Amitriptyline revitalizes ICB response via dually inhibiting Kyn/Indole and 5-HT pathways of tryptophan metabolism in ovarian cancer

Summary

Reprogramming tryptophan metabolism (TRP) may be able to overcome immunosuppression and restore the immune checkpoint blockade (ICB) response in patients with epithelial ovarian cancer (EOC) resistant to ICB therapy because TRP metabolism is involved in the kynurenine/indole and serotonin pathways of tryptophan metabolism. Herein, employing amitriptyline (AMI), an antagonist of TLR4 and serotonin transporter (SERT), we revealed that AMI remodels the immunological landscape of EOC. In particular, AMI lowered the expression of IDO1, IL-4I1, and PD-L1, the quantity of KYN and indoles, and the level of immunosuppressive immune cells MDSC, Tregs, and CD8⁺CD39+/PD-1+ T cell. AMI boosted the killing potential of anti-PD-1-directed CD8+T cells and worked in concert with PD-1 inhibitors to suppress tumor growth and to prolong the survival of EOC-bearing mice. This work highlights AMI as an effective regulator of ICB response by manipulating EOC cell TRP metabolism, indicating it could be a potential strategy for improving EOC ICB therapy.

Graphical abstract

Highlights

• •TRP/5-HT metabolic checkpoints IDO1, IL-4I1 and TGM2 mediate T cell dysfunction
• •Kyn/Indole and 5-HT pathways contribute to the resistance of EOC to ICB response
• •AMI down-regulates TRP/5-HT metabolic checkpoints, reviving suppressed CD8⁺ T cell
• •Dual targeting of ICs and the metabolic checkpoints improves immunotherapy for EOC

Cancer; Immune response; Microenvironment

Introduction

Epithelial ovarian cancer (EOC) is characterized by high lethality.¹ Patients with EOC, accounting for about 90% of all cases, are often diagnosed with advanced-stage disease (FIGO stage III-IV), and the prognosis for patients is poor, culminating in a 5-year overall survival of approximately 40%–45%.^2,3 Although primary cytoreductive surgery followed by platinum-based chemotherapy effectively induces initial remission and eventually by maintenance with bevacizumab and PARP inhibitors, most women will ultimately succumb to recurrent and therapy-resistant disease.^4,5 Immune checkpoint blockade (ICB) can unleash antitumor immunity and mediate durable cancer regressions in patients with various malignancies. However, EOC has benefited very little from this success story. In the multiple clinical trials of ICB therapy completed to date, the objective response rates (ORR) to PD-1/PD-L1 checkpoint inhibitors have been disappointingly low at only 8%–15% in EOC.^6,7,8 These clinical trials targeting ICB in recurrent EOC failed. So, EOC remains a cancer with no ICB specific approvals. These negative results showed a lack of activity of ICB in EOC. The response of ICB therapy to EOC has exceedingly limited, as the targets and mechanisms of resistance to ICB response by EOC-induced tumor immune suppressive microenvironment (TIME) remain unclear. EOC not only undergoes genetic and epigenetic changes but also generates a unique TIME that can regulate the interaction between the tumor and the host’s immune system^9,10,11 and leads to persistent systemic immunosuppression. Tumor development dynamically reshapes the composition and function of the immune microenvironment in which all the T cells in the tumor microenvironment (TME) and peripheral T cells are exhausted and irreversibly dysfunctional, thus, T cell-targeted interventions alone are not sufficient, and systemic immunosuppression can not be completely reversed with surgical tumor resection as well.^12,13,14

It is critical to find new strategies to overcome barriers to effective immunotherapy and revitalize ICB response. In recent years, it has been evidenced that tumors use tryptophan (TRP) metabolites to their advantage by promoting the mobility of tumor cells and weakening the immune response against tumors and that most free TRP is degraded into several biologically active compounds through the kynurenine (KYN) pathway (KP) or serotonin pathway. TRP metabolism immunosuppressive checkpoints (TMICs), or TRP degrading enzymes indoleamine 2,3-dioxygenase 1 (IDO1), tryptophan-2,3-dioxygenase (TDO), interleukin 4 induced protein 1 (IL-4I1) and TRP hydroxylase-1 (THP1), as well as TRP metabolites, have been shown to be important in causing immunosuppression in both TME and systemic immune.^{15,16,17,18,19} TRP metabolite and its derivatives as mediators of suppression of immune cells can impair the body’s defense against the tumor. Several studies indicated that the expression and activation of IDO1 and IL-4I1 via Toll-like receptor 4 (TLR4)/Myeloid differentiation primary response protein 88 (MYD88) pathway can increase the degradation of TRP to form derivatives such as Kynurenine (KYN)/Kynurenic acid (KYNA), and Indole-3-pyruvate (I3P)/Indole-3-acetaldehyde (I3A)/Indole-3-lactic acid (ILA)/Indole-3-acetic acid (IAA), leading to the aryl hydrocarbon receptor (AHR) activation-mediated impairment of immune function^{15,17,20,21,22,23,24} and that TRP hydroxylase-1 (TPH1) can catalyze or initiate the chemical reaction of converting TRP to serotonin (5-hydroxytryptamine, 5-HT), which upregulates expression of IDO1 and PD-L1 respectively through melanin/Forkhead Box Protein O1 (FoxO1) pathway²⁵ and Transglutaminase 2 (TGM2)-mediated serotonylation mechanism.^26,27 Serotonin upregulates IDO1 transcription through melatonin-mediated FoxO1, while TGM2 mediates 5-HT metabolism through Na+ dependent serotonin transporter (SERT/SLC6A4) to produce a "Serotonylation" effect and upregulate PD-L1 expression, inducing endogenous PD-1/PD-L1 inhibitor resistance.^25,28,29 These findings suggest that IDO1/KYN/AHR and IL-4I1/I3P/AHR axes, as well as the serotonin signaling pathway collectively drive immunosuppression in cancer. Accordingly, dual blocking of KP and serotonin signaling pathways of TRP metabolism might be a novel strategy to improve ICB therapy for EOC.³⁰ Interestingly, the antidepressant amitriptyline (AMI) is an antagonist of both TLR4 and SERT (serotonin transporter).^31,32 AMI is a widely used antidepressant drug to treat depression and is used in combination therapy with typical anti-tumor drugs which show a synergic effect,³³ and also used in the depression-related impaired regulation of immune system activity in patients with cancer .^34,35

In this work, we aimed to determine whether AMI modulates TRP metabolism-mediated immunosuppression in EOC. We found that AML down-regulated TMIC expression and TRP metabolism derivatives, and improved the response of ICB therapy to EOC.

Results

Amitriptyline down-regulates LPS-induced increased MYD88, Ido1, and IL4-I1 expression

LPS, as an agonist ligand of TLR4, has been documented to induce activation and expression of the MYD88/NF-κВ pathway and its downstream molecular targets in EOC cells, leading to paclitaxel resistance and IDO1-mediated immunosuppression.^36,37 To identify whether the blocking of TLR4/MYD88 pathway could down-regulate TLR4-mediated expression of MYD88, IDO1 and IL-4I1, we used AMI, a TLR4 antagonists, to investigate the expression of MYD88, IDO1 and IL-4I1 by EOC cells exposed to LPS. As shown (Figures 1A–1I), RT-qPCR and Western blot or ELISA results showed that treatment with indicated concentrations of AMI for 12 or 24 h significantly reduced LPS-induced expression of MYD88, IDO1 and IL-4I1 by OVCAR3 and ID8, not A2780 cells compared their corresponding control cells (p < 0.05), MyD88 and IDO1 were found to be lowly expressed in A2780, and these alterations did not significantly alter upon stimulation with LPS and AMI. Our findings suggest that AMI could inhibit the expression of IDO1 and IL-4I1 by EOC cells with positive TLR4/MYD88.

Figure 1.

Expression of MYD88, IDO1, and IL4-I1 in EOC cells after LPS-induced combination with AMI

(A−C) RT-qPCR and Western blot verification of MYD88 and IDO1 expression in A2780, OVCAR3, and ID8 EOC cells after LPS-induction (1 μg/ml).

(D−F) mRNA expression of MYD88 and IDO1 by the combination of LPS with AMI in A2780, OVCAR3 (AMI: 50, 100μm) and ID8 (AMI: 30, 60μm) EOC cells.

(G−H) MYD88 and IDO1 expression was not detected in A2780 EOC cells according to the pre-WB results. Therefore, only the expression of MYD88 and ID8 protein levels in OVCAR3 and ID8 were performed in the combined treatment. (Protein expression of MyD88 and IDO1 relative to Tubulin in EOC cells treated with different concentrations of drugs).

(I) mRNA expression of IL4I1 on A2780, OVCAR3, and ID8 EOC cells after LPS combined with AMI treatment was analyzed by RT-qPCR, in which IL4I1, as a secreted protein, was detected in the levels in the EOC cells culture supernatant with Elisa. Each assay was performed in triplicate, and data are presented as mean ± SD (n = 3).

Amitriptyline downregulates Ido1 and 5-HT/transglutaminase 2-mediated PD-L1 expression

5-HT signaling has been reported to promote expression of IDO1 and PD-L1 via SERT/5-HT/MT/FOXO1 and SERT/5-HT/TGM2/serotonylation pathways (Figure 2A). It has been shown that AMI is a TLR4 and SERT inhibitor, which can block TLR4/MyD88 signaling-mediated IDO1 expression,^31,32 and respectively reduce 5-HT-dependent melatonin and TGM2-mediated IDO1 and PD-L1 expression.^25,26,29 To identify the effect of AMI on the expression of IDO1 and PD-L1, we determined the expression of IDO1 and PD-L1 by EOC cells treated with or without 5-HT. RT-qPCR and WB analysis indicated that the levels of IDO1 and PD-L1 were elevated in A2780, OVCAR3, and ID8 EOC cells following exposure to 5-HT (100μM) (Figures 2B–2F). Although PD-L1 expression is generally low in OVCAR3 and ID8 cells, an increase in its protein expression can still be observed after 5-HT induction. IDO1 and PD-L1 protein expressions were not detected in A2780 cells, and A2780 cells lack MyD88 expression, however, in EOC cells, the TLR4/MyD88 signaling pathway induces the expression of IDO1 via an NF-κB dependent mechanism. Furthermore, EOC cells treated with 5-HT were subsequently exposed to AMI and TGM2 inhibitors. The administration of AMI led to a reduction in the expression levels of IDO1 and PD-L1, whereas the TGM2 inhibitor specifically decreased the expression of PD-L1 (Figures 2G–2J), suggesting that TGM2-mediated serotonylation upregulates PD-L1 in the presence of increased intracellular 5-HT that activates TGM2. Our findings suggest that increased intracellular 5-HT can induce the conversion of 5-HT into the melatonin and activation of TGM2 to upregulate IDO1 and PD-L1 via 5-HT/melatonin/FoxO1²⁵ and 5-HT TGM2/serotonylation pathways,²⁹ respectively. AMI targeting SERT-mediated 5-HT reuptake is conducive to reducing IDO1- and PD-L1-mediated ICB resistance in EOC.

Figure 2.

5-HT-mediated expression of IDO1 and PD-L1 after treatment with AMI and TGM2 inhibitors

(A) Schematic diagram of the tryptophan-kynurenine metabolism pathway.

(B and C) mRNA expression of IDO1 and PD-L1 in A2780, OVCAR3, and ID8 EOC cells exposed to different concentrations of 5-HT (50/100/200μM).

(D−F) Protein expression of IDO1 and PD-L1 in A2780, OVCAR3, and ID8 EOC cells exposed to 5-HT (100μM).

(G) AMI and TGM2 inhibitor (LDN-27219) were utilized respectively to inspect the mRNA expression of IDO1 and PD-L1 by A2780, OVCAR3, and ID8 EOC cells exposed to 5-HT (100μM).

(H−J) Protein level expression of IDO1 and PD-L1 was detected in A2780, OVCAR3, and ID8 EOC cells exposed to 5-HT (100μM) using AMI and TGM2 inhibitor (LDN-27219), respectively. (protein expression of IDO1 and PD-L1 relative to GAPDH in EOC cells treated with different drugs) Each assay was performed in triplicate, and data are presented as mean ± SD (n = 3).

Amitriptyline reduces the generation of tryptophan metabolism metabolites that activate aryl hydrocarbon receptor

As the above results indicate, AMI could suppress the expression or activity of TMICs IDO1 and IL-4I1 that yield TRP metabolites such as KYN and I3P, and so forth, and metabolic derivatives of KYN and I3P can induce immunosuppressive phenotypes of T cells by directly activating AHR on T cells. In view of this, we further determined the effect of AMI on the yield of tryptophan metabolites. Based on the results of PCR and WB, the highest dose of AMI showed the greatest downward trend, therefore the highest dose of AMI concentration was selected for tryptophan metabolite determination. As shown in (Figures 3A–3D), LC-MS/MS results demonstrated that AMI significantly reduced the amount of KYN, KYNA, IAA, I3A, and ILA, and the ratio of KYN/TRP in the supernatants of OVCAR3 and ID8 rather than A2780 cells in a time-dependent manner following exposure to LPS. TRP did not change significantly after 12 h of AMI treatment but decreased after 24 h of treatment, which was considered to be due to AMI inhibiting IDO1 and IL4I1, leading to compensatory increases in TDO and TPH1 to decompose TRP.³⁸ TRP is also required for the enhancement of immune cells and for the body’s requirement of essential amino acids. KYN/TRP ratio presents as a more robust biomarker, showing a decreasing trend in our experimental results, reflecting the fact that AMI may lead to a reduction in tryptophan metabolites through the TLR4/Myd88 pathway. However, I3P failed to be detected in the supernatant of EOC cells. Our findings suggest that AMI could intervene in TMICs and TRP metabolites that contribute to EOC cell-induced CD8⁺ T cell exhaustion.

Figure 3.

The effect of AMI on the yield of tryptophan metabolites and to study the role of AMI on T cells

(A−C) LC-MS/MS was used to detect the amount of KYN, KYNA, and the ratio of KYN/TRP in the culture supernatants of A2780, OVCAR3, and ID8 EOC cells after exposure to AMI induced by LPS.

(D) LC-MS/MS was used to detect the amount of AHR ligands (I3P, IAA, I3A, and ILA) in the culture supernatants of A2780, OVCAR3, and ID8 EOC cells after exposure to AMI induced by LPS.

Amitriptyline decreases epithelial ovarian cancer cell supernatant-induced CD8⁺/CD39⁺ T cells and reinvigorates the killing activity of T cells against epithelial ovarian cancer cells

Since T cell exhaustion is considered to be a biomarker for response to immunotherapy,^39,40 and an important factor affecting the efficacy of ICB therapy in patients with EOC. We examined whether the intervention of AMI in TRP metabolites could reduce T cell exhaustion and reinvigorate the cytotoxicity of T cells against EOC cells. To conduct cytotoxicity assay and flow cytometry analysis, we produced peripheral blood lymphocytes from humans and mice and treated them with various cell supernatants (see Figure S1). As shown in (Figure 4), our results showed that the percentage of CD8+/CD39+ exhausted T cells decreased (Figure 4A), the killing activity of T cells against EOC cells (Figure 4B) as well as the levels of TNF-α and IFN-γ (Figure 4C) increased significantly in healthy female lymphocytes co-incubated with the supernatants of OVCAR3, A2780 and ID8 cells treated with AMI, compared to that of the control cells without AMI treatment (p < 0.01). These findings indicate that AMI may alter TRP metabolites-mediated immunosuppression to reinvigorate T cells and benefit EOC to ICB therapy.

Figure 4.

The killing activity of T lymphocytes against EOC cells and expression of CD39 after AMI stimulation

(A) Expression of CD8⁺CD39⁺ in T lymphocytes after exposure to different concentrations of AMI-treated (AMI concentrations refer to the preceding) EOC cells culture supernatants for 12 or 24 h as analyzed by FCM.

(B) The killing activity of corresponding T lymphocytes (effector cells) to parental EOC cells (target cells) or AMI-treated EOC cells (target cells) shown are from the effector/target ratio at 1:1, 5:1, and 10:1. Supernatant 0h, 24h, and 48h represent the stimulation of T lymphocytes by supernatants collected at different times from EOC cells treated with AMI, respectively.

(C) Measure the levels of TNF-α and IFN-γ in lymphocyte culture supernatants stimulated with AMI-treated EOC cell culture supernatants by Elisa assays. Each assay was performed in triplicate, and data are presented as mean ± SD (n = 3). In Elisa results, ∗ indicates compared to the lymphocytes without any treatment (LN), # indicates compared to the lymphocytes stimulated with DMSO-treated EOC cell culture supernatant. #, ##, ### and #### indicates p < 0.05, p < 0.01, p < 0.001, and p < 0.0001 respectively.

Amitriptyline improves the efficacy of PD-1 inhibitors against epithelial ovarian cancer in vivo

Based on the above considerations, we further evaluated the tumor growth and survival of ID8 EOC-bearing female C57BL/6JGpt mice that underwent 3 cycles of chemotherapy with paclitaxel plus cisplatin (TP) followed by the treatment of PD-1 inhibitor combined with AMI (Figure 5A). We observed that the treatment of PD-1 inhibitor combined with AMI significantly repressed the tumor progression or even promoted the tumor regression following TP chemotherapy (Figures 5B–5E), Quantification of tumor volumes of different groups in day 30: Control, TP, TP + PD-1 inhibitor, TP + AMI and TP+AMI+PD-1 inhibitor groups were 939.65 ± 52.83, 621.24 ± 66.32, 574.00 ± 82.03, 458.47 ± 52.40 and 370.24 ± 22.72, respectively. And prolonged the median survival time with a decreased mortality of EOC-bearing mice (Figure 5F), PD-1 inhibitor combined with AMI has the longest survival period: 65.83 ± 5.27 days, compared to PD-1 inhibitor or AMI alone, suggesting that AMI could enable PD-1 inhibitor more effectively to exert ICB therapeutic response to EOC with less adverse effects (see Figures 5G and S2). Overall, these results showed that AMI potently enhances the in vivo antitumor effect of PD-1 inhibitor therapy.

Figure 5.

The therapeutic efficacy of combining TP and PD-1 inhibitors with AMI in xenograft models of EOC

(A) Timeline of experiments on animals.

(B) Body weight (g) and (C) Tumor volume (mm³) were recorded every 3 days, during treatment with TP + PD-1 inhibitor and/or AMI.

(D) Control, TP, TP + PD-1 inhibitor, TP + AMI, and TP+AMI+PD-1 inhibitor groups sizes of tumors.

(E) Quantification of tumor volumes of different groups in day 30.

(F) Survival curves of tumor-bearing mice treated in five groups.

(G) Level of alanine aminotransferase (ALT), aspartate aminotransferase (AST), urea and creatinine (Crea) in C57BL/6 mice’s serum. Data are presented as mean ± SD (n = 9 in B, C and E; n = 6 in F, n = 3 in G).

Amitriptyline reshapes tumor immune suppressive microenvironment and reduces the serum level of IL-4I1 and tryptophan metabolism metabolites of epithelial ovarian cancer-bearing mice

To verify whether the TMICs in tumor tissue and the serum level of TRP metabolites were altered in EOC-bearing mice receiving PD-1 inhibitor-combined AMI therapy, we checked the expression of IDO1, MYD88 and AHR in the tumor tissues, the serum level of IL-4I1 and TRP metabolites of EOC-bearing mice by using IHC staining, RT-qPCR, ELISA, and LC-MS/MS assays, respectively. As shown in (Figures 6A–6F), PD-1 inhibitor-combined AMI therapy significantly reduced the expression of IDO1, MYD88, and AHR in the tumor tissues. Control, TP, TP + PD-1 inhibitor, TP + AMI, and TP+AMI+PD-1 inhibitor groups % DAB Positive Tissue (MyD88): 2.27 ± 0.15, 7.76 ± 0.90, 3.51 ± 0.43, 1.33 ± 0.32, 0.80 ± 0.36, respectively (Figure 6A). % DAB Positive Tissue (AHR): 1.51 ± 0.73, 6.45 ± 0.43, 3.44 ± 0.35, 1.27 ± 0.13, 0.53 ± 0.08, respectively (Figure 6B). % DAB Positive Tissue (IDO1): 4.79 ± 0.43, 11.00 ± 2.04, 7.82 ± 0.45, 4.39 ± 0.57, 2.94 ± 1.05, respectively (Figure 6C). Comparison of serum tryptophan metabolites in unimplanted tumor mice and implanted tumor mice (Figures 7A and 7B), with slightly elevated serum tryptophan metabolites, KYN, KYNA, and KYN/TRP ratios in implanted tumor mice, and insignificant changes in TRP (some TRP was ingested in the food). (Figures 7C and 7D) show the comparison of tryptophan metabolites in the serum of normal mice (without implanted tumors) before and after the use of AMI, and the results showed that the changes in tryptophan metabolites were not significant after the use of AMI. The serum level of KYN, KYNA, TRP, IL-4I1, I3P, IAA, I3A, and ILA as well as KYN/TRP ratio, compared to PD-1 inhibitor alone or the control (Figures 7E–7J). A transient decrease in KYN was observed in a short-term evaluation of the efficacy of the addition of a PD-1 inhibitor in animal experiments, which may be attributed to alterations in the immunosuppressive microenvironment caused by the IDO-KYN-AHR metabolic circuit, which enhances the catabolism of the KYN pathway by TRP.⁴¹ These findings reflect that inhibition by AMI of TMICs expression and activity in EOC could lead to a reduction of TRP metabolites, thereby reshaping TIME and attenuating TRP metabolites-mediated systemic immunosuppression of EOC-bearing mice.

Figure 6.

Expression of MyD88, AHR, and IDO1 in tumor tissues

(A−C) Expression of MyD88, AHR, and IDO1 in tumor tissues of ID8 tumor-bearing mice after the above treatment by immunohistochemistry (The scale bar is 20μm, and the lower panel shows a 6× magnification view).

(D−F) Expression of MyD88, AHR, and IDO1 in tumor tissues of ID8 tumor-bearing mice after the above treatment by RT-qPCR. Data are presented as mean ± SD (n = 3).

Figure 7.

AMI reshapes TIME, the serum level of IL-4I1 and TRP metabolites peripheral blood of EOC-bearing mice

(A and B) Tryptophan metabolites in the serum of unimplanted tumor-bearing mice (normal mice) and compared with implanted tumor-bearing mice (controls).

(C and D) Tryptophan metabolites in the serum of normal mice (without implanted tumors) before and after the use of AMI.

(E−H) EOC-bearing mice receiving PD-1 inhibitor-combined AMI therapy, LC-MS/MS was used to detect the amount of KYN, KYNA, TRP, and the ratio of KYN/TRP in the serum of EOC-bearing mice.

(I) IL4I1 in EOC-bearing mice serum by Elisa.

(J) LC-MS/MS was used to detect the amount of AHR ligands (I3P, IAA, I3A, and ILA) in the serum of EOC-bearing mice. Data are presented as mean ± SD (n = 3).

AMI reduces the expression of MDSC, PD-1+T cell, and Treg subsets in the peripheral blood of EOC-bearing mice

Since AMI reduces the serum level of TRP metabolites associated with systemic immunosuppression,^42,43 we also investigated proportions and the phenotype of Tregs and MDSC in peripheral blood of EOC-bearing mice receiving PD-1 inhibitor-combined AMI therapy using FCM assay. As shown in (Figures 8A–8G), we observed that PD-1 inhibitor-combined AMI therapy showed a significantly reduced percentage of Tregs, MDSC, and CD8+/PD-1+ T cells subsets in the peripheral blood of EOC-bearing mice, compared to PD-1 inhibitor alone or the control. Control, TP, TP + PD-1 inhibitor, TP + AMI and TP+AMI+PD-1 inhibitor groups CD25⁺FoxP3⁺ of CD4⁺ T cells (%): 6.08 ± 0.23, 6.97 ± 0.19, 6.10 ± 0.24, 3.79 ± 0.45 and 1.30 ± 0.33, respectively (Figures 8A and 8D). M-MDSC: 23.35 ± 2.97, 21.72 ± 1.71, 18.66 ± 1.52, 17.28 ± 1.65 and 13.95 ± 1.54, respectively (Figures 8B and 8E). G-MDSC: 6.23 ± 0.98, 3.87 ± 0.46, 4.50 ± 0.21, 5.12 ± 0.82 and 4.95 ± 0.73, respectively (Figures 8B and 8F). PD-1⁺ of CD3⁺ CD8⁺ T cells (%): 68.49 ± 3.46, 75.68 ± 3.43, 63.31 ± 2.12, 64.71 ± 2.27 and 54.42 ± 4.19, respectively (Figures 8C and 8G). The results indicated that AMI could modulate TRP metabolites-mediated immunosuppression to improve the T cell function of EOC-bearing mice.

Figure 8.

Expression of MDSC, PD-1+T cell, and Treg subsets in peripheral blood of EOC-bearing mice

(A and D) The percentage of CD25⁺Foxp3 in CD4⁺ T cells of ID8 tumor-bearing mice in each treatment group was analyzed by flow cytometry.

(B, E, and F) FACS analysis of the percentage of CD11b⁺Ly6C^hi cell (M-MDSC) and CD11b⁺Ly6G^hi cell (G-MDSC).

(C and G) Detection of PD-1 expression on CD8⁺ T cells by flow cytometry. Data are presented as mean ± SD (n = 5).

Tumor-infiltrating lymphocytes' expression is changed by amitriptyline

In addition to examining MDSCs and T cells in peripheral blood, we also looked for tumor-infiltrating lymphocyte expression in tumor tissues. Using the immunofluorescence technique, the expression of cytotoxic T lymphocytes (CD8⁺T cells) and exhaustion T lymphocytes (CD8⁺CD39⁺T cells) was examined. The findings demonstrated a substantial increase in CD8⁺T cell expression in each treatment group as compared to the control group, and a significant decrease in CD8⁺CD39⁺T cell expression (Figures 9A–9C). The expression of CD8⁺lymphocytes was significantly increased in the combination therapy group compared to the monotherapy group. The PD-1 inhibitor monotherapy group had a smaller decline in CD8⁺CD39⁺T cells than the PD-1 inhibitor-combined AMI treatment group. Nonetheless, there was no statistically significant distinction in the expression of CD8⁺CD39⁺T cells between the PD-1 inhibitor monotherapy group and the PD-1 inhibitor-combined AMI therapy group. The results indicate that PD-1 inhibitor-combined AMI treatment can increase the expression of tumor infiltrating lymphocytes in tumor tissues.

Figure 9.

Expression of tumor infiltrating lymphocytes in tumor tissues of mice in each group

(A) Immunofluorescence images of CD8 and CD39 expression in the tumors of various mice groups: CD8+T cells are shown by red and red arrows, CD39+T cells by green and green arrows, and CD8⁺CD39+T cells by orange and orange arrows. The figure’s scale is 50μm.

(B and C) Bar graphs displaying the average CD8 expression fluorescence intensity and the number of CD8CD39/CD8 cells in each mouse group’s tumors. Data are presented as mean ± SD (n = 3).

Discussion

TRP metabolic disorder leading to immune suppression of TME and systemic immunity remains an unavoidable contributor to the failure of ICB therapy against the PD-1/PD-L1 system for EOC.^20,22 Reshaping TIME and systemic immunity by modulating TRP metabolism checkpoints might benefit ICB therapy for EOC, as shown by TRP metabolism checkpoint inhibitors-induced invigoration of the anti-tumor activity of T cells and ICB therapy in several cancers.^{17,18,19,21,26,30,44} However, the underlying mechanism by which TRP metabolism checkpoint inhibitors reshape the TME and systemic immunity to revitalize ICB response in EOC remains uncertain. Upregulation of TRP catabolism by IDO1/IL-4I1 represents one of the most studied and clinically validated pathways for immune suppression in tumors. IDO1 might constitute a potential escape mechanism from anti-PD1 ICB, however, due to their convergence on AHR, IL4I1 could constitute a resistance mechanism against IDO1 inhibition, indicating that IL4I1 could mediate IDO1 inhibitor resistance in the context of ICB, and vice versa.^20,45

Tryptophan metabolism involves three main pathways, KYN (IDO/TDO), 5-hydroxytryptamine (TPH1), and indole (IL4I1). The KYN pathway is the major tryptophan metabolic pathway, IDO expression was positively correlated with KYN, and the KYN/TRP ratio was a highly valuable marker of IDO activity.⁴⁶ TDO is the major TRP-degrading enzyme in the tryptophan metabolic pathway. In vivo, TDO is expressed primarily in the brain and liver.^47,48 In the whole animal in vivo, the role of TDO cannot be ignored，TDO is responsible for most of the tryptophan catabolism under normal conditions.

In this study, focused on the TLR4/MyD88 pathway in ovarian cancer, Approximately 70% of ovarian cancer cells have abnormally high expression of TLR4/MyD88, and its high expression is not only an independent factor in the poor prognosis of ovarian cancer,^28,49 and it was closely associated with T cell immunosuppression induced by high EOC expression of IDO1, IL4I1, PD-L1 and T cell AHR activation and secretion of immunosuppressive factors. IDO1 and IL4I1 in the tryptophan metabolic pathway are mainly mediated through the TLR4/MyD88 pathway and are expressed in ovarian cancer.⁵⁰

TRP metabolism regulates immunity through TLR4/MYD88/IDO1/KYN/AHR and IL-4I1/I3P/AHR pathways^37,51,52 and SERT/TPH1/TGM2-mediated serotonin signaling pathway.^26,27,30 KYN/KYNA and indoles as ligands of AHR induce an inhibitory phenotype of immune cells via binding to and activating AHR.²¹ AHR is a key sensor allowing immune cells to adapt to environmental conditions and changes in AHR activity have been associated with cancer.⁵³

To reverse IDO1, IL-4I1, and serotonin-mediated EOC immune suppression by the systemic depletion of TRP metabolic derivatives a dual inhibitor, AMI combined with PD-1 inhibitor was used to modulate the TMICs in EOC cells and tumor tissues as well as TRP metabolic derivatives in serum of EOC-bearing mice after exposure to cisplatin. Our findings showed that AMI significantly down-regulated the expression of MYD88 and IDO1 and reduced the serum IL-4I1, KYN/KYNA and I3P derivatives and immune cell subsets including MDSC, CD8+/CD39+, CD8+/PD-1+T cell and Treg in the peripheral blood, indicating that TRP metabolic reprogramming closely interacts to the phenotypic and functional shift of immune cells in the TME and systemic immune.^54,55,56 We further discovered that AMI enhanced CD8⁺ T cell cytotoxicity, revealing its immunoregulatory potential. Moreover, we found that AMI suppressed EOC cell uptake of serotonin via the inhibition of SERT and reduced serotonin-induced IDO1 expression, and also showed that the inhibition of TGM2 significantly reduced serotonylation-mediated PD-L1 expression in EOC. More recently, serotonin has demonstrated carcinogenic properties. This has sparked further research into its potential role at different stages of tumor progression and the utility of SERT inhibitors to prevent cancer growth. Serotonin is a growth factor for several human carcinomas and is also involved in cancer cell migration, metastasis, and progression.^57,58,59

Our results reinforced the notion that TLR4/MYD88/IDO1-IL-4I1 and serotonin signaling pathways jointly mediate immune suppression and AMI serves as a potentiator of antitumor immunity to revitalize the response of ICB therapy to EOC. The state of immune infiltration in cancer is associated with patient prognosis and the efficacy of ICB therapy for EOC.

Therefore, the modification effects of AMI on the immunosuppression in the TME and systemic immune might contribute to better prognosis in patients with EOC and improve the efficacy of ICB therapy for EOC. Our results showed that AMI had a marked effect on the TMICs of EOC, with a remarkable change in TRP metabolism. In addition, AMI combined with PD-1 inhibitor repressed tumor progression and extended the survival time of EOC-bearing mice. Several studies have suggested that TRP metabolites inhibit proliferation and activation, and even cause CD8+T cell exhaustion. Therefore, therapeutic approaches targeting tryptophan metabolism could represent a specific strategy to improve the immunosuppressive state of EOC. Current strategies to inhibit TRP metabolism mainly target the KYN pathway, the major catabolic pathway of TRP, to reduce the production of immunosuppressive TRP catabolites. However, this strategy has some limitations partly because of IL-4I1/I3P/AHR^51,52 and serotonin signaling.^26,27,30 Therefore, we focused on the KYN and I3P pathways as well as serotonin signaling between EOC cells and immune cells. Interestingly, AMI inhibited the production of KYN and I3P derived from TRP metabolism and reduced serotonin uptake-mediated increased PD-L1 and IDO1. According to the multi-target effects, AMI restores CD8⁺ T cell activity and thus promotes CD8⁺ T cell function. We emphasize that increased TRP degradation in EOC cells could impair the immune response to ICB, and highlight AMI as a safe and efficient strategy to target the TRP degradation process, thereby blocking immune evasion. Unlike IDO1 inhibitors and ICIs, AMI reduces the expression of IDO1, IL-4I1, and PD-L1 by dually blocking TLR4/MYD88/and serotonin signaling pathways. Our work has not only generated insights into TRP metabolism regulation by AMI but also identified TMICs as a mechanism of resistance of EOC to ICB therapy.

Compared to the general population, patients with tumor have a noticeably increased incidence of depression. Depression will worsen the psychological suffering that patients with tumor experience, lower their quality of life and adherence to treatment, and have an impact on their prognosis.^60,61 In addition, this study included patients with stage III or IV ovarian epithelial cancer. The experimental group received standard chemotherapy combined with AMI (Initiate dosage one day before to chemotherapy and continue for three days) and PD-1 inhibitors for 6 cycles, while the control group received only standard chemotherapy. Following treatment, there was no discernible difference in the occurrence of adverse events between the experimental and control groups: 62.3% for the former and 64% for the latter. This part of the data is not yet published in this article due to ongoing clinical study. The adverse effects of using AMI just during or after chemotherapy are within an acceptable range, according to the experimental evidence that is currently available, whether it is from mice or clinical patient data.

However, the combination of multiple drugs may produce cumulative adverse effects due to long-term antitumor therapy in patients with tumor and the combination of antidepressants. The use of TCA drugs during or after chemotherapy may produce, for example: nausea and vomiting, excessive sedation, insomnia, tachycardia and postural hypotension, and other adverse reactions, and may even produce anticholinergic side effects.^62,63,64 Although current research suggests that TCA based drugs can improve depression and even prolong the survival of patients with cancer with depression,⁶⁵ the accumulated adverse reactions are undoubtedly detrimental to patients. In the context of cancer treatment, it is crucial to find antidepressants with higher safety, fewer side effects, and better tolerability, as well as to detect the blood concentration of antidepressants to determine safe and effective drug doses, maximize anti-tumor effects, and minimize adverse reactions. To reduce the incidence of side effects, the interaction between antidepressants and concurrent anti-cancer treatment should be studied. Different methods of combination therapy should be selected for patients with different types of cancer, and basic research evidence is needed to improve the understanding of the mechanism of action, its possible efficacy, and interactions, in order to develop personalized and optimal combination treatment plans. The cumulative adverse reactions caused by long-term use of antidepressants such as AMI or SSRIs remain our research focus and require close observation over the long term. Future studies will examine the serotonin pathway’s mechanism in vivo.

In summary, we showed that EOC cells highly expressed TMICs via TLR4/MYD88 and serotonin signaling pathways, thereby suppressing immune cell function. We also showed that AMI is able to reshape EOC TME and revitalize the cytotoxicity of CD8⁺ T cells by diminishing TMICs expression by EOC cells, and immunosuppressive TRP metabolites. Thus, AMI, especially in combination with ICB therapy, might be an effective therapeutic intervention to target aberrant TRP metabolism and promote anti-tumor immunity in patients with EOC.

Limitations of the study

In the study, we show that AMI inhibits TLR4 signaling-mediated expression of IDO1 and IL-4I1 specific for TRP metabolism in EOC cells, leading to reducing kynurenine/indole pathway-induced immunosuppression and that AMI inhibits SERT-mediated intake of serotonin in EOC cells, leading to downregulating serotonin signaling-mediated IDO1 and PD-L1. However, the underlying mechanism by which serotonin/TGM2/serotonylation upregulates PD-L1 remains unclear, even though the inhibition of TGM2 reduces PD-L1 expression, Therefore, further identification or characterization of serotonylated proteins or transcriptional factors is needed to better understand how serotonylation induces the persistent expression of PD-L1.

It has been documented that IDO1 or IL-4I1 inhibitor combined ICIs do not benefit ICB therapy for EOC due to the convergence of IDO1 and IL-4I1 in TRP metabolism. So, a better understanding of the advantages of AMI combined with ICB in treating EOC is needed in preclinical trials with multiple combination scenarios.

Resource availability

Lead contact

Further information and requests for resources and reagents should be directed to and will be fulfilled by the lead contact, Jianming Huang (wesleyhuangcn2002@163.com).

Materials availability

This study did not generate new unique reagents.

Data and code availability

• •Data: All data reported in this article will be shared by the lead contact upon request.
• •Code: This article does not report the original code.
• •Additional information requests: Any additional information required to reanalyze the data reported in this article is available from the lead contact upon request.

Acknowledgments

This research was funded by the Scientific Research Foundation of Department of Science and Technology of Sichuan, China (Grant No. 2019YFS0036, Guonan Zhang, 2023YFS0093, Jianming Huang, 23NSFSC1143, Dengfeng Wang), Beijing Kanghua Foundation for the Development of Traditional Chinese and Western Medicine (Le Fund KH-2020-LJJ-07, Dengfeng Wang). We also thank Figdraw as the Graphic Abstract in this article was drawn by Figdraw.

Author contributions

Conceptualization, J.H. and G.Z.; methodology, J.H. and G.Z.; validation, J.L., B.M., L.F., and X.W.; formal analysis, J.L., B.M., and L.F.; data curation, X.W. and D.W.; writing-original draft preparation, J.L., B.M., and L.F.; writing-reviewing and editing, J.H. and G.Z.; supervision, J.H. and G.Z.; project administration, J.H. and G.Z.; funding acquisition, G.Z., J.H., and D.W. All authors have read and agreed to the published version of the article.

Declaration of interests

The authors declare no competing interests.

STAR★Methods

Key resources table

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
Antibody-MyD88 (Human)	abcam	Cat# ab219413; RRID: AB_3083083
Antibody-MyD88 (Mouse)	Affinity	Cat# AF5195; RRID: AB_2837681
Antibody- IDO1	Cell Signaling Technology	Cat# 86630; RRID: AB_2864407
Antibody- Tubulin	ZEN-BIOSCIENCE	Cat# 380628;
Antibody- CD39− APC (Human)	BD Biosciences	Cat# 560239; RRID: AB_1645459
Antibody- CD39− PE (Mouse)	BD Biosciences	Cat# 567104; RRID: AB_2870021
Antibody- MyD88	R&D Systems	Cat# AF3109; RRID: AB_2146703
Antibody- AHR	R&D Systems	Cat# AF6697; RRID: AB_10891869
Antibody- IDO1	Bioss	Cat# bs-15493R;
Antibody- PD-L1 (Human)	abcam	Cat# ab205921; RRID: AB_2687878
Antibody- PD-L1 (Mouse)	R&D Systems	Cat# AF1019; RRID: AB_354540
Antibody- CD8	Biolegend	Cat# 372902; RRID: AB_2650657
Antibody- CD39	abcam	Cat# ab223842; RRID: AB_2889212
Antibody- CD3-FITC	BD Biosciences	Cat# 553061; RRID: AB_394594
Antibody- CD4-APC	BD Biosciences	Cat# 553051; RRID: AB_398528
Antibody- CD25− PE-CY7	BD Biosciences	Cat# 552880; RRID: AB_394509
Antibody- Foxp3-PE	BD Biosciences	Cat# 563101; RRID: AB_2738006
Antibody- CD45-APC-CY7	BD Biosciences	Cat# 557659; RRID: AB_396774
Antibody- CD11b-PE	BD Biosciences	Cat# 557396; RRID: AB_396679
Antibody- Ly-6G and Ly-6C -FITC	BD Biosciences	Cat# 553128; RRID: AB_394644
Antibody- Ly6G-PE-CY7	BD Biosciences	Cat# 560601; RRID: AB_1727562
Antibody- Ly6C-APC	BD Biosciences	Cat# 560595; RRID: AB_1727554
Antibody- CD3-FITC	BD Biosciences	Cat# 553061; RRID: AB_394594
Antibody- CD8-P-PerCP5.5	BD Biosciences	Cat# 561094; RRID: AB_2034012
Antibody- PD-1-PE	BD Biosciences	Cat# 561788; RRID: AB_10895570
Chemicals, peptides, and recombinant proteins
RPMI 1640 medium	Gibco	31800022
High-glucose DMEM medium	Gibco	12100046
McCoy’s 5A medium	Sigma Aldrich	M4892
FCS	ExCell Bio	FSP500
TRIzol	Invitrogen	15596026
Eastep RT Master Mix kit	Promega	LS2052
Eastep qPCR Master Mix Kit	Promega	LS2062
RIPA buffer	EpiZyme	PC101
BCA protein assay Kit	Solarbio	PC0020
SDS-PAGE loading buffer	Beyotime	P0015L
10% SDS-PAGE gels	EpiZyme	PG112
PVDF membrane	Merck Millipore	0207
Amitriptyline	Sigma-Aldrich	A8404
LDN 27219	aladdin	312946
LPS	Sigma-Aldrich	SMB00704
Lymphocyte Separation Medium (Human)	Solarbio	P8610
Mouse splenic lymphocyte isolation kit	Solarbio	P8860
ELISA kit- IL4I1 (Human)	Jianglaibio	JL19484
ELISA kit- IL4I1 (Mouse)	Jianglaibio	JL20730
ELISA kit- TNF-α (Human)	Novus Biologicals	VAL105G
ELISA kit- TNF-α (Mouse)	Novus Biologicals	VAL609
ELISA kit- IFN-γ (Human)	Novus Biologicals	VAL104C
ELISA kit- IFN-γ (Mouse)	Novus Biologicals	VAL607
LDH assay kit	Promega	CytoTox 96®-G1780
Paclitaxel	Med Chem Express	HY-B0015
Cisplatin	Sigma-Aldrich	232120
HRP-goat anti-rabbit antibody	Servicebio	GB23303
FITC	Servicebio	GB22303
Cy3-fluorescein	Servicebio	GB21301
PD-1 inhibitor	This paper	N/A
Critical commercial assays
Tryptophan metabolites	Shanghai Zhongke New Life Biotechnology Co., Ltd.	N/A
Experimental models: Cell lines
A2780	CTCCCAS	N/A
OVCAR3	CTCCCAS	N/A
ID8	CTCCCAS	N/A
Experimental models: Organisms/strains
Mouse: female C57BL/6JGpt	Dossy Animals CO.	N/A
Oligonucleotides
GAPDH-F (Human)	This paper	5′-AACGGATTTGGTCGTATTG-3′
GAPDH-R (Human)	This paper	5′-GGAAGATGGTGATGGGATT-3′
MyD88-F (Human)	This paper	5′-GTCTGACCGCGATGTCC-3′
MyD88-R (Human)	This paper	5′-AGGCGAGTCCAGAACCA-3′
AHR-F (Human)	This paper	5′-GTCAAATCCTTCCAAGCGGC-3′
AHR-R (Human)	This paper	5′-CAGTTATCCTGGCCTCCGTT-3′
IDO1-F (Human)	This paper	5′-TCTCATTTCGTGATGGAGACTGC-3′
IDO1-R (Human)	This paper	5′-GTGTCCCGTTCTTGCATTTGC-3′
PD-L1(CD274）-F (Human)	This paper	5′-GGACAAGCAGTGACCATCAAG-3′
PD-L1(CD274）-R (Human)	This paper	5′-CCCAGAATTACCAAGTGAGTCCT-3′
IL4I1-F (Human)	This paper	5′-GCCAAGACCCCTTCGAGAAAT-3′
IL4I1-R (Human)	This paper	5′-CCGATCCTGTTATCTGCCTCC-3′
GAPDH-F (Mouse)	This paper	5′-AACGGATTTGGTCGTATTG-3′
GAPDH-R (Mouse)	This paper	5′-GGAAGATGGTGATGGGATT-3′
MyD88-F (Mouse)	This paper	5′-TCCCACAAACAAAGGAACTG-3′
MyD88-R (Mouse)	This paper	5′-TCAGAAACAACCACCACCAT-3′
AHR-F (Mouse)	This paper	5′-CCCTACCAATACGCACCA-3′
AHR-R (Mouse)	This paper	5′-AGGGCTTGAAGGAGGACA-3′
IDO1-F (Mouse)	This paper	5′-TGGAACCGAGGGGATGACG-3′
IDO1-R (Mouse)	This paper	5′-GGATACAGTGGGGATTGCTTTGA-3′
PD-L1(CD274）-F (Mouse)	This paper	5′-CTGCTTGCGTTAGTGGTGT-3′
PD-L1(CD274）-R (Mouse)	This paper	5′-CGTGATTCGCTTGTAGTCC-3′
IL4I1-F (Mouse)	This paper	5′-AACACTTGTTGGTGGAAACGA-3′
IL4I1-R (Mouse)	This paper	5′-TCCTTGCGATTAGGAGTGGTC-3′
Software and algorithms
ImageJ	NIH	https://imagej.nih.gov/ij/
GraphPad Prism 10.0	GraphPad	https://www.graphpad.com/
Primer 5.0	Premier	N/A
Other
CFX96 C1000 Touch PCR system	Bio-Rad	N/A
Infinite Tecan M200 PRO microplate reader	Switzerland	N/A
BD FACS Calibur	BD Biosciences	N/A
Cell Quest software	BD Biosciences	N/A
Halo 101-WL- HALO-1 software	Indicator labs	N/A

Experimental models and study participant details

Ethics approval and consent to participate

The study was conducted according to the guidelines of the Declaration of Helsinki, and approved by the Ethics Committee of Sichuan Cancer Hospital (KY-2019-035-01, Chengdu, China).

Cell lines

Human EOC cell lines (A2780 and OVCAR3) and mouse EOC cell lines (ID8) were purchased from the Committee on Type Culture Collection of the Chinese Academy of Sciences (CTCCCAS, Shanghai, China). A2780 and ID8 were cultivated in RPMI 1640 medium (Gibco, USA) and high-glucose DMEM (Gibco, USA) respectively, with 10% FCS supplemented. OVCAR3 cells were cultured in McCoy’s 5A (Sigma Aldrich, USA), containing 1% human insulin and 15% fetal calf serum (FCS). 1% penicillin-streptomycin (100 U/ml penicillin and 100 μg/ml streptomycin) was added to the mediums for these three cells. All cells were maintained in a humidified atmosphere at 37°C and 5% CO₂. Routine testing confirmed cell lines to be mycoplasma-negative. Relevant literature indicates that A2780, OVCAR3, and ID8 cells express TLR4 positively,^37,49,66,67 and The Human Protein Atlas databases (https://www.proteinatlas.org/) shows positive expression of SLC6A4 in ovarian cancer cells.

Animals

Specific pathogen-free, 6–8-week-old female C57BL/6JGpt littermate mice (20 ± 2g) (Chengdu Dossy Experimental Animals CO., Chengdu, China) were used for studies. Animal studies were performed according to ethical guidelines approved by the Ethics Committee of Sichuan Cancer Hospital (approval No. KY-2019-035-01).

Method details

Real-time quantitative PCR

Following the manufacturer’s instructions, total RNA was extracted from EOC cells and tissues using TRIzol reagent (Invitrogen, 15596026, USA) according to the manufacturer’s protocol. A total of 2μg RNA was reverse transcription using an Eastep RT Master Mix kit (Promega, LS2052, USA) with random primers, and 2μL cDNA was used for qPCR using Eastep qPCR Master Mix Kit (Promega, LS2062, USA) with the corresponding primers on CFX96 C1000 Touch PCR system according to manufacturers' protocols. Primers were designed by Primer5.0 according to the published sequences in GenBank (see Table S1).

Western Blotting

EOC cells were lysed in RIPA buffer (EpiZyme Biotechnology, PC101-016A1140, China) and sonicated on ice for 30 s. Cell lysate was centrifuged at 12000 rpm for 15 min at 4°C, the protein concentration was measured by BCA protein assay Kit (Solarbio, PC0020, China). SDS-PAGE loading buffer (Beyotime Biotechnology, P0015L, China) was used to denature the sample thermally. All proteins were resolved in 10% SDS-PAGE gels (EpiZyme Biotechnology, PG112, China), and then transferred on polyvinylidene difluoride (PVDF) membrane (Merck Millipore, 0207, Cork Ireland).

After blocking with buffer (5% non-fat dry milk, 0.2% tween in PBS buffer pH 7.4) for 2 h, the membrane was incubated with primary antibody MyD88 (Human: abcam, ab219413, UK; Mouse: Affinity, AF5195, USA), IDO1 (Cell signaling technology, 86630, USA), PD-L1 (Human: abcam, ab205921, UK; Mouse: R&D Systems, AF1019, USA) and Tubulin (ZEN-BIOSCIENCE, 380628, China) at 4°C overnight. After 3 times washes, the membranes were incubated with secondary antibody (Absin Biotechnology, abs20002, China) at room temperature for 1 h, and then the membrane developed with BeyoECL Plus kit (Beyotime Biotechnology, P0018S, China). Western blot bands were quantified with Tanon 4200 (Biotanon, China). Tubulin antibody as loading control was used to normalize the levels of protein detected.

Drug stimulation of cells

For A2780, OVCAR3, and ID8 EOC cells, we used AMI platform doses below IC50, and displays the concentration profiles of these cells (see Figure S3). A2780 and OVCAR3 cells select two drug concentrations of 50μM and 100μM, while ID8 cells choose two drug concentrations of 30μM and 60μM. AMI was added to A2780 (120×10⁴ cells), OVCAR3 (90×10⁴ cells), and ID8 (90×10⁴ cells) for 2 and 6 h, respectively, in order to stimulate them for qRT-PCR detection. For 12 and 24 h, respectively, stimulation was carried out in Western Blotting studies. We employed 5-HT and TGMi medicines to operate on the aforementioned cells (with the same number of cells as previously) in order to identify serotonin pathway-related indicators. We chose 5-HT drug doses of 50, 100 and 200μM and TGM2i drug concentrations of 5μM.

Preparation of EOC cells supernatants

EOC A2780, OVCAR3, and ID8 cells were seeded into 6-well plates and preincubated for 24 h before treatment. Then replaced the culture medium and added the specified concentration of Amitriptyline (AMI, Sigma-Aldrich, A8404, USA) and LPS (Sigma-Aldrich, SMB00704, USA) incubated at 37°C in a 5% CO₂ for 24 or 48 h. Finally, the EOC cells culture supernatants (abbreviated throughout as EOC cells CS) were harvested by filtering with a diameter of 0.22μm sterile filter and stored at −80°C until use for subsequent analysis.

Preparation of monocyte-free lymphocytes

Extract peripheral blood from healthy adult female and anti-coagulate with heparin. The peripheral blood mononuclear cells (PBMCs) were isolated by Ficoll–Hypaque (1.077 g/mL) density-gradient centrifugation, centrifuge at 450 g at room temperature for 15 min. Collect mononuclear cells, wash twice with RPMI-1640 complete culture medium, and then incubate in a 37° C, 5% CO₂ incubator for 4 h. Lastly, the nonadherent cells devoid of contaminating monocytes, i.e., lymphocytes, were collected to be used for this study. C57BL/6JGpt mice lymphocytes from the spleens were isolated according to the instructions of the mouse splenic lymphocyte isolation kit (Solarbio, P8860, China), to be used for this study.

Elisa

Enzyme-linked immunosorbent assay (ELISA) was used to determine cytokine IL4I1((Human: Jianglaibio, JL19484, China; Mouse: Jianglaibio, JL20730, China) expression in supernatants of EOC cells treatment with indicated concentrations of AMI, and cytokine TNF-α (Human: Novus Biologicals, VAL105G, USA; Mouse: Novus Biologicals, VAL609, USA) and IFN-γ (Human: Novus Biologicals, VAL104C, USA; Mouse: Novus Biologicals, VAL607, USA) in culture supernatants from T lymphocytes stimulated with EOC cells CS (healthy adult female T lymphocytes stimulated with A2780 and OVCAR3 cells CS, C57BL/6JGpt mice T lymphocytes stimulated with ID8 cells CS). In vivo experiments detection of cytokine IL4I1 expression in the peripheral serum of ID8 tumor-bearing mice in each experimental group.

According to the manufacturer’s protocols, the serum samples were diluted in a sample diluent provided with the ELISA kit. Briefly, 100μL of the samples, the standard or the control was added to each well in 96-well ELISA plate and incubated at room temperature for 2 h. Each well was aspirated and washed four times with Wash Buffer, then 100μL of the antibody specific for cytokine conjugated to horseradish peroxidase was added to each well, incubated at room temperature for 2 h, aspirated and washed four times with Wash Buffer. 100μL of the substrate solution to each well, and incubated in the dark at room temperature for 30 min, and then 100μL of the stop solution was added to each well. Plates were analyzed with an Infinite Tecan M200 PRO microplate reader (Switzerland) at 450 nm, with the correction at wavelength set 540 or 570 nm. All samples were measured in repeatedly.

Flow cytometry for the CD39 expression by T lymphocytes

After being treated with EOC cells culture supernatants (CS) for 12 and 24 h, 1×10⁶ lymphocytes were incubated away from light at room temperature with corresponding quantity of CD39 antibody (Human: BD Biosciences, APC560239, USA; Mouse: BD Biosciences, PE567104, USA) for 30 min. After this, the lymphocytes were washed and resuspended in staining buffer, and fluorescence labeling was measured by BD FACS Calibur (BD Biosciences). Data were analyzed using Cell Quest software (BD Biosciences).

Cytotoxicity assay

Cytotoxicity of T lymphocytes were measured by the LDH assay kit (Promega, CytoTox 96-G1780, USA). Briefly, T lymphocytes stimulated with EOC cells CS, and harvested the T lymphocytes (effector), and EOC cells were treated with or without amitriptyline (target). The detailed method is as follows: As stated in the previous method, gather the EOC cells supernatant and carry out sterile filtering for later use after stimulating EOC cells (A2780, OVCAR3, and ID8 cells) with AMI and LPS for 24 and 48 h (supernatant 24h and supernatant 48h), and make the supernatant of EOC cells without using any drugs (supernatant 0h). Separately add EOC supernatant to T lymphocytes, then let them co-culture for 24h (as effector cell). And prepare two different target cell types: those that are stimulated by LPS and AMI for 24h or not. T lymphocytes (effector cell) and EOC cells (target cell) were mixed by the ratio of effect and target at 1:1, 5:1, and10:1 in 1% FCS RPMI 1640 medium in a 96-well V-bottom at 37°C for 4 h. The subsequent operations were performed according to the manufacturer’s procedure. The percentage of cytotoxicity for each effector: target cell ratio is calculated as:

$$\%\text{Cytotoxicity}=\left(\frac{\text{Experimental}-\text{Effector}\text{Spontaneous}-\text{Target}\text{Spontaneous}}{\text{Target}\text{Maximum}-\text{Target}\text{Spontaneous}}\right)\times 100$$

Measurement of metabolites tryptophan

Freeze-dry 1000ul of EOC cells culture supernatants and ID8 EOC-bearing mice serum, add 500ul methanol-acetonitrile-water solution (2:2:1, v/v), add 10μL internal standard (2 μg/ml), and add ceramic beads. Homogenize 3 times (each time for 20s); vortex mix for 30s; subject to low temperature ultrasonication for 30min. Centrifuge at 14000rcf for 15 min at 4°C and collect 400μL of supernatant for liquid chromatography-tandem mass spectrometry/Mass spectrometry (LC-MS/MS). Tryptophan metabolites were analyzed by Shanghai Zhongke New Life Biotechnology Co., Ltd.

In vivo therapeutic effects, and safety and toxicity profiles

Based on the results of in vitro experiments, we hypothesize that Amitriptyline might be able to improve the immunosuppressive microenvironment and enhance the efficacy of ovarian cancer immunotherapy. To further verify whether Amitriptyline enhances the efficacy of immunosuppressive agents in vivo, we used ID8 (MyD88⁺) EOC cells, which were derived from an ovary tumor of C57BL/6JGpt female mice, and could be implanted into it without suffering from immune rejection. 1×10⁶ of ID8 cells were subcutaneously inoculated in the right Hip back of mice, the tumor size was measured every three days, and the volume was calculated as (length × width²) × 0.52.

We mimicked the clinical treatment regimen for in vivo experimental validation. Paclitaxel (Med Chem Express, HY-B0015, USA) + Cisplatin (Sigma-Aldrich, 232120, USA) as a standard first-line chemotherapy (TP) for ovarian cancer. When tumor volume reached 100–150 mm³, ID8 tumor-bearing C57BL/6JGpt female mice were randomly divided into five groups (nine mice per group) as follows: (1) Control (saline); (2) TP; (3) TP + PD-1 inhibitor; (4) TP + AMI; and (5) TP + PD-1 inhibitor + AMI. Treated with TP (paclitaxel 10 mg/kg, i.p. and cisplatin 5 mg/kg, i.p.) every 3 days for a total of three times. 5 days apart, treatment with PD-1 inhibitor (sintilimab, 10 mg/kg, i.p.) every 3 days for a total of five times, and combined with amitriptyline (AMI 10 mg/kg, i.p.) once a day for 15 days. All mice were sacrificed 30 days later (Figure 5A).

Samples were collected after euthanasia, and the toxicity of different treatment strategies was assessed by measuring the change in plasma levels of alanine aminotransferase (ALT), aspartate aminotransferase (AST), urea and creatinine (CREA). The organs (heart, liver, lung, and kidney) were harvested and fixed in 4% paraformaldehyde, followed by paraffin embedding. Histopathological analysis was performed by using hematoxylin and eosin (H&E) staining.

Immunohistochemistry (IHC)

Briefly, tissue sections underwent microwave antigen retrieval for 20 min in citric acid buffer (pH 6.5) and the endogenous peroxidase was inhibited with 3% H₂O₂. After blocking, the tissue sections were mixed with the primary antibody MyD88 (R&D Systems, AF3109, USA), AHR (R&D Systems, AF6697, USA) and IDO1 (Bioss, bs-15493R, China) incubated overnight at 4°C. Next, sections were exposed to biotinylated HRP-goat anti-rabbit antibody (Servicebio, GB23303, China) for 30 min in an incubator at 37°C. Then, stain with diaminobenzidine (DAB) and evaluate the staining intensity using Halo 101-WL-HALO-1 software (Indicator labs, USA).

Fluorescent microscopy

Sections of mice ID8 tumor tissue that had undergone antigen repair were immunostained using primary antibodies against murine CD8 (Biolegend, 372902, USA) and CD39 (abcam, ab223842, UK). Secondary antibodies against primary antibodies were conjugated with FITC (Servicebio, GB22303, China) and Cy3-fluorescein (Servicebio, GB21301, China) to enable fluorescent microscopy imaging and stain intensity analysis using ImageJ (Fiji) software.

Flow cytometry for the phenotype of T lymphocytes

Lymphocytes were purified from C57BL/6JGpt female mice spleens were assessed by flow cytometry (FCM). 100μL of lymphocyte suspension containing 1×10⁶ cells were incubated away from light at 2°C–8°C with corresponding quantity CD3-FITC/CD4-APC/CD25-PE-CY7 antibodies (BD Biosciences, 553061/553051/552880, USA) for 30 min, centrifuged once with 2 mL of Stain Buffer for cleaning. Subsequently, intracellular staining was performed in accordance with the Transcription Factor Buffer Set guidelines for fixing membrane-breaking and Foxp3-PE antibody (BD Biosciences, 563101, USA) incubation in the dark for 40–50 min at 2°C–8°C.

Similarly, lymphocyte surface staining was performed for flow-through detection of MDSCs and PD-1. CD45-APC-CY7/Gr-1-PE/CD11b-FITC/Ly6G-PE-CY7/Ly6C-APC antibodies (BD Biosciences, 557659/557396/553128/560601/560595, USA) incubation in the dark for 30 min at 2°C–8°C for MDSCs. CD3-FITC/CD8-P-PerCP5.5/PD-1-PE antibodies (BD Biosciences, 553061/561094/561788, USA) incubation in the dark for 30 min at 2°C–8°C for PD-1.

After this, the lymphocytes were washed and resuspended in staining buffer, and fluorescence labeling was measured by BD FACS Calibur (BD Biosciences). Cell Quest software (BD Bioscience) was used to determine the percentage of CD25⁺/Foxp3⁺ Tregs in the CD3⁺/CD4⁺ cells, the percentage of CD11b⁺Ly6C^hiLy6G^lo (M-MDSCs) cells and CD11b⁺Ly6G^hiLy6C^lo (G-MDSCs) cells in the CD45⁺/Gr-1⁺ cells, and the percentage of PD-1 in the CD3⁺/CD8⁺ cells.

Quantification and statistical analysis

GraphPad Prism 10.0 software was used for data analysis and plotting. All data， otherwise indicated, are presented as mean ± SD of at least three biological experiments unless otherwise indicated. Statistical analysis was performed using the Student’s t test (Two-group comparisons) or by one-way ANOVA (multiple-group comparisons). Nonparametric tests, such as the Mann–Whitney test for comparisons between two groups or the Kruskal–Wallis with Dunn post hoc test for multiple comparisons, were used when data were not normally distributed. Survival rates of different treatment groups were estimated by the Kaplan-Meier curves and compared by the Log rank test. For all statistical tests, nd, ns, ∗, ∗∗, ∗∗∗ and ∗∗∗∗ indicates not detected, not significant, p < 0.05, p < 0.01, p < 0.001 and p < 0.0001, respectively.

Published: November 28, 2024