The impact of CCR8+ regulatory T cells on cytotoxic T cell function in human lung cancer

Regulatory T cells (Tregs) suppress the host immune response and maintain immune homeostasis. Tregs also promote cancer progression and are involved in resistance to immune checkpoint inhibitor treatments. Recent studies identified selective CCR8 expression on tumor-infiltrating Tregs; CCR8+ Tregs have been indicated as a possible new target of cancer immunotherapy. Here, we investigated the features of CCR8+ Tregs in lung cancer patients. CCR8+ Tregs were highly activated and infiltration of CCR8+ Tregs in tumors was associated with poor prognosis in lung cancer patients. We also investigated their immune suppressive function, especially the influence on cytotoxic T lymphocyte cell function. The Cancer Genome Atlas analysis revealed that CD8 T cell activities were suppressed in high CCR8-expressing tumors. Additionally, depletion of CCR8+ cells enhanced CD8 T cell function in an ex vivo culture of lung tumor-infiltrating cells. Moreover, CCR8+ Tregs, but not CCR8− Tregs, induced from human PBMCs markedly suppressed CD8 T cell cytotoxicity. Finally, we demonstrated the therapeutic effect of targeting CCR8 in a murine model of lung cancer. These findings reveal the significance of CCR8+ Tregs for immunosuppression in lung cancer, especially via cytotoxic T lymphocyte cell suppression, and suggest the potential value of CCR8-targeted therapy for cancer treatment.


CCR8+ Treg infiltration is associated with poor prognosis in lung cancer.
To determine the clinical significance of CCR8+ Tregs in lung cancer, we evaluated the frequency of CCR8+ Tregs in tumor-infiltrating CD45+ cells in 50 lung cancer patients by flow cytometry and analyzed the correlation with clinical stage and outcome. There was no significant relationship between CCR8+ Treg infiltration and clinical cancer stage ( Fig. 2A). We next divided 50 lung cancer patients into the CCR8-high and low groups based on the median value of the percentage of CCR8+ Tregs in CD45+ cells and found that the disease-free survival (DFS) of the CCR8-high group was significantly shorter than that of the CCR8-low group (p = 0.011) (Fig. 2B). The clinical data of both groups are shown in Supplementary Table 1. There was no significant difference in characteristics between the CCR8-high group and CCR8-low group except for age.
Expression profiles of CCR8+ Tregs in lung cancer. We next analyzed the profile of CCR8+ Tregs in lung tumors. Treg-related molecules that are associated with suppressive functions, such as FOXP3, CD25, CTLA4 and CD39, were highly expressed in CCR8+ Tregs compared with CCR8− Tregs as confirmed by flow cytometry (n = 7-22) (Fig. 3A, Supplementary Fig. 2). We performed RNA-sequencing of CCR8+ Tregs and CCR8− Tregs sorted from tumor-infiltrating cells (TICs) obtained from 3 patients and identified differentially expressed genes (DEGs) between CCR8+ Tregs and CCR8− Tregs. Gene Ontology (GO) enrichment analysis of the top 100 DEGs with the lowest p-value among the upregulated genes in CCR8+ Tregs was performed using DAVID. The enriched immune-related terms categorized in biological process are shown in Fig. 3B. The upregulated genes in CCR8+ Tregs were mainly associated with adapted immunity and inflammatory signaling. Furthermore, the chemokine-related pathway was also enriched. Significantly upregulated chemokine genes (p < 0.05) in CCR8+ Tregs are shown in Fig. 3C (n = 3). We then analyzed the expression of chemokine receptors by flow cytometry (n = 10-13) and found that CCR8+ Tregs expressed higher levels of various chemokine receptors compared with levels in CCR8− Tregs (Fig. 3D).
CD8 T cell activities are suppressed in high CCR8-expressing tumors. Because CCR8+ Tregs highly express Treg-related molecules that are associated with suppressive functions, we hypothesized that tumors with high CCR8 expression might exhibit a suppressed immune profile. To investigate this hypothesis, we examined mRNA expression data of lung adenocarcinoma (n = 510) and lung squamous cell carcinoma (n = 484) obtained from the cBioPortal The Cancer Genome Atlas (TCGA) database. We found a correlation between the gene expression of CCR8 and receptor-type tyrosine-protein phosphatase C (PTPRC), which encodes the lymphocyte marker CD45 (Supplementary Fig. S3). To estimate the contribution of CCR8-expressing cells to other immune cells, we calculated the CCR8/PTPRC expression ratio and divided patients into the CCR8-high and low groups based on the median value of the CCR8/PTPRC ratio. We then examined the differences of immune-related gene www.nature.com/scientificreports/ expression between the two groups. We found that gene related T cell and antigen-presenting cell (APC) were significantly reduced in CCR8-high tumors compared with levels in CCR8-low tumors in both lung adenocarcinoma and lung squamous cell carcinoma (Supplementary Table S2). Because CTLs are main players in anticancer immunity, we focused on CD8 T cells. To characterize gene expression signatures and pathway activation associated with CD8 T cell function, we performed Gene Set Variation Analysis (GSVA) with selected gene sets from the literature 24 . According to Guo et al. 24 , the CD8-C1-LEF1 fraction, CD8-C3-CXCR1 fraction and CD8-C6-LAYN fraction were used as naïve, effector and exhausted CD8 T cell gene sets, respectively. Naïve, effector, and exhausted CD8 T cell scores were calculated by GSVA, and we compared the scores between CCR8-high and -low tumors (Fig. 4). The CCR8-high group showed significantly lower naïve and effector CD8 T cell scores compared with the CCR8-low group. In particular, the difference in the effector CD8 T cell score between the two groups was remarkable in both lung squamous cell carcinoma and lung adenocarcinoma.

Depletion of CCR8+ cells enhances CD8 T cell function in lung tumors.
We next examined whether depletion of CCR8+ Tregs could eliminate immunosuppression, leading to activation of CD8 T cells. We confirmed that CCR8+ cell depletion using MicroBeads reduced FOXP3+ Tregs but not FOXP3− conv CD4 T cells ( Supplementary Fig. S4). To evaluate the influence of CCR8+ cell depletion on CD8 T cell functions, TICs with or without CCR8+ cell depletion were cultured with IL-2 and CpG for 5 days, and the expression of functional molecules in CD8 T cells was measured by flow cytometry (n = 15) (Fig. 5A, B). Depletion of CCR8+ cells led  www.nature.com/scientificreports/ to an increased frequency of GZMB-expressing CD8 T cells and IFNγ-producing CD8 T cells upon PMA/ionomycin stimulation. Frequency of tumor necrosis factor-α (TNFα) -producing CD8 T cells was increased, too (data not shown). In contrast, the depletion of CCR4+ cells had no effect on IFNγ-producing CD8 T cells (n = 5) (Fig. 5C). To investigate the involvement of MHC class I in the suppression mechanism by CCR8+ Tregs, we added anti-HLA-A,B,C antibody in TIC culture after CCR8+ cell depletion. The results showed that the effect of CCR8+ cell depletion was canceled by treatment with anti-HLA-A,B,C antibody (n = 4) (Fig. 5D).
Induced CCR8+ Tregs highly suppress CD8 T cell cytotoxicity. Next, we directly evaluated the suppressive potential of CCR8+ Tregs on CD8 T cell function. To obtain sufficient amounts of Tregs for functional assay, CCR8+ Tregs were induced and expanded from healthy donor PBMCs by anti-CD3/CD28 stimulation using Treg Expander Beads. We confirmed that expanded Tregs expressed FOXP3 and had suppressive function against effector T cell proliferation ( Supplementary Fig. S5). Approximately 40% of the expanded Tregs expressed CCR8 (data not shown) and Treg-related molecules were expressed at higher levels in the CCR8+ fraction than the CCR8− fraction (n = 8) (Fig. 6A) as with tumor-infiltrating Tregs. We sorted CCR8+ Tregs and CCR8− Tregs and examined the effects on CD8 T cell cytotoxicity using two models. To examine antigen non-specific cytotoxicity, sorted CCR8+ Tregs or CCR8− Tregs were co-cultured with CD8 T cells from healthy donor PBMCs and stimulated by anti-CD3 antibody in the presence of APCs. CD8 T cells co-cultured without Tregs served as controls. After 4 to 6 days of culture, we observed downregulation of GZMB and perforin expression in CD8 T cells co-cultured with CCR8+ Tregs (n = 7) (Fig. 6B). Antigen non-specific cytotoxicity of CD8 T cells was evaluated using membrane-bound anti-CD3scFv-expressing BALL-1 (aCD3-BALL1) cells (n = 6) (Fig. 6C). The results showed that the cytotoxicity against aCD3-BALL1 cells was suppressed only in CD8 T cells co-cultured with CCR8+ Tregs but not those cultured with CCR8− Tregs.
Targeting CCR8 has anti-tumor effects in a murine model of lung cancer. We finally evaluated the potential anti-tumor activity of CCR8-targeted therapy in a murine model of lung cancer. MHC class I-expressing Lewis lung carcinoma (H-2Kb-LLC) cells were established and inoculated them into C57BL/6 mice. To completely eliminate CCR8+ cells, anti-CCR8 antibody was administered twice at early time points, day 3 and day 8 after tumor inoculation. In the model mice, we observed infiltration of CCR8+ Tregs in tumor tissue, and treatment of anti-CCR8 antibody resulted in Treg reduction ( Supplementary Fig. S6). Tumor growth was inhibited by anti-CCR8 antibody administration (n = 10) (Fig. 7). Tumors completely regressed in 5 of 10 mice treated with anti-CCR8 antibody.

Discussion
In this study, we performed a detailed investigation of CCR8+ Tregs in human lung cancer and demonstrated that they have highly suppressive functions against CD8 T cells. As has been reported in other cancer types 21-23 , we found that CCR8 was selectively expressed on tumor-infiltrating Tregs in human lung cancer, and CCR8 expression on other immune cells was very low, indicating that adverse events may not be a concern for CCR8targeted therapy. Prognostic analysis revealed that high infiltration of CCR8+ Tregs, but not CCR8− Tregs (data not shown), was significantly associated with poor prognosis (p = 0.011). www.nature.com/scientificreports/ Previous studies reported shorter overall survival in cancer patients with high CCR8 expression by RNA-seq analysis of breast cancer and lung cancer, and by immunohistochemical and flow cytometric analysis of bladder cancer 23 . Our study showed a similar result through flow cytometric analysis of CCR8-expressing Tregs in lung tumors. These findings indicate that cancer patients with a high infiltration of CCR8+ Tregs may be more likely to experience recurrence. We recently reported that mice that showed tumor regression with anti-CCR8 therapy rejected re-inoculated tumors because of the formation of memory T cells, whereas mice in which tumors were surgically removed failed to form memory T cells and thus showed tumor regrowth 19 . We speculated that patients with a high infiltration of CCR8+ Tregs in tumors may not form an immunological memory and are thus more liable to show tumor metastasis and a shorter DFS. However, the small sample size is a limitation of this study, and our findings should be confirmed in larger cohorts. www.nature.com/scientificreports/ In the present study, expression profiling of lung tumor-infiltrating CCR8+ Tregs revealed three features. First, we observed increased expression of Treg-related molecules in CCR8+ Tregs compared with CCR8− Tregs. Second, the adaptive immunity, inflammatory response, and cytokine receptor pathways were upregulated in CCR8+ Tregs. Third, CCR8+ Tregs showed higher expression of various chemokines and chemokine receptors compared with CCR8− Tregs. High expression of Treg-related molecules in CCR8+ Tregs has been reported in several studies [18][19][20]23 . We found that CCR8+ Tregs in human lung cancer exhibited a similar suppressive expression profile to previous studies, and further revealed that they are in a strongly activated state and enhance chemokine signaling. CCR8+ Tregs may recruit effector T cells to their surroundings by producing various chemokines, leading to efficient suppression of this cell activation. Interestingly, CCR8+ Tregs highly express CCL22 and its receptor CCR4, as well as CCL20 and its receptor CCR6 25 , and thus CCR8+ Tregs might form a Treg accumulation site through the autocrine activity of the chemokine and its receptor, leading to the induction of an immunosuppressive environment.
In addition, we showed that CCR8+ Tregs have strong suppressive functions against other immune cells, especially CTLs, through three approaches. First, using public lung cancer mRNA expression databases, we revealed that CCR8-high tumors showed reduced T cell-and APC-related genes and significantly lower naïve and effector CD8 T cell scores compared with CCR8-low tumors. Although there was no significant difference in prognosis between the high and low CCR8 expression groups, probably because of the use of multicenter TCGA data, the results suggest that CCR8+ Tregs might suppress mobilization, differentiation, or activation of naïve/ effector CD8 T cells. Effector T cell fractions were characterized by high expression of genes associated with cytotoxicity 24 , and thus these findings suggest that CTL activation might be suppressed, and that the anti-tumor activity might be decreased in cancers with high infiltration of CCR8+ Tregs. Second, we experimentally showed that depletion of CCR8+ cells from TIC culture enhanced CD8 T cell function such as GZMB and IFNγ expression, while depletion of CCR4+ cells did not induce CD8 T cell activation. CCR4 is known to be expressed on both Tregs and conv CD4 T cells 26 , and in our experiment, depletion of CCR4+ cells not only reduced Tregs but also conv CD4 T cells, which might play a positive role in anti-tumor immunity. Hence, CCR8+ cell depletion may have activated CD8 T cells more efficiently than CCR4+ cell depletion. In this assay, blocking MHC class I Figure 4. Naïve, effector and exhausted CD8 T cell enrichment scores in CCR8-high and CCR8-low lung tumors. Whole tumor mRNA expression data of patients with adenocarcinoma (n = 510) and squamous cell carcinoma (n = 484) were obtained from TCGA cBioPortal database. Patients were divided into high and low CCR8 groups based on the median CCR8/PTPRC ratio. Gene Set Variation Analysis (GSVA) was performed with naïve, effector and exhausted CD8 T cell signature genes, as described by Guo et al., and the calculated GSVA enrichment scores in CCR8-high and low group are shown. Statistical analysis significance was determined by unpaired t-test with Welch's correction (*p ≤ 0.05; **p < 0.01; ***p < 0.001). www.nature.com/scientificreports/ canceled the CD8 T cell activation by CCR8+ cell depletion. Tumor antigen is taken up and processed by APCs and presented to CTLs by MHC class I 11,27 . Blocking of MHC class I inhibits this process. Thus, CCR8+ Tregs were considered to suppress CD8 T cell activation via APCs. Third, we directly evaluated the immune suppressive function of CCR8+ Tregs in CTL cytotoxicity using CCR8+ Tregs induced from PBMCs. In the co-culture with CD8 T cells, only CCR8+ Tregs dramatically suppressed the cytotoxic ability of CD8 T cells in both the antigen non-specific model and the antigen-specific model. Interestingly, CCR8− Tregs did not suppress cytotoxicity in either model. We speculated that CCR8+ Tregs may be responsible for most Treg suppressive functions against CTL cytotoxicity. In these experiments, it took 4 days for CCR8+ Tregs to suppress CD8 T cell cytotoxicity, and CCR8+ Tregs did not suppress CD8 T cell cytotoxicity at earlier time points (data not shown). CCR8+ Tregs are likely to suppress the differentiation process of CD8 T cells to acquire CTL functions over the 4-day period. On the basis of the results from these three approaches, we concluded that CCR8+ Tregs strongly suppressed CD8 T cell effector functions for cancer immunity.

Scientific Reports
There have been several reports showing that human CCR8+ Tregs are highly suppressive on the basis of their expression profiles, but only a few reports directly demonstrated their suppressive function against CD8 T cell activity. One report by Wang et al. 23 revealed that CD8 T cells in muscle-invasive bladder cancer with high levels of CCR8+ Tregs displayed decreased expression of effector molecules (IFNγ and TNFα) and elevated expression of inhibitory receptors (PD-1 and TIGIT). Although we did not analyze such a detailed CD8 T cell profile by flow cytometry in the lung tumors of the high and low CCR8+ Treg groups, our GSVA analysis using TCGA dataset of lung cancer had similar results to the study by Wang et al. in terms of effector CD8 T cells 23 . Most likely, the results in exhausted CD8 T cells differed depending on whether the analysis was performed on whole tumors or only in CD8 T cells. Wang et al. also showed that blockade of CCR8 with CCR8 neutralizing antibody in ex vivo tumor culture for 12 h promoted Treg destabilization and led to significant upregulation of IFNγ and TNFα in CD8 T cells 23 . Our study demonstrated that CCR8+ Treg depletion using MicroBeads markedly enhanced not only IFNγ and TNFα expression but also that of the cytotoxicity molecule GZMB via APCs www.nature.com/scientificreports/ during 5-day ex vivo culture of lung TICs. These results estimated the effect of anti-CCR8-depleting antibody in humans. Furthermore, we successfully induced, isolated, and co-cultured CCR8+ Tregs with CD8 T cells and demonstrated that only CCR8+ Tregs exhibited the ability to suppress the cytotoxic activity of CTLs, whereas CCR8− Tregs could not suppress it. This is first report to reveal the direct suppressive functions of CCR8+ Tregs against CD8 T cells using human lung cancers and induced human Tregs. However, the effects on APCs, conv CD4 T cells, and other effector cells were not evaluated, and this is an issue for future study.  www.nature.com/scientificreports/ Several mechanisms of Treg suppressive functions in CTLs have been identified, such as the production of suppressive cytokines, IL-2 consumption, the APC-mediated pathway via CTLA-4, TIGIT, and LAG3, the metabolite-related mechanism via CD39/CD73, contact-dependent suppression by PD-L1, and other mechanisms 9,10,12,[28][29][30] . In this study, the results of CCR8+ cell depletion among TICs showed the requirement of APCs for the suppression of CD8 T cells by CCR8+ Tregs. We observed higher expression of CTLA-4, TIGIT, and LAG3 in CCR8+ Tregs than in CCR8− Tregs, and therefore we speculated that the CCR8+ Treg suppressive functions exerted via APCs by these molecules are important for suppressing CD8 T cell function. In addition, a Treg suppressive pathway independent of APCs was also observed. Induced CCR8+ Tregs suppressed the antigen-specific cytotoxicity of CTLs, even in the absence of IL-2 and APCs. We surmised that the production of inhibitory cytokines such as TGFβ and IL-10 or the expression of CD39 and PD-L1 are important for suppressing CTL cytotoxicity under these conditions. Further analysis is necessary to investigate the most important mechanism of CCR8+ Tregs for suppressing CTL functions and why CCR8− Tregs do not show these effects.
Finally, we revealed the therapeutic potential of targeting CCR8 for cancer treatment using a murine model of lung cancer. Treatment with anti-CCR8 antibody reduced Tregs in tumor tissue and showed a remarkable anti-tumor effect including complete response. These results suggest that CCR8-targeted therapy is effective for lung cancer in vivo. From the results of ex vivo culture of human lung TICs (Fig. 5), CD8 T cell activation was observed with the removal of small amounts of Tregs. This indicates that even if the depletion efficiency is slightly low, it can still be effective. In this study, we used a subcutaneous transplantation model because of the difficulty in conducting stable orthotopic tumor transplantation, but the evaluation of CCR8-targeted therapy in an orthotopic lung cancer model is another issue for future study.
In conclusion, we revealed the pathophysiological characteristics of CCR8+ Tregs in human lung cancers and their inhibitory effect on CD8 T cells. We found that CCR8+ Tregs are responsible for the suppressive function against CTL cytotoxicity. We also revealed the anti-cancer efficacy of CCR8-targeted therapy using a murine lung cancer model. We provide evidence that CCR8-targeted therapy may be effective for the treatment of lung cancer. Several CCR8-targeted drugs are currently in development, and we are also planning a clinical study of anti-human CCR8 antibody for human cancer. Together, these findings may help to provide evidence and a foundation for the development of CCR8-targeted immunotherapy for cancers including lung carcinoma.

Materials and methods
Human samples. We obtained 50 fresh lung tumor tissues and 11 peripheral blood samples from lung cancer patients during primary surgical treatment from May 2017 to November 2020 at Osaka University Hospital. The clinical characteristics of the patients are summarized in Supplementary Table S1. Patients did not receive any neoadjuvant therapy. Peripheral blood samples from 7 healthy donors were also obtained. All participants provided written informed consent before sampling. This study was conducted in accordance with the Declaration of Helsinki and was approved by the Institutional Ethics Committee of Osaka University Hospital (#13266-15, #8226-10).
Blood and tissue sample preparation. PBMCs were purified by gradient density centrifugation using Lymphoprep (Axis Shield, Dundee, UK). To prepare tissue-infiltrating cells in tumor tissues, tumor-adjacent normal tissues and lymph nodes, patients tissue samples were minced using surgical scalpels and further enzymatically dissociated using the human Tumor Dissociation Kit (Miltenyi Biotec, Bergisch Gladbach, Germany) and gentleMACS Dissociator (Miltenyi Biotec) according to the manufacturer's protocol. Cell suspensions were filtered through a 70-mm cell strainer and isolated by Percoll (GE Healthcare, Tokyo, Japan) gradient centrifugation.
Flow cytometry. Cells  RNA-sequencing and data processing. One million CCR8+ or CCR8− Tregs (CD45+ CD3+ CD4+ CD45RA-CD25 high ) cells were sorted from three lung tumors by FACS Aria II (BD Bioscience). RNA was extracted and purified using TRIzol Reagent (Thermo Fisher Scientific) and miRNeasy Micro Kit (Qiagen, Germany). Library preparation was performed using RNA with the Ion Total RNA-Seq Kit v2 (Thermo Fisher Scientific) and sequencing was performed by Ion Proton in triplicate. Sequencing data were mapped to mm9 with TopHat2 (version 2.0.11). The count data are shown in Supplementary Table S4. For analysis of DEGs between CCR8+ Tregs and CCR8− Tregs, tag counts obtained by HT-seq (version 0.6.1) and the data were normalized using DESeq2 package in R software (version 4.1.0). We performed GO enrichment analysis for biological process terms in the SOM bidimensional space using DAVID (https:// david. ncifc rf. gov/ home. jsp). The top 100 DEGs with the lowest p-value among CCR8+ Treg upregulated genes were used for terms ranking and selection of DAVID.
The Cancer Genome Atlas (TCGA) analysis. We downloaded TCGA Pan-Cancer Atlas mRNA expression data including lung adenocarcinoma and lung squamous cell carcinoma data from cBioPortal (https:// www. cbiop ortal. org). The mRNA expression data were normalized by Expectation-Maximization method. GSVA was performed according to the method by Hänzelmann et al. 30  www.nature.com/scientificreports/ All tumor-bearing mice were euthanized according to institutional animal care guidelines on the basis of tumor size, body weight or general condition. Euthanasia was performed by isoflurane inhalation followed by cervical dislocation.
Statistical analysis. GraphPad Prism Version 8 and Excel software were used for statistical analysis.
ANOVA was used for group comparisons, using a Dunn post hoc multiple comparison test. The difference between the two groups was assessed using Student's t-test or two-tailed paired t-test. The disease-free survival DFS rate were estimated using the Kaplan-Meier method and compared by the log-rank test. The Mann-Whitney U test was used for results from the in vivo study. A value of p ≤ 0.05 were considered significant.

Data availability
The human RNA-sequence data have been deposited in the DNA Data Bank of Japan under accession number JGAS000454. Data generated during this study are available from the corresponding author on reasonable request. www.nature.com/scientificreports/