Inactivation of adenomatous polyposis coli (APC) is common across many cancer types and serves as a critical initiating event in most sporadic colorectal cancers. APC deficiency activates WNT signaling, which remains an elusive target for cancer therapy, prompting us to apply the synthetic essentiality framework to identify druggable vulnerabilities for APC-deficient cancers. Tryptophan 2,3-dioxygenase 2 (TDO2) was identified as a synthetic essential effector of APC-deficient colorectal cancer. Mechanistically, APC deficiency results in the TCF4/β-catenin–mediated upregulation of TDO2 gene transcription. TDO2 in turn activates the Kyn–AhR pathway, which increases glycolysis to drive anabolic cancer cell growth and CXCL5 secretion to recruit macrophages into the tumor microenvironment. Therapeutically, APC-deficient colorectal cancer models were susceptible to TDO2 depletion or pharmacologic inhibition, which impaired cancer cell proliferation and enhanced antitumor immune profiles. Thus, APC deficiency activates a TCF4–TDO2–AhR–CXCL5 circuit that affects multiple cancer hallmarks via autonomous and nonautonomous mechanisms and illuminates a genotype-specific vulnerability in colorectal cancer.

Significance:

This study identifies critical effectors in the maintenance of APC-deficient colorectal cancer and demonstrates the relationship between APC/WNT pathway and kynurenine pathway signaling. It further determines the tumor-associated macrophage biology in APC-deficient colorectal cancer, informing genotype-specific therapeutic targets and the use of TDO2 inhibitors.

This article is highlighted in the In This Issue feature, p. 1599

Colorectal cancer is the second leading cause of cancer-related death in developed countries, causing more than 600,000 deaths globally each year. The evolution of colorectal cancer from adenoma to adenocarcinoma and ultimately invasive and metastatic disease is governed by the acquisition of signature genetic alterations, most prominently inactivation of adenomatous polyposis coli (APC) and p53 tumor suppressors and activation of the KRAS oncogene (1). Loss of APC is considered the critical initiating event, occurring in the vast majority (∼90%) of sporadic colorectal cancers. Consistent with its gatekeeper role, ApcMin/+ mice harboring a mutated Apc gene develop adenomatous polyps throughout the intestine (2). Colorectal cancer mouse models have also established an essential role for APC deficiency in tumor maintenance (3).

APC loss occurs frequently across many cancer types (4–7), motivating efforts to identify key APC signaling surrogates essential for tumor maintenance. In normal cells, APC activates glycogen synthase kinase 3β (GSK3β), which in turn phosphorylates N-terminal serine/threonine residues of β-catenin, mediating β-catenin degradation through ubiquitination. Thus, APC-deficient cancers accumulate β-catenin, which then translocates to the nucleus to bind and derepress the T-cell factor/lymphoid enhancer factor (TCF/LEF) transcription factor complex (8), enabling the activation of the canonical WNT signaling network.

Despite its importance in cancer, the therapeutic targeting of this WNT/APC signaling cascade remains an elusive goal for cancer therapy. Currently, agents targeting the WNT pathway include inhibitors of WNT ligands, the β-catenin degrading complex, TCF/LEF, and Notch and Sonic Hedgehog signaling, which cross-talk with WNT. To date, these WNT-targeting programs have yet to produce meaningful clinical results, motivating us to adopt an orthogonal strategy to identify key downstream effectors of APC deficiency needed for tumor maintenance. To that end, we adopted the synthetic essentiality (SE) approach, which begins with a search for genes that can be occasionally mutated/deleted in cancers but are never or rarely deleted in cancers harboring loss of a specific tumor suppressor gene. These mutually exclusive patterns in the cancer genome might merely belie an epistatic relationship or indicate that the SE gene serves as an essential effector of the specific tumor suppressor gene deficiency in supporting tumorigenesis. The first validated example of SE was the chromatin helicase DNA-binding protein 1 (CHD1), which serves as an essential effector of PTEN deficiency in prostate and breast cancers (9).

In the current study, the SE approach (10) identified tryptophan 2,3-dioxygenase 2 (TDO2) as a key downstream effector specifically in APC-deficient colorectal cancer. TDO2 mediates the first and rate-limiting step of the kynurenine pathway (KP), the major tryptophan (Trp) catabolism pathway in mammals, and converts Trp into N-formylkynurenine (Kyn). TDO2 is highly expressed and constitutively active in diverse cancers, which results in the accumulation of Kyn in the tumor microenvironment (TME) to suppress antitumor immunity. Trp depletion and Kyn accumulation promote the differentiation of monocytes into immunosuppressive tumor-associated macrophages (TAM) and inhibit T-cell proliferation/activation (11). TDO2–Kyn also mediates cancer cell–intrinsic pathways through Kyn action as an agonist for the aryl hydrocarbon receptor (AhR), which upregulates protumorigenic genes in glioblastoma and triple-negative breast cancer (12, 13). However, the genotypic context in which TDO2 (and by extension other KP enzymes, IDO1 and IDO2) might serve critical rate-limiting roles in specific cancers is not known. Our studies establish that the TDO2–Kyn–AhR axis serves a critical role in promoting APC-deficient tumor growth via cancer cell–autonomous (metabolism and proliferation) and nonautonomous mechanisms (tumor immunity).

Identification of TDO2 as a Downstream Effector for APC Deficiency in Cancer

To identify synthetic essential effectors of APC deficiency, we first searched for genes showing mutually exclusive mutation/deletion patterns with APC in The Cancer Genome Atlas (TCGA) database (Supplementary Table S1). To overcome the limitation that only a small fraction of colorectal cancer cases are intact for APC, we conducted a pan-cancer analysis that showed consistent retention of TDO2 in APC-deleted/mutated cancers including colorectal cancer, breast cancer, prostate cancer, lung cancer, head and neck squamous cell carcinoma, and sarcoma (Supplementary Fig. S1A). Recognizing the limited sample size and the low frequency of these genomic events, we triangulated these genomic results with (i) hits from genome-wide loss-of-function screens designed to identify genes that are consistently retained in cancer cells bearing APC loss-of-function mutations (14) and (ii) unbiased transcriptomic analyses to identify genes with positive correlations of WNT pathway activation signature (15) and HALLMARK_WNT_BETA_CATENIN_SIGNALING (16). These intersections yielded five potential SE genes for APC-deficient tumors (TDO2, C3, MAFB, CAB39L, and PPFIA2), with TDO2 as the top hit (Fig. 1A).

Analysis of human TCGA colorectal cancer data sets (COAD and READ) revealed that TDO2 gene expression indeed correlated positively with WNT pathway activation (Fig. 1B; Supplementary Fig. S1B–S1D). Correspondingly, the tissue microarray (TMA) analysis of human colorectal cancer samples showed coincident increased signals for TDO2 and for nuclear β-catenin and c-Myc, which are indicative of WNT pathway activation (Fig. 1C and D; Supplementary Fig. S1E–S1H). In murine models, colorectal cancer tumors of iAP mice (Apcmut/Trp53mut) and iKAP mice (inducible Krasmut with Apcmut/Trp53mut; ref. 17) showed that TDO2 expression tracks closely with β-catenin and Ki-67 signals in tumors (Fig. 1E; Supplementary Fig. S1I). Finally, ApcMin/+ organoids and CRISPR/Cas9-mediated APC-knockout (APC-KO) organoids showed significantly increased TDO2 expression compared with APC-wild-type (APC-WT) organoids (Fig. 1F and G).

In contrast, another KP enzyme, IDO1, did not exhibit mutually exclusive patterns with APC and CTNNB1 mutations in TCGA colorectal cancer or correlate with WNT pathway activation. IDO2 expression exhibited a correlation with WNT signaling, although its baseline expression level was extremely low (Supplementary Fig. S1J–S1L). Collectively, APC deficiency correlates with increased TDO2 expression in normal and malignant intestinal epithelium in humans and mice.

APC Deficiency Upregulates TDO2 Expression via TCF4

Increased Tdo2 mRNA levels upon APC deletion in isogenic cells (Fig. 2A and B; Supplementary Fig. S2A) prompted an examination of the human and mouse TDO2 gene promoter region for transcription factors using JASPAR and ECR Browser transcription factor binding profile databases. A conserved WNT pathway transcription factor binding element for TCF4/TCF7L2 was identified immediately upstream of the human and mouse TDO2 transcription start sites (Fig. 2C). Chromatin immunoprecipitation sequencing (ChIP-seq) confirmed TCF4 binding in the Tdo2 promoter of APC-KO but not APC-WT MC38 cells (Fig. 2D). In the human APC-null DLD-1 cell line, ChIP-PCR also documented TCF4 binding to the promoters of TDO2 and the classic WNT target genes AXIN2 and MYC but not the GAPDH promoter, which served as a negative control (Fig. 2E). Furthermore, a luciferase reporter driven by the human TDO2 promoter showed increased reporter activity upon the transduction of constitutively active β-catenin (CTNNB1 Δ90), which mimics WNT pathway activation (Fig. 2F). Conversely, dominant-negative TCF4 expression or TCF4 binding motif mutation abrogated reporter activity (Fig. 2G and H). Finally, TCF4 depletion or WNT inhibitor XAV-939 treatment, which destabilizes β-catenin, decreased TDO2 levels in multiple independent WNT-activated cells (Fig. 2I; Supplementary Fig. S2B–S2I). Thus, APC loss activates WNT–β-catenin, resulting in the TCF4-mediated upregulation of TDO2 gene transcription.

TDO2 Depletion Specifically Impairs the Growth and Survival of APC/WNT-Mutated Colorectal Cancer Cells

To assess TDO2 essentiality as a function of APC status, the biological impact of TDO2 depletion or pharmacologic inhibition was tested across multiple murine and human models. Using validated short hairpin RNAs (shRNA; Supplementary Fig. S3A), we discovered that TDO2 depletion had no impact on the colony formation of human APC-WT RKO cells yet impaired the colony formation of isogenic CRISPR/Cas9-generated APC-null RKO controls (Supplementary Fig. S3B and S3C). Similarly, multiple human APC/CTNNB1-mutant colorectal cancer lines (DLD-1, LS180, HT-29, and Caco-2; Supplementary Fig. S3D) showed markedly reduced colony formation upon TDO2 depletion (Supplementary Fig. S3B and S3C). Correspondingly, TDO2-specific inhibitor 680C91 (18) treatment impaired the growth and survival of APC-deficient but not APC-WT cancer cells, including primary CCD-841-CoN colon epithelial cells (Supplementary Fig. S3E and S3F). In murine cell models, both TDO2 depletion and pharmacologic inhibition impaired the growth and cell death of APC-null MC38 cells but not the parental APC-WT controls (Fig. 3AC; Supplementary Fig. S3G–S3I). Similarly, shRNA-mediated TDO2 depletion in cultured ApcMin/+ intestinal organoids induced cell death, and pharmacologic inhibition of TDO2 reduced the growth of APC-KO intestinal organoids but not APC-WT controls (Fig. 3D and E).

In tumor models, TDO2 depletion decreased the growth of orthotopic APC-null RKO tumors in immune-deficient NSG mice (Supplementary Fig. S4A). Similarly, TDO2 depletion decreased the growth of APC-null DLD-1 tumors, which was rescued by the enforced expression of a hairpin-resistant TDO2 open reading frame (ORF; Supplementary Fig. S4B). A pathologic analysis of these TDO2-depleted tumors revealed decreased cancer cell proliferation (Ki-67) and increased apoptosis (cleaved caspase-3), indicating that TDO2 drives these cancer cell–intrinsic hallmarks (Supplementary Fig. S4C and S4D). Similarly, using immune-competent mice, TDO2 depletion impaired orthotopic tumor growth, decreased cancer cell proliferation, and increased cancer cell apoptosis (Fig. 3F; Supplementary S4E and S4F), which resulted in prolonged overall survival specifically in murine APC-KO MC38 but not in APC-WT controls (Fig. 3G). In immune-deficient mice, APC-KO MC38 tumors produced similar survival curves to those in immune-competent mice but showed reduced survival benefit from induction of TDO2 depletion, consistent with cancer cell–intrinsic and immune-modulatory roles for TDO2 specifically in APC-null cancers (Supplementary Fig. S4G).

The recent failure of IDO inhibitors in colorectal cancer trials (19) prompted us to compare the impact of TDO2 and IDO inhibition in our model system. The recently developed validated TDO2 inhibitor PF06845102/EOS200809 (20) was administrated by oral gavage to mice bearing APC-WT or APC-KO MC38 orthotopic tumors. TDO2 inhibitor treatment improved the survival of mice bearing APC-KO MC38 tumors but not APC-WT controls (Fig. 3H). Histopathology showed that TDO2 inhibitor treatment decreased Ki-67 and increased cleaved caspase-3 signals specifically in the APC-KO MC38 tumors (Supplementary Fig. S4H). Consistent with the IDO inhibitor failures, the IDO inhibitor epacadostat did not exhibit antitumor activity in mice bearing either APC-WT or APC-KO MC38 colorectal cancer orthotopic tumors (Supplementary Fig. S4I). Finally, we confirmed the survival benefit of TDO2 inhibitor treatment in autochthonous established tumors arising in the iAP mouse model of colorectal cancer. Specifically, tumor-bearing iAP mice treated with the TDO2 inhibitor 3 weeks following 4-OHT injection into the colon wall showed significant survival benefits compared with vehicle-treated mice (Fig. 3I). Together, these data support the view that TDO2 supports tumor growth specifically in APC-null colorectal cancer.

The TDO2–Kyn–AhR Axis Supports APC-Deficient Cancer Cell Proliferation, Survival, and Tumorigenic Potential

As noted, TDO2 metabolizes Trp to produce Kyn, which in turn activates AhR to upregulate genes governing myriad cellular functions. Gene set enrichment analysis (GSEA) of isogenic APC-KO and APC-WT MC38 cell lines showed that the doxycycline (Dox) induction of inducible shTDO2 decreased signatures of tryptophan metabolism as well as xenobiotic metabolism—patterns consistent with the main functions of the AhR pathway (Supplementary Fig. S5A). Correspondingly, the expression of AhR and its target gene CYP1B1 correlated positively with TDO2 levels in the TCGA COAD data set (Supplementary Fig. S5B). Colorectal cancer tumors from iAP and iKAP also showed that AhR expression strongly tracks with nuclear β-catenin and Ki-67 (Supplementary Fig. S5C). Moreover, the APC–KP connection was verified in the APC-KO MC38 model system via ELISA, which documented elevated Kyn secretion relative to APC-WT controls (Supplementary Fig. S5D) and that TDO2 depletion in APC-KO cells and ApcMin/+ organoids reduced Kyn levels (Supplementary Fig. S5D and S5E). Finally, gene expression analysis showed upregulated AhR and its downstream genes in APC-KO MC38 cells compared with APC-WT MC38 cells, which was reversed upon TDO2 depletion in APC-KO MC38 cells and DLD-1 cell lines (Supplementary Fig. S5F and S5G).

To validate Kyn and AhR in mediating TDO2-regulated biology, we assayed the impact of Kyn treatment or AhR depletion in colony formation assays using the APC-KO MC38 ishTDO2 cell lines and APC-KO MC38 shAhR cell lines. In APC-KO MC38 cells, reduced colony formation upon induction of TDO2 depletion or pharmacologic inhibition (680C91) was partially rescued by Kyn treatment (Supplementary Fig. S5H–S5L). In the iKAP model system, Kyn treatment also decreased 680C91-induced cell death (Supplementary Fig. S5M). Finally, AhR depletion in APC-KO MC38 tumors resulted in increased survival with corresponding decreased proliferation (Ki-67) and survival (caspase-3) in the cancer cells (Supplementary Fig. S5N and S5O). Together, these findings are consistent with a key role for Kyn and AhR as mediators of TDO2 in APC-null cancer cell proliferation, survival, and tumorigenic potential.

TDO2 Promotes Cancer Cell Glycolysis and TAM Recruitment

To discern the cancer hallmarks regulated by TDO2, GSEA was conducted on APC-KO MC38 cell lines and derivative tumors following TDO2 depletion. Consistent with known cancer cell–intrinsic functions of the APC/WNT pathway (21), hypoxia and glycolysis pathways were upregulated in APC-KO cells (Fig. 4A; Supplementary Fig. S6A). Correspondingly, APC-KO MC38 cells exhibited higher sensitivity to the GLUT1 inhibitor STF-31 than APC-WT controls (Supplementary Fig. S6B) and showed increased glucose uptake and lactate secretion, which were reversed by TDO2 depletion (Supplementary Fig. S6C and S6D). Glycolytic flux Seahorse analysis showed that enforced TDO2 expression increased the key parameters of glycolytic flux, which are glycolysis, glycolytic capacity, glycolytic reserve, as well as nonglycolytic acidification, relative to the MC38 empty vector controls, reinforcing the role of TDO2 in promoting glycolysis (Supplementary Fig. S6E–S6G). Metabolite analysis of cell lysates and conditioned media (CM) from APC-KO MC38 cells showed decreased levels of glycolysis pathway–related metabolites upon TDO2 depletion (Supplementary Fig. S6H). To reinforce the link between TDO2 and the regulation of metabolic pathways, we examined multiple elements in the GCN2 and mTOR pathways in MC38 APC-WT and APC-KO cells containing an inducible shTDO2 construct. APC deletion increased the level of phosphorylated eIF2, and this increase was reversed upon TDO2 depletion in APC-null cells. In addition, TDO2 depletion decreased phosphorylated mTOR only in the APC-null cells (Supplementary Fig. S6I). Finally, RT-PCR analysis confirmed the upregulation of key glycolysis genes (SLC2A1, HK1/2, and PFKL), which were downregulated upon TDO2 or AhR depletion (Supplementary Fig. S6J and S6K). Together, these experimental data show that TDO2–AhR signaling plays a key role in promoting cancer cell glycolysis.

In addition to cancer cell–intrinsic processes, we observed that APC status (APC-KO vs. APC-WT MC38) or TDO2 depletion in APC-deficient cancer cells and tumors resulted in the prominent representation of immune signaling signatures such as TNFA signaling, inflammatory response, IL6_JAK_STAT, allograft rejection, and complement (Fig. 4A and B). These in silico observations prompted the immunoprofiling of orthotopic tumors generated from isogenic APC-KO and APC-WT MC38 cells with and without TDO2 depletion. Visualization of t-distributed stochastic neighbor embedding (viSNE) plots of CyTOF data showed that APC deficiency resulted in significantly increased macrophage abundance, which decreased upon TDO2 depletion (Fig. 4C). Polyps in ApcMin/+ mice also showed increased F4/80+ macrophage infiltration (Supplementary Fig. S7A). Quantification of CD11b+F4/80+ macrophages and CD11b+F4/80+ CD206hi M2-like macrophages in CD45+ population confirmed the enrichment of macrophages in APC-KO tumors and their reduction upon TDO2 depletion (Fig. 4D). IHC staining of F4/80 and CD163 in these tumors, as well as orthotopic tumors treated with the TDO2 inhibitor PF06845102/EOS200809, aligned with the aforementioned CyTOF data (Supplementary Fig. S7B and S7C). In contrast, epacadostat treatment did not increase the infiltration of total and M2-like macrophages. (Supplementary Fig. S7C). A comparative transcriptomic analysis of TAMs isolated from APC-WT and APC-KO MC38 tumors expressed higher levels of multiple classic M2-like markers (CD163, CCL22, and YM1) compared with APC-WT tumors (Supplementary Fig. S7D and S7E). Interestingly, IHC analysis of APC-WT and APC-KO MC38 tumors confirmed that (i) loss of APC results in decreased CD8+ cells, (ii) TDO2 inhibition increased the number of infiltrating CD8+ cells in APC-KO tumors relative to APC-WT controls, and (iii) IDO inhibition is unable to increase CD8+ cells in the APC-KO tumors (Supplementary Fig. S7F). Evaluation of the activity state of CD8+ T cells in APC-WT and APC-KO MC38 tumors by immune costaining of CD8 and activation marker granzyme B showed a higher number of activated T cells in APC-KO MC38 tumors compared with APC-WT tumors (Supplementary Fig. S7G).

To corroborate TDO2-mediated TME modulation, TCGA colorectal cancer data sets were examined for the expression of macrophage (total and M2) as well as regulatory T-cell and myeloid-derived suppressor cell markers, revealing strong positive correlations between the degree of WNT activation and TDO2 expression levels (Fig. 4E; Supplementary Fig. S7H and S7I). This WNT–macrophage correlation was further validated by human colorectal cancer TMA analyses that showed that cancer cells with a nuclear β-catenin signal exhibited higher CD163 expression in the TME (Fig. 4F and G). Together, these findings support the model that the activated WNT-driven upregulation of TDO2 expression, in turn, activates the AhR network, which functions to recruit immune-suppressive TAMs into the TME.

TDO2–AhR–CXCL5 Promotes Tumor Growth by Recruiting TAMs into the Colorectal Cancer TME

To identify WNT–TDO2–AhR-regulated factors that may recruit TAMs, we performed cytokine array profiling of CM from APC-KO MC38 ishTDO2 cells. Induction of TDO2 depletion reduced the secretion of classic macrophage cytokines, including G-CSF, GM-CSF, CXCL2 (Supplementary Fig. S8A), and other cytokines (see below). Correspondingly, transwell migration assays using bone marrow–derived macrophages (BMDM) showed that CM from APC-KO MC38 ishTDO2 cultures increased macrophage migration, which was nullified upon TDO2 depletion (Supplementary Fig. S8B and S8C).

Next, to more fully vet the most highly regulated cytokines in our system, we identified and qRT-PCR validated the top-ranked genes in the RNA sequencing (RNA-seq) data set and found that CXCL5, CXCL7 (PPBP), CSF3 (G-CSF), CXCR2, CXCL2, CXCL10, CCL2, and CXCL1 showed the most significant expression changes associated with APC deletion or TDO2 depletion (Fig. 5A). To further identify the target cytokines of TDO2, cell lines that express ORFs of the top three genes from RNA-seq data—CXCL5, CXCL7 (PPBP), and CSF3—were generated in APC-KO MC38 ishTDO2 cells and monitored for tumor growth to identify genes that rescue the impaired proliferation by TDO2 knockdown. Enforced expression of CXCL5, which showed the highest fold changes, was most active in rescuing the decreased tumor growth mediated by TDO2 depletion (Fig. 5B). Moreover, CyTOF analysis of CXCL5-overexpressing APC-KO tumors showed increased TAMs in the presence of shTDO2 (Fig. 5CE).

Migration assays showed the rescue of macrophage recruitment when CM from APC-KO TDO2-depleted MC38 cells was supplemented with CXCL5, whereas cotreatment with a CXCR1/2 inhibitor (SX-682), to which CXCL5 binds, abrogated the rescue by CXCL5 supplementation (Fig. 5F). In addition, AhR inhibitor treatment (CH223191) profoundly decreased CXCL5 expression in APC-KO MC38 cells (Supplementary Fig. S8D). Moreover, BMDMs cocultured with Kyn or CXCL5 showed increased M2 macrophage marker expression, supporting a role for the TDO2–AhR axis in promoting TAM polarization (Supplementary Fig. S8E).

IHC analysis of macrophage markers showed increased infiltration of macrophages in tumors with enforced CXCL5 expression (Supplementary Fig. S8F). To validate the roles of CXCL5 in promoting tumor growth in vivo, we coinjected the CT26 cell line and Raw264.7 macrophage cells that were pretreated with recombinant CXCL5 proteins and CXCL5-treated macrophages promoted the growth of CT26 significantly (Supplementary Fig. S8G). Finally, allograft mice with APC-KO MC38 cells showed increased survival upon TDO2 or macrophage depletion (Fig. 5G). CXCL5 overexpression in APC-KO MC38 cell lines significantly shortened the survival of mice, which was reversed by depleting macrophages (Fig. 5G). APC-KO MC38 cells treated with the anti-CXCL5 neutralizing antibody also showed prolonged survival (Supplementary Fig. S8H).

To further validate the relationship between the TDO2–AhR–CXCL5 axis and TAM abundance in human colorectal cancer, the TCGA colorectal cancer (COAD and READ) data set was clustered based on CXCL5 expression and was analyzed for immune populations. These analyses revealed that TAM abundance correlated positively with high CXCL5 expression (Supplementary Fig. S8I and S8J). In addition, CXCL5 expression correlated positively with increased tryptophan metabolism (TDO2 as the top pathway signature gene) and xenobiotic metabolism in TCGA colorectal cancer (Supplementary Fig. S8K–S8M). Together, these data establish that TDO2–AhR signaling upregulates CXCL5, which recruits TAMs to promote tumor growth; conversely, neutralization of the TDO2–AhR–CXCL5 pathway is a validated antitumor strategy in APC-null colorectal cancer.

In this study, we identified TDO2 as a synthetic essential effector in the maintenance of APC-deficient cancers. Increased TDO2 activates the KP to generate excessive Kyn, which activates the AhR network. Genetic and pharmacologic interventions established that this TDO2–Kyn–AhR axis increases APC-deficient colorectal cancer cell glycolysis, promotes cancer cell proliferation and survival, and upregulates CXCL5 to recruit TAMs into the TME (Fig. 6). In preclinical models, APC-deficient colorectal cancer exhibited hypersensitivity to TDO2 inhibition but not to the IDO1 inhibitor, providing a responder hypothesis for further testing of these immune-modulatory agents in colorectal cancer clinical trials. Importantly, iAP mice, engineered with conditional null alleles of Apc and Trp53, were induced to develop colorectal cancer; subsequent administration of a TDO2 inhibitor increased survival. Together with the correlative clinicopathologic profiles of human colorectal cancer, these experimental findings establish TDO2 as a potential therapeutic target for APC-null colorectal cancer.

Recent studies have revealed TDO2 overexpression in multiple cancer types and its role in facilitating tumorigenic signaling via KP (12, 22, 23). Another key KP enzyme, IDO1, is also highly expressed in various tumors and is known to suppress antitumor immunity. However, these functionally related KP enzymes, IDO1 and TDO2, appear to operate in nonredundant, context-specific settings and are differentially regulated. Specifically, AhR can regulate IDO1 but not TDO2 expression (24). In contrast, transcriptional regulatory mechanisms governing TDO2 but not IDO1 expression include hemes and glucocorticoid hormones (25), as well as the WNT transcription factor TCF4 specifically in APC-deficient colorectal cancer cells (this study).

With respect to tumor biology, the TDO2–Kyn–AhR axis regulates glycolysis as a cancer cell–intrinsic mechanism, a finding that aligns with previous work showing the AhR-mediated regulation of metabolism genes controlling lipid and cholesterol synthesis (26). In APC-deficient colorectal cancer, we further document that AhR also regulates glucose uptake and overall glycolytic flux by modulating multiple glycolysis genes including SLC2A1, HK1/2, and PFKL. Experimentally, TDO2 or AhR depletion resulted in the downregulation of these metabolic genes and anabolic processes in APC-deficient cancer cells. In addition to metabolism, KP and AhR signaling is also known to regulate immunity in both physiologic and pathologic conditions. In mice, AhR plays a critical role in the maintenance and function of innate T cells in the gastrointestinal tract (27). In stress conditions, mice with whole-body knockout of AhR exhibit impaired differentiation and function of T helper 17 cells and regulatory T cells to environmental toxins (28). In cancer, previous studies support both pro- and antitumorigenic roles for AhR. Whole-body knockout of AhR in ApcMin/+ mice causes increased cecal tumors (29), underscoring the highly context-specific actions of AhR in cancer. Further study is needed to define AhR actions in this setting, which may relate to nonligand-dependent roles of AhR such as degradation of β-catenin, effects of AhR on noncancer cell types, tissue-specific biology, and/or presence of additional oncogenic mutations. In different cancer types, regulatory mechanisms for AhR by its modulators such as ARNT, HSP90, XAP2, diverse agonists/antagonists, and direct immune modifying roles of AhR both in cancer cells and immune cells could further account for its contrasting impact on cancer. In contrast to the impact of AhR deletion in the ApcMin/+ model, multiple reinforcing lines of evidence establish the newly identified APC–TCF4–TDO2–AhR pathway in driving cancer cell–intrinsic and tumor microenvironmental processes to maintain APC-deficient colorectal cancer tumors.

With respect to translational relevance, our human colorectal cancer profiles mirrored our murine findings, showing a positive correlation between TAM abundance and TDO2 expression levels. Protumorigenic TAMs are known to support tumor progression and limit the efficacy of immunotherapy (30, 31). In glioblastoma, Kyn produced by glioma cells has been shown to recruit TAMs by binding to AhR and to promote CD8 T-cell dysfunction via expression of CD39 in TAMs (32). Our TDO2 and IDO1 inhibitor study also highlights the context-specific TAM biology by the TDO2 inhibitor in APC-deficient MC38 tumors. In a recent clinical trial in melanoma targeting IDO1 with epacadostat in combination with an anti–PD-1 antibody (ECHO-301), the basis for failure may relate to the upregulation of IDO1 expression provoked by immune-checkpoint blockade, BRAF inhibitors, or chemotherapy, resulting in inadequate target inhibition with the selected dosing of the IDO1 inhibitor (33). IDO1 inhibitor studies also showed that cancer cells upregulate ABC transporters, which might further reduce the availability of the IDO1 inhibitor in the TME (34). Given the frequent coexpression of IDO1 and TDO2 in melanoma (19), our study also encourages the assessment of ECHO-301 posttreatment specimens, which includes activation of WNT/β-catenin signaling, APC status, intratumoral Kyn concentration, and expression of TDO2 and CXCL5. Functional redundancies between IDO1 and TDO2 may also reveal a possible compensatory mechanism involving TDO2 upregulation, which would serve to sustain Trp metabolism and the KP–AhR pathway despite IDO1 inhibition. Encouragingly, however, we did not observe compensation by IDO1 upon TDO2 depletion, further underscoring the importance of understanding the common and distinct tumorigenic roles of IDO1 versus TDO2 as well as the genotypic context in which they operate in order to rationalize dual inhibition of IDO1/TDO2 and/or inhibition of downstream effectors such as CXCL5 (33). In conclusion, the identification of TDO2 as a synthetic essential effector of APC deficiency in colorectal cancer may serve as a promising precision treatment for this intractable cancer.

Mice

Mice were grouped by five animals in large plastic cages and were maintained under pathogen-free conditions. All animal experiments were performed with the approval of MD Anderson Cancer Center's Institutional Animal Care and Use Committee. ApcMin/+, NSG (NOD.Cg-PrkdcscidIl2rgtm1Wjl/SzJ, RRID:IMSR_JAX:005557), and C57BL/6J (RRID:IMSR_JAX:000664) mice were purchased from The Jackson Laboratory (stock #: 005557, 000664, and 000651). Colorectal orthotopic xenograft tumor models were established following a previously published protocol (35). After orthotopic injection of cells, mice that exhibited successful tumor formation were randomized before starting Dox, antibody, or inhibitor treatment for each cell line.

iAP mice were established as described by Boutin and colleagues (17). Briefly, the tamoxifen-inducible Villin-Cre-ERT2 allele was crossed with the Apc Lox allele and the Trp53 Lox allele and backcrossed to C57BL/6. Cre expression driven by the Villin promoter was detected throughout the gastrointestinal tract. To limit Cre activity to the colon, we used the tamoxifen-inducible Villin-Cre-ERT2 and delivered tamoxifen directly to the colon by injecting 4-OTH into the distal colon. The sample size was determined based on previous similar experiments performed in our lab.

Cell Culture

The colorectal cancer cell line MC38 (RRID:CVCL_B288) and its isogenic cells, as well as BMDM and HEK 293T (RRID:CVCL_0063) cells, were cultured in Dulbecco's Modified Eagle Medium (DMEM). CCD-841-CoN (ATCC; cat. #CRL-1790, RRID:CVCL_2871), RKO (RRID:CVCL_0504), HT-29 (RRID:CVCL_0320), Caco-2 (RRID:CVCL_0025), and LS180 (RRID:CVCL_0397) cells were cultured in Eagle's Minimum Essential Medium. HT-29 cells were cultured in McCoy's 5A medium. DLD-1 (RRID:CVCL_0248), CT26 (RRID:CVCL_7256), and Raw264.7 macrophage cell lines (RRID:CVCL_0493) were cultured in RPMI 1640 medium (RPMI). All cell lines were cultured in the indicated medium containing 10% Tet System Approved FBS (Clontech) and 100 U/mL ampicillin/penicillin. All human cell lines were validated through fingerprinting by the MD Anderson Cell Line Core Facility. All cells were confirmed to be Mycoplasma free and were maintained at 37°C and 5% CO2. BMDMs from C57BL/6 mice (RRID:IMSR_JAX:000664) were cultured as previously described (36). CM were collected from treated or untreated cells as indicated after culturing for 24 hours in the FBS-free culture medium. Inducible shTDO2 MC38 cell lines were treated with 1 μmol/L of Dox (Sigma-Aldrich; cat. #D9891) for indicated periods to induce TDO2 knockdown. Inhibitors and supplements used included CH223191 (Sigma-Aldrich; cat. #C8124, CAS: 301326-22-7), XAV-939 (Selleck Chemicals; cat. #S1180, CAS: 284028-89-3), recombinant mouse CXCL5 (LIX; R&D Systems; cat. #433-MC-025), and L-Kynurenine, ≥98% (HPLC; Sigma-Aldrich; cat. #K8625).

CRISPR/Cas9 Transfection

sgRNA plasmids targeting the human APC gene (cat. #sc-400374) were purchased from Santa Cruz Biotechnology. For the mouse Apc gene, an sgRNA target sequence of TTGAGCGTAGTTTCACTCCG was cloned into pCas-Guide-EF1a-GFP plasmids (Origene Technologies, Inc.; cat. #GE100018). Human RKO and mouse MC38 cells were maintained in 6-well plates to 70% to 80% confluency in culture media supplemented with 10% heat-inactivated FBS and 100 U/mL ampicillin/penicillin. The plasmids with sgRNA were transiently transfected into cells using Lipofectamine 2,000 according to the manufacturer's protocol. Cells were harvested 72 hours later, and GFP+ cells were sorted into each well of a 96-well plate as single cells by flow cytometry. At day 10 after cell sorting, the grown cell colonies were expanded in 24-well plates. Knockout of the APC gene in each colony was confirmed by RT-PCR and Western blot for APC and β-catenin.

Mouse Colon Organoid Culture and Genome Engineering

To isolate colonic crypts for organoid culture, a 2-cm piece of the distal large intestine was incubated in PBS containing 5 mmol/L EDTA and 0.2% FBS at 4°C for 45 minutes on a shaker. Incubated colon pieces were shaken vigorously to release crypts. Crypts were washed and spun down sequentially at 300 × g, 200 × g, and 100 × g to enrich for intact crypts. Crypts were resuspended in Matrigel and plated in 24-well plates containing 50 μL Matrigel per well. Organoid culture medium (500 μL) containing Wnt3a, R-spondin, Noggin, and EGF was added and changed every 2 days.

Knockout of APC was performed via the transient transfection of a plasmid expressing Cas9 and an sgRNA targeting APC (APC sgRNA-LentiCRISPRv2; sgRNA sequence: APC-G0-1—CGCTTGTCTAGATAAGCACG). APC-KO organoids were selected by the removal of Wnt and R-spondin from the media.

Mouse ApcMin/+ Organoids

Intestinal polyps from an 18-week-old, male ApcMin/+ mouse were harvested, and the cut tissue was treated with a complete chelating solution containing 30 mmol/L EDTA for 30 minutes at 4°C. The tissue pieces were then pipetted gently to dissociate the crypts. These crypts were then seeded in Matrigel (Corning; cat. #47743-722) in the presence of high WNT organoid media in the presence of the ROCK inhibitor Y-27632 (STEMCELL Technologies Inc.; cat. #72302) for 7 to 10 days.

Human Samples

Human colorectal cancer TMA slides were obtained from the Department of Pathology at the University of Texas MD Anderson Cancer Center. Studies related to human specimens were approved by the MD Anderson Institutional Review Board under protocol Lab09-0373.

Mutual Exclusivity Analysis

For the analysis of mutual exclusiveness for APC in colorectal cancer, genetic alteration data of 220 TCGA colorectal cancer samples with copy-number alterations and sequencing data were downloaded from cBioPortal (RRID:SCR_014555); the gene expression data set was downloaded from the Broad GDAC website (http://gdac.broadinstitute.org/runs/stddata__2016_01_28/data/COAD/20160128/). The detailed method for estimating mutation exclusivity was previously described (8). Briefly, the rank score (odds ratio score) was calculated to indicate mutual exclusiveness between gene A and gene B deletion. The mean values of gene B expression in all 220 samples and that in gene A–deleted samples were calculated and analyzed with a Student t test. For APC mutations in colorectal cancer data sets, only deletion and mutations with known significance (annotated by OncoKB, RRID:SCR_014782) cases were considered. The list of mutually exclusive genes to APC is given in Supplementary Table S1.

TCGA Data Computational Analysis

For the analysis of human colorectal cancer and BRCA data, we downloaded the gene expression and copy-number data of TCGA data sets or other available data sets from cBioPortal (RRID:SCR_014555). Correlation analysis of TDO2, AhR, and CYP1B1 expression in colorectal cancer was performed with the R2 platform (https://r2.amc.nl/).

Gene-Stable shRNA/siRNA Knockdown and Inducible shRNA Knockdown

Mission shRNA hairpins targeting mouse TDO, AhR, and TCF4 were purchased from Santa Cruz; GIPZ shRNA hairpins targeting human TDO were purchased from Horizon Discovery. For inducible TDO2 knockdown, SMARTvector Inducible Lentiviral shRNA for mouse TDO2 was purchased from Horizon Discovery. The sequences that reduced mRNA and/or protein levels by >70% were chosen. For in vivo bioluminescence imaging, luciferase vector EF1-RFP-T2A-Luciferase (System Biosciences; cat. #BLIV502MC-1) and D-Luciferin (PerkinElmer; cat #NC0921725) were used. Recombinant lentiviral particles were produced by the transient transfection of plasmids into HEK 293T cells (RRID:CVCL_0063). In brief, 8 μg of shRNA plasmid, 4  μg of psPAX2 plasmid (RRID:Addgene_12260), and 2  μg of pMD2.G plasmid (RRID:Addgene_12259) were transfected using Lipofectamine 3000 into 293T cells plated in 100-mm dishes. Viral supernatant was collected 48 and 72 hours after transfection and filtered. Cells were infected twice in 48 hours with viral supernatant containing 8  μg/mL polybrene and then selected using 2  μg/mL puromycin. The expression of TDO2, AhR, and TCF4 was measured by RT-qPCR. The following shRNA sequences were used:

  • Human shTDO2 #3: NM_005651: 5′-AATCTGATTCATCACTGCT-3′

  • Human shTDO2 #6: NM_005651: 5′-AAATCTACAAATACCTTGT-3′

  • Mouse shTDO2 #2: NM_019911: 5′-CGGCCAAAGATGAATCCGATCATTCTCGAGAATGATCGGATTCATCTTTGGTTTTTG-3′

  • Mouse shTDO2 #4: NM_019911: 5′-GGGCGCAAGAACTTCAGAGTGAACTCGAGTTCACTCTGAAGTTCTTGCGCTTTTTG-3′

  • Mouse ishTDO2 #3: NM_019911.2: 5′-GGATTTAATTTCTGGGGAA-3′

  • Mouse shAhR #1: NM_013464: 5′-CGGCATCGACATAACGGACGAAATCTCGAGATTTCGTCCGTTATGTCGATGTTTTTG-3′

  • Mouse shAhR #2: NM_013464: 5′- GTACCGGGTCAAGCCTGTTAGCTATATTCTCGAGAATATAGCTAACAGGCTTGACTTTTTTG-3′

  • Human shTCF4 #1: NM_030756: 5′-CCGGCCTTTCACTTCCT CCGATTACCTCGAGGTAATCGGAGGAAGTGA AAGGTTTTTG-3′

  • Human shTCF4 #2: NM_030756: 5′-CCGGAGAGAAGAGCAAGCGAAATACCTCGAGGTATTTCGCTTGCTCTTCTCTTTTTTG-3′

For siRNA experiments, the Lipofectamine RNAiMAX Transfection Reagent (Thermo Fisher; cat. #13778030) was used, and the assay was performed following the manufacturer's protocol. Transfected cells were maintained for three days, and knockdown efficiency for TCF4 was measured by western blotting. The following siRNAs (Sigma-Aldrich) were used:

  • Human siTCF4 #1: NM_030756: SASI_Hs01_00197690

  • Human siTCF4 #2: NM_030756: SASI_Hs01_00197691

  • Human siTCF4 #3: NM_030756: SASI_Hs01_00197692

  • Mouse siTCF4 #1: NM_009333: SASI_Mm01_00142189

    Mouse siTCF4 #2: NM_009333: SASI_Mm02_00315891

  • Mouse siTCF4 #3: NM_009333: SASI_Mm01_00142190

Western Blot

Cell lysates were prepared with RIPA lysis buffer (Roche) with Halt Protease and Phosphatase Inhibitor Single-Use Cocktail (Thermo Fisher Scientific; cat. #78442). Immunoblotting was performed following the standard protocol. Antibodies were purchased from the indicated companies: β-actin (Sigma-Aldrich; cat. #A1978, RRID:AB_476692), APC (Santa Cruz Biotechnology; cat. #sc-896, RRID:AB_2057493), tubulin (Sigma-Aldrich; cat. #T9026; RRID:AB_477593), vinculin (Millipore; cat. #05-386, RRID:AB_309711), TDO2 (Origene; cat. #TA504730, RRID:AB_2622554), β-catenin (Cell Signaling Technology; cat. #9587, RRID:AB_10695312), TCF4 (Santa Cruz Biotechnology; cat. #sc-166699, RRID:AB_2199823), phospho-eIF2 (Cell Signaling Technology; cat. #9721, RRID:AB_330951), eIF2 (Cell Signaling Technology; cat. #9722, RRID:AB_2230924), ATF4 (Cell Signaling Technology; cat. #11815, RRID:AB_2616025), phospho-mTOR (Cell Signaling Technology; cat. #2971, RRID:AB_330970), mTOR (Cell Signaling Technology; cat. #2972, RRID:AB_330978), GCN2 (Santa Cruz Biotechnology; cat. #sc-374609, RRID:AB_10986130), and cleaved caspase-3 (Cell Signaling Technology; cat. #9661, RRID:AB_2341188).

ORF and Hairpin-Resistant ORF Expression

To construct the hairpin-resistant hTDO2 ORF expression vector to shTDO2 #3, site-directed mutagenesis was performed on the human TDO2 ORF gene in pcDNA3.1+/C-(k)DYK vector (GenScript; cat. #OHu09674D). Nucleotide mutation was targeted for (i) 1,272 T to C, (ii) 1,275 T to C, (iii) 1,278 A to G, and (iv) 1,281 A to G, and no amino acid was altered. The mutated TDO2 ORF gene insert was subcloned into PS100102 (pLenti-C-mGFP-P2A-BSD Tagged Cloning Vector; Origene; cat. #PS10094).

For mutagenesis, the following primers were used:

  • F: 5′-CCTACTTCAGCAGCGAC GAGTCGGATTAAAATCG-3′

  • R: 5′-CGATTTTAATCCGACTCG TCGCTGCTGAAGTAGG-3′

Lentiviral ORF-expressing vectors for blank, CXCL5, CXCL7, and CSF3 were purchased from ABM (cat. #LV587, LV407122, LV395200, and LV455866).

For Seahorse glycolytic flux assay in TDO2-overexpressing cells, mouse Tdo2 ORF (NM_019911)–expressing vector was purchased from GenScript (cat. #OMu17612D).

Luciferase Assay

HEK 293T cells were seeded in 24-well plates and transfected with luciferase reporter vectors of pGL3-Basic (Promega; cat. #E1751), pGL3-hTDO2 promoter, or pGL3-hTDO2 promoter with a mutated TCF4 binding site with pRL Renilla Luciferase Control Reporter Vector (Promega; cat. #E2261) and pLV-beta-catenin ΔN90 (RRID:Addgene_36985) using Lipofectamine 2,000 reagent (Thermo Fisher Scientific; cat. #11668019). The pcDNA/Myc DeltaN TCF4 expression vector (RRID:Addgene_16513) was transfected to express dominant-negative TCF4. Luciferase activity was measured with Dual-Luciferase reagent (Promega; product code E1910) according to the manufacturer's instructions.

Glycolytic Flux Measurement

The Agilent Seahorse XF Glycolysis Stress Test Kit (Agilent; cat. #103020–100) was used to measure glycolytic flux according to the manufacturer's instructions. In brief, MC38 cells expressing Blank-ORF or Tdo2-ORF 96-well XF were plated at 2 × 104 cells per well in a 96-well Seahorse plate in DMEM and were incubated overnight. The next day, the culture medium was removed and changed with the Seahorse XF DMEM assay medium (Agilent; cat. #103680-100) containing 2 mmol/L L-glutamine. Cells were incubated in the assay medium for 1 hour, and sensor cartridges incubated in calibrant solution (Agilent; cat. #100840–000) were loaded with glucose, oligomycin, and 2-DG (final working concentrations: 1 mmol/L glucose, 1 μmol/L oligomycin, and 50 mmol/L 2-DG). Glycolytic flux was measured using an Agilent Seahorse XFe Analyzer (Agilent Technologies). Raw data were analyzed with Wave software (Agilent Technologies).

Cytokine Array

For cytokine array, colorectal cancer orthotopic tumors established with ishTDO2 APC-WT and APC-KO MC38 cells were incubated in RIPA buffer with protease/phosphatase inhibitor cocktail and homogenized. A cytokine array was performed with the mCytokine Array Kit, Panel A (R&D Systems; cat. #ARY006) following the manufacturer's protocol. For the phospho-RTK array, ishTDO2 APC-WT and APC-KO MC38 cells were treated with Dox for 48 hours, and the lysates were used for the array.

IHC and Immunofluorescence

IHC was performed using a standard protocol we previously described (37). Antibodies were TDO2 (Abnova Corporation; cat. #H00006999-B01P, RRID:AB_1138993), AhR (Santa Cruz Biotechnology; cat. #sc-133088, RRID:AB_2273721), β-catenin (Cell Signaling Technology; cat. #9587, RRID:AB_10695312), Ki-67 (Thermo Fisher Scientific; cat. #MA1-90584, RRID:AB_2314700), cleaved caspase-3 (Cell Signaling Technology; cat. # 9661, RRID:AB_2341188), F4/80 (Cell Signaling Technology; cat. #70076, RRID:AB_2799771), CD163 (Abcam; cat. #ab182422, RRID:AB_2753196), and CD206 (BioLegend; cat. #141705, RRID:AB_10896421). For immuno­fluorescence staining, CD8 (Cell Signaling Technology; cat. #98941, RRID:AB_2756376) and Granzyme B (Thermo Fisher Scientific; cat. #MA1-80734, RRID:AB_931084) antibodies were used. For nuclei staining, DAPI (Thermo Fisher Scientific; cat. #D1306, RRID:AB_2629482) was used. The human and mouse tumor tissue sections were reviewed and scored.

Migration Assay

Macrophages (1  ×  104 for Raw264.7 and BMDM) were suspended in a serum-free culture medium and seeded into 24-well Transwell inserts (5.0 μm, Corning; cat. #CLS3422). Medium with indicated factors or CM were added to the remaining receiver wells. The CXCR1/2 inhibitor SX-682 was obtained from Syntrix Biosystems. After 24 hours, the migrated macrophages were fixed, stained with crystal violet (0.05%, Sigma), and counted with ImageJ (RRID:SCR_003070).

Colony Formation Assay

Colorectal cancer cell proliferation in vitro was assayed through colony formation. Cells (1 × 103) were seeded in 6-well plates and cultured for 5 to 7 days, and then fixed and stained with 0.5% crystal violet in 25% methanol for 1 hour. These experiments were performed in triplicate.

In Vivo TDO2 Inhibitor Drug Treatment

For C57BL/6J mice with APC-WT and APC-KO MC38 tumors, the TDO2 inhibitor (200 mg/kg, synthesized in-house) was dissolved in 0.5% HPMC before each injection and administered orally twice daily by oral gavage. Epacadostat (MedChemExpress; cat. #HY-15689) was dissolved in 10% DMSO and further diluted in 90% corn oil and administered twice daily at 100 mg/kg by oral gavage.

For iAP mice, APC and TP53 deletions were induced by injecting 4-OHT into the distal colon. Three weeks after induction, 0.5% HPMC or TDO2 inhibitor (100 mg/kg) was administered daily by oral gavage to randomized mice.

In Vivo Neutralizing Antibody Treatment and Macrophage Depletion

For the CXCL5 neutralizing experiment, Rat IgG2b Isotype Control (cat. #BE0090) was purchased from Bio X Cell, and anti-Mouse CXCL5 (Clone 61905) neutralizing antibody was purchased from Leinco Technologies (cat. #C1414). For the macrophage depletion study, the Standard Macrophage Depletion Kit (Clodrosome + Encapsome) (Encapsula NanoSciences; cat. #CLD-8901) was used following the manufacturer's protocol.

Mass Cytometry (CyTOF)

CyTOF analysis was performed as described previously (37). Briefly, tumors were digested, and single cells blocked with FcR were incubated with a surface antibody. Cells were then incubated with Cell-ID Cisplatin (Fluidigm; cat. #201064) and permeabilized for FOXP3 intracellular staining. For nuclei staining, cells were incubated with Cell-ID Intercalator-Ir (Fluidigm; cat. #201192A) during fixing. Samples were analyzed with a CyTOF instrument (Fluidigm) in the Flow Cytometry and Cellular Imaging Core Facility at MD Anderson Cancer Center. Cell numbers and percentages of each cell population were analyzed with FlowJo (Tree Star, RRID:SCR_008520) and GraphPad Prism 6 software (RRID:SCR_002798). CyTOF data were visualized using a dimensionality reduction method, viSNE (38), which was implemented using the Cytobank (RRID:SCR_014043; ref. 39).

Kyn, 2-DG Uptake, and Lactate Secretion Measurement

Kyn concentration was measured following the manufacturer's protocol for the Kyn ELISA measurement kit (ImmuSmol; cat. #BA-E-2200). The 2-DG uptake assay was performed according to the manufacturer's protocol for the 2-Deoxyglucose Uptake Measurement Kit (Cosmo Bio; cat. #CSR-OKP-PMG-K01TE). For secreted lactate measurement, the Lactate Colorimetric/Fluorometric Assay Kit (BioVision; cat. #10186–852) was used, and the assay was performed following the manufacturer's protocol.

LC-MS/MS–Based Targeted Metabolomics

Media from cultured cells were harvested and quickly placed into dry ice or a −80°C freezer. Cells were washed twice with ice-cold PBS and snap-frozen using liquid nitrogen. Frozen cells were scraped into 1 mL of −70°C-cooled 80% methanol and quickly stored at −80°C. LC-MS/MS analyses were performed on an AB SCIEX QTRAP 6500 LC-MS/MS system by the Karmanos Cancer Institute Pharmacology Core. Analyst 1.6 software was used for system control and data acquisition, and MultiQuant 3.0 software was used for data processing and quantitation. For statistical analysis, Metaboanalyst (RRID:SCR_015539) was used.

ChIP-seq and ChIP-PCR

ChIP was performed as we described recently (9). Briefly, chromatin from PFA-fixed cells was cross-linked with 1% PFA, and then reactions were quenched using 0.125 M glycine. Cells were lysed with ChIP lysis buffer [10 mmol/L Tris-HCl (pH 8.0), 140 mmol/L NaCl, 1 mmol/L EDTA (pH 8.0), 1% Triton X-100, 0.2% SDS, and 0.1% deoxycholic acid] for 30 minutes on ice. Chromatin fragmentation was performed using a Diagenode BioruptorPico sonicator (30 seconds on and 30 seconds off, 45 cycles) and incubated with the appropriate mixture of antibody and Dynabeads (Thermo Fisher Scientific; cat. #10003D) overnight. Immune complexes were washed with RIPA buffer (three times), once with RIPA-500 (RIPA with 500 mmol/L NaCl), and once with LiCl wash buffer [10 mmol/L Tris-HCl (pH 8.0), 1 mmol/L EDTA (pH 8.0), 250 mmol/L LiCl, 0.5% NP-40, and 0.5% deoxycholic acid]. Elution and reverse-crosslinking were performed in direct elution buffer [10 mmol/L Tris-Cl (pH 8.0), 5 mmol/L EDTA, 300 mmol/L NaCl, 0.5% SDS] containing proteinase K (20 mg/mL) at 65°C overnight. Eluted DNA was purified using AMPure beads (Beckman Coulter; cat. #A63880) and then used to generate libraries using the NEBNext Ultra DNA library kit (New England BioLabs Inc.; cat. #E7370) or to perform qPCR. Sequencing was performed using an Illumina HiSeq 2500 instrument to generate the data set. ChIP-PCR primers used were as follows:

mRNA Expression Analysis, Microarray, and RNA-seq

Cells were pelleted, and RNA was isolated with the RNeasy Mini Kit (Qiagen; cat. #74104). RNA was reverse-transcribed into cDNA following the SuperScript III First-Strand Synthesis SuperMix (Invitrogen; cat. #18080400). qRT-PCR was performed using the SYBR Green PCR Master Mix (Thermo Fisher Scientific) in a 7500 Fast Real-Time PCR instrument (Applied Biosystems). qRT-PCR primers used were as follows: hAPC; F: TCTTGGCGAG CAGATGTAAA- R: TCCACAAAGTTCCACATGC-, hTDO2; F: GGGAACTACCTGCATTTGGA- R: GTGCATCCGAGAAACAACCT-, hAhR; F: ATTGTGCCGA GTCCCATATC- R: AAGCAGGCGTGCATTAGACT-, hCyp1A1; F: CTTGGACCTCTTTGGAGCT- R: GACCTGCCAATCACTGTG-, hCyp1B1; F: GACGCCTTTATCCTCTCTGCG- R: ACGACCTGATCCAATTCTGCC-, hGAPDH; F: GTCTCCTCTGACTTCAACAGCG- R: ACCACCCTGTT GCTGTAGCCAA-, mAPC; F: CTTGTGGCCCAGTTAAAATCTGA- R: CGCTTTTGAGGGTTGATTCCT-, mTDO2; F: ATGAGTGGGTGCCCGTTTG- R: GGCTCTGTTTACACCAGTTTGAG-, mAhR; F: AGCCGGTGCAGAAAACAGTAA- R: AGGCGGTCTAACTCTGTGTTC-, mCyp1A1; F: GACACAGTGATTGGCAGAG- R: GAAGGTCTCCAGAATGAAGG-, mCyp1B1; F: CACCAGCCTTAGTGCAGACAG- R: GAGGACCACGGTTTCCGTTG-, mHK1; F: CGGAATGGGGAGCCTTTGG- R: GCCTTCCTTATCCGTTTCAATGG-, mHK2; F: TGATCGCCTGCTTATTCACGG- R: AACCGCCTAGAAATCTCCAGA-, mSLC2A1; F: GCAGTTCGGCTATAACACTGG- R: GCGGTGGTTCCATGTTTGATTG-, mPFKL; F: GGAGGCGAGAACATCAAGCC- R: CGGCCTTCCCTCGTAGTGA-, mLDHA; F: GCTCCCCAGAACAAGATTACAG- R: TCGCCCTTGAGTTTGTCTTC-, mALDOA; F: CGTGTGAATCC CTGCATTGG- R: CAGCCCCTGGGTAGTTGTC-, mCXCL5; F: TCCAGCTCGCCATTCATGC- R: TTGCGGCTATGACTGAGGAAG-, mCXCL7; F: CTCAGACCTACATCGTCCTGC- R: GTGGCTATCACT TCCACATCAG-, mCSF3; F: ATGGCTCAACTTTCTGCCCAG- R: CTGACAGTGACCAGGGGAAC-, mCXCR2; F: ATGCCCTCTATTCTGCCAGAT- R: GTGCTCCGGTTGTATAAGATGAC-, mCSF1; F: ATGAGCAGGAGTATTGCCAAGG- R: TCCATTCCCAATCATGTGGCTA-, mCD206; F: CTCTGTTCAGCTATTGGACGC- R: CGGAATTTCTGG GATTCAGCTTC-, mYM1; F: CAGGTCTGGCAATTCTTCTGAA- R: GTCTTGCTCATGTGTGTAAGTGA-, mYM2; F: TCCACTTTGAACCACATTCCAA- R: CCAGCACTAACAGTAGGGTCA-, mArg1; F: CTCCAAGCCAAAGTCCTTAGAG- R: AGGAGCTGTCATTAGGGACATC-, miNOS; F: GTTCTCAGCCCAACAATACAAGA- R: GTGGACGGGT CGATGTCAC-, mbeta-actin; F: AAATCTGGCACCACACCTTC- R: GGGGTGTTGAAGGTCTCAAA-, mGAPDH; F: GCCTTCCGTGTTCCTACCC- R: CAGTGGGCCCTCAGATGC-

The expression of each gene was normalized to that of GAPDH or Actin. For microarray, tumors established with ishTDO2 APC-WT, and APC-KO MC38 cells were harvested (biological triplicates for control and APC-KO MC38 tumors). RNAs were isolated using TRIzol (Invitrogen; cat. #15596-026) and further purified with the RNeasy Mini Kit. Samples were analyzed at the MD Anderson Microarray Core facility using the GeneChip Mouse Clariom D array (Affymetrix) to generate the data set. Genes that were differentially expressed between control and APC-depleted MC38 cells were subjected to GSEA. For RNA-seq, RNAs were isolated from ishTDO2 APC-WT and APC-KO MC38 cells with and without Dox treatment (biological triplicates per group) using the RNeasy Mini Kit. The Illumina TrueSeq ChIP library was used for Illumina Next Seq 500 sequencing.

Tumor-Infiltrated Macrophage RNA Isolation and Sequencing

Tumor-infiltrated macrophage isolation and sorting from MC38 tumors were performed by following a published protocol (40) with slight modifications. Briefly, APC-WT and APC-KO MC38 tumors (two per group) implanted in C57BL/6J mice were digested with Liberase DL (Roche) and Liberase TL. After red blood cell lysis, cells were blocked with the CD16/CD32 blocking antibody (BD Biosciences; cat. #553142, RRID:AB_394657) for 30 minutes and stained with antibodies CD45 APC 30-F11 (BioLegend; cat. #103111, RRID:AB_312976), CD11b eFluor 605 M1/70 (Thermo Fisher Scientific; cat. #69-0112-82, RRID:AB_2637406), F4/80 PerCp-Cy5.5 BM8 (BioLegend; cat. #123127, RRID:AB_893496), CD206/MMR 169tm C068C2 (Fluidgm; cat. #3169021B, RRID:AB_2832249), and SYTOX Green (Thermo Fisher Scientific; cat. #R37168). Stained cells were sorted as CD45+, CD11b+, F4/80+, and CD206+. Total RNA was isolated from sorted cells and was sequenced by ultra-low-input RNA-seq (Illumina Nextera XT) with six replicates of each sample. Comparisons between groups were hampered by the low input of total RNA isolated from tumor macrophages manifesting as inconsistent raw/normalized read counts for housekeeping genes in the different groups. To enable comparisons between APC-WT and APC-KO tumors, we renormalized the read counts of multiple M2 macrophage markers with housekeeping gene read counts (GAPDH and ACTB; ref. 41).

Quantification and Statistical Analysis

The analysis of TAM IHC staining for correlation with nuclear β-catenin, TDO2, and CD163 was performed using the chi-squared test. Mouse survival analysis was performed using the log-rank (Mantel–Cox) test (GraphPad Prism 9, RRID:SCR_002798). All other statistical analyses were performed with the Student t test and represented as mean ± SD. The P values were designated as *, P < 0.05; **, P < 0.01; and ***, P < 0.001; n.s., nonsignificant (P > 0.05).

Data Availability

RNA-seq, ChIP-seq, and microarray data have been deposited in the NCBI Gene Expression Omnibus with the accession numbers GSE200910, GSE201414, and GSE201415, respectively. Additional data, reagents, and materials generated in this study can be obtained from the corresponding authors upon request.

R. Lee reports grants from the NIH during the conduct of the study. S. Jiang reports grants from the NIH during the conduct of the study. R.A. DePinho reports grants from the NIH/NCI and MD Anderson during the conduct of the study; personal fees and other support from Tvardi Therapeutics, Asylia Therapeutics, Stellanova Therapeutics, and Nirogy Therapeutics and other support from Sporos Bioventures outside the submitted work; and a patent for methods and compositions for treatment of APC-deficient cancer pending. No disclosures were reported by the other authors.

R. Lee: Conceptualization, formal analysis, investigation, methodology, writing–original draft. J. Li: Formal analysis. J. Li: Formal analysis. C.-J. Wu: Formal analysis. S. Jiang: Investigation. W.-H. Hsu: Investigation. D. Chakravarti: Investigation. P. Chen: Investigation. K.A. LaBella: Investigation. J. Li: Formal analysis. D.J. Spring: Writing–review and editing. D. Zhao: Formal analysis. Y.A. Wang: Conceptualization, supervision, methodology. R.A. DePinho: Conceptualization, formal analysis, supervision, funding acquisition, methodology, writing–original draft.

This study is dedicated to the memory of Alvaro DePinho who succumbed to colorectal cancer and continues to serve as an inspiration to R.A. DePinho. The authors thank Dr. Scott Kopetz, Dr. Guillermina Lozano, and Dr. Trevor Hart for scientific discussion and advice; Dr. Dipen Maru for human colorectal cancer TMA samples; the Institute for Applied Cancer Science (IACS) for inhibitor synthesis and distribution; and Dr. Jing Li for metabolomics analysis. The results shown here are in whole or part based upon data generated by the TCGA Research Network (https://www.cancer.gov/tcga). This work was supported by the MD Anderson SPORE in Gastrointestinal Cancer (R.A. DePinho), NIH/NCI R01 CA231360 (R.A. DePinho), and NIH/NCI 1R01 CA231349 (Y.A. Wang). R. Lee was supported by an NIH T32 Training Grant in Cancer Biology (T32 CA186892; R. Kalluri). Jiexi Li and W.-H. Hsu were supported by the Cancer Prevention & Research Institute of Texas (CPRIT) Research Training Program (RP170067). K.A. LaBella was supported by a training fellowship from the UT Health Science Center at Houston Center for Clinical and Translational Sciences TL1 Program (TL1 TR003169). D. Zhao was supported by the CPRIT Recruitment of First-Time Tenure-Track Faculty Award RR190021 (CPRIT Scholar in Cancer Research). The Flow Cytometry and Cellular Imaging Core at the MD Anderson Cancer Center is partially funded by NCI Cancer Center Support Grant P30 CA16672. The metabolomic profiling was done in the Pharmacology and Metabolomics Core at Karmanos Cancer Institute, which is supported, in part, by the United States Public Health Service Cancer Center Support Grant P30 CA022453.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1.
Vogelstein
B
,
Papadopoulos
N
,
Velculescu
VE
,
Zhou
S
,
Diaz
LA
Jr
,
Kinzler
KW
.
Cancer genome landscapes
.
Science
2013
;
339
:
1546
58
.
2.
Su
LK
,
Kinzler
KW
,
Vogelstein
B
,
Preisinger
AC
,
Moser
AR
,
Luongo
C
, et al
.
Multiple intestinal neoplasia caused by a mutation in the murine homolog of the APC gene
.
Science
1992
;
256
:
668
70
.
3.
Dow
LE
,
O'Rourke
KP
,
Simon
J
,
Tschaharganeh
DF
,
van Es
JH
,
Clevers
H
, et al
.
Apc restoration promotes cellular differentiation and reestablishes crypt homeostasis in colorectal cancer
.
Cell
2015
;
161
:
1539
52
.
4.
Fang
DC
,
Luo
YH
,
Yang
SM
,
Li
XA
,
Ling
XL
.,
Fang L. Mutation analysis of APC gene in gastric cancer with microsatellite instability
.
World J Gastroenterol
2002
;
8
:
787
91
.
5.
Furuuchi
K
,
Tada
M
,
Yamada
H
,
Kataoka
A
,
Furuuchi
N
,
Hamada
J
, et al
.
Somatic mutations of the APC gene in primary breast cancers
.
Am J Pathol
2000
;
156
:
1997
2005
.
6.
Horii
A
,
Nakatsuru
S
,
Miyoshi
Y
,
Ichii
S
,
Nagase
H
,
Ando
H
, et al
.
Frequent somatic mutations of the APC gene in human pancreatic cancer
.
Cancer Res
1992
;
52
:
6696
8
.
7.
Ohgaki
H
,
Kros
JM
,
Okamoto
Y
,
Gaspert
A
,
Huang
H
,
Kurrer
MO
.
APC mutations are infrequent but present in human lung cancer
.
Cancer Lett
2004
;
207
:
197
203
.
8.
MacDonald
BT
,
Tamai
K
,
He
X
.
Wnt/beta-catenin signaling: components, mechanisms, and diseases
.
Dev Cell
2009
;
17
:
9
26
.
9.
Zhao
D
,
Lu
X
,
Wang
G
,
Lan
Z
,
Liao
W
,
Li
J
, et al
.
Synthetic essentiality of chromatin remodelling factor CHD1 in PTEN-deficient cancer
.
Nature
2017
;
542
:
484
8
.
10.
Zhao
D
,
DePinho
RA
.
Synthetic essentiality: Targeting tumor suppressor deficiencies in cancer
.
Bioessays
2017
;
39
:10.1002/bies.201700076.
11.
Hu
C
,
Pang
B
,
Lin
G
,
Zhen
Y
,
Yi
H
.
Energy metabolism manipulates the fate and function of tumour myeloid-derived suppressor cells
.
Br J Cancer
2020
;
122
:
23
9
.
12.
D'Amato
NC
,
Rogers
TJ
,
Gordon
MA
,
Greene
LI
,
Cochrane
DR
,
Spoelstra
NS
, et al
.
A TDO2-AhR signaling axis facilitates anoikis resistance and metastasis in triple-negative breast cancer
.
Cancer Res
2015
;
75
:
4651
64
.
13.
Opitz
CA
,
Litzenburger
UM
,
Sahm
F
,
Ott
M
,
Tritschler
I
,
Trump
S
, et al
.
An endogenous tumour-promoting ligand of the human aryl hydrocarbon receptor
.
Nature
2011
;
478
:
197
203
.
14.
Rosenbluh
J
,
Nijhawan
D
,
Cox
AG
,
Li
X
,
Neal
JT
,
Schafer
EJ
, et al
.
beta-Catenin-driven cancers require a YAP1 transcriptional complex for survival and tumorigenesis
.
Cell
2012
;
151
:
1457
73
.
15.
Van der Flier
LG
,
Sabates-Bellver
J
,
Oving
I
,
Haegebarth
A
,
De Palo
M
,
Anti
M
, et al
.
The intestinal Wnt/TCF signature
.
Gastroenterology
2007
;
132
:
628
32
.
16.
Subramanian
A
,
Tamayo
P
,
Mootha
VK
,
Mukherjee
S
,
Ebert
BL
,
Gillette
MA
, et al
.
Gene set enrichment analysis: a knowledge-based approach for interpreting genome-wide expression profiles
.
Proc Natl Acad Sci U S A
2005
;
102
:
15545
50
.
17.
Boutin
AT
,
Liao
WT
,
Wang
M
,
Hwang
SS
,
Karpinets
TV
,
Cheung
H
, et al
.
Oncogenic Kras drives invasion and maintains metastases in colorectal cancer
.
Genes Dev
2017
;
31
:
370
82
.
18.
Pilotte
L
,
Larrieu
P
,
Stroobant
V
,
Colau
D
,
Dolusic
E
,
Frederick
R
, et al
.
Reversal of tumoral immune resistance by inhibition of tryptophan 2,3-dioxygenase
.
Proc Natl Acad Sci U S A
2012
;
109
:
2497
502
.
19.
Muller
AJ
,
Manfredi
MG
,
Zakharia
Y
,
Prendergast
GC
.
Inhibiting IDO pathways to treat cancer: lessons from the ECHO-301 trial and beyond
.
Semin Immunopathol
2019
;
41
:
41
8
.
20.
Schramme
F
,
Crosignani
S
,
Frederix
K
,
Hoffmann
D
,
Pilotte
L
,
Stroobant
V
, et al
.
Inhibition of tryptophan-dioxygenase activity increases the antitumor efficacy of immune checkpoint inhibitors
.
Cancer Immunol Res
2020
;
8
:
32
45
.
21.
Pate
KT
,
Stringari
C
,
Sprowl-Tanio
S
,
Wang
K
,
TeSlaa
T
,
Hoverter
NP
, et al
.
Wnt signaling directs a metabolic program of glycolysis and angiogenesis in colon cancer
.
EMBO J
2014
;
33
:
1454
73
.
22.
van Baren
N
,
Van den Eynde
BJ
.
Tryptophan-degrading enzymes in tumoral immune resistance
.
Front Immunol
2015
;
6
:
34
.
23.
Ott
M
,
Litzenburger
UM
,
Rauschenbach
KJ
,
Bunse
L
,
Ochs
K
,
Sahm
F
, et al
.
Suppression of TDO-mediated tryptophan catabolism in glioblastoma cells by a steroid-responsive FKBP52-dependent pathway
.
Glia
2015
;
63
:
78
90
.
24.
Puccetti
P
.
On the non-redundant roles of TDO2 and IDO1
.
Front Immunol
2014
;
5
:
522
.
25.
Badawy
AA
.
Kynurenine pathway of tryptophan metabolism: regulatory and functional aspects
.
Int J Tryptophan Res
2017
;
10
:
1178646917691938
.
26.
Gabriely
G
,
Wheeler
MA
,
Takenaka
MC
,
Quintana
FJ
.
Role of AHR and HIF-1alpha in glioblastoma metabolism
.
Trends Endocrinol Metab
2017
;
28
:
428
36
.
27.
Stange
J
,
Veldhoen
M
.
The aryl hydrocarbon receptor in innate T cell immunity
.
Semin Immunopathol
2013
;
35
:
645
55
.
28.
Quintana
FJ
,
Basso
AS
,
Iglesias
AH
,
Korn
T
,
Farez
MF
,
Bettelli
E
, et al
.
Control of T(reg) and T(H)17 cell differentiation by the aryl hydrocarbon receptor
.
Nature
2008
;
453
:
65
71
.
29.
Kawajiri
K
,
Kobayashi
Y
,
Ohtake
F
,
Ikuta
T
,
Matsushima
Y
,
Mimura
J
, et al
.
Aryl hydrocarbon receptor suppresses intestinal carcinogenesis in ApcMin/+ mice with natural ligands
.
Proc Natl Acad Sci U S A
2009
;
106
:
13481
6
.
30.
Liu
Y
,
Cao
X
.
The origin and function of tumor-associated macrophages
.
Cell Mol Immunol
2015
;
12
:
1
4
.
31.
Pathria
P
,
Louis
TL
,
Varner
JA
.
Targeting tumor-associated macrophages in cancer
.
Trends Immunol
2019
;
40
:
310
27
.
32.
Takenaka
MC
,
Gabriely
G
,
Rothhammer
V
,
Mascanfroni
ID
,
Wheeler
MA
,
Chao
CC
, et al
.
Control of tumor-associated macrophages and T cells in glioblastoma via AHR and CD39
.
Nat Neurosci
2019
;
22
:
729
40
.
33.
Opitz
CA
,
Patterson
LFS
,
Mohapatra
SR
,
Dewi
DL
,
Sadik
A
,
Platten
M
, et al
.
The therapeutic potential of targeting tryptophan catabolism in cancer
.
Br J Cancer
2020
;
122
:
30
44
.
34.
Zhang
Q
,
Zhang
Y
,
Boer
J
,
Shi
JG
,
Hu
P
,
Diamond
S
, et al
.
In vitro interactions of epacadostat and its major metabolites with human efflux and uptake transporters: implications for pharmacokinetics and drug interactions
.
Drug Metab Dispos
2017
;
45
:
612
23
.
35.
Tseng
W
,
Leong
X
,
Engleman
E
.
Orthotopic mouse model of colorectal cancer
.
J Vis Exp
2007
;
10
:
484
.
36.
Chen
P
,
Zhao
D
,
Li
J
,
Liang
X
,
Li
J
,
Chang
A
, et al
.
Symbiotic macrophage-glioma cell interactions reveal synthetic lethality in PTEN-null glioma
.
Cancer Cell
2019
;
35
:
868
84
.
37.
Liao
W
,
Overman
MJ
,
Boutin
AT
,
Shang
X
,
Zhao
D
,
Dey
P
, et al
.
KRAS-IRF2 axis drives immune suppression and immune therapy resistance in colorectal cancer
.
Cancer Cell
2019
;
35
:
559
72
.
38.
Amir
ED
,
Davis
KL
,
Tadmor
MD
,
Simonds
EF
,
Levine
JH
,
Bendall
SC
, et al
.
viSNE enables visualization of high dimensional single-cell data and reveals phenotypic heterogeneity of leukemia
.
Nat Biotechnol
2013
;
31
:
545
.
39.
Chen
TJ
,
Kotecha
N
.
Cytobank: providing an analytics platform for community cytometry data analysis and collaboration
.
Curr Top Microbiol Immunol
2014
;
377
:
127
57
.
40.
Cassetta
L
,
Noy
R
,
Swierczak
A
,
Sugano
G
,
Smith
H
,
Wiechmann
L
, et al
.
Isolation of mouse and human Tumor-associated macrophages
.
Adv Exp Med Biol
2016
;
899
:
211
29
.
41.
Abrams
ZB
,
Johnson
TS
,
Huang
K
,
Payne
PRO
,
Coombes
K
.
A protocol to evaluate RNA sequencing normalization methods
.
BMC Bioinf
2019
;
20
(
Suppl 24
):
679
.

Supplementary data