Targeting the IDO1 pathway in cancer: from bench to bedside

IDO1 is not expressed in most tissues in adult humans under physiological conditions but is constitutively expressed in many types of cancer cells, stromal cells, and immune cells in the tumor microenvironment (Fig. 3a). IDO1 is activated by diverse inflammatory molecules, such as IFN-γ, tumor necrosis factor α (TNF-α), transforming growth factor β (TGF-β), pathogen-associated molecular patterns (PAMPs), damage-associated molecular patterns (DAMPs), and prostaglandin E2 (PGE2), through canonical and non-canonical NF-κB and JAK–STAT pathways [21, 23, 25‐27]. Constitutive IDO1 expression in human cancers is driven by cyclooxygenase 2 (COX-2) and PGE2 via the PKC and PI3K pathways [28]. Moreover, autocrine TGF-β sustains the activation of IDO1 in a tolerogenic subpopulation of CD8⁺ dendritic cells (DCs), while exogenous TGF-β converts immunogenic CD8⁻ DCs into tolerogenic cells in conjunction with induction of IDO1 [29]. IDO1 is under genetic control of the cancer-suppression gene bridging integrator 1 (Bin1), and Bin1 knockout results in tumor growth and immune suppression in mice by upregulating STAT1- and NF-κB-dependent expression of IDO1 [20]. Ras and Kit also upregulate IDO1 expression in cancer cells [22, 27]. Importantly, several immune checkpoints (e.g., PD-1 and CTLA4) on the T cell surface modulate IDO1 expression in antigen-presenting cells. Recently, Wang et al. [30] demonstrated that the IL-6-inducible proto-oncogene protein intestine-specific homeobox (ISX) gene induces IDO1 and TDO expression, which increases Kyn and AhR and thereby promotes the tumorigenic potential and immunosuppression of hepatocellular carcinoma cells expressing CD86 and PD-L1. Others report that hypoxia enhances IDO1 production in monocyte-derived dendritic cells, yet the underlying mechanisms remain elusive [31].

IDO1 transduces signaling through three major effectors: GCN2, mTOR, and AhR. The depletion of Trp by IDO1 leads to the accumulation of uncharged Trp–tRNA, which binds and activates GCN2, a stress-response kinase. Activated GCN2 phosphorylates and inhibits eukaryotic initiation factor 2α kinase, resulting in attenuation of RNA transcription and protein translation. Specific to T effector cells, GCN2 activation mediated by Trp deprivation leads to cell cycle arrest and/or apoptosis [25, 32]. Another signaling molecule inhibited by Trp deprivation is the mammalian target of rapamycin (mTOR) [23]. IDO1-mediated suppression through mTORC1 triggers autophagy, leading to anergy in T cells in the tumor microenvironment [23, 33]. Most relevant publications report that mTOR inhibition suppresses effector T cell function and promotes Treg cell function although others dispute these findings [34], and the precise role of IDO1 in mTOR signaling and immune regulation continues to be debated. Both GCN2 kinase activation and mTOR inhibition are immunosuppressive. During CD4⁺ T-cell differentiation, GCN2 kinase activation suppresses only Th2 differentiation, whereas mTOR inhibition induces Treg cell differentiation and suppresses differentiation of the Th1, Th2, and Th17 lineages [35]. These findings suggest that Trp deprivation-mediated GCN2 kinase activation and mTOR inhibition have differential impacts on the function of distinct CD4⁺ T cell populations. AhR is a cytosolic ligand-activated transcription factor involved in embryogenesis, adaptive immunity, mucosal barrier function, and malignancy. Its most potent ligand, 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD), activates AhR signaling to upregulate IDO1 expression [36]. A breakthrough was reported in 2011 in which Kyn was found to be an endogenous ligand of AhR [37]. The binding of Kyn to AhR promotes naive CD4⁺ T cell differentiation into Treg cells, which contributes to an immunosuppressive tumor microenvironment [38].

Increased IDO1 and accumulating Trp metabolites prevent the activation of CD8⁺ and CD4⁺ effector T cells, inhibit NK cell function, stimulate the activation of Treg cells, and promote the expansion and activation of DCs and MDSCs [25, 44‐46]. Treg cells are a major suppressive cell type in the tumor microenvironment, as their recruitment is induced by high IDO1 levels and correlates with poor prognosis in several tumor types [47]. Furthermore, IDO1 presence leads to activation of the phosphatase and tensin homolog (PTEN) pathway in Treg cells to maintain their immunosuppressive phenotype in vitro [48] .

IDO1 is also essential for the recruitment and sustenance of MDSCs through a Treg-dependent mechanism [49]. MDSCs can suppress antitumor immune responses through STAT3-dependent IDO1 expression in breast cancer [50]. Increased STAT3 activation in MDSCs was found to be correlated with activation of the non-canonical NF-κB pathway, in which RelB–p52 dimers directly bind to the IDO1 promoter, leading to IDO1 expression in MDSCs [51]. In a Kras mutant lung cancer model, Ido1-deficient mice had impaired MDSC function [27].

Cancer-associated fibroblasts recruit and educate DCs to become IDO1-producing regulatory DCs through IL-6-mediated STAT3 activation [52]. Mesenchymal stem/stromal cells in the tumor microenvironment promote tumor growth through IDO1-mediated immune suppression [53]. Expression of IDO1 in tumor endothelial cells is negatively associated with long-term survival in patients with renal cell carcinoma [54]. In addition, the IDO1 expression level in tumor endothelial cells is associated with the tumor’s response to checkpoint inhibitor treatment in metastatic renal cell carcinoma [55].

Antibodies targeting cytotoxic T cell antigen 4 (CTLA-4), programmed cell death 1 (PD-1), and programmed cell death ligand 1 (PD-L1) have been approved by the Food and Drug Administration (FDA) for multiple cancers [56, 57]. Recent research further revealed that IDO1 is closely linked to both CTLA-4 and PD-1–PD-L1 via complex pathways. Treg cell-expressed CTLA-4 induced IDO1 expression in DCs [46], and PD-1/PTEN signaling in IDO1-activated Treg cells was required to maintain immune suppression of the Tregs [58]. IDO1 expression in DCs is induced by the interaction of PD-1 with PD-L1 on the surface of mast cells and/or with PD-L2 on the surface of DCs [59]. Moreover, IDO1 participates in the intracellular signaling events responsible for long-term tolerance by DCs [29]. Based on its upstream effects on DCs, this IDO1 activity is non-redundant with that of the more distal T cell checkpoints, and combined therapy with IDO1 inhibitors and CTLA-4 or PD-1 inhibitors should confer additional benefit.

IDO1 has also been reported to participate in a resistance mechanism to checkpoint inhibitors, and the combination of CTLA-4 blockade and an IDO1 inhibitor resulted in more effective antitumor immunity in a melanoma mouse model [13]. In a phase 2 trial, excessive infiltration of IDO1⁺ macrophages and a significant increase in the kynurenine-to-Trp ratio were observed in the majority of patients who had poor response to metronomic cyclophosphamide and pembrolizumab combination treatment [60]. In addition, Liu and colleagues identified a Kyn–AhR pathway-dependent mechanism that promoted tumor-repopulating cell immune escape by increasing PD-1 binding to CD8⁺ T cells. Local IFN-γ produced by tumor-infiltrating CD8⁺ T cells stimulates tumor-repopulating cell release at high levels of Kyn, which then activates AhR on the T cell surface to promote PD-1 upregulation. Thus, targeting the Kyn–AhR pathway (such as by inhibiting IDO1 in tumor-repopulating cell or AhR in CD8⁺ T cells) may enhance the efficacy of adoptive T cell therapy. Taken together, these results suggest that combination strategies targeting multiple immune checkpoints might be the preferred weapon of the future.

Neovascularization, characterized by excessive and disorganized growth of blood vessels, is critical for tumor development, progression, and metastasis. In mice, Ido1 deficiency inhibited lung tumor growth and improved animal survival with reduced density of the underlying pulmonary blood vessels [27]. Recent work also showed a new role for IDO1 in supporting pathologic neovascularization [11]. In 4T1 breast cancer pulmonary metastases and oxygen-induced retinopathy mouse models, IDO1 ablation was sufficient to reduce pathological neovascularization in both lung metastases and reinopathy. IL-6 is regarded as a pro-angiogenic cytokine potentiated by IDO1, and IL-6 deletion results in IFN-γ-dependent reduction in neovascularization and increases resistance to metastasis [11]. Administration of the IDO1 inhibitor epacodostat (INCB024360) in mice with tumors or vascularized metastases significantly reduced neovascularization [11]. High IDO1 expression positively correlates with microvessel density and poor prognosis in breast cancer patients [61]. In summary, IDO1-mediated effects on neovascular development and established neovascular networks broaden the potential effectiveness of IDO1 inhibitors in clinical trials and in practice.

The gut microbiome has been found to regulate cancer initiation, progression, and response to therapies [62]. Preclinical mouse models suggest that resistance in melanoma patients to anti-PD-1 immunotherapy can be attributed to abnormal gut microbiome composition [63]. Bacteria, including Bifidobacterium longum, Collinsella aerofaciens, and Enterococcus faecium, are more abundant in treatment responders, whereas the efficacy of immune checkpoint blockade therapies is diminished with administration of antibiotics [64‐66]. As an essential nutrient in mammals, Trp and its IDO1-catalyzed endogenous metabolites play a key role in the gut microbiota and immune homeostasis. Recent discoveries have underscored the modulation of host immune systems by changes in the microbiota that affect Trp metabolism [43]. Activation of toll-like receptors by microbial components has been identified as a key factor in initiating Kyn metabolism and gut microbial homeostasis, as reduced toll-like receptor stimulation in germ-free mice resulted in decreased Trp metabolism [53, 67].

Indoximod (1-methyl-D-tryptophan, 1MT, NLG-8189) is the most-studied IDO1 inhibitor and has been granted orphan-drug designation by the US FDA for the indication of stage IIb to stage IV melanoma. Recent results suggest that indoximod is not only a valid inhibitor of IDO1 enzymatic activity but also acts as a high-potency Trp mimetic in reversing mTORC1 inhibition [23]. mTORC1 is a central integrator of cell growth signals that monitors levels of essential amino acids that are needed to activate cell growth [68]. Clinical results indicated that indoximod exerted little antitumor efficacy as a single agent, but efficacy was markedly enhanced when it was combined with other therapies, such as PD-1 checkpoint inhibitors (in advanced melanoma), cancer vaccines (in metastatic prostate cancer), and chemotherapy (in pancreatic cancer and acute myeloid leukemia). A phase 1 dose-escalation trial aiming to evaluate the safety, dosing, pharmacokinetics, and immunologic effects of indoximod found indoximod to be safe at doses up to 2000 mg orally twice daily [69]. Another phase 1 dose-escalation trial designed to study co-administration of docetaxel and indoximod reported that this combination is well tolerated, with no increase in expected toxicities in patients with metastatic solid tumors [70]. Administering indoximod with the PD-1 antibody pembrolizumab (Keytruda) led to a 61% overall response rate, including 10 complete responses (20%) and 21 partial responses (41%), in patients with advanced melanoma. The median progression-free survival under combinatorial treatment was 12.9 months, with a 1-year rate of 56%, suggesting a synergistic antitumor therapeutic effect [71].

Previous studies have found that IDO1 peptides elicit specific CD8⁺ T cells that recognize and kill IDO1-expressing tumor cells and DCs and simultaneously enhance other T-cell responses [85]. In a phase 1 trial to evaluate the efficacy and safety of IDO1 vaccines, 15 patients with advanced NSCLC were injected with an IDO1-derived human leukocyte antigen A2-restricted epitope. The median overall survival was 25.9 months with no grade 3 or 4 toxicities. Furthermore, all treated patients had a significantly reduced Treg cell population after the sixth dose of vaccine [86]. Based on these promising results and distinct mechanisms of action, an additional phase 1 trial combining IDO1 vaccines with CTLA-4 inhibitor ipilimumab (Yervoy) for stage III or IV melanoma patients and a phase 1/2 clinical trial that tests a combination treatment with a PD-1 monoclonal antibody (nivolumab) and a PD-L1–IDO1 peptide vaccine, have been initiated. Other potential strategies, such as combining IDO1 inhibitors with IDO1 vaccines may also produce synergistic anti-tumor effects [87].

Gene expression is often repressed by microRNAs. Our group analyzed IDO1 downregulation by microRNA-153 (miR-153) in colon cancer cells and the association of IDO1 and miR-153 expression with colorectal patient survival [18]. We found that IDO1 is highly expressed in colorectal tumors and is inversely associated with patient survival. miR-153 directly inhibits IDO1 expression by targeting its 3′ untranslated region in colon cancer cells, yet miR-153 overexpression does not affect colon cancer cell survival, apoptosis, or colony formation. When colon cancer cells are targeted by chimeric antigen receptor (CAR) T cells, miR-153 overexpression within tumor cells significantly enhances T cell killing in vitro and suppresses xenograft tumor growth in mice. These findings indicate that miR-153 is a tumor-suppressive miRNA that enhances CAR T cell immunotherapy and supports the combinatorial use of IDO1 inhibitors and CAR T cells in treating solid tumors [18].