A comprehensive analysis of IDO1 expression with tumour‐infiltrating immune cells and mutation burden in gynaecologic and breast cancers

Abstract Gynaecologic and breast cancers share some similarities at the molecular level. The aims of our study are to highlight the similarities and differences about IDO1, an important immune‐related gene in female cancers. The NGS data from TCGA of cervical squamous cell carcinoma (CESC), ovarian serous cystadenocarcinoma (OV), uterine corpus endometrial carcinoma (UCEC), uterine carcinosarcoma (UCS) and breast invasive carcinoma (BRCA) were analysed to identify molecular features, and clinically significant and potential therapeutic targets of IDO1. We found IDO1 was significantly up‐regulated in four gynaecologic cancers and breast cancer. According to breast cancer PAM50 classification scheme, IDO1 expression was higher in tumours of basal than other subtypes and showed better survival prognosis in BRCA and OV. Through immune infiltration analysis, we found a strong correlation between IDO1 and immune cell populations especially for dendritic cells and T cells. In addition, we investigated the association between IDO1 and tumour mutation burden (TMB) and found that IDO1 was significantly correlated with TMB in BRCA and CESC. GSVA revealed that hallmarks significantly correlated with IDO1 were involved in interferon gamma response, allograft rejection and inflammatory response. We also found PD‐L1 and LAG3 were highly positive related to IDO1 in gynaecologic cancers when comparing with their corresponding normal tissues. Our results indicated that IDO1 participated in anti‐tumour immune process and is correlated with mutation burden. These findings may expand our outlook of potential anti‐IDO1 treatments.


| INTRODUC TI ON
Gynaecologic and breast cancers are considered as global public health problems. They share a variety of characteristics, such as similar embryonic origins in the Mullerian ducts and influenced by female hormones. 1,2 Gynaecologic and breast cancers have an estimated incidence/deaths of more than 773 000/237 700 and 380 270/75 360 cases in Europe (2018) and the United States (2019). 3,4 Breast cancer is the first and uterine corpus cancer is the fourth most common cancer in women. 4 Therefore, it is necessary to discover and identify the novel and specific strategies to treat and prevent carcinogenesis of those cancers in women.
Cancer progression is a comprehensive process depending on the interaction between individual cells in the tumour, the microenvironment and the immune system, all of which can act to promote or suppress tumour growth and metastasis. 5 Increasing evidence revealed that microenvironment plays a critical role in supporting the progression of tumours. Several studies reported an increased tumour-infiltrating lymphocytes (TILs) in different tumours and correlations between specific immune subsets and prognosis. 6,7 Consequently, immune therapy alone or in combination with radiation and chemotherapeutic therapy is the current meaningful neoadjuvant treatment.
Immuno-antibodies that target immune check point such as CTLA-4, PD1 and PD-L1 have significantly improved survival in several tumours. 8 Immune evasion is one of the hallmarks of cancer, and researchers clarified complicated mechanisms to enable cancer cells escaping from the host immune response. One of the mechanisms is that cancer cells utilize the indoleamine 2,3-dioxygenase (IDO) pathway to suppress immune surveillance. 9-11 IDO1 is a haeme-containing enzyme which catalyses the breakdown of tryptophan to kynurenine.
Try is an essential amino acid obtained from the diet that is a fuel required by the body to build proteins needed for cellular growth as well as immune function. In healthy status, IDO1 removes the extra fuel needed for immune activity, and it ensures the balance of immune system and acts to inhibit the immune response through the following mechanisms: (a) inhibition of effector T cell activity to suppress the immune response; and (b) promotion of T regulatory cell activity to suppress the immune response. Namely, elevated IDO activity leads the consumption of tryptophan in the tumour microenvironment. Without tryptophan to fuel the immune cells, cytotoxic T cells starve and immunosuppressive Tregs are up-regulated leading to a failure of the immune system to respond appropriately to the cancer. 12 Human IDO1 has an evolutionary paralog (indolamine-2,3-dioxygenase 2, IDO2) and a functional ortholog (tryptophan-2,3-dioxygenase, TDO2) that catalyse the same biochemical reaction. IDO1 also has high enzyme activity for Trp (Km ~20 μmol/L). 13 The expression of IDO1 and IDO2 is restricted to eukaryotes, but IDO2 has almost 1000-fold lower enzyme activity (Km ~6.8 mmol/L) 14 compare to IDO1. TDO2 is a tetrameric haeme-containing complex and conserved across different species including both prokaryotes and eukaryotes with lower enzyme activity for Trp The function of IDOs as a checkpoint used by tumours to escape immune surveillance was a focus of research and drug discovery efforts, as well as efforts to understand whether it could be used as a biomarker for prognosis. The aim of this paper was to provide a comprehensive analysis of IDOs in gynaecologic and breast cancers, especially focusing on the molecular features of IDOs in order to improve the efficacy of currently available immune therapeutic strategies.

| Expression analysis
Expression analysis of IDO1, IDO2 and TDO2 based on TCGA and GTEx samples was conducted in GEPIA (http://gepia.cance r-pku.cn/ index.html). 16 The data of IDO1 expression in TCGA BRCA subtypes were downloaded in UCSC Xena (http://xena.ucsc.edu) and analysed by R software and GraphPad.

| Survival analysis
Overall survival (OS) and disease-free survival (DFS) were analyzed via GEPIA database. GEPIA uses log-rank test for the hypothesis evaluation. The Cox proportional hazard ratio is based on Cox PH model.

| Gene signatures and ssGSEA
The marker gene set for immune cell types was obtained from Bindea et al. 17 We used single-sample gene set enrichment analysis (ssGSEA) in R package gsva to calculate gene set enrichment. The deconvolution approach used in our study included 24 immune cells,

| Tumour mutation burden data
Tumour mutation burden (TMB) data for the TCGA cohort were obtained from cBioPortal. 18 Tumour data were selected samples from TCGA projects which include breast invasive carcinoma (BRCA) and four projects to represent the gynaecologic cancers: endocervical adenocarcinoma (CESC), high-grade serous ovarian cystadenocarcinoma (OV), uterine corpus endometrial carcinoma (UCEC) and uterine carcinosarcoma (UCS).

| Statistical analysis
In this study, statistical analysis was mainly performed using R language (https://www.r-proje ct.org/) with several publicly available packages (R x64 3.5.1). Multivariate survival analysis was performed by using Cox proportional hazards model. A probability value (P) <.05 was considered to be significant in this study.

| IDO1 expression status and clinical outcome
Considering not only IDO1 but also IDO2 and TDO2 could catalyse the first and rate-limiting step of tryptophan (Trp) catabolism, we analysed both of their expression status with GEPIA, a web-based tool to deliver fast and customizable functionalities based on the cancer genome atlas (TCGA) and genotype-tissue expression (GTEx) data. As shown in Figure 1A and Figure S1, higher IDO1 expression was found in both breast and gynaecologic cancers, while TDO2 was higher expressed in BRCA, OV and UCS. However, IDO2 had no significant difference between normal and cancer samples in breast and gynaecologic cancers ( Figure S1). When taking breast cancer subtypes into account, we found that IDO1 was highly expressed in ER-negative, PR-negative and Her2-negative samples ( Figure 1B

| IDO1 expression and tumour-infiltrating immune cells
Through the single-sample gene set enrichment analysis, we evaluated the association between IDO1 and tumour-infiltrating immune cells from transcriptomic data. The landscape of IDO1 and tumour-infiltrating immune cells correlation is shown in Figure 2A.  Figure 2C, Tables S1 and S2). A strong correlation between IDO1 and aDC in breast and four gynaecologic cancers was discovered.
IDO1 was also correlated with CD8 + T cells, NK CD56dim cells, Th1 cells and macrophages in more than three different cancers ( Figure 2D and Table S3).

| IDO1 expression and TMB
Considering tumour mutations can result in immunogenic neo-antigens, which have also been correlated with responsiveness to a more effective response to immune therapy, we analysed IDO1 expression with TMB status and grouped patients into IDO1-high and IDO1-low and analysed their TMB status. As shown in Figure 3A, IDO1 had a higher expression level in TMB-high group in BRCA and CESC. In addition, UCEC contained a higher TMB level and IDO1 expression than other four female cancers ( Figure 3B). It suggests that IDO1 inhibitor may be more effective in TMB-high patients. Then, we combined group analysis with IDO1 and TMB, and it showed that patients with low IDO1 expression and high TMB have the worse OS than IDO-high/TMB-low (P = .0347) and IDO-high/TMB-high (P = .0085) groups in BRCA. In CESC, IDO1-low/TMB-low group has the worse OS than IDO1-high/TMB-high group (P = .0269). In OV, IDO1-low/TMB-low group has the worse OS than IDO-low/ TMB-high (P = .0139) and IDO-high/TMB-high (P = .0139) groups.
In UCEC, IDO1-high/TMB-low group have the worse OS than other three groups and there was no statistic difference in UCS ( Figure 3C).
Taken together, these data suggest that both IDO1-and TMB-high group may have a better outcome.
These hallmarks were further analysed by GSVA (R packages). IDO1 was divided into two groups: IDO-low (the lowest 30% expression of IDO1) and IDO1-high (the highest 30% expression of IDO1). We found that these hallmarks with higher IDO1 level in five female cancers were mainly involved in immune-related hallmarks, including in-

| IDO1-related genes and pathway
To further investigate the role of IDO1 in gynaecologic and breast cancers, we evaluated its co-expression related genes. We found IDO1  Table S4). Interestingly, we found that PD-L1 and LAG3 were highly positive related to IDO1 in gynaecologic and breast cancers, but their  Table S1. Favour for overall survival is indicated in yellow, and risk for overall survival is indicated in black. D, IDO1 and immune cells correlation in each individual cancer types. See also Table S3 correlation was reduced in normal breast tissue, ovary and uterus/cervix ( Figure 5C,D). It suggests that IDO1, PD-L1 and LAG3 were partially regulated by the same mechanism or inducer in tumour.

| D ISCUSS I ON
The immune escape function of IDO1 in cancer was first reported in 2002 by Friberg et al. 19 Currently, more and more studies have demonstrated that IDO1 is associated with immune escape by suppressing T cell activity and enhancing Treg in different tumour types.
However, its role in female tumours lacks a detailed report, so we aimed to provide some new insights into understanding of IDO1 in the regulatory mechanism of female tumours and suggest that IDO1 may be a potential therapy target for female. Our study is the most comprehensive study characterizing the expression pattern of IDO1 together with its related genetic features and prognostic values in gynaecologic and breast cancers. Although there is a difference between these tumours regarding their correlation with IDO1 expression level and clinical outcome, both investigations appear to have a positive correlation between the presence of tumour-infiltrating T cells and increased IDO1 expression. Consistent with previous studies, we found a strong correlation between IDO1 and immune cell populations especially for dendritic cells and T cells. Studies reported that IDO1 could be induced by IFN-γ which mainly produced by activated infiltrating T cells and seems to be a key player in the innate immune system. 32 Our results also showed that IDO1 was significantly correlated with T cells and IFN-γ. Furthermore, our results also showed the correlation of IDO1 and antigen-presenting cells (APCs), like dendritic cells and macrophages, and other studies investigated those APCs could mediate immune response via IDO1 through cell cycle arrest in T cells. Evidence has shown that IFN-γ and tumour infiltration by activated CD8 + cytotoxic T lymphocytes correlate with better survival. 36 Due to the positive correlation between IDO1 and TILs/ IFN-γ, IDO1 can be subsequently related to a better prognosis.
Since tumours with a higher mutation burden have been hypothesized to have more chance to produce neoantigens and these neoantigens can be recognized by the immune system. To learn more about the potential relation between tumour mutation and IDO1, we investigated the association between IDO1 and TMB, and found that IDO1 was significantly correlated with TMB in BRCA and CESC. We also observed that high IDO1 expression and high TMB had a better clinical outcome than other IDO1/ TMB expression patterns in female cancer except uterine carcinosarcoma. This point may be contrary to the routine study and conclusions, in which patients with high TMB usually display poor clinical outcome, but nothing is absolute. There is another view that TMB-high cancer may exhibit an active tumour immune microenvironment, which is due to the recognition of a large number of tumour-specific neoantigens. The tumour-specific neoantigens produced by a TMB-high tumour are presented on APCs and stimulate T cell activation, leading to an active Th1/CTL microenvironment which exhibits a better outcome for patients. Not only in the present study, but also in Liu et al 37 study, they showed that BLCA exhibited a correlation between higher TMB and improved overall survival, which indicated the existence of inter-tumour heterogeneity of TMB as a prognostic biomarker. Moreover, we should take MSI status into consideration. The majority of MSI-high tumours were TMB-high; however, only few of TMB-high tumours were MSI-high. 37 In the previous studies, 38 MSI-high tumours have shown a better prognosis than microsatellite stable (MSS), where their higher mutational load may contribute to better survival. Therefore, the MSI and immune infiltrating status increased complexity and diversity between TMB status and prognosis.
Our analysis also found the correlation between increasing IDO1 levels with other immune checkpoints, including PD-L1, LAG3, CD86, IRFs and HLAs in female cancers. What is more, we found that immune-related genes, such as PD-L1 and LAG3, were highly positive related to IDO1 in gynaecologic cancers when comparing with their corresponding normal tissues. PD-L1 and LAG3 can be induced by IFN-γ, the same as IDO1, and this may partially explain their positive correlation in cancer samples. Up-regulation of immune checkpoint proteins has been linked to cancer immune escape, but in our study, they showed a better prognosis and we explained the mechanism of IDO1 in the above. PD-L1 also has the contradictory phenomenon.
Previous studies indicated the correlations between PD-L1 and reduced prognosis, but some other studies [38][39][40] found that PD-L1 expression is paradoxically associated with improved patients' survival. Different studies reported that combined with other immunotherapeutic agent, such as CTLA-4 or PD-1/PD-L1 inhibitors, this objective response rates range from 10% to 57% among different tumour types. These findings lead to promising opportunities for the combined synergistic treatment of female cancers.
In conclusion, the involvement of IDO1 in different cancers appears to be highly complex. Despite this complexity, this is the first study exploring the expression pattern, clinical outcome, tumour-infiltrating, TMB and biological processes in gynaecologic and breast cancers. Our results revealed the role of the IDO1 in tumour progression and immune responses that may lead to the development of IDO1 targeting therapy for assessing the efficacy and receptiveness in female cancer treatment.

CO N FLI C T O F I NTE R E S T
All authors declare no conflicts of interest.

AUTH O R CO NTR I B UTI O N
Xu Feng, Ranran Tang and Runjie Zhang conceived the study. Xu

E TH I C A L A PPROVA L
All analyses were based on previously published TCGA data; thus, no ethical approval and patient consent are required.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data used to support the findings of this study are available from the corresponding author upon request.