregs at the late stage of CRC tumor (Supplementary

information, Fig. S4a). In parallel, ILC3s decreased during CRC tumor progression, whereas ILCregs gradually increased (Fig. 5a). At day 120, ILC3s dramatically decreased and accompanied

with a large number of ILCregs inside CRC tumors (Fig. 5a). In addition, tumor infiltrating ILC3s mainly produced IL-22, while ILCregs majorly secreted IL-10 (Fig. 5b). In order to trace the

fate of lost ILC3s in CRC tumors, we generated _Rosa26-STOP-tdTomato;Rorc-Cre;IL-10-GFP_ lineage tracing mice and followed by AOM/DSS treatment. We found that Lin−tdTomato+RORγt− cells used

to be ILC3 (called exILC3) appeared in late stage CRC tumors, but not in early stage tumors (Fig. 5c and Supplementary information, Fig. S4b). Moreover, exILC3s expressed IL-10 and Id3

(Fig. 5c), which are signature markers of ILCregs.22 Moreover, tdTomato+RORγt- ILCregs only appeared at late stage tumors (Supplementary information, Fig. S4c), suggesting conversion of

ILC3s to ILCregs over CRC progression. Consistently, exILC3s displayed in late stage tumors, but not in early stage tumors (Fig. 5d). In addition, tumor infiltrating ILC3s overtly decreased

and ILCregs increased in human advanced CRC tumors (Fig. 5e, f). Notably, ILCregs in CRC tumors did not express signature markers of other ILCs (Supplementary information, Fig. S4d).

Altogether, ILC3s transdifferentiate into ILCregs during CRC progression. TGF-Β INDUCES THE CONVERSION OF ILC3 TO ILCREG To explore the role of ILCregs in CRC progression, we isolated tumor

infiltrating ILCregs from CRC tumors and transferred them together with tumor cells into B-NSG mice (Fig. 6a). We observed that ILCregs apparently promoted tumor growth (Fig. 6a, b). We

previously showed that the fate decision factor Id3 drives progenitor CHILP cells toward ILCregs.22 We noticed that Id3 also existed in exILC3s (Fig. 5c). We established

_Id3__flox/flox__;Id2-CreERT2_ mice to deplete ILCregs, followed by AOM/DSS treatment (Supplementary information, Fig. S5a). Deletion of Id3 indeed depleted ILCregs in CRC tumors, whereas

ILC3s slightly increased (Supplementary information, Fig. S5a, b). By contrast, other ILCs and Tregs were not apparently changed (Supplementary information, Fig. S5b). Of note, depletion of

ILCregs dramatically suppressed tumor growth (Fig. 6c–d). ILCregs are able to produce large amounts of IL-10 that participates in immunosuppression of tumor progression. We next used

CRISRP/Cas9 technology to delete IL-10 in ILCregs as previously described,22 followed by adoptive transfer assays with IL-10-deleted ILCregs (Supplementary information, Fig. S5c). Notably,

deletion of IL-10 in ILCregs dramatically suppressed tumor development (Supplementary information, Fig. S5d), suggesting that ILCregs promote tumor progression in an IL-10 dependent manner.

These data indicate that ILCregs promote CRC tumor growth. We next wanted to explore the mechanism by which ILC3s converted into ILCregs over CRC progression. Through analysis of scRNA seq

data, we noticed that TGF-β signaling was elevated in ILCregs (Fig. 6e). We established _Tgfbr2__flox/flox__;Id2-CreERT2_ mice to delete TGF-β receptor on ILCs and treated with AOM/DSS. Of

note, ILCregs were almost undetectable in CRC tumors, whereas ILC3s slightly increased (Fig. 6f). In addition, deletion of TGF-β receptor on ILCs remarkably suppressed tumor growth (Fig. 

6g). Moreover, deletion of TGF-β receptor on ILC3s also suppressed tumor growth and decreased ILCreg numbers (Supplementary information, Fig. S5e, f). Consistently, TGF-β inhibitor treatment

abrogated the conversion of ILC3s to ILCregs (Fig. 6h; Supplementary information, Fig. S5g), and consequently inhibited patient derived xenograft (PDX) tumor growth (Fig. 6i, j). Notably,

TGF-β inhibitor treatment also blocked the conversion of ILC3s to ILCregs in PDX tumors (Fig. 6k). We conclude that TGF-β signaling is involved in the transdifferentiation of ILC3s to

ILCregs over CRC progression. Altogether, TGF-β induces the conversion of ILC3 to ILCreg, and blocking the conversion of ILC3s to ILCregs can suppress tumor growth both in AOM/DSS-induced

CRC tumors and in PDX tumors. DISCUSSION Colorectal cancer is a common malignancy worldwide, which is highly related to chronic inflammation. Inflammatory microenvironment affects the fate

and function of immune cells leading to immune evasion and tumor development. ILCs play critical roles in early antigen defence, allergy and inflammation. However, the physiological roles

and regulation of ILCs in colorectal cancer progression are still unclear. Here we used scRNA-seq to profile colorectal tumor infiltrating ILCs and found ILCs are dynamically regulated

during CRC progression. ILC1s undergo functional conversion during CRC progression, which exert inhibitory functions at the late stage of CRC. Three ILC2 subsets were identified inside CRC

tumors and ILC2-Cs promote tumor development. HS3ST1 and PD1 are highly expressed in tumor filtrating ILC2s of late stage CRC tumors and deletion of HS3ST1 and PD1 in ILC2s suppress tumor

growth. ILC3s convert to ILCregs under the stimulation of TGF-β during CRC progression. Importantly, ILCregs can promote tumor development and progression. Blocking TGF-β signaling can

abrogate the transdifferention of ILC3s to ILCregs and suppress tumor development as well. The roles of ILCs in regulation of tumorigenesis and progression still remain elusive. It has been

reported that ILC1-like cells inside mammary tumors exert immunosurveillance function at an early stage.27 However, under TGF-β-rich tumor environment, NK cells convert to ILC1s leading to

tumor escape from immunosurveillance.35 Herein we found that tumor infiltrating ILC1s express activating receptors at an early stage and inhibitory receptors at a late stage. Production of

IFN-γ is suppressed in ILC1s of late stage CRC tumors, suggesting ILC1s undergo functional conversion during CRC progression. IFN-γ is considered as a critical cytokine in tumor

immunosurveillance,30 and production of IFN-γ by innate lymphocytes such as NK cells apparently inhibit tumor development.36 It is possible that ILC1s and NK cells produce initial levels of

IFN-γ and inhibit tumor development. However, inflammatory environment might affect ILC1 function and disarm their immunosurveillance function leading to CRC development. The mechanism for

ILC1 conversion and regulation is needed to be further investigated. Several reports show that ILC2s are associated with pulmonary and hepatic fibrosis, suggesting a potential role of ILC2s

in the inflammation of lung and liver.37,38 In this study, we showed that tumor filtrating ILC2s are heterogeneous inside CRC tumors. The ratio of tumor cell/ILC2-C is 10:1 that can

efficiently promote tumor growth. We tested tumor/ILC2-C ratios from 20:1 to 4:1 and successfully induced tumor growth (data not shown). ILC2-Cs predominate in the late stage of CRC and

markedly promote tumor growth. Heparan sulfate 3-O-sulfotransferase 1 (HS3ST1) and PD1 are highly expressed in tumor filtrating ILC2s. HS3ST1 is a rate-limiting enzyme for heparan sulfate

production. However, HS3ST1 deficient mice are lethal but do not display an obvious procoagulant phenotype.31 Heparan sulfate has been reported to be implicated in tumor proliferation,

metastasis, angiogenesis.32 In T cells, PD1 recruits SHP2 and inhibits ZAP70 activation leading to T cell suppression.39 Deficiency of HS3ST1 or PD1 in ILC2s apparently suppresses tumor

growth, but whose deletion does not affect cell death of ILC2s. In addition, conditional deletion of HS3ST1 remarkably inhibits AOM/DSS-induced CRC tumor development. These data suggest that

targeting HS3ST1 and PD1 might be ideal strategies for CRC immunotherapy. However, how HS3ST1 and PD1 regulate ILC2 function requires to be further investigated. Besides HS3ST1 and PD1, we

also identified several genes in tumor filtrating ILC2s, including _Ltb4r1_ (encoding leukotriene B4 receptor, BLT1), _Hilpda_ (encoding HILPDA), and _Tnfrsf18_ (encoding

Glucocorticoid-induced TNFR-related protein, GITR). The roles of these molecules in the regulation of ILC2s for CRC development are worthy of further investigation. ILC3s in gut are related

to chronic inflammation that might cause gut tumor.40 IL-23 signaling and high levels of IL-22 promote gut tumorigenesis.41,42 We previously identified a regulatory subpopulation of ILCs

called ILCregs, which promote resolution of gut inflammation by secretion of IL-10.22 IL-10 is considered to be a negative indicator for clinical outcome in many cancers.43 We showed that

ILC3s transdifferentiate into ILCregs during CRC progression by stimulation of TGF-β. Moreover, ILCregs can promote colon tumor growth and depletion of ILCregs dramatically inhibits tumor

growth. Our results suggest that the conversion of ILC3s to ILCregs plays a critical role in the progression of CRC. The transdifferentiation of ILC3 to ILCreg is worthy to be analyzed by

RNA velocity in our further analysis.44 In addition, TGF-β signaling induces the transdifferentiation of ILC3s to ILCregs during CRC progression. Blocking TGF-β signaling remarkably

suppresses tumor growth of CRC, suggesting inhibitors against TGF-β signaling might be ideal agents for CRC treatment. In summary, we define six clusters of tumor infiltrating ILCs during

CRC progression. ILC2-Cs and ILCregs promote tumor development and growth of CRC. Our findings reveal that tumor infiltrating ILCs undergo conversions of their functions and subsets during

CRC progression. The conversions of ILCs are related to CRC development and tumor progression. Blocking ILCreg transdifferentiation can suppress tumor development. Therefore intervention of

ILC conversions may be a potential immunotherapy for CRC patient treatment. MATERIALS AND METHODS ANTIBODIES AND REAGENTS Antibodies used for flow cytometry: anti-mouse CD3(17A2), anti-mouse

CD8a (53-6.7), anti-mouse CD19 (1D3), anti-mouse NK1.1 (PK136), anti-mouse CD11b (M1/70), anti-mouse CD11c (N418), anti-mouse Gr1 (RB6-8C5), anti-mouse F4/80 (BM8), anti-mouse Ter119

(TER-119), anti-mouse CD45 (30-F11), anti-mouse CD127 (A7R34), anti-mouse IL-10 (JES5-16E3), anti-mouse CD117 (2B8), anti-mouse NKp46 (29A1.4), anti-mouse ST2 (RMST2-33), anti-mouse KLRG1

(2F1), anti-mouse RORγt (AFKJS-9), anti-mouse Ly49G2 (4D11), anti-human IL-10 (JES3-9D7), anti-human CD127 (eBioRDR5), anti-human CD45 (2D1), Human Hematopoietic Lineage Monoclonal Antibody,

anti-human T-bet (4B10), anti-human CD56 (TULY56), anti-human CRTH2 (BM16) were purchased from eBioscience; anti-mouse Id3 (S30-778) was from BD Pharmingen; anti-human/mouse IL-12RB2 was

from R&D System; anti-mouse NKG2I was from Invitrogen; anti-human KIR2DL2/L3(DX27) was from Biolegend; anti-KIR2DL1 (REA284) was from Miltenyi Biotec. Anti-human PD1 antibody (nivolumab)

was from invivogen. Recombinant murine IL-12, mIL-18, and mTGF-β1 were purchased from PeproTech. Brefeldin A and 7-AAD was from eBioscience. LY2109761 was from MedChemExpress. MOUSE AND CRC

PATIENT SAMPLES _Id2-CreERT2_, _Rorc_-_Cre_, and _Tgfbr2__flox/flox_ mice were from Jackson laboratory. _Rosa26-STOP-tdTomato_ mice were from Shanghai Research Center for Model Organisms.

B-NSG (NOD_-Prkdc__scid__IL2rg__tm1__/_Bcgen) mice were from Beijing Biocytogen Co., Ltd. IL-10-GFP mice were from Dr. Flavell (Yale University). _Id3__flox/flox_ mice were from Drs. Yuan

Zhuang (Duke University) and Chen Dong (Tsinghua University). Lineage tracing mice _Rosa26-STOP-tdTomato;Rorc-Cre_ were obtained by crossing _Rosa26-STOP-tdTomato_ mice with _Rorc-Cre_ mice.

Both female and male mice were used in experiments. Age- and sex-matched littermates between 8 and 28 weeks of age were used. Mice were assigned randomly to experimental groups. To generate

inducible TGF-β1 receptor deletion mouse model, _Tgfbr2__flox/+_ mice were crossed with _Id2-CreERT2_ mice. To generate Id3 deletion mouse model, _Id3__flox/+_ mice were crossed with

_Id2-CreERT2_ mice for several rounds to obtain _Id3__flox/flox__;CreERT2_. _Hs3st1__flox/+_ mice were generated by inserting loxp sites flanking exon 2 of _Hs3st1_ gene by CRISPR/Cas9

technology (Supplementary information, Fig. S3f) as described previously.45 In short, a targeting construct containing loxP-flanked exon 2 of _Hs3st1_ was generated with CRISPR mediated

long-chain single-stranded oligodeoxynucleotides (lsODNs). This construct was cloned into the pLSODN-1 plasmid and introduced by homologous recombination into the endogenous _Hs3st1_ gene.

Targeted alleles were identified by PCR screening and DNA sequencing. Upstream sgRNA: 5’-CAGACAATAGACAGAGACAGGGG-3’; downstream sgRNA: 5’-CAGGGGTTAAACCCACATGTGGG-3’. All the mice are C57BL/6

background and maintained under specific pathogen-free conditions with approval by the Institutional Committee of Institute of Biophysics, Chinese Academy of Sciences. The study is

compliant with all relevant ethical regulations regarding animal research. Human resected colon tissues of CRC patients were obtained from West China Hospital, Sichuan University (Chengdu,

China) and Cancer Hospital Chinese Academy of Medical Sciences (Beijing, China) with informed consents, and approved by the Institutional Review Board of the Institute of Biophysics, Chinese

Academy of Sciences. The study is compliant with all relevant ethical regulations regarding research involving human participants. AOM/DSS-INDUCED COLITIS-ASSOCIATED COLORECTAL CANCER 8–12

week-old mice of body weight > 20 g were i.p. injected of azoxymethane (AOM, 10 mg/kg body weight). After 4 days, mice were treated with 1.5% dextran sodium sulfate (DSS) in drinking

water for six days. Then, DSS was replaced by drinking water. After two weeks, mice were treated with 1.5% DSS for another six days. Three cycles of DSS treatment were needed before tumor

evaluation. Colon tumors were analyzed at early stage (60 days after AOM injection) or late stage (120 days after AOM injection). ISOLATION OF TUMOR INFILTRATING ILCS Colon tumors from mice

colons or CRC patient samples were isolated and cut into fine pieces. Tumor samples were digested twice for 40 min each at 37 °C with digestion buffer containing 5% FBS, Collagenase II and

III (1 mg/mL; Worthington), DNase I (200 μg/mL; Roche) and dispase (4U/mL; Sigma). Mononuclear cells were isolated with 40%–80% Percol gradient, and washed twice with PBS. Isolated cells

were blocked with anti-CD16/32 antibody for 30 min on ice and then stained with antibodies against CD45, CD127, lineage cocktail (Lin = CD3e,CD8a,CD19,CD11b,CD11c,Gr1,F4/80,Ter119) and ILC

markers on ice for 1 h followed by 7AAD staining. Tumor infiltrating ILCs were isolated by flow cytometer (FACS Aria III, BD) for functional study or PDC experiments. Purity of ILC cells was

over 95% for each assay that was determined by post sorting analysis of flow cytometry. FLOW CYTOMETRY Tumor infiltrating cells were isolated and blocked with anti-CD16/32 antibody for 30 

min on ice. Surface markers were then stained for 1 h on ice. For intracellular cytokine staining, cells were cultured in completed RPMI1640 media in the presence of Brefeldin A at 37 °C for

4 h. Next, cells were harvested for surface marker staining and fixed and permeablized by Intracellular Fixation & Permeablization buffer set (eBioscience), followed by intracellular

antigen staining. Cell samples were analyzed by flow cytometer (FACS Aria III, BD) as described.46 IMMUNOFLUORESCENCE ASSAY For in situ immunofluorescence of ILCs, mouse colons were treated

as previously described.22 Briefly, colon tumors or intestines with longitudinally opened were fixed in 4% paraformaldehyde (PFA) (Sigma) fixative for 8 h, and rehydrated in 30% sucrose

solution for 24 h and frozen in OCT for sectioning. Tumor or intestinal sections were rehydrated in PBS and blocked in 10% donkey serum and anti-CD16/32 antibody. Primary antibodies were

used for section staining for 2 h at room temperature (RT). Next, we used Alex488-conjugated and Alex594-conjugated donkey secondary antibodies for further staining. Then, we used

APC-conjugated lineage cocktails to stain slides at RT for 1 h. After washing for 3 times, nuclei were stained by DAPI if needed and the sections were subjected to dehydrating in EtOH

gradient: 70%, 85%, 95%, and 100%. Mounted sections were analyzed by confocal microscopy (ZIESS LSM700). Antibodies for immunofluorescence staining were anti-mouse KLRG1 (2F1; eBioscience),

anti-human/mouse HS3ST1 (Sigma), anti-human/mouse CD279 (PD1) (Invitrogen), anti-human CRTH2 (BM16, eBioscience), anti-GFP (GF28R, Invitrogen), anti-tdTomato (Biorbyt), anti-human/mouse

RORγt (AFKJS-9, eBioscience). For 5-color staining, OpalTM 7-color fIHC kit (PerkinElmer) was used according to the manufacturer’s instructions. Briefly, paraffin sections of colon tumors

were sequentially stained with primary antibodies and HRP conjugated secondary antibody. Microwave treatment was used after each Opal dye staining. Anti-GFP, anti-tdTomato, anti-RORγt, and

anti-Lin antibodies were used as primary antibodies. Opal520, Opal570, Opal620, Opal690 dyes were used for secondary staining. Samples were observed using Vectra Automated Quantitative

Pathology Imaging System (PerkinElmer). DROPLET-BASED SCRNA-SEQ CDNA LIBRARY PREPARATION AND SEQUENCING Tumor infiltrating ILCs from early (D60) or late (D120) stage of colon tumors were

harvested and sorted by flow cytometry (n = 5 for each group). These mice shared the same background and similar distribution of ILCs according to our flow cytometry analysis (data not

shown) and we merged them together for scRNA-seq. Tumor infiltrating ILCs (Lin−CD45+CD127+) (Lin = CD3e,CD8a,CD19,CD11b,CD11c,Gr1,F4/80,Ter119) were isolated and cell suspensions (300-500

living cells per microliter) were loaded on a Chromium Single Cell Controller (10x Genomics) to generate single-cell Gel Bead-In-Emulsions (GEM) by using Single Cell 3’ Library and Gel Bead

Kit V2 (10x Genomics, 120237). mRNAs of captured cells were barcoded through reverse transcription. Barcoded cDNAs were pooled and cleaned up by using DynaBeads® MyOne Silane Beads

(Invitrogen, 37002D). Single-cell RNA-seq cDNA libraries were prepared using Single Cell 3’ Library Gel Bead Kit V2 (10x Genomics, 120237) according to the manufacture’s protocol followed by

sequencing on an Illumina HiSeq X Ten with pair end 150 bp (PE150). GSE number of scRNA sequencing is GSE142694. DROPLET-BASED SCRNA-SEQ DATA PREPROCESS Raw sequencing data was checked by

FastQC software. Next, fastq sequences were trimmed by FASTX-Toolkit. Gene counts were obtained by aligning reads to the mouse mm10 genome by using CellRanger software (version 2.0.1) to

generate the single cell information. We abandoned cells with apparent lineage markers such as _Cd3e_, _Cd19_, _Ly6g_, _Ly6c_ and _Adgre1_ which might be caused by a contamination during

FACS-based cell separation. Next, we removed doublets and poor-quality cells by Seurat (v2.2.1) based on the number of unique molecular identifiers (UMIs) and percentage of mitochondrial

genes. Briefly, raw cell counts were filtered in Seurat CreateSeuratObject function with the following criteria: one gene was expressed in at least 3 cells and at least 200 genes were

detected in one cell. Mitochondrial genes inside one cell were also calculated with a percentage under 20%. Cells were further filtered with the following requirements: gene numbers inside

on cell were no more than 8000 and no less than 200. “FindVariableGenes” function of Seurat was used to get the variable genes with the following parameters: x.low.cutoff 0.05, x.high.cutoff

8 and y.cutoff 0.5. 1226 variable genes were found as the input genes for PCA or DiffusionMap analysis. PCA analysis was conducted using the “RunPCA” function of Seurat and PCs 1–12 were

chosen for dimension reduction analysis with Seurat function “RunTSNE” and DiffusionMap analysis with Seurat function “RunDiffusion”. For the clustering analysis, the first twelve PCs were

used to calculate clusters with a resolution of 0.7 using Seurat function “FindClusters”. Differently expressed genes were calculated through Seurat “FindAllMarkers”. For heatmap and scatter

plots, scaled data ([email protected]) from Seurat pipeline were used. For violin plots, normalized data (scRNA@data) from Seurat pipeline were used. Diffusion map was drawn using the Seurat

package command RunDiffusion with the following parameters: max.dim = 5, dims.use = 1:12 using top 10 marker genes of ILC3 and ILCreg. DM1, DM3, DM4 were used to plot the three-dimensional

diffusion map using the R package “rgl” with the default parameters. GENE KNOCKOUT BY CRISPR/CAS9 TECHNOLOGY ILC2-C cells from stage IV CRC patient samples were isolated and infected with

lentivirus (lentiCRISPR-v2) carrying sgRNA against _HS3ST1_ or _PDCD1_ gene (Supplementary information, Table S2). Infected cells were selected with 0.5 μg/mL puromycin (Invivogen) and

subcutaneously injected into B-NSG mice together with human colon tumor cells. After 3 days, expression of HS3ST1 or PD1 in ILC2-C cells was analyzed by flow cytometry. _HS3ST1__−/−_ or

_PDCD1__−/−_ ILC2-C cells together with primary tumor cells from stage IV CRC patients were subcutaneously injected into B-NSG mice. CRC PATIENT-DERIVED XENOGRAFT (PDX) AND PATIENT-DERIVED

CELL (PDC) MODELS CRC PDX and PDC models were established as previously described.47 For PDX models, tumor samples from different stages of CRC patient samples were cut into 25–30 mm3 pieces

in PBS with 100 mg/mL penicillin and 100 U/mL streptomycin. Tumor sample pieces were coated in Matrigel (BD Biosciences), followed by subcutaneous implantation into 6-week-old B-NSG mice

(Biocytogen Co.). While the established tumors achieved 400 mm3, the mice were intratumorally injected with TGF-β inhibitor LY2109761. Tumor volumes were determined every 5 days. For PDC

models, tumor samples from different stages of CRC patient samples were cut into small pieces and subjected with digestion by Collagenase II and III (1 mg/mL; Worthington). Tumor cells

(EPCAM+) and ILC2 cells were isolated by flow cytometry and subcutaneously injected into 6-week-old B-NSG mice. Isolated ILC2 cells were transfected with virus carrying CRISPR/Cas9 elements

against _HS3ST1_ or _PDCD1_ and subcutaneously injected into B-NSG mice together with tumor cells. For anti-PD1 antibody treatment, 1 mg/kg anti-human PD1 antibody (nivolumab, Invivogen) was

injected into tumors at day 25 after tumor injection once a week. PBS was used as negative control. Tumor volumes were examined every 5 days. GENERATION OF MOUSE MODEL WITH IL-10-DEFICIENT

ILCREGS ILCregs were isolated from _Rosa26-STOP-Cas9;Id2-ERT2-Cre;IL-10-GFP_ mice and infected with lentivirus (lentiGuide-Puro-sgRNA-_Il10_) carrying sgRNA against _Il10_ gene (#1:

5’-TATTGTCTTCCCGGCTGTAC-3’; #2: 5’-GCATGTGGCTCTGGCCGACT-3’). Infected cells were selected with 1 μg/mL puromycin (Invivogen) and adoptively transferred into _Rag1__−/−__Il2rg__−/−_ mice

together with respective WT (CD45.1) lymphocytes. Expression of IL-10 in ILCregs (CD45.2) after treatment with TMX (50 mg/kg i.p. for five consecutive days) was detected prior to

transplantation. ELISA ASSAY Tumor-infiltrating ILCs from colon tumors were isolated and cultured in complete media for 24 h. Cytokines in supernatants were analyzed by ELISA kit

(eBioscience) by manufacturer’s instructions. Detection thresholds of different cytokines were 50 pg/mL for IFN-γ, 50 pg/mL for IL-17A, 50 pg/mL for IL-22, and 50 pg/mL for IL-10. Values

below detection thresholds were shown as ND (not detectable). ELISA data were normalized to total cell numbers and shown as means ± SD per 5 ×103 cells. REAL TIME PCR ASSAY Tumor

infiltrating ILCs were isolated from colon tumors. mRNA was extracted by using Neasy Micro Kit (Qiagen) according to the manufacturer’s instructions. mRNA quality was determined by A260/A280

(between 1.8 and 2.2) and A260/230 (> 1.7) ratios. RNA integrity was assessed by agarose gel electrophoresis. cDNAs were synthesized by using oligo dT followed by real time PCR. Primers

for real-time PCR in this study are as shown in Supplementary information, Table S1. STATISTICAL ANALYSIS For statistical analysis, data were analyzed by using GraphPad Prism 7.0. Two-tailed

unpaired Student’s _t_ test, One way ANOVA and Non parametric Mann–Whitney U-test were used according to the type of experiments. _P_-values  <0.05 were considered significant (*_P_ <

 0.05; **_P_ < 0.01; _P_ < 0.001); _P_ > 0.05, non-significant (NS). All flow cytometry data were analyzed with FlowJo (Treestar). No statistical methods were used to predetermine

sample sizes. CHANGE HISTORY * _ 15 JUNE 2020 An amendment to this paper has been published and can be accessed via a link at the top of the paper. _ * _ 25 MAY 2020 An amendment to this

paper has been published and can be accessed via a link at the top of the paper. _ REFERENCES * Grivennikov, S. I., Greten, F. R. & Karin, M. Immunity, inflammation, and cancer. _Cell_

140, 883–899 (2010). Article  CAS  PubMed  PubMed Central  Google Scholar  * Lasry, A., Zinger, A. & Ben-Neriah, Y. Inflammatory networks underlying colorectal cancer. _Nat. Immunol._

17, 230–240 (2016). Article  CAS  PubMed  Google Scholar  * Shalapour, S. & Karin, M. Immunity, inflammation, and cancer: an eternal fight between good and evil. _J Clin. Invest._ 125,

3345–55 (2015). Article  Google Scholar  * Feagins, L. A., Souza, R. F. & Spechler, S. J. Carcinogenesis in IBD: potential targets for the prevention of colorectal cancer. _Nat. Rev.

Gastroenterol. Hepatol._ 6, 297–305 (2009). Article  CAS  PubMed  Google Scholar  * Terzic, J., Grivennikov, S., Karin, E. & Karin, M. Inflammation and colon cancer. _Gastroenterology_

138, 2101–2114 (2010).  Article  CAS  PubMed  Google Scholar  * Arthur, J. C. et al. Intestinal inflammation targets cancer-inducing activity of the microbiota. _Science_ 338, 120–123

(2012). Article  CAS  PubMed  PubMed Central  Google Scholar  * Schreiber, R. D., Old, L. J. & Smyth, M. J. Cancer immunoediting: integrating immunity’s roles in cancer suppression and

promotion. _Science_ 331, 1565–1570 (2011). Article  CAS  PubMed  Google Scholar  * Shankaran, V. et al. IFNgamma and lymphocytes prevent primary tumour development and shape tumour

immunogenicity. _Nature_ 410, 1107–1111 (2001). Article  CAS  PubMed  Google Scholar  * Sun, J. C. & Lanier, L. L. NK cell development, homeostasis and function: parallels with CD8(+) T

cells. _Nat. Rev. Immunol._ 11, 645–657 (2011). Article  CAS  PubMed  PubMed Central  Google Scholar  * Vesely, M. D., Kershaw, M. H., Schreiber, R. D. & Smyth, M. J. Natural innate and

adaptive immunity to cancer. _Annu. Rev. Immunol._ 29, 235–271 (2011). Article  CAS  PubMed  Google Scholar  * Zou, W. Regulatory T cells, tumour immunity and immunotherapy. _Nat. Rev.

Immunol._ 6, 295–307 (2006). Article  CAS  PubMed  Google Scholar  * Hashimoto, M. et al. CD8 T cell exhaustion in chronic infection and cancer: opportunities for interventions. _Annu. Rev.

Med._ 69, 301–318 (2018). Article  CAS  PubMed  Google Scholar  * Artis, D. & Spits, H. The biology of innate lymphoid cells. _Nature_ 517, 293–301 (2015). Article  CAS  PubMed  Google

Scholar  * Serafini, N., Vosshenrich, C. A. & Di Santo, J. P. Transcriptional regulation of innate lymphoid cell fate. _Nat. Rev. Immunol._ 15, 415–428 (2015). Article  CAS  PubMed 

Google Scholar  * Fuchs, A. et al. Intraepithelial type 1 innate lymphoid cells are a unique subset of IL-12- and IL-15-responsive IFN-gamma-producing cells. _Immunity_ 38, 769–781 (2013).

Article  CAS  PubMed  PubMed Central  Google Scholar  * Bernink, J. H. et al. Human type 1 innate lymphoid cells accumulate in inflamed mucosal tissues. _Nat. Immunol._ 14, 221–229 (2013).

Article  CAS  PubMed  Google Scholar  * Hoyler, T. et al. The transcription factor GATA-3 controls cell fate and maintenance of type 2 innate lymphoid cells. _Immunity_ 37, 634–648 (2012).

Article  CAS  PubMed  PubMed Central  Google Scholar  * Gerbe, F. et al. Intestinal epithelial tuft cells initiate type 2 mucosal immunity to helminth parasites. _Nature_ 529, 226–230

(2016). Article  CAS  PubMed  PubMed Central  Google Scholar  * Wallrapp, A. et al. The neuropeptide NMU amplifies ILC2-driven allergic lung inflammation. _Nature_ 549, 351–356 (2017).

Article  CAS  PubMed  PubMed Central  Google Scholar  * Buonocore, S. et al. Innate lymphoid cells drive interleukin-23-dependent innate intestinal pathology. _Nature_ 464, 1371–1375 (2010).

Article  CAS  PubMed  PubMed Central  Google Scholar  * Ibiza, S. et al. Glial-cell-derived neuroregulators control type 3 innate lymphoid cells and gut defence. _Nature_ 535, 440–443

(2016). Article  CAS  PubMed  PubMed Central  Google Scholar  * Wang, S. et al. Regulatory innate lymphoid cells control innate intestinal inflammation. _Cell_ 171, 201–216 (2017).  Article

  CAS  PubMed  Google Scholar  * Neufert, C., Becker, C. & Neurath, M. F. An inducible mouse model of colon carcinogenesis for the analysis of sporadic and inflammation-driven tumor

progression. _Nat. Protoc._ 2, 1998–2004 (2007). Article  CAS  PubMed  Google Scholar  * Man, S. M. et al. Critical role for the DNA sensor AIM2 in stem cell proliferation and cancer. _Cell_

162, 45–58 (2015). Article  CAS  PubMed  PubMed Central  Google Scholar  * Gury-BenAri, M. et al. The spectrum and regulatory landscape of intestinal innate lymphoid cells are shaped by the

microbiome. _Cell_ 166, 1231–1246 (2016). e1213. Article  CAS  PubMed  Google Scholar  * Eberl, G., Colonna, M., Di Santo, J. P. & McKenzie, A. N. Innate lymphoid cells. Innate lymphoid

cells: a new paradigm in immunology. _Science_ 348, aaa6566 (2015). Article  PubMed  PubMed Central  Google Scholar  * Dadi, S. et al. Cancer immunosurveillance by tissue-resident innate

lymphoid cells and innate-like T ccells. _Cell_ 164, 365–377 (2016). Article  CAS  PubMed  PubMed Central  Google Scholar  * Robinette, M. L. et al. Transcriptional programs define molecular

characteristics of innate lymphoid cell classes and subsets. _Nat. Immunol._ 16, 306–317 (2015). Article  CAS  PubMed  PubMed Central  Google Scholar  * Vivier, E., Tomasello, E., Baratin,

M., Walzer, T. & Ugolini, S. Functions of natural killer cells. _Nat. Immunol._ 9, 503–510 (2008). Article  CAS  PubMed  Google Scholar  * Kaplan, D. H. et al. Demonstration of an

interferon gamma-dependent tumor surveillance system in immunocompetent mice. _Proc. Natl. Acad. Sci. USA_ 95, 7556–7561 (1998). Article  CAS  PubMed  PubMed Central  Google Scholar  *

HajMohammadi, S. et al. Normal levels of anticoagulant heparan sulfate are not essential for normal hemostasis. _J Clin. Invest._ 111, 989–999 (2003). Article  CAS  PubMed  PubMed Central 

Google Scholar  * Knelson, E. H., Nee, J. C. & Blobe, G. C. Heparan sulfate signaling in cancer. _Trends Biochem. Sci._ 39, 277–288 (2014). Article  CAS  PubMed  PubMed Central  Google

Scholar  * Chen, L. & Han, X. Anti-PD-1/PD-L1 therapy of human cancer: past, present, and future. _J Clin. Invest._ 125, 3384–3391 (2015). Article  PubMed  PubMed Central  Google Scholar

  * Sharma, P. & Allison, J. P. The future of immune checkpoint therapy. _Science_ 348, 56–61 (2015). Article  CAS  PubMed  Google Scholar  * Gao, Y. et al. Tumor immunoevasion by the

conversion of effector NK cells into type 1 innate lymphoid cells. _Nat. Immunol._ 18, 1004–1015 (2017). Article  CAS  PubMed  Google Scholar  * Martin-Fontecha, A. et al. Induced

recruitment of NK cells to lymph nodes provides IFN-gamma for T(H)1 priming. _Nat. Immunol._ 5, 1260–1265 (2004). Article  CAS  PubMed  Google Scholar  * Hams, E. et al. IL-25 and type 2

innate lymphoid cells induce pulmonary fibrosis. _Proc. Natl. Acad. Sci. USA_ 111, 367–372 (2014). Article  CAS  PubMed  Google Scholar  * McHedlidze, T. et al. Interleukin-33-dependent

innate lymphoid cells mediate hepatic fibrosis. _Immunity_ 39, 357–371 (2013). Article  CAS  PubMed  PubMed Central  Google Scholar  * Okazaki, T., Chikuma, S., Iwai, Y., Fagarasan, S. &

Honjo, T. A rheostat for immune responses: the unique properties of PD-1 and their advantages for clinical application. _Nat. Immunol._ 14, 1212–1218 (2013). Article  CAS  PubMed  Google

Scholar  * Ebbo, M., Crinier, A., Vely, F. & Vivier, E. Innate lymphoid cells: major players in inflammatory diseases. _Nat. Rev. Immunol._ 17, 665–678 (2017). Article  CAS  PubMed 

Google Scholar  * Chan, I. H. et al. Interleukin-23 is sufficient to induce rapid de novo gut tumorigenesis, independent of carcinogens, through activation of innate lymphoid cells. _Mucosal

Immunol._ 7, 842–856 (2014). Article  CAS  PubMed  Google Scholar  * Jiang, R. et al. IL-22 is related to development of human colon cancer by activation of STAT3. _BMC Cancer_ 13, 59

(2013). Article  CAS  PubMed  PubMed Central  Google Scholar  * Moore, K. W., Malefyt, deWaal, Coffman, R. & O’Garra, R. L. A. Interleukin-10 and the interleukin-10 receptor. _Annu. Rev.

Immunol._ 19, 683–765 (2001). Article  CAS  PubMed  Google Scholar  * La Manno, G. et al. RNA velocity of single cells. _Nature_ 560, 494–498 (2018). Article  PubMed  PubMed Central  Google

Scholar  * Xia, P. et al. The ER membrane adaptor ERAdP senses the bacterial second messenger c-di-AMP and initiates anti-bacterial immunity. _Nat. Immunol._ 19, 141–150 (2018). Article 

CAS  PubMed  Google Scholar  * Wang, S. et al. Natural killer-like B cells prime innate lymphocytes against microbial infection. _Immunity_ 45, 131–144 (2016). Article  CAS  PubMed  Google

Scholar  * Zhu, P. et al. LncGata6 maintains stemness of intestinal stem cells and promotes intestinal tumorigenesis. _Nat. Cell. Biol._ 20, 1134–1144 (2018). Article  CAS  PubMed  Google

Scholar  Download references ACKNOWLEDGEMENTS We thank Junying Jia, Shuang Sun, Xiaoxiao Zhu, Yihui Xu, Shu Meng, Dongdong Fan and Yan Teng for technical support. We thank Drs Flavell (Yale

University), Yuan Zhuang (Duke University) and Chen Dong (Tsinghua University) for providing genetic mouse strains. We thank Jing Li (Cnkingbio Company Ltd., Beijing, China) for technical

support. We thank CapitalBio Technology Inc. for technical support. This work was supported by the Strategic Priority Research Programs of the Chinese Academy of Sciences (XDB29020000,

XDB19030203), the National Natural Science Foundation of China (81722023, 81922031, 81921003, 31930036, 91640203, 31671531, 91940305, 31770939), Key Research Program of Frontier Sciences of

Chinese Academy of Sciences (ZDBS-LY-SM025), National Key R&D Program of China (2019YFA0111800, 2019YFA0508501), CAMS Innovation Fund for Medical Sciences (2017-I2M-1-006), Beijing

Natural Science Foundation (7181006) and Youth Innovation Promotion Association of CAS to S.W. AUTHOR INFORMATION Author notes * These authors contributed equally: Shuo Wang, Yuan Qu,

Pengyan Xia, Yi Chen. AUTHORS AND AFFILIATIONS * CAS Key Laboratory of Infection and Immunity, CAS Center for Excellence in Biomacromolecules, Institute of Biophysics, Chinese Academy of

Sciences, 100101, Beijing, China Shuo Wang, Yuan Qu, Pengyan Xia & Zusen Fan * CAS Key Laboratory of Pathogenic Microbiology and Immunology, Institute of Microbiology, Chinese Academy of

Sciences, 100101, Beijing, China Shuo Wang * Center for Biosafety Mega-Science, Chinese Academy of Sciences, Wuhan, 430071, Hubei, China Shuo Wang * University of Chinese Academy of

Sciences, 100049, Beijing, China Yuan Qu & Zusen Fan * Department of Immunology, School of Basic Medical Sciences, NHC Key Laboratory of Medical Immunology, Peking University, 100871,

Beijing, China Pengyan Xia * Department of Gastrointestinal Surgery, State Key Laboratory of Biotherapy, West China Hospital, Sichuan University, Chengdu, 610041, Sichuan, China Yi Chen 

& Guan Wang * CAS Key Laboratory of RNA Biology; Institute of Biophysics, Chinese Academy of Sciences, 100101, Beijing, China Xiaoxiao Zhu & Yong Tian * Department of Pathology,

National Cancer Center/National Clinical Research Center for Cancer/Cancer Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College, 100021, Beijing, China Jing Zhang 

& Jianming Ying Authors * Shuo Wang View author publications You can also search for this author inPubMed Google Scholar * Yuan Qu View author publications You can also search for this

author inPubMed Google Scholar * Pengyan Xia View author publications You can also search for this author inPubMed Google Scholar * Yi Chen View author publications You can also search for

this author inPubMed Google Scholar * Xiaoxiao Zhu View author publications You can also search for this author inPubMed Google Scholar * Jing Zhang View author publications You can also

search for this author inPubMed Google Scholar * Guan Wang View author publications You can also search for this author inPubMed Google Scholar * Yong Tian View author publications You can

also search for this author inPubMed Google Scholar * Jianming Ying View author publications You can also search for this author inPubMed Google Scholar * Zusen Fan View author publications

You can also search for this author inPubMed Google Scholar CONTRIBUTIONS S.W. designed and performed experiments, analyzed data and wrote the paper; Y.Q., P.X., and Y.C. performed

experiments and analyzed data; X.Z. and Y.T. generated animal models. G.W., J.Z. and J.Y provided human CRC samples. Z.F. initiated the study, organized, designed, and wrote the paper.

CORRESPONDING AUTHORS Correspondence to Shuo Wang, Yong Tian or Jianming Ying. ETHICS DECLARATIONS COMPETING INTERESTS The authors declare no competing interests. SUPPLEMENTARY INFORMATION

SUPPLEMENTARY FIGURE S1 SUPPLEMENTARY FIGURE S2 SUPPLEMENTARY FIGURE S3 SUPPLEMENTARY FIGURE S4 SUPPLEMENTARY FIGURE S5 SUPPLEMENTARY TABLE S1 SUPPLEMENTARY TABLE S2 RIGHTS AND PERMISSIONS

Reprints and permissions ABOUT THIS ARTICLE CITE THIS ARTICLE Wang, S., Qu, Y., Xia, P. _et al._ Transdifferentiation of tumor infiltrating innate lymphoid cells during progression of

colorectal cancer. _Cell Res_ 30, 610–622 (2020). https://doi.org/10.1038/s41422-020-0312-y Download citation * Received: 18 October 2019 * Accepted: 16 March 2020 * Published: 04 May 2020 *

Issue Date: July 2020 * DOI: https://doi.org/10.1038/s41422-020-0312-y SHARE THIS ARTICLE Anyone you share the following link with will be able to read this content: Get shareable link

Sorry, a shareable link is not currently available for this article. Copy to clipboard Provided by the Springer Nature SharedIt content-sharing initiative