The kynurenine pathway presents multi-faceted metabolic vulnerabilities in cancer

The kynurenine pathway (KP) and associated catabolites play key roles in promoting tumor progression and modulating the host anti-tumor immune response. To date, considerable focus has been on the role of indoleamine 2,3-dioxygenase 1 (IDO1) and its catabolite, kynurenine (Kyn). However, increasing evidence has demonstrated that downstream KP enzymes and their associated metabolite products can also elicit tumor-microenvironment immune suppression. These advancements in our understanding of the tumor promotive role of the KP have led to the conception of novel therapeutic strategies to target the KP pathway for anti-cancer effects and reversal of immune escape. This review aims to 1) highlight the known biological functions of key enzymes in the KP, and 2) provide a comprehensive overview of existing and emerging therapies aimed at targeting discrete enzymes in the KP for anti-cancer treatment.


Introduction
The kynurenine pathway (KP) has been recognized as a key mediator of tumor immune escape (1)(2)(3).Three enzymes, indolamine 2,3-dioxygenase 1 and 2 (IDO1, IDO2) and tryptophan 2,3-dioxygenase (TDO) initiate the first steps in the KP, converting tryptophan (Trp) to N-formyl-kynurenine, which can be further metabolized by arylformamidase (AFMID) yielding the immunosuppressive catabolite kynurenine (Kyn).Whereas TDO is predominately expressed in the liver and accounts for the majority of Trp metabolism, IDO-mediated metabolism predominately occurs secondary to inflammation (4,5).Upregulation of IDO in tumor cells or antigen-presenting cells leads to Trp depletion and accumulation of its downstream catabolite Kyn in the local tumor microenvironment (TME), resulting in immunosuppression by inducing T cell anergy and apoptosis and suppressing T cell differentiation (1)(2)(3)(6)(7)(8)(9).Building upon this underlying biological mechanism, substantial efforts have been dedicated to the development of small molecule inhibitors and agonists targeting the KP as anti-cancer agents to circumvent tumor immune suppression.Among the most investigated IDO inhibitors are epacadostat, navoximod, and indoximod, each of which have been explored as a monotherapy and in combination with immune check-point inhibitors (ICI) and/or chemotherapy (10)(11)(12)(13)(14)(15)(16)(17)(18)(19)(20)(21).Despite early success in Phase I/II clinical trials, ECHO-202/ KEYNOTE-037 failed to demonstrate additional benefit of epacadostat in combination with pembrolizumab (10) and similar findings have since been reported for navoximod with atezolizumab (19), raising uncertainty about the benefits of targeting IDO1 for anti-cancer treatment.
Yet, it is increasingly recognized that, beyond the well covered Trp-IDO-Kyn axis, other enzymes in the KP along with their associated catabolites can modulate the TME to evade immune surveillance and enable tumor progression (22)(23)(24)(25)(26).Moreover, emerging data suggests that KP-related enzymes, such as IDO1, can promote tumor progression through functions that are independent of enzymatic activity (27).The purpose of this review is to 1) highlight the known tumor promotive functions of key enzymes in the KP and 2) provide a comprehensive overview of existing and emerging therapies aimed at targeting discrete enzymes in the KP for anti-cancer treatment.
Seminal studies by Munn and colleagues demonstrated that reduced bioavailability of Trp due to IDO overactivation in dendritic cells (DCs) leads to accumulation of uncharged Trp-tRNAs in T cells and that the accumulation of Trp-tRNAs induces integrated stress response kinase (GCN2)-mediated suppression of the translation initiation factor 2a (eIF2a) resulting in subsequent cell growth arrest (6).Tryptophan starvation-mediated activation of GCN2 has also been shown to down-regulate the T cell receptor (TCR) zeta-chain in naïve T cells, resulting in diminished T cell function (32).Another study demonstrated a GCN2-independent mechanism whereby IDO-mediated catabolism of Trp inhibits immunoregulatory kinases mTOR and PKC resulting in induction of autophagy and immunosuppression (18).Subsequent studies demonstrated that IDO-mediated catabolism of Trp results in the accumulation of Kyn (7) that stimulates ligand-activated transcription factor aryl hydrocarbon receptor (AhR) signaling and promotes the generation of immune-tolerant DCs and regulator T cells (Tregs) (1,7,8).To this end, Liu and colleagues demonstrated that interferon-gamma (IFNg) produced by cytotoxic CD8+ T cells stimulates Kyn release by tumor-repopulating cells (TRCs) and that TRC-derived Kyn is subsequently taken up by CD8+ T cells through TCR-mediated upregulation of solute carrier family 7 member 8 (SLC7A8) and Solute Carrier Family 36 Member 4 (SLC36A4, also known as PAT4), resulting in Kyn-AhR-mediated upregulation of programmed cell death protein 1 (PD-1) (33).Inhibition of IDO1 in CD8+ T cells promoted Trp accumulation and downregulation of PD-1 (34).Another study found that IFNg secreted by acute myeloid leukemia (AML) cells upregulated IFNgdependent genes related to Treg induction, including IDO1, in mesenchymal stromal cells (MSCs).Genetic ablation of IFNg production by AML cells reduced MSC IDO1 expression and Treg infiltration, thereby attenuating AML engraftment (35).
The IDO1-Kyn-AhR axis has also been shown to impair natural killer (NK) cell and macrophage function.Specifically, Kyn inhibits expression of the natural cytotoxicity triggering receptor 1 (NCR1, also known as NKp46) and killer cell lectin like receptor K1 (KLRK1, also referred to as NKG2D) on NK cells resulting in impaired tumor-killing functions (36,37).Recently, Fang and colleagues demonstrated that IDO1 regulates expression of ADAM metallopeptidase domain 10 (ADAM10) via IDO1-Kyn-AhR pathway in non-small cell lung cancer (NSCLC) cells and that the induction of ADAM10 downregulates the NKG2D Ligand (NKG2DL).Targeting of IDO1 via RY103, a novel IDO1 inhibitor, improved NK cell-mediate anti-tumor activity in xenograft mouse models of lung cancer (38).Takenaka and colleagues reported that increased IDO-mediated release of Kyn by glioblastoma cells promotes AhR activation in tumor-associated macrophages (TAMs); activation of AhR signaling in TAMs increased expression of the ectonucleotidase CD39 and accumulation of adenosine in the tumor microenvironment (TME) leading to suppression of CD8+ T cell function (39).Interestingly, studies by Campesato and colleagues reported that IDO-Kyn-AhR-mediated immunosuppression depends on the interplay between Tregs and TAMs (40).Using a preclinical mouse model of B16-F10 melanoma overexpressing IDO1 (B16 IDO ), authors demonstrated that IDO-expressing tumors have marked increases in Treg frequency and upregulated expression of FoxP3 in tumor infiltrating CD4+ T cells compared to B16 WTtumor bearing mice.Concomitant to Tregs was an increased abundance of TAMs (CD11b+F4/80highLy6G−) that accumulated during tumor progression.Gene expression analyses of B16 IDOderived Tregs and TAMs revealed prominent upregulation of AhR and AhR-responsive genes.Depletion of macrophages with aCSF1-R or clodronate liposomes delayed the progression of IDO-expressing tumors but not wild-type tumors, an effect that could be reversed upon CD8+ T cell depletion.Similarly, specific depletion of Tregs using Foxp3DTR mice significantly abrogated the myeloid-enriched phenotype found in B16 IDO tumors, with reduced infiltration of CD11b+ cells and M2-like TAMs after treatment with DT.Co-culture experiments further demonstrated that Treg inhibitor function and expansion of M2-like macrophages was dependent on Kyn-AhR signaling (40).
Whereas IDO1 has largely been studied in the context of Trp-IDO1-Kyn-AhR signaling, IDO1 may also promote tumor progression through non-enzymatic functions (27).To this end, IDO1 can act as a signal-transducing molecule via two immunoreceptor Tyrosine-based inhibitory motifs (ITIMs) in the non-enzymatic domain of the enzyme.Phosphorylation of the ITIMs promotes interaction with SHP family tyrosine phosphatases (41,42).Moreover, IDO1 can directly bind class IA PI3K regulatory subunits, resulting in IDO1 anchoring to early endosomes and activation of ITIM-mediated signaling (43).
In the context of cancer, using B16-F10 mouse melanoma cells transfected with either the wild-type Ido1 gene (Ido1 WT ) or a mutated variant lacking the catalytic, but not signaling activity (Ido1 H350A ), Orecchini and colleagues demonstrated that IDO1 promote tumor growth by increasing SHP-2-mediated activation of Ras and Erk signaling in vitro and in vivo.Furthermore, B16 H350A tumor bearing mice exhibited markedly enhanced tumor growth, and worse overall survival compared to B16 wt or B16 mock control mice.Immunophenotyping of tumors indicated that B16 H350A tumors had lower IFNg-producing CD4+ and CD8+ T lymphocytes and a significantly higher percentage of CD4 + Fox3p + Tregs compared to respective controls (27).Zhai and colleagues established IDO-deficient glioblastoma (GBM) cell lines reconstituted with IDO wild-type or IDO enzyme-null cDNA to assess tumor promotive roles of IDO1 independent of its catalytic activity.They found that nonenzymatic tumor cell IDO1 activity increased expression of complement factor H (CFH) and its isoform, factor H like protein (FHL-1) in human GBM, resulting in increased intratumoral Tregs and myeloid-derived suppressor cells, accelerated tumor growth and poor survival (44).These findings highlight an increasingly appreciated role of IDO1 in signal transduction, and, importantly, underscore an alternative mechanism of IDO1-mediated immune suppression that may not be impeded by small molecule IDO inhibitors.
With respect to cancer cell functions, small interfering RNA (siRNA)-mediated knockdown of IDO2 inhibited cancer cell proliferation, induced cell cycle arrest in G1 and apoptosis, and reduced cell migration of B16-BL6 melanoma cells in vitro.Mechanistically, IDO2 knockdown increased generation of reactive oxygen species (ROS), and decreased generation of NAD + (50).
In the context of the tumor immunophenotype, a pan-cancer analysis of IDO2 in the TCGA transcriptomic datasets revealed IDO2 mRNA expression to be strongly associated with high infiltration of immune cells in the tumor microenvironment, as well as tumor mutational load (TMB), microsatellite instability (MSI), mismatch repair (MMR), and immune checkpoint related genes (51).
IDO2 may promote tumor immune suppression through the Trp-Kyn-AhR axis.Studies by Yamasuge and colleagues demonstrated that Ido2 knockout (KO) tumors had higher intratumor Trp and reduced Kyn levels compared to wild-type Ido2 tumors in a Lewis lung carcinoma mouse model, indicating Ido2 Trp catalytic activity (52).Moreover, Ido2 KO tumors had significantly higher CD3+ and CD8+ TILs compared to wild-type Ido2 tumors (52).
Yet, it is unlikely that the biological function(s) of IDO2 are completely redundant with that of IDO1.This is reinforced by the fact that the K m of IDO2 for Trp is markedly lower than IDO1 (K m 500-to 1000-fold lower), despite both enzymes containing residues necessary for Trp catalytic activity (53, 54).Indeed, studies in autoimmune disease have identified several functions of IDO2 that are independent of IDO1 and Trp catalytic activity.For instance, Metz and colleagues found that IDO1-dependent T regulatory cell generation is defective in Ido2 -/-mice.Mechanistically, Ido2 deficient mice, but not wild-type Ido2 mice, exhibited lower expression of immune regulatory cytokines, including IFN-g, TNF-a, IL-6, and MCP-1/CCL2 (55).Merlo and colleagues reported that IDO1 mediates T cell suppressive effects whereas IDO2 mediates autoreactive B cell responses to promote inflammation (56).Using a catalytically inactive IDO2 knock-in mouse model, authors demonstrated that IDO2, but not IDO1, can directly interact with GAPDH, Runx1, RANbp10, and Mgea5 to potentiate an inflammatory response in autoimmune arthritis (57).
The above studies lend to an intriguing question as to whether a concomitant upregulation of tumoral IDO1 and IDO2 may synergize to suppress an anti-cancer immune response, which may lend to additional opportunities for cancer interception (see Section 2 below).
Like IDO1, TDO2-mediated Kyn promotes tumor immune suppression through AhR-signaling in various cancer types (63, 64).To this end, single-cell RNA sequencing (scRNA-seq) of 13 cancerous tissues identified a subset of myofibroblasts that exclusively expressed TDO2 and that clustered with CD4+ and CD8+ T cells distal to tumor nests.Functional experiments demonstrated that TDO2+ myofibroblasts induce transformation of CD4+ T cells into Tregs and caused CD8+ T cell dysfunction.In a murine model of oral squamous cell carcinoma (OSCC), small molecule inhibition of TDO2 via LM10 attenuated the inhibitory states of T cells, restored the T cell antitumor response, and prevented malignant transformation (65).Interestingly, Schramme and colleagues demonstrated that C57BL/6 TDO-KO mice (66) engrafted with MC38 CRC cells were more sensitive to anti-CTLA4 or anti-PD1 treatment compared to MC38-tumor bearing wild-type C57BL/6 mice.Enhanced efficacy of ICI in C57BL/6 TDO-KO tumor bearing mice was attributed to higher systemic Trp levels, which could be reversed through a Trp-low diet (67).These studies implicate host metabolism Trp independent of cancer TDO2 status as being a determinant of response to ICI.
In addition to its role in immune suppression, TDO2 has also been demonstrated to promote migration and invasion of liver cancer cells via Wnt5a signaling activities (59).Another study demonstrated that TDO2-Kyn-AhR activation in liver cancer cells induces autocrine interleukin 6 (IL-6)-mediated Signal transducer and activator of transcription 3 (STAT3) and Nuclear factor kappa beta (NFkB)/T Cell Immunoglobulin And Mucin Domain Containing 4 (TIM4) signaling to promote tumor progression (68).Inhibition of AhR signaling by PDM2 attenuated tumor growth in a xenograft model of liver cancer (68).In the context of CRC, deficiency in adenomatous polyposis coli (APC) results in TCF4/b-catenin-mediated upregulation of the TDO2-Kyn-AhR axis to increase glycolysis and drive anabolic cancer cell growth (69).Knockdown of TDO2 in LoVo and HCT116 CRC cells attenuated cell growth, and reduced migration, invasion and colony formation potential through inactivation of TDO2kynureninase (KYNU)-AhR signaling (60).

Arylformamidase
Arylformamidase (AFMID) is another downstream KP enzyme that is frequently overexpressed in various malignancies (70), and that catalyzes the hydrolysis of N-formyl-Kynurenine into Kyn and formate (Figure 1) (71).Although the functional role of AFMID upregulation in cancer remains poorly understood, Venkateswaran and colleagues demonstrated that oncogenic MYC increases expression of the Trp importers SLC1A5 and SLC7A5 as well as AFMID in colon cancer.Increased uptake and catabolism of Trp results in the accumulation of Kyn that enables cancer cell proliferation in part through AhR-signaling (72).
Splicing changes in AFMID have also been associated with survival and relapse in patients with hepatocellular carcinoma (HCC) (73).Specifically, the switch of AFMID isoforms was found to be an early event in HCC development and was associated with driver mutations in TP53 and AT-rich interactive domain-containing protein 1A (ARIDIA).Authors further found that overexpression of the full-length AFMID isoform leads to higher NAD+ levels, lower DNA-damage response, and slower cell growth in HepG2 cells (73).To this end, Tummala and colleagues found that loss of AFMID expression in SNU-449 HCC cells resulted depletion of NAD+, indicating that Trp catabolism in HCC cells supports de novo NAD+ synthesis (74).Additional studies are warranted to better elucidate the functional relevance of AFMID in tumorigenesis and progression; however, one may surmise that the elevation of AFMID underlies increased accumulation and secretion of Kyn.
Recently, our group demonstrated selective enrichment of KYNU, but not other downstream KP enzymes, in lung adenocarcinomas (LUAD) harboring a loss-of-function KEAP1 mutation (22).Metabolomic analyses confirmed that KYNU was enzymatically functional, as evidenced by the accumulation of AA in conditioned medium of LUAD cell lines.Activation of the Nuclear factor erythroid 2-related factor 2 (NRF2) pathway through siRNA-mediated knockdown of KEAP1 or chemical induction with the NRF2-activator CDDO-Me upregulated KYNU at both the mRNA and protein levels with concomitant increases in AA production.Moreover, elevated tumoral KYNU expression was found to be associated with a tumor suppressive immunophenotype and was prognostic for poor overall survival in a tissue microarray (TMA) of LUAD as well as multiple independent LUAD transcriptomic datasets (22).A pan-cancer analyses further revealed upregulation of KYNU to be a prominent feature of NRF2activated cancers and is associated with tumor immunosuppression and poor prognosis (23).Interestingly, a recent study by Liu and colleagues demonstrated that KYNY-derived 3HA has anti-oxidant properties and that elevations in cancer cell intracellular 3HA pools infer resistance to ferroptosis (82).
Studies by Heng and colleagues similarly found that KYNU was highly upregulated in HER2-enriched and triple-negative (TN) breast cancers (BrCa), leading to increased production of AA and 3HA (81).Interestingly, stimulation of TNBC cell lines with IFNg resulted in a pronounced 286-fold increase in 3HA and higher serum levels of 3HA was able to distinguish TNBC from the other BrCa molecular subtypes (81).These findings are notable given that 3HA has been shown to enhance the percentage of Tregs, inhibit Thelper (Th)-1 and Th17 cells (24) and suppress antigenindependent proliferation of CD8+ T cells (83).To this end, a large-scale transcriptomic analysis of 2,994 BrCa tumors found a strong correlation between tumoral KYNU with inflammatory and immune responses; elevated KYNU was also linked to BrCa grade and poorer clinical outcomes (78).
In addition to cancer cell-associated upregulation of KYNU, a recent study by Itoh and colleagues found that KYNU is also elevated in cancer-associated fibroblast (CAF)-educated fibroblasts (CEFs).Specifically, authors demonstrated that CAFs induce CEF generation from normal fibroblasts (NFs) via reactive oxygen species (ROS)-induced activation of NFkB signaling, resulting in induction of a pro-inflammatory loop and secretion of asporin (ASPN), a small leucine-rich proteoglycan.ASPN in turn promoted upregulation of IDO1 and KYNU in CEFs, which induced cytocidal effects against CD8+ T-cells and promoted tumor spreading (79).
In the context of TNBC, elevated tumoral KMO is associated with worse overall survival and invasive breast cancers have among the highest rates of KMO copy number amplification (81,89).In vitro ectopic overexpression of KMO in TNBC cell lines promoted increased cell growth, colony and mammosphere formation, migration and invasion, as well as increased expression of mesenchymal markers (85).In xenograft models, mice harboring CRISPR KMO-KD MDA-MB-231 TNBC cells had reduced tumor growth, attenuated capacity for lung metastases, and prolonged overall survival compared to respective control mice.Mechanistically, KMO prevented degradation of b-catenin, thereby enhancing the transcription of pluripotent genes, including CD44, Nanog, Oct4, and SOX2 (85).In support of the aforementioned finding, knockdown of KMO in CRC cells was also reported to decrease expression of cancer stem cell markers CD44 and Nanog, and reduce sphere formation as well as migration and invasion (86).Inhibition of KMO enzymatic activity, using the small molecule inhibitor UPF648, similarly attenuated sphere formation and cell motility of CRC cell lines (86).Interestingly, using immunohistochemistry, flow cytometry, immunofluorescence assay, and transmission electron microscopy, Lai and colleagues demonstrated that KMO is highly expressed on the cell membranes of breast cancer tissues and the cancer cell surfaceome of cell lines (90).Treatment of MDA-MB-231 TNBC cell lines with anti-KMO antibodies reduced the cell viability and inhibited the migration and invasion of the triplenegative (90).Collectively, these findings demonstrate a cancer cell intrinsic function of KMO that promotes tumor aggressiveness and pluripotency.
Notably, induction of KMO in immune cell subtypes has also been linked to dysfunctional anti-tumor activity.Using melanomaderived cell lines and primary CD4+ CD25-T cell co-cultures, Rad Pour and colleagues demonstrated that activation of CD4+ T cells results in increased production of IFNg with concomitant increases in Kyn and kynurenic acid that is attributed to reduced KMO expression in CD4+ T cells (25).Accumulation of Kyn and kynurenic acid suppresses CD4+ T cell expansion and viability (25).In the context of multiple myeloma (MM), Ray and colleagues demonstrated that KMO is upregulated in plasmacytoid dendritic cells (pDCs), resulting in immune suppression (26).Specifically, using co-culture models of patient autologous pDC-T-NK-MM cells, authors showed that pharmacological blockade of KMO activates pDCs and triggers both MM-specific cytotoxic T-cell lymphocytes (CTL) and NK cells`cytolytic activity against tumor cells.Combination treatment with R0-61-8048, a potent competitive inhibitor of KMO, and anti-PD-L1 antibody yielded superior anti-cancer efficacy compared to either treatment alone (26).

Kynurenine aminotransferase
Kynurenine aminotransferase family members (KATI-KATIV) catalyze the transamination of Kyn and 3HK into kynurenic acid (KA) (91) and xanthurenic acid (XANA), respectively (Figure 1) (4).Although limited information exists regarding the functional role of KAT in tumorigenesis, KAT-derived KA is readily detectable in tumor tissues, as well as in blood and urine of cancer patients (92).
Early studies demonstrated that KAT-derived KA exerts antiproliferative effects by modulating key pathways associated with proliferation, survival, apoptosis, and migration (93-96).Yet, KA is documented to exert immunosuppressive effects via the G-protein-coupled receptor 35 (GPR35) (97), which may potentiate tumor progression (98).Moreover, KA has also been reported to be a potent agonist of the AhR that synergistically induces IL-6 (99) with potential pro-tumoral effects (100).To this end, a recent pan-tissue study revealed that interleukin-4-induced-1 (IL4I1) more frequently associates with AhR activity than does IDO1 or TDO2.Mechanistically IL4I1-mediated generation of indole metabolites and KAT-derived KA that activated AhR signaling to promote cancer cell motility and suppression of adoptive immunity (101).The above findings underscore the enigmatic role of KAT and KA in tumor progression.Further investigations are needed.
2.8 3-hydroxyanthranilate 3,4-dioxygenase 3-hydroxyanthranilate 3,4-dioxygenase (HAAO) is involved in the synthesis of quinolinic acid (QA) (Figure 1).Expression patterns of HAAO are variable among different cancer types (5,22,23); however, studies of endometrial carcinomas revealed promotor hypermethylation of HAAO is prominent and to be associated with microsatellite instability and poor clinical outcomes (102).Hypermethylation of HAAO has also been reported to be higher in prostate cancer tissues compared to adjacent control tissue, and it is associated with a more aggressive phenotype (103).

Quinolinic acid phosphoribosyl transferase
Quinolinic acid phosphoribosyl transferase (QPRT) is a ratelimiting enzyme in de novo NAD+ biosynthesis and catalyzes the conversion of QA to nicotinate ribonucleotide (Figure 1).The cancer cell intrinsic effects of QPRT upregulation have been studied in a variety of cancer types (104-109).For instance, neoplastic transformation in astrocytes is associated with a QPRT-mediated switch in NAD+ metabolism by exploiting microglia-derived quinolinic acid as an alternative source of replenishing intracellular NAD+ pools.Mechanistically, the increase in QPRT expression increases resistance to oxidative stress, enabling disease progression (104).Increased expression of QPRT in invasive breast cancer promotes cell migration and invasion through purinergic signaling (105).Another study found down syndrome cell adhesion molecule antisense RNA 1 (DSCAM-AS1) increases QPRT expression in breast cancer cells via competitively binding miRNA-150-5p and miRNA-2467-3p, resulting in increased cell growth, migration, and invasion of estrogen-receptor breast cancer cells (106).Upregulation of QPRT has also been shown to confer resistance to chemotherapy in leukemic cells and ovarian cancer (107,108).In this regard, Thongon and colleagues evaluated the mechanisms of cancer cell resistance to nicotinamide phosphoribosyltransferase (NAMPT), the rate-limiting enzyme in NAD+ biosynthesis from nicotinamide.In their study, FK866-resistant CCRF-CEM (T cell acute lymphoblastic leukemia) cells were found to have exceptionally high QPRT activity and exhibited an addiction to exogenous Trp to maintain NAD+ pools under stress conditions (109).
Despite the above studies, QPRT has also been shown to be inversely associated with other KP pathway enzymes, suggesting reduced expression in some cancer types (5,22,23).For example, studies in renal cell carcinoma revealed QPRT to be downregulated in tumors and loss of QPRT expression led to anchorageindependent growth of RCC cells (110).Thus, the relevance of QPRT in promoting tumor progression is likely to be context dependent.

Targeting the kynurenine pathway for anti-cancer treatment
Based on intrinsic malignant properties as well as tumor immune suppression, targeting of the kynurenine pathway has garnered considerable attention, with several small molecule inhibitors being explored for the treatment of solid malignancies as monotherapies or in combination with immunotherapy (Table 1).The below sections highlight existing small molecule inhibitors as well as emergent therapies that target the KP pathway for anti-cancer treatment.3.1 Indoleamine 2,3-dioxygenase 1
Pharmacokinetic and pharmacodynamic studies have demonstrated that, as monotherapy, epacadostat is generally well tolerated, with a maximal inhibition of IDO1 activity (based on reductions in circulating kynurenine levels) achieved at doses ≥ 100mg BID.Adverse events (AEs) associated with such treatment included fatigue, nausea, decreased appetite, vomiting, constipation, abdominal pain, diarrhea, dyspnea, back pain, and cough (11).In a Phase I clinical study evaluating efficacy of epacadostat in patients with advanced solid malignancies, 18 out of 52 patients achieved stable disease, with 7 patients (13.5%) having stable disease lasting ≥16 weeks (11).
A seminal study, ECHO-202/KEYNOTE-037, evaluated the efficacy of epacadostat in combination with the PD-1 inhibitor pembrolizumab in advanced melanoma, with reported objective response rates (ORR) of 56% and a disease control rates (DCR; complete response (CR) + partial response (PR) + stable disease (SD)) of 78% (111).Despite early success in Phase I and II clinical trials, a Phase III clinical trial (ECHO-301/KEYNOTE-252) did not demonstrate additional benefit of epacadostat + pembrolizumab compared to placebo + pembrolizumab alone in patients with unresectable or metastatic melanoma (10).
ECHO-202/KEYNOTE-037 also reported results from Phase I/ II studies evaluating combination treatment of epacadostat + pembrolizumab in sixty-two patients with various advanced solid tumors (13).Antitumor activity was observed in several cancer types, with ORR being observed in 25 (40.3%) of 62 patients.
Epacadostat 100 mg twice a day plus pembrolizumab 200 mg every 3 weeks was recommended for Phase II evaluation.In the Phase II study, an objective response (OR) occurred in 55% (12 of 22) patients with melanoma and in those with non-small-cell lung cancer, renal cell carcinoma, endometrial adenocarcinoma, urothelial carcinoma, and head and neck squamous cell carcinoma (13).For NSCLC, combination treatment of epacadostat + pembrolizumab resulted in a ORR (CR + PR) of 35% (14 out of 40 patients) and a DCR (CR+PR+SD) of 60% (24 out of 40 patients) (112).Similar results were reported for renal cell carcinoma (15) and urothelial carcinomas (16).ECHO-207/ KEYNOTE-723, a Phase I/II study of epacadostat plus pembrolizumab and chemotherapy for advanced solid tumors, also demonstrated ORRs of 31.4% across all treatment groups (113).However, a Phase II study evaluating epacadostat + pembrolizumab in 30 patients with selective sarcoma subtypes failed to achieve meaningful antitumor activity, with ORRs of <5% (114).
Epacadostat has also been explored in combination with other ICIs, including ipilimuab, an anti-cytotoxic T-lymphocyteassociated protein 4 (CTLA4) antibody, and nivolumab, an anti-PD-1 antibody (14,115).To this end, Gibney and colleagues reported the results of a Phase I/II study of epacadostat in combination with ipilimuab in patients with unresectable or metastatic melanoma (14).In this study, immunotherapy-naïve patients (n= 39) achieved an ORR of 26% according to the immune-related response criteria and 23% by the Response Evaluation Criteria in Solid Tumors (RECIST 1.1); no ORR was observed in patients who received prior immunotherapy (14).Combination treatment of epacadostat plus nivolumab in advanced solid tumors achieved respective DCR (CR+PR+SD) of 24%, 28%, and 100% for colorectal cancer, ovarian cancer, and melanoma (115).

Navoximod
Navoximod is an orally available non-competitive inhibitor of IDO1 ( 17).An open-label Phase Ia study by Nayak-Kapoor and colleagues assessed the safety, pharmacokinetics, pharmacodynamics, and preliminary anti-tumor activity of navoximod as a monotherapy in 22 patients with recurrent/advanced tumors (116).Trialassociated adverse events, regardless of causality, included fatigue (59%), cough, decreased appetite, and pruritus (41% each), nausea (36%), and vomiting (27%).Grade ≥3 AEs occurred in 14 of the 22 patients; however, only 2 were related to navoximod.Navoximod was found to be rapidly absorbed (Tmax ~1 h) with a half-life of approximately 11 hours, and transiently decrease plasma kynurenine levels.No objective responses were met based on RECIST v1.1 criteria; however, stable disease was observed in 36% of efficacy-evaluable patients (116).
Combination of navoximod plus the anti-PD-L1 inhibitor atezolizumab has also been explored for anti-cancer treatment of solid malignancies (19)(20)(21).Results from an open-label Phase Ib trial of 52 patients treated with navoximod plus atezolizumab demonstrated acceptable toxicity profiles; efficacy data available for 45 patients included 4 (9%) patients with a partial response and 11 (24%) patients with stable disease (21).Another Phase I study of navoximod plus atezolizumab in 20 Japanese patients with advanced solid tumors achieved stable disease in 65% (13 out of 20) patients.No dose-limiting toxicities were observed.The recommended dose of navoximod monotherapy was determined as 1000 mg orally BID, and 1000 mg orally BID could be considered in combination with atezolizumab (20).Despite these findings, a larger Phase I study of navoximod plus atezolizumab in 157 patients with advanced solid tumors failed to show benefit of adding navoximod to atezolizumab (19).

Indoximod
Indoximod, also known as 1-methyl tryptophan (1-MT), is an indirect inhibitor of IDO1/IDO2 and is reported to serve as a highpotency Trp mimetic that reverses mTORC1 inhibition and the accompanying autophagy that is induced by Trp depletion in cells (17,18).Indoximod has been explored for anti-cancer treatment in several Phase I/II clinical trials, both as a monotherapy as well as in combination with chemotherapy (117)(118)(119)(120).
Recently, Zakharia and colleagues reported results from a singlearm Phase II clinical trial evaluating the addition of indoximod to standard of care ICI (pembrolizumab, nivolumab, or ipilimumab) approved for melanoma (121).Indoximod was administered continuously (1200 mg orally two times per day), with concurrent ICI dosed per US Food and Drug Administration (FDA)-approved label.A total of 131 patients were enrolled; pembrolizumab was the most frequently used ICI (114 out of 131; 87%).Efficacy was evaluable in 89 patients from the Phase II cohort with non-ocular melanoma who received indoximod combined with pembrolizumab.The ORR for the evaluable population was 51%, with confirm CR of 20% and DCR of 70%.ORR were highest among PD-L1 positive patients (70% compared to 46% for PD-L1 negative patients).Median progression-free survival was 12.4 months (95% CI 6.4 to 24.9) (121).Another recently completed Phase I/II trial (NCT01042535) evaluated an adenovirus-p53 transduced DC vaccine together with indoximod for treatment of metastatic breast cancer.A total of 44 patients were recruited, of which 36 completed the study.The results of this trial are pending.
In addition to the above inhibitors, structural elucidation studies have also led to the discovery of imidazopyridines as potent IDO1 inhibitors (132) and substituted oxalamides as novel heme-displacing IDO1 inhibitors (133), as well as the development of first in class IACS-9779 and IACS-70465 inhibitors that bind the IDO1 apoenzyme (134).Anti-cancer efficacy of the mentioned above agents has been demonstrated in preclinical models; however, they have not been explored yet in clinical trials.
Another alternative approach has been to degrade intratumor IDO1.Bollu and colleagues reported the results of a novel IDO1 degrader using proteolysis targeting chimera (PROTAC) technology, which utilizes an E3 ligase complex to ubiquitinate targets for proteasome-mediated degradation.The IDO1-PROTAC resulted in potent cereblon-mediated proteasomal degradation of IDO1, and prolonged survival in mice with established brain tumors (135).Interestingly, studies by Shi and colleagues suggest that IDO1 expression may be regulated post-transcriptionally.Specifically, authors found that overexpression of USP14, a proteosome-associated deubiquitinating enzyme, in CRC cells deubiquitinates the IDO1 at the K48 residue, thus preventing proteasomal degradation (136).Knockdown or pharmacological targeting of USP14 decreased IDO1 expression, independent of AhR signaling, and improved anti-tumor immunity in a MC38 orthotopic syngeneic mouse model of CRC.Combination treatment of IU1, a selective small molecule inhibitor of USP14, and anti-PD-1 resulted in improved anti-killing effects compared to either treatment alone (136).Targeting IDO1 protein stability, rather than enzymatic activity, may thus provide an alternative approach to circumvent IDO1-mediated anti-cancer immunity.

Indoleamine 2,3-dioxygenase 2 inhibitors
Although the exact crystal structure of IDO2 remains incomplete, its high homology with IDO1 has enabled development of IDO2 inhibitors, including several diaryl hydroxylamines and 1,2,3-triazoles.For example, diaryl hydroxylamine compounds with a 3,4-dichloro aryl ring substitution exerted potent pan (IDO1/IDO2/TDO) inhibitory function whereas a 2-chloro aryl ring substitution inhibited IDO1 and IDO2 activity with respective IC 50 values of 18 and 25mM (137).In cellular assays, 4-Aryl-1,2,3-triazole derivatives also found to inhibit IDO2 function, with the best compound having an IC 50 value of 50mM for mouse IDO2, and a twofold higher selectivity over human IDO1.Functionally, the 4-Aryl-1,2,3-triazole compounds were found to occupy both the A and B pockets of the IDO active site, which inhibited IDO2 more strongly than IDO1.Notably, the mM activity of the 4-Aryl-1,2,3-triazole compounds were similar to L-1MT (138).Another screening study identified the proton pump inhibitor, tenatoprazole, to exhibit an IC 50 value of 1.8mM for IDO2 with no observed inhibition of IDO1 or TDO.Similar findings were found with other proton pump inhibitors.Proton pump inhibitors, such as tenatoprazole, were predicted to have one heteroatom coordinating to the heme iron in the active site of IDO2 (139).
A more recent study by He and colleagues reported in vitro and in vivo findings of their novel dual IDO1/IDO2 inhibitor, 4t.Structurally, 4t is an 1,2,3-triazole with a hydroxyethyl ether substitution with excellent inhibitory activity (IC 50 values of 28 nM and 144nM for IDO1 and IDO2, respectively).Functionally, the hydroxyethyl ether chain of 4t formed two hydrogen bonding interactions with Tyr244 and Lys366 of IDO2, contributing to improved activity against IDO2.Pharmacokinetic experiments demonstrated favorable profiles, with adequate membrane permeability, high plasma protein binding, and safety.Moreover, 4t exhibited potent anti-tumor activity in a CT26 colorectal xenograft mouse model (140).Although promising, it remains to be determined whether the above mentioned IDO2 inhibitors will yield satisfactory anti-tumor activity in the clinical setting.

Tryptophan 2,3-dioxygenase inhibitors
There has been a considerable interest in the development of small molecule inhibitors of TDO2 given recent evidence that TDO2 is upregulated in several cancer types and, like IDO, mediates the catabolism of Trp and promote tumor immune suppression (7).To-date, small molecule inhibitors with high affinity to TDO2 are limited (141,142).
Recently, W-0019482 was identified to be a potent inhibitor of IDO1, resulting in pronounced reductions in plasma and intratumoral ratios of Kyn-to-Trp and delayed growth of subcutaneous GL261-hIDO1 tumors in mice.Synthetic modification of W-0019482 yielded several analogues with either dual or TDO2-selective profiles.Four analogues, SN36458, SN36896, SN36499, and SN36704, were found to be TDO2 selective, exhibiting IC50 values 5.8 to 8.1-fold lower than that measured against IDO1 (143).The utility of these newer generation TDO2 inhibitors for anti-cancer treatment remains to be determined.Studies using HepG2 liver cancer cells demonstrated that TDO2-inhibitors 680C91 or LM10 significantly reduced Trp degradation and that TDO2, but not IDO1, was the primary source of Kyn production in these cells (144).680C91 has also been explored in other cancer types including melanoma (145)(146)(147), colon (147), leiomyomas (148), and gliomas (149).
Interestingly, Zhang and colleagues reported that sodium tanshinone IIA sulfonate (STS), a sulfonate derived from tanshinone IIA (TSN), reduced the enzymatic activities of IDO1 and TDO2 with a half inhibitory concentration (IC50) of less than 10 mM using enzymatic assays (150).STS markedly decreased Kyn synthesis in IDO1-or TDO2-overexpressing cell lines and reduced the percentage of Forkhead Box P3 (FOXP3) T cells.In vivo, STS delayed tumor growth and combination treatment of STS with anti-PD-1 yielded superior anti-cancer efficacy compared to either treatment alone (150).
With regards to dual IDO/TDO2 targeting therapies, preclinical studies demonstrated that RG70099, potent dual small molecule inhibitor (IC 50 <100nM) (151), efficiently reduced Kyn levels in plasma by ~90%, and >95% in tumors and draining lymph nodes.Notably, compared to conventional IDO inhibitors, RG70099 was able to reduce tumor Kyn levels by >90% in a TDO+ tumor xenograft, reinforcing that RG70099 exerting TDO2 inhibitory activity (151).EPL-1410, a fused heterocycle-based analogue, showed IDO1/TDO2 inhibitor activity in biochemical assays and demonstrated a significant dose dependent pharmacological efficacy in reducing the tumor volume in the syngeneic cancer models of CT-26 colon carcinoma and B16F10 melanoma (152).Recently, a Phase I clinical trial of M4112, an oral small molecule dual inhibitor of IDO/TDO2, in 15 patients with advanced solid tumors reported tolerable toxicity profiles (153).Treatment-emergent adverse events included fatigue, nausea, and vomiting.Dose-limiting toxicities (DLTs) were observed in one patient (grade 3 allergic dermatitis), and one grade 2 QT prolongation that resulted in a dose reduction.Neither the maximum tolerated dose (MTD) nor recommended Phase II dose was achieved (153).

Kynurenine aminotransferase inhibitors
To date, there exist several selective inhibitors of KAT, however, their applications in the context of anti-cancer treatment has not been readily explored (154)(155)(156).Despite this, several cancer types overexpress KAT (5), and, as mentioned in the above sections, KAT-derived KYNA may serve as a ligand for AhR signaling, resulting in tumor immune suppression (101), Evaluation of KAT inhibitors for anti-cancer treatment and potential reversal of immune suppression warrants further exploration.
However, targeting of KYNU should be met with caution.KYNU mediates the catabolism of Kyn to AA, which is thought to be immunologically inert.Administration of PEGylated-KYNU in immune-competent models of melanoma, breast, and colon cancer resulted in drastic reductions in tumor KYN levels with concomitant elevations in KYNU-derived AA, resulting in increases in tumor-infiltrating CD8+ T cells and subsequent tumor reduction (8).Moreover, anti-cancer efficacy was further enhanced when combining PEG-KYNase with ICIs (8).Overexpression (OE) of KYNU in modified Chimeric Antigen Receptor (CAR)-T cells yields superior anti-cancer efficacy in a NALM6 acute lymphoblastic leukemia mouse model (164).Compared to KYNU-knockout (KO) CAR T cells, KYNU-OE CAR T cells had higher glucose uptake, increased proliferation in carboxifluorescein diacetate succinimidyl ester (CSFE) assays, and exhibited a higher percentage of effector memory cell and effector CAR T-cells.In cell killing assays, KYNU-OE CAR T cells had higher lytic granules of granzyme B and enhanced cytokine production of IFNg, compared to KYNU-KO CAR T cells.Importantly, KYNU-OE CAR T cells retained anti-cancer efficacy in the presence of high KYN levels (164).In another study, an organic polymer nanoenzyme (SPNK) conjugated with kynureninase (KYNase) via PEGylated singlet oxygen (1O2) cleavable linker with near-infrared (NIR) photoactivatable immunotherapeutic effects was developed for photodynamic immunometabolic therapy (165).Upon NIR photoirradiation, SPNK generates 1O2 to induce the immunogenic cell death of cancer and released KYNase to degrade KYN in the TME.In vivo, SPNK treatment elicited effector T cell tumor infiltration and expansion, enhanced systemic antitumor T cell activity, and prolonged survival of tumor-bearing mice (165).
The potential of KYNU as a direct target of therapy or as the therapy itself (i.e.pegylated-KYNU) remains an active area of exploration.

Kynurenine 3-Monooxygenase inhibitors
To date, several inhibitors of KMO have been developed, the majority of which share a common pharmacophore containing both an acidic moiety and a mono or 1,2-dichloro substitution of the core phenyl ring (84,(166)(167)(168)(169).More recent inhibitors, GSK065 and GSK366, trap the catalytic flavin in a previously unobserved tilting conformation, resulting in picomolar affinities, increased residence times, and an absence of the peroxide production observed with prior inhibitors (166).
Evaluation of KMO inhibitors for anti-cancer treatment remains limited; however, studies by Ray and colleagues demonstrated that Ro 61-8048, a potent, selective inhibitor of KMO (170), activates pDCs and enhances pDC-triggered T cell proliferation and enhanced anti-cancer activity (26).Specifically, in co-culture models of patient pDCs, T cells, or NK cells with autologous multiple myeloma (MM) cells, blockade of KMO via Ro 61-8048 resulted in a robust MM-specific CD8+ CTL activity as well as significantly increased NK cell cytolytic activity (26).Treatment of CRC cell lines with Ro 61-8048 or UPF 648, another KMO inhibitor, resulted in reduced viability, decreased sphere formation and attenuated cell migration and invasion (86).
The therapeutic potential of KMO inhibitors remains to be determined; however, a prior clinical trial of the KMO inhibitor GSK3335065 was conducted among healthy volunteers.While GSK3335065 rapidly increases kynurenine levels suggesting partial inhibition of KMO activity, the trial was terminated after a serious adverse event (SAE) occurred on the study and a relationship with the study drug could not be excluded (171).
Regarding QPRT, existing small molecule inhibitors include phthalic acid (PA) as well as the P2Y11 antagonist (NF340) (105,175,176).To this end, phthalic acid and NF340 were able to attenuate QPRT-enhanced invasiveness pCMV6-QPRT-transfected MDA-MB-231 TNBC cell lines (105).Given that NAD is essential for metabolic and immune homeostasis, QPRT inhibitors for cancer therapy warrants investigation, although should be met with some caution due to potential off target toxicities (177-179).

Perspectives
Over the past few decades, there has been considerable advancement in our understanding of the cancer cell intrinsic and complex extrinsic functions of the KP in tumor development and progression (Figure 3).This progress has led to the development of several small molecule inhibitors that exploit multiple facets of the KP to achieve anti-tumor effects and the reversal of tumor-derived immune suppression (Figure 3; Table 1).
Despite these advances, clinical benefit resulting from targeting the KP pathway has been less than anticipated.Limited clinical success is likely to be multifactorial.For instance, advanced stage tumors have already undergone immunoediting to enable immune escape.Benefit of IDO inhibitors, or other KP inhibitors, may be most efficacious during early stages of disease when the immune system iteratively selects and/or promotes the generation of tumor cell variants with increasing capacities to survive immune attack, also referred to as the equilibrium phase (180).
Lack of clinical benefit may also be attributed to patient selection.Tumors that co-express IDO2 and TDO2 may be less responsive to selective IDO1 inhibitors.Consequently, molecular profiling of tumors with specific evaluation of IDO1, IDO2, and TDO2 expression may better select for patients that would benefit from selective IDO1 inhibitors or from broader pan-IDO/TDO2 inhibitors (Table 1).Recent findings implicating a nonenzymatic role of IDO1 in signaling and immune suppression (44) necessitates existing or emerging inhibitors should also be evaluated for their effects on catalytic activity in addition to signaling functions to maximize potential for clinical efficacy.
It is also important to consider additional factors that may influence KP activity such as diet (67) and environmental contributors (181,182).Liang and colleagues demonstrated that nicotine-derived nitrosaminoketone (NNK) upregulates IDO1 via a7 nicotinic acetylcholine receptor (a7nAChR)-mediated activation of the factor c-Jun in lung tumors.In A/J mice, NNK reduced CD8+ T cells and increased Tregs.Moreover, individuals with non-small cell lung cancer who smoked had higher tumoral IDO1 levels and lower Trp/Kyn ratios compared patients who never smoked, further suggesting that smoking impacts KP activity (181).
Exposure to other environmental toxicants can also activate AhR signaling (182), which may upregulate IDO expression resulting in immune suppression (48).
of PD-L1, macrophage colony-stimulating factor (M-CSF) and KYNU, and accelerating catabolism of cytotoxic-labile heme Upregulation of tumoral KYNU, coupled with NRF2 activation may thus synergize to yield a TME that is highly immune suppressed.Whether therapeutic targeting of KYNU alone or in combination with ICI in NRF2-activated KYNU-high tumors reverses tumor immune suppression remains an area of ongoing inquiry.
In summary, the KP remains an active area of investigation, and insights gained from this study are expected reveal additional promising points actionable metabolic vulnerability within the pathway that can be targeted for anti-cancer treatment.

FIGURE 2 Forest
FIGURE 2Forest plot illustrating association between IDO1 and IDO2 mRNA expression in the TCGA PanCancer Atlas transcriptomic datasets.Nodes represent spearman correlation coefficients, corresponding 95% confidence intervals, and 2-sided p-values.

TABLE 1
Existing and emerging therapies that target the Kynurenine Pathway.