IDO1- mediated Trp- kynurenine- AhR signal activation induces stemness and tumor dormancy in oral squamous cell carcinomas

The non- proliferative residual tumor cells in cancer therapeutics are considered tumor- repopulating cells (TRCs) and a major issue for tumor recurrences and distant

late co-suppressive signaling via the receptors of cytotoxic Tlymphocyte-associated antigen 4 (CTLA-4) and programmed cell death protein 1 (PD-1) expressed on the surface of activated antitumor T cells. 7 ICIs are currently being used in several solid tumors, including malignant melanoma, 8,9 lung cancer, 10 and head and neck cancer 11,12 to evade the immune surveillance of cancer cells. However, the response rate by the ICI treatment only is approximately 10%-40%, 13 and this is why it is necessary to elucidate new mechanisms of cancer immune tolerance and develop more effective cancer immunotherapies.
In cancer therapeutics, non-proliferative residual tumor cells are considered one of the major problems for the tumor recurrence and distant metastasis. 14,15 These residual tumor cells are considered tumor-repopulating cells (TRCs) which are stem cell-like cancer cells repopulating tumors and associating with immune-mediated tumor dormancy, [16][17][18] and tumor dormancy and TRC are thought to be inextricably linked. 19 TRCs are a subpopulation of self-renewing and highly neoplastic cancer cells that play an important role in the initiation, promotion, and progression of tumorigenesis. In TRCs, cell cycle arrest is also coupled to a tumor-initiating or pluripotent capacity, namely stemness features. Tumor dormancy was originally defined by Willis and then redefined by Hadfield as a temporary mitotic arrest 20 and a growth arrest. 6 Dormancy was now divided into three categories: (1) cellular dormancy, where tumor cells enter quiescence, (2) angiogenic dormancy, where dormancy is kept by a balance between sufficient and insufficient vascularization for tumor cells, and (3) immune-mediated dormancy. 21 The "immune-mediated dormancy" is conceptually attributed to the interaction between the immune system and the tumor, and is considered a major reason why further success of immunotherapies cannot be achieved by such advanced methods of ICIs therapy and oncolytic virus or adoptive T-cell transfer. 22,23 TRCs have the same potential for self-renewal and differentiation as normal tissues stem cells, but they are characterized by an immune escape mechanism. 24,25 Even if TRCs escape killing by cytotoxic immune cells, it is possible that the anti-tumor immune response may generate a specific mechanism to inhibit TRC proliferation and shift to a dormant state. 26,27 However, it is currently unclear how the major immune response is induced when the tumor is going dormant and stemness is coupled to the tumor cell arrest.
In cancer immune-surveillance according to the concept of "cancer immune-editing" developed as 3E theory, 28,29 dormancy of the tumor mass indicates equilibrium between tumor growth and immune-mediated tumor death, which determines the outcome of tumor escape or elimination. Recently, it is known that metabolizing system of tryptophan (Trp), one of the essential amino acids, is playing an important role in tumor pathophysiology. Indoleamine 2, 3-dioxygenase 1 (IDO1) is a metabolic enzyme to produce kynurenine (KYN) from Trp and is expressed in cancer cells and dendritic cells in cancer immune system. IDO1 plays an important role in the mechanism for cancer cells to evade the attack from the cancer antigenprimed and activated cytotoxic T lymphocytes (CTLs) in the late phase of the cancer immune cycle. 30 In addition, the signaling from the interaction between KYN and aryl hydrocarbon receptor (AhR), a dioxin receptor, has been reported to induce cancer cell dormancy. 31 INFγ, a cytokine with an anti-tumor effect, also acts as an inhibitory effect on the cancer immune response via the induction of PD-L1 on the surface of cancer cells and macrophages. It is also apparent that INFγ is expressed in CTLs and natural killer (NK) cells and promotes the induction of IDO1, then activates Trp-KYN-AhR signaling.
In this study, we revealed that IDO1 induced by IFNγ activated Trp-KYN-AhR signaling and induced stemness in OSCC cells contributing to the entry of tumor into immunological dormancy. In addition, the induction of cancer stemness was inversely correlated with STAT1 phosphorylation suggesting that STAT1 was a key regulator between tumor differentiation and tumor dormancy.

| Antibodies
Details of antibodies are described in supporting information.

| Immunostaining for tissues and cells
Formalin-fixed and paraffin-embedded tissue blocks were cut into 4μm thick sections for HE and immunohistochemical staining. For immunocytochemistry, the same immunostaining procedure described above was applied for cells after fixation with 4% paraformaldehyde. Details are described in supporting information.

| Immunohistochemical assessment
The degree of positivity of immunoreaction in each lesion was determined according to the modified method of the one originally described by Allred et al 33 Briefly, we randomly chose three areas at the lesion of OTSCC section and counted the number of immunoreactive atypical cells for IDO1 and leucine-rich repeatcontaining G-protein-coupled receptor 6 (LGR6) in their cytoplasm or nuclei at least 300 atypical cells. The percentage of immunoreactive atypical cells was described as proportion score (PS); IDO1 (scored on a scale of 0-3; 0:0%, 1: less than 10%, 2: less than 30%, 3: equal to or more than 50%) and LGR6 (scored on a scale of 0-3; 0:0%, 1: less than 30%, 2:30%-50%, 3: equal to or more than 50%).
Staining intensity was also described as intensity score (IS) (scored on a scale of 0-3; 0: negative, 1: weak positive, 2: intermediate positive, 3: strong positive). The proportion and intensity scores were summed to produce total score (TS = PS + IS). Then, the mean score of TS was statistically compared for analyzing the correlation between IDO1 or LGR6 expression and clinicopathological indices.

| Kynurenine measurement
To confirm whether the Trp metabolite KYN was actually produced by the activation of IDO1 after the IFNγ stimulation on HSC-4 cells, we measured the concentration of KYN in the medium. After the treatment with IFNγ (0, 10, and 50 ng/mL) for 72 hours, the supernatant was collected and the kynurenine concentration ( ng/mL) was measured with an enzyme-linked immunosorbent assay (ELISA) kit (ImmuSmol, Bordeaux, France).

| Western blotting analysis
Western blotting analysis was performed as previously described. 34 Details are described in supporting information.

| RT-qPCR analysis
Details are described in supporting information.

| Wound healing assay
Wounds were prepared using the Culture-Insert 2 well (ibidi, with that in the IFNγ untreated condition (control).

| Statistical analysis
All data were expressed as the mean ± standard error of the mean (SEM). Student's t test and Mann-Whitney U test were applied for the comparison between two groups. Statistical significance was set as *P < .05, **P < .01 and ***P < .001.

| IDO1 was expressed in human OSCC cells and infiltrating immune cells
Immunohistochemically, IDO1 expression was recognized in the OSCC cells and infiltrating immune cells, especially dendritic cells.
IDO1 expression in cancer cells was apparent in the poorly differentiated OSCCs and at the front of tumor invasion ( Figure 1A).
To discriminate the expression of IDO1 in between parenchymal tumor and stromal immune cells, we performed the dual immunostaining of IDO1 and ∆Np63, a representative marker of OSCC, against moderately differentiated OSCCs. We confirmed that IDO1 was expressed in cancer cells as well as infiltrating immune cells in and around the cancer cell nests ( Figure 1B). In the correlation of Total Score (TS) (mean) between IDO1 expression in cancer cells and histopathological tumor differentiation, IDO1 TS was gradually increased in accordance with the poor tumor differentiation with a statistical significance between W and MP (**P = .002) ( Figure 1C).

| LGR6 was expressed in human OSCC cells and became distinct in accordance with the poor tumor differentiation
Leucine-rich repeat-containing G-protein-coupled receptor 6 (LGR6) was reported as a putative stem cell marker in the human lung. 35 However, it has not been well known how LGR6 is expressed in human OSCCs. We revealed that LGR6 was immunohistochemi-

| IFNγ upregulated IDO1, AhR and Lgr6 expressions but downregulated p27 expression
IFNγ is known to upregulate IDO-1 expression in the tumor microenvironment. 31,36,37,38 However, it has not been well known whether IDO1 upregulated by IFNγ activates Trp-KYN-AhR signaling and consequently stemness is induced in cancer cells or not. In our in vitro study using OSCC cell lines HSC-4 and SAS, we first confirmed that the concentration of KYN, a Trp metabolite, was actually increased by the IFNγ stimulation of both a low (10 ng/mL) and a high (50 ng/mL) concentration with statistical significances in comparison with that of IFNγ unstimulated state (control) ( Figure 3A).
Western blotting analysis revealed that IDO1 expression was not apparent in the IFNγ unstimulated state but was remarkably induced by the IFNγ (10 ng/mL) stimulation. Conversely, p27 was apparently expressed in the IFNγ unstimulated state but remarkable repressed by the IFNγ (10 ng/mL) stimulation ( Figure 3B). TNFα (100 ng/mL) stimulation did not affect any significant changes on the IDO1 and p27 expressions ( Figure 3B). At the beginning of this study, we considered that p27 was one of the critical factors in the  Figure 3C). Immunohistochemically, AhR expression was recognized in the nuclei and apparently increased at 72h after IFNγ (10 ng/mL) stimulation ( Figure 3D, green). These results suggested that IFNγ, even in the low concentration as 10 ng/mL, upregulated IDO1 and activates Trp-KYN-AhR signaling, and then, consequently a stem cell marker LGR6 was upregulated in HSC-4 cells.

| Stem cell marker expression was upregulated by the IFNγ stimulation in HSC-4 cells
In our experiment, expressions of stem cell makers LGR6 and CD133 were both upregulated by the IFNγ (10 ng/mL and 200 ng/mL) stimuli but both markers expression patterns by the IFNγ (200 ng/mL) stimulus showed a lower expression tendency compared to those by the IFNγ (10 ng/mL) stimulus ( Figure 4A). By the RT-qPCR analyses, IDO1 mRNA expression was gradually upregulated in accordance with the higher concentration of INFγ ( Figure 4B). AhR, LGR6, and p27 mRNAs were also upregulated by the IFNγ stimulus with a low concentration of 10 ng/mL. However, these upregulations were downregulated by the IFNγ stimulus with a high concentration of 200 ng/mL ( Figure 4C-E).  Figure 4E). To clarify whether this discrepancy in between p27 protein and mRNA expressions came due to a post-translational protein regulation through the ubiquitin proteasome system, we performed the inhibition analysis of proteasome activity by the treatment of MG132, a well-known proteasome inhibitor. In the IFNγ-stimulated condition, p27 protein expression was gradually downregulated without MG132 treatment but was gradually increased with MG132 treatment in the time-course-dependent manner ( Figure 4F). From these findings, we confirmed that p27 was post-translationally regulated by its degradation through the ubiquitin proteasome system.

| Upregulation of stem cell marker expression by IFNγ stimulation was dependent on the Trp-KYN-AhR signaling cascade in HSC-4 cells
To clarify whether the upregulation of stem cell marker LGR6 and

| Analyses of the behavioral effect of the IFNγ stimulation on the HSC-4 cell proliferation
We evaluated how IFNγ stimulation affected on the HSC-4 cell be- LGR6 TS is increased in accordance with the poor tumor differentiation. There is a statistical significance between W and M, and, W and MP. The statistical significance assessed by Mann-Whitney U test is indicated by *P < .05

| D ISCUSS I ON
In our clinicopathological studies, IDO1 was expressed in OSCC cells and the expression was significantly increased in accordance with prognosis in many solid tumors and hematopoietic tumors. [36][37][38] Our clinicopathological data also showed that IDO1 expression was significantly correlated with poor tumor differentiation and tumor front invasion, that was one of the poor prognostic factors.
In cancer therapeutics, non-proliferative residual tumor cells are considered TRCs and one of the major problems for the tumor recurrence and distant metastasis. 14,15 TRCs are stem cell-like cancer cells with a highly neoplastic potentials such as repopulation of tumors and association with immune-mediated tumor dormancy. [16][17][18]42,43 In TRCs, cell cycle arrest is also coupled to a tumorinitiating or pluripotent capacity. It was recently reported that the signaling from the interaction between KYN and AhR, a dioxin receptor, induced cancer cell dormancy. 31  LGR6 has been reported as a putative stem cell marker in the human lung. 35 LGR6 is an epithelial stem cell marker and its expression increases with tumor progression and metastasis. 35,44,45 In addition, LGR6 is enriched in advanced tumor cells in non-small-cell lung cancer and LGR6-positive cancer cells have self-renewal, differentiation, and high oncogenic potential. 45 CD133 has been reported as a cancer stem cell marker in several tumors including oral cancer. 46,47 However, LGR6 has so far not been analyzed in the head and neck cancers as a stem cell marker. In our study, LGR6 was co-expressed with IDO1 in OSCC cells and LGR6 and CD133 expressions were Recently, it was reported that p27/CDKN1B (Cyclin-dependent kinase inhibitor 1B), a well-known inhibitor at Gap 1 (G1) of the cell cycle, 48 was involved in slow cell cycling and in the Trp-KYN-AhR signaling pathway, leading to tumor dormancy. 31 Unexpectedly, in our study, p27 protein expression was decreased, which was an opposite result compared to that of the previous report. 31 However, p27 mRNA expression was upregulated even by the low concentration of IFNγ (10 ng/mL) stimulation revealing a discrepancy between p27 protein and mRNA expression. We clarified that p27 protein was post-translationally regulated through the 1%, respectively. These values were two times or much lower than that of the HSC-4 cells stimulated with a higher IFNγ concentrate (200 ng/mL) as 11.0% (F, G). Statistical significance was set as **P < .01 and ***P < .001 the reduction of the p27 expression in the Trp-KYN-AhR signal activation in our result might be related to tumor grade rather than inhibition of cell cycling. In addition, we found that upregulation of LGR6 and CD133 expressions was inversely correlated with STAT1 phosphorylation but not correlated with p27 expression. These data suggested that STAT1 was a key regulator between tumor differentiation and tumor dormancy.
In BrdU and cleaved caspase 3 assays, we confirmed that In summary, we revealed that IDO1 induced by IFNγ activated

Trp-KYN-AhR signal pathway and induced stemness in OSCC cells
contributing to the entry of tumor into immunological dormancy ( Figure 7E). In addition, the induction of cancer stemness was inversely correlated with STAT1 phosphorylation suggesting that STAT1 was a key regulator between tumor differentiation and tumor dormancy. From these findings, the regulation of IDO1-mediated Trp-KYN-AhR signal activation could be expected to provide a new strategy for the cancer treatment especially by inhibiting recurrences and distant metastases.

ACK N OWLED G M ENTS
This work was supported in part by Grants-in-Aid for Scientific Research (C) (18K07033 to SH), Grant-in-Aid for Young Scientists (B) (17K18301 and 20K18490 to SY) from Japan Society for the Promotion of Science (JSPS)/ KAKENHI.

CO N FLI C T O F I NTE R E S T
All authors have no financial disclosures and no conflict of interest.