Human embryonic kidney 293T cells were cultured in DMEM. The mouse colorectal carcinoma cell line CT-26 and the acute leukemia cell line Jurkat were cultured in RPMI-1640 medium. Both media were supplemented with 10% fetal bovine serum and 1% penicillin/streptomycin in a humidified 5% CO₂ incubator at 37°C.

A 6X His tag for purification and a peptide of adenylate kinase (AK sequence: RIILLGAPGAGKGTQAQFIMEKYGIPQISTGDMLRAAVKSGSELGKQAKDIMDAGKLVTDELVIALVKERIAQEDCRNGFLLDGFPRTIPQADAMKEAGINVDYVLEFDVPDELIVDRIVGRRVHAPSGRVYHVKFNPPKVEGKDDVTGEELTTRKDDQEETVRKRLVEYHQMTAPLIGYYSKEAEAGNTKYAKVDGTKPVAEVRADLEKILG) to improve the yield were added at the N-terminus of proteins IDO1 and TDO2 (His-AK-IDO1 or His-AK-TDO2) (Luo et al., 2016). The primers (F1: cgc​ata​tgg​gca​gca​gcc​atc​atc​atc; R1: cca​tag​ctg​gag​ctt​tat​cat​cat​cat​cac​cag​aac​cac) were used for PCR of His-AK from the vector provided kindly by Dan Luo. The primers (F2: tag​tgg​ttc​tgg​tga​tga​tga​tga​taa​agc​tca​cgc​tat​gga​aaa​ctc​ttg; R2: cgc​agc​tct​taa​cct​tct​ttc​agc​aga​gat​ttt​tc) were used for clone IDO1. For clone His-AK-TDO2, the primers (F1: cgc​ata​tgg​gca​gca​gcc​atc​atc​atc; R1: ttt​atc​atc​atc​atc​acc​aga​acc​act​a) were used for PCR for His-AK, and the primers (F2: tag​tgg​ttc​tgg​tga​tga​tga​tga​taa​aag​cgg​gtg​tcc​gtt​ttt​agg​g; R2: cgg​agc​tct​taa​tcg​ctt​tca​tcg​ctg​cta​aaa​tag) were used for clone TDO2. Two fusion fragments were cloned into PET25b digested with NdeI and Sacl. After 0.5 mM isopropyl-β-D-thiogalactopyranoside (IPTG) induction at 16°C for 8 h, the cells were harvested by centrifugation at 8,000 rpm/min for 5 min. To increase the expression of TDO2, 40 μM hemin was added to the Luria-Bertani (LB) medium when induced with IPTG. After centrifugation, the cell pellet was resuspended in an appropriate volume of lysis buffer [50 mM Tris-HCl, pH = 8.0, 100 mM NaCl, 5 mM EDTANa₂, 2 mM DTT, cell pellet mass (g): buffer volume (mL) was 1: 20] and lysed by sonication. Then, 13,500 rpm/min for 15 min was implemented at 4°C to remove cell debris, and the protein was solubilized in the supernatant. Protein purification was carried out using the Ni column (Bestchrom, Shanghai, China).

The enzymatic assays were carried following the methods described in the reported procedures (Malachowski et al., 2016). Briefly, the 100 μL reaction mixture contained 0.05 M potassium phosphate buffer (PBS), pH = 6.5, 40 mM vitamin C, 0.2 mg/mL H₂O₂, 0.02 mM methylene blue, and 0.08 mM tryptophan, and with 100 μg/mL IDO1 or 200 μg/mL TDO2. First, the reaction solution and compounds were incubated for 30 min at 37°C. Then, the reaction was stopped with 80 μL of 30% (w/v) trichloroacetic acid at 65°C for 15 min. The solution was then centrifuged at 12,000 rpm/min for 10 min, and the supernatant was mixed with an equal volume of 2% (w/v) p-dimethylaminobenzaldehyde in acetic acid. Finally, the kynurenine in the yellow product was measured using a multifunction microplate reader (Synergy H1, BioTek, United States) at 490 nm.

To overexpress IDO1 and TDO2 in cells, the primers (F1: atg​gca​cac​gct​atg​gaa​a; R1: tta​acc​ttc​ctt​caa​aag​gga; F2: cac​cga​ctc​tag​aac​tag​tga​tgg​cac​acg​cta​tgg​aaa; R2: gcc​agt​aac​gcg​atc​gaa​ttt​taa​cct​tcc​ttc​aaa​agg​ga) were used to clone IDO1 into the CPPT-puro vector by digestion with BamHI and EcoRI. The primers (F1: atgagtg ggtgcccattttta; R1: tta​atc​tga​ttc​atc​act​gct​ga; F2: cac​cga​ctc​tag​aac​tag​tga​tga​gtg ggtgcccatt ttta; R2: gcc​agt​aac​gcg​atc​gaa​ttt​taa​tct​gat​tca​tca​ctg​ctg​a) were used to clone TDO2. The cell-based assay procedure were performed under the protocol described in the literature (Malachowski et al., 2016). Briefly, IDO1-or TDO2-overexpressing 293T cells were harvested and plated at a density of 2.5 × 10⁴ cells/well in a 96-well culture plate. Simultaneously, a serial dilution of STS in 50 μL culture media containing 100 μM tryptophan was added to the wells. After 24 h, 75 μL of supernatant was transferred to a new 96-well plate and mixed with 35 μL of 50% trichloroacetic acid in each well, and the plate was incubated at 65°C for 15 min turning formylkynurenine to kynurenine. The reaction mixture was centrifuged for 10 min at 13,000 g to remove the deposits. Then, 50 μL supernatant from each well was transferred to another 96-well plate and mixed with an equal volume of 2% p-dimethylaminobenzaldehyde in acetic acid. A microplate reader was used to measure the yellow reactions at 490 nm. Additionally, the cell activity was detected by the MTT assay following the transfer of supernatant.

To investigate the effect of STS on cell proliferation, 20 μL of 2.5 g/L MTT solution was added to each well of a 96-well plate and incubated at 37°C for 4 h. After lysing formazan in 100 µL of 10% (w/v) SDS solution, the absorbance at 570 nm was measured using a multidetection microplate reader.

IDO1- or TDO2- overexpressing and normal 293T cells were washed with precooled PBS and lysed in RIPA solution with a protease inhibitor cocktail for 30 min on ice. After centrifuging at 13,500 g at 4°C for 15 min, the supernatant was quantified by BCA (Biyuntian, China) and denatured at 100°C for 10 min. SDS-PAGE was used to separate the protein samples. Then, the samples were transferred to a 0.22 μm PVDF membrane at 100 V for 90 min. After blocking the membrane with TBST containing 5% BSA for 1 h at room temperature, it was incubated with the primary antibodies overnight at 4°C and then with secondary antibodies for 1 h. Proteins were visualized using ECL reagent (Millipore, United States). Finally, the gel imaging system (Bio-Rad, United States) was used for imaging. The antibodies used for WB are listed below: rabbit monoclonal antibodies to TDO2 (Abcam, Cat # 259359, 1:1000 dilution), rabbit monoclonal antibodies to IDO1 (Abcam, Cat # 211017, 1:1000 dilution), and mouse monoclonal antibodies to GAPDH (Zhongshan Jinqiao, 1:1000 dilution) as the reference.

ChemBio3D Ultra 14.0 (Cambridge Soft, Cambridge, MA, United States) was adopted to construct the structures of IDO1 and TDO2. Then, energy majorization of the molecular structures was performed using the MM2 force field such that the convergence condition of the root mean square (RMS) was less than 0.0001 kcal mol⁻¹ Å⁻¹. Following an initial minimization of the molecule, the restrained electrostatic potential (RESP) charges of the molecule and precise optimization [at the b3lyp/6-311++G (d,p) level] calculations were performed using the Gaussian 09 software combined with the multiwfn software (Frisch et al., 2009; Lu and Chen, 2012). After optimization, the optimized molecules were docked to the binding pockets of the receptor by the AutoDock4.2 software package based on the Lamarckian genetic algorithm (LGA), which evaluated docking results by searching for conformations and the semiempirical free energy scoring function (Hou et al., 2013). In this work, we built a 50 Å × 50 Å × 50 Å rectangular box with a grid space of 0.375 Å, and the box centers of IDO1 (PDB entry: 2D0T) and TDO2 (PDB entry: 2NW8) were set as x: 58.506/40.593, y: 52.978/−48.557 and z: 16.523/30.523, respectively. The GA-LS running set as 128 in all docking and the structure with the lowest energy in the largest cluster was taken as the final result.

Six-week-old female BALB/c mice were purchased from Weitong Lihua Experiment Animal Technology (Beijing, China). All animal experiments were performed according to the protocols approved by the Ethics Review Committee of Animal Experimentation of Sichuan University. CRC cells CT26 (5 × 10⁵) in 100 μL D’-HANKS were inoculated subcutaneously into one flank of each mouse to examine the therapeutic potential of STS and STS combined anti-mouse PD1 [Gifted by Conmed Biosciences Inc. (Chengdu, China)] on antitumor immune responses in vivo. When the tumor size reached ≈80 mm³, mice were randomized into four treatment groups (n = 5 per group). The mice received intraperitoneal injections of the following drugs: 1) saline; 2) STS (20 mg/kg, q.d.); 3) anti-PD1 (3 mg/kg, q.w.); and 4) STS (20 mg/kg, q.d.) + anti-PD1 (3 mg/kg, q.w.). Tumor volume was measured every 2 days using a Vernier caliper and computed using the formula: volume = (tumor length) × (tumor width) ²/2. On day 27, tumors were collected for flow cytometry and IF analysis. ELISA was used to measure kynurenine in plasma and IFN-γ in tumors.

The tumor fragments were fixed in 10% formalin before being embedded in paraffin. Paraffin slides were then deparaffinized with xylene and ethanol. Tumor tissues were cut into 3 μm pieces and incubated at 62°C for 4 h. Heat-mediated antigen retrieval was performed in EDTA buffer (pH = 9.0) for 16 min, followed by 15 min in 3% hydrogen peroxide, and then overnight in a wet container at 4°C with primary CD8 antibody (Abcam, Cat# 217344, diluted 1:2000). Furthermore, the paraffin sections were treated with CD4 antibody (Abcam, Cat# 183685, diluted 1:1000) and FOXP3 antibody (Invitrogen, Cat# 14577380, diluted 1:100) after being permeabilized with 0.3% Triton X-100 for 15 min. Relevant secondary antibodies conjugated with Alexa Fluor 488 (Invitrogen, Cat# A21206, diluted 1:400) and Alexa Fluor 594 (Invitrogen, Cat# A48264, diluted 1:400) were added and incubated for 1 h in darkness at room temperature. Hoechst 33342 (Biyuntian, Cat# C1011) was added to a final concentration of 10 μg/mL (Biyuntian, Cat# C1011) and incubated for 10 min at room temperature in the dark to stain the nucleus. Nikon Ni-E was used to capture immunofluorescence images.

To study the effect of STS on the proliferation of lymphocytes, the spleen was harvested from a normal BALB/c mouse after being sacrificed and processed with a 70 μm Nylon cell strainer (Biofil, Guangzhou, China). According to the manufacturer’s instructions, lymphocytes from the spleen were isolated with Mouse Lymphocyte Separation Medium (Dakewe Biotech Company Ltd., Shenzhen, China). Then, the lymphocytes were stained with 2.5 μM and 100 μL CFSE for 8 min at room temperature. Following staining, a 500 μL RPMI-1640 medium was added to block the CFSE-labeling reaction, and the cells were washed twice with a 500 μL RPMI-1640 medium. Next, 3 × 10⁶ lymphocytes per well were seeded into a 12-well plate that had been precoated with anti-CD3 antibody (Biolegend, Cat# 100359, 10 μg/mL). After cell seeding, the STS formulated with medium containing CD28 antibody (Biolegend, Cat# 102121) and IL-2 (Peprotech, Cat# 212-12) was added. After culturing for 48 h, the CFSE-positive cells were detected utilizing a flow cytometer.

To study the effect of STS on Tregs, the CT26 cells (1 × 10⁵/well) were seeded into a 12-well plate with 1 mL of medium containing 100 ng/mL of mouse IFN-γ (Peprotech. Cat# 315-05). In the meantime, the cells were treated with different concentrations of STS (0, 10, 25, 50 μM) and incubated at 37°C. After 24 h, the medium was withdrawn, and new lymphocytes were introduced with anti-CD3, anti-CD28, IL-2, and concentration gradient STS (0, 10, 25, 50 μM). After 48 h of coculture, the cells were harvested. Before analysis, cells were treated for 15 min at room temperature with the Zombie UV Fixable Viability Kit (BioLegend, Cat# 423108) to exclude dead cells. After washing, anti-CD45-APC-Cy7 (BioLegend, Cat# 103116) and anti-CD4-FITC (BD, Cat# 557307) were applied for a 30 min incubation at room temperature. The cells were then fixed and permeabilized using the True-NuclearTM Transcription Factor Buffer Set (BioLegend, Cat# 424401). Finally, the cells were incubated at room temperature for 30 min with anti-FOXP3-PE (Invitrogen, Cat# 12-5773-80). Following washing, the percentage of Zombie- CD45+ CD4+ FOXP3+ T cells was analyzed by flow cytometry.

To study the effect of STS on the percentage of cytotoxic T cells in tumors, tumor samples were cut into pieces with scissors and digested for 30 min at 37°C in RPMI 1640 containing 0.2 mg/mL collagenase type I (Gibco, Cat# 12-5773-80) and type IV (Gibco, Cat# 17104019). Tissues that had been digested were gently filtered through a 70 μm cell strainer. Single-cell suspensions were centrifuged for 5 min at 600 g/min and washed twice with D’-HANKS buffer. Then, anti-CD3-FITC (Biolegend, Cat# 100204, 1:100) and anti-CD8-PE (Biolegend, Cat# 553032, 1:100) antibodies were employed for cell staining at room temperature for 30 min. The cells were then stained with 7AAD Viability Staining Solution (Biolegend, Cat# 420404) for 5 min at room temperature before being analyzed by flow cytometry.

To detect the concentration of kynurenine in plasma, we employed a mouse kynurenine ELISA kit (Yanjin Biology Company, Cat# F07945) for examination. First, after collecting approximately 600 μL of blood from mouse eyeballs, serum was collected by centrifugation at 500 g at 4°C for 5 min. Purer serum was obtained after an additional 15 min of centrifugation at 4000 rpm/min. Serum samples were determined after being diluted five times.

To measure the concentration of IFN-γ in tumor tissue, 0.1 g of the tissue block was weighed and placed in 1.5 mL EP tubes before adding PBS [tissue weight (g): PBS volume (mL) = 1:9]. Next, a small scissor was used to cut the tissues as soon as possible on ice. After grinding with a high-throughput tissue grinder (SCIENTZ-48, Zhejiang) at 70 Hz for 30 s at 10 s/time, the samples were centrifuging at 3000 rpm/min for 15 min. An ELISA kit (Biolegend, #1210002) was used to detect IFN-γ in the supernatant.

The hematoxylin and eosin (H&E) staining was performed according to standard procedures. Briefly, internal organ paraffin sections were dewaxed to water. Hematoxylin (Thermo, Cat# 7211) was used to counterstaining for 30 s, followed by rinsing with running water for 5–10 min to remove the blue color. After rinsing, 10 s with 1% hydrochloric acid alcohol before rinsing for 5 min with running water. Eosin (Thermo, Cat # 7111) staining was performed for 10 s, washed twice with running water for 5 min, dehydrated with graded alcohol, rendered transparent with xylene, and mounted with neutral gum. Nikon Ni-E was used to capture images.

One-way ANOVA analysis on GraphPad Prism 9.0.0 was used for statistical analysis in this work. The results were all reported as the mean ± SD. Statistical significance thresholds were set at *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001.

Considering the importance of IDO1 and TDO2 inhibitors in tumor immunotherapy, we developed an extracellular inhibitor screening model. First, we employed a prokaryotic protein expression system to produce the full-length IDO1 and TDO2 proteins. To facilitate purification, IDO1 and TDO2 carried an N-terminal His-tag. To improve the yield of IDO1 and TDO2, a peptide of Escherichia coli adenylate kinase (AK) was added between the His-tag and IDO1 or TDO2 proteins. Then, the proteins were purified using Ni-affinity chromatography. After that, the activities of IDO1 and TDO2 to catalyze the degradation of tryptophan were initially tested. Despite the presence of His-tag and AK peptide, the purified protein retained evident activity. Using the inhibitor screening approach described in the procedures (Figure 1A), we discovered that several compounds, including mangostin and androstenedione, could inhibit the activity of IDO1 (Supplementary Figures S1A,B), and tanshinone IA inhibited TDO2 activity (Supplementary Figure S1C). The IDO1 inhibitory activities of mangostin and androstenedione were firstly found by us. Remarkably, STS (Figure 1B), a derivative of TSN extracted from the dried roots of Danshen, could simultaneously inhibit the enzymatic activities of IDO1 and TDO2 (IC₅₀ < 10 μM). Moreover, it was interesting to know that TSN only reduced the activity of TDO2 but not IDO1 (Figure 1C).

To investigate the molecular basis of the inhibitory activity, molecular docking analyses were conducted. STS was docked into the binding pocket of the crystal structure of IDO1 (PDB access code: 2D0T) (Sugimoto et al., 2006) or TDO2 (PDB access code: 2NW8) (Forouhar et al., 2007). As illustrated in Figure 1D, the carbonyl of STS interacted with the heme iron in the tryptophan-IDO1 binding pocket, which was critical for IDO1’s enzymatic activity (Basran et al., 2021). Furthermore, STS could interact with PHE-163 and SER-167, which may compete with tryptophan binding to IDO1 (Figure 1D). For TDO2, STS was only linked to ARG-117 (Figure 1E), which also influenced the coupling of tryptophan and TDO2 (Rafice et al., 2009).

To detect the effect of STS on IDO1 and TDO2 activities at the cellular level, we established the IDO1- or TDO2-overexpressing cell lines IDO1-293T and TDO2-293T by the lentiviral introduction of vectors to further validate the dual inhibitory activity of STS against IDO1 and TDO2. After 48 h lentiviral transfection, the expression levels of IDO1 and TDO2 were increased compared with the control. We also noted that the primary expression levels of IDO1 and TDO2 in the 293T cells were weak compared with those in overexpressing cells by WB (Figures 2A,B). The presence of kynurenine in the culture medium further confirmed this conclusion (Supplementary Figure S2). When kynurenine was detected in the medium, we discovered that STS had a significant promoting effect on the proliferation of IDO1- or TDO2-overexpressing 239T cells, particularly at high concentrations (Figures 2C,D). In the IDO1-293T and TDO2-293T cell culture medium, the relative level of kynurenine was impeded by 24 h of 100 μM STS treatment after considering the impact of STS on cell proliferation when we measured the level of kynurenine in the medium (Figures 2E,F).

It has been proven that kynurenine from tryptophan metabolism in cancer cells can induce the expression of FOXP3, a marker of regulatory T cells (Tregs), by activating the AhR nuclear transform to suppress tumor immunotherapy (Roncador et al., 2005). We wondered whether STS could weaken the effect of tumor cells on Treg production through the IDO-kynurenine-AhR axis in vitro. To be clear, we utilized a cocultured system of CT26 and mouse spleen lymphocytes to preliminary mimic the effect of tumor cells on T cells. Prior to this, we used the MTT assay to examine the growth of CT26 cells after 72 h of STS administration and observed that STS had little effect on CT26 cell growth (Supplementary Figure S3). Then, after 48 h of STS treatment, flow cytometric analysis showed that STS (50 μM) could lower the percentage of FOXP3+ CD4+ T cells in the cocultured system (Figures 3A,B).

TSN has been reported to have antioxidative, anti-inflammatory, and cytotoxic properties in multiple types of human cancer cells (Su et al., 2008). In addition, TSN can decrease lymphocyte proliferation (Zhang et al., 2016). As a result, it is worth considering whether STS can inhibit lymphocyte cell proliferation similar to TSN. To make it clear, we first assessed the effect of STS on Jurkat cell proliferation by MTT assay. According to our findings, even at high concentrations (100 μM), STS treatment for 72 h had only a minor effect on Jurkat cell proliferation (Figure 3C). Furthermore, the lymphocyte CFSE staining assay showed a similar result: after treatment for 48 h, STS scarcely affected the proliferation of the lymphocytes (Figure 3D).

Several publications have reported that TSN inhibits tumor growth in vivo, including in CRC (Wang et al., 2005; Su and Lin, 2008; Chiu et al., 2013). However, there is no evidence indicating that STS inhibits tumor development in vivo. To investigate the effect of STS on tumor growth in vivo, we developed a murine model of CRC via subcutaneous injection of CT26 cells. When the tumor volume reached ≈80 mm³, the mice were randomly assigned to four therapy groups (n = 5): PBS; STS; anti-PD1; and STS + anti-PD1. The dosage of anti-PD1 (3 mg/kg, once a week) was according to the library (Liu et al., 2019) with a minor modification. All the treatments were intraperitoneal injections (Figure 4A), and the tumor volumes were measured every 2 days. The mice were euthanized on day 27, and the excised tumors were photographed, measured in volume, and weighed. The tumor growth curves showed that STS monotherapy better suppressed tumor development than PBS and enhanced therapeutic effect of anti-PD1 (Figure 4B). As expected, the curative effect of combined therapy with STS plus anti-PD1 was significantly better than other therapies in retarding tumor development in terms of tumor volume and weight (Figures 4C–E).

After anti-PD1 therapy, the plasma kynurenine level of melanoma and renal cancer patients are increased (Li et al., 2019a). This phenomenon was also observed in our results. In plasma, the level of kynurenine from anti-PD1 therapy was partially increased compared with PBS (Figure 4F). As predicted, compared with anti-PD1 monotherapy, comnination with STS had a lower amount of kynurenine in plasma (Figure 4F).

Kynurenine raises the quantity of Tregs, which harms the effective of immune response, as is generally accepted (Cheong and Sun, 2018). Therefore, decreasing Tregs may be commendably benefit to tumor immunotherapy (Munn et al., 2018; Tanaka and Sakaguchi, 2019). We accordingly investigated the number of Tregs in tumors after therapy. Compared with anti-PD1 monotherapy, STS decreased the number of FOXP3+ CD4+ T cells in combination therapy (Figures 4G,H).

The number of tumor-infiltrating CD8+ T cells (TILs) is closely related to the degree of response to immunotherapy (Dahlin et al., 2011; Paijens et al., 2021). We intended to explore whether STS therapy could increase the number of CD8+ T cells in tumors. Immunofluorescence assay of tumor tissues revealed that the combination therapy group had the maximum numbers of CD8+ T cells compared with the other three groups (Figures 5A,B). Flow cytometry also showed consistent results (Figures 5C,D). Furthermore, IFN-γ in tumor tissues also plays an important role in tumor immunotherapy. Therefore, we also detected the level of IFN-γ in tumor tissues by ELISA. A higher IFN-γ concentration was observed in STS and anti-PD1 combination therapy (Figure 5E). The findings mentioned demonstrated that STS could enhance immunotherapy for CRC.

In addition, we stained and inspected the main organs (heart, liver, spleen, lung, and kidney) of mice using hematoxylin and eosin (H&E) to evaluate the toxicity in vivo. There was no indication that STS, either alone or in combination with anti-PD1, produced obvious organ damage (Figure 5F).