Th7R Cells: CD4⁺ T-Cell Partners That Drive Tpex-Mediated Antitumor Immunity in the Tumor Microenvironment

Th7R Cells: CD4⁺ T-Cell Partners That Drive Tpex-Mediated Antitumor Immunity in the Tumor Microenvironment | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Article Th7R Cells: CD4 ⁺ T-Cell Partners That Drive Tpex-Mediated Antitumor Immunity in the Tumor Microenvironment Hiroshi Kagamu, Shota Takei, Satoshi Yamasaki, Ou Yamaguchi, Atsuto Mouri, and 10 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-6764582/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted You are reading this latest preprint version Abstract The priming, expansion, and function of CD8⁺ T cells depend on CD4⁺ T-cell help via dendritic cells. Precursor exhausted T cells (Tpex) maintain self-renewal and supply cytotoxic CD8⁺ T cells in the tumor microenvironment (TME), but their CD4⁺ T-cell partners were unclear. We performed scRNA-seq, scTCR-seq, and mass cytometry on peripheral blood, tumor, and lymph nodes from lung cancer patients and identified that IL-7R high CCR6⁺ Th1-like CD4⁺ T cells (Th7R) are numerically and spatially partnered with Tpex. Th7R expressed lymphotoxin-β and CXCL13, correlated with high endothelial venules, and co-localized with Tpex in tertiary lymphoid structures. Th7R abundance correlated with Tpex numbers in TME and lymph nodes, and adoptive transfer of Th7R increased Tpex in a mouse model. In patients, Th7R and Tpex in TME were associated with better response to neoadjuvant PD-1 blockade therapy. These results suggest that Th7R act as partners of Tpex to support sustained antitumor T-cell immunity. Biological sciences/Immunology/Tumour immunology Biological sciences/Immunology/Translational immunology tumor microenvironment immune checkpoint inhibitors T-cell immunity Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Introduction Precursor exhausted CD8⁺ T cells (Tpex) were identified by scRNA-seq as a subset of effector T cells with self-renewal capacity that support long-term antitumor immunity in the tumor microenvironment (TME) 1–3 . Unlike cytotoxic T lymphocytes (CTLs), which lose TCF7 expression and die after tumor killing, Tpex maintain TCF7 and IL-7 receptor expression, infiltrate tumors via high endothelial venules (HEVs), and localize within tertiary lymphoid structures (TLSs) 4,5 . Tpex are thought to be the source of T-cell expansion following PD-1 blockade therapy 1 . CD4⁺ T-cell help is essential for inducing effective CD8⁺ T-cell responses, including Tpex 6 . It has been shown that CD4⁺ T-cell-mediated signals, particularly in the presence of antigen-presenting dendritic cells, prevent deep exhaustion of CD8⁺ T cells and promote TCF7 and IL-7 receptor expression 7,8 . CD4⁺ T cells are also involved in forming ectopic HEVs, crucial for Tpex entry into tumors 9 . However, the specific CD4⁺ T-cell subset responsible for supporting Tpex is not well defined. Upon priming, CD4⁺ T cells differentiate into functionally distinct subsets 10 . While Th1 cells are traditionally linked with antitumor immunity through IFN-γ and T-bet expression, recent analyses of immune checkpoint inhibitor (ICI) responses have identified Th1-like, non-canonical CD4⁺ T cells in tumors. We previously described a unique CD4⁺ T-cell population, Th7R, which expresses TCF7 and IL-7R and is associated with better outcomes in lung cancer patients treated with ICIs or surgery 11, 12 . Based on their shared molecular features with Tpex, we hypothesized that Th7R cells may function as CD4⁺ T-cell partners of Tpex. In this study, we analyzed CD4⁺ and CD8⁺ T cells from lung cancer patients using scRNA-seq, scTCR-seq, and imaging mass cytometry to define the relationship between Th7R cells and Tpex in antitumor immunity. Results Cells with the same TCR clonotypes and gene expression signature as Th7R cells in the peripheral blood are present in the tumor microenvironment. Th7R is a novel Th1-like CD4 + T-cell cluster identified by scRNA-seq analysis of peripheral blood from advanced-stage lung cancer patients who responded to immune checkpoint inhibitor (ICI) therapy 11 . To clarify whether Th7R cells are also present in the tumor microenvironment (TME), we analyzed the scRNA-seq and scTCR-seq data of CD4 + T cells derived from the peripheral blood and tumor tissues of five lung cancer patients who had undergone surgery and examined whether CD4 + T-cell clusters with the same gene expression patterns and TCR clonotypes as peripheral blood Th7R cells are present in lung cancer tissues. A total of 58,687 CD3D + CD4 + cells (50,608 PB-CD3D + CD4 + cells and 8,079 TIL-CD3D + CD4 + cells) were analyzed, and unsupervised clustering was performed on the basis of the expression of 2,000 highly variable genes (Fig. 1 A). Since Th7R cells were discovered in a cluster of effector T cells with low SELL expression in the peripheral blood, we first focused on the SELL low-expression clusters 12, 16, 19, and 20 (Fig. 1 B, Fig. S1 ). The SELL low-expression clusters contained many clonally expanded T cells (Fig. 1 C, Table S1 ). Cluster 8, which contained many multicellular clonotypes with high SELL expression in the TME, was thought to include regulatory T cells (Tregs) with high FoxP3 expression (Fig. 1 D). Next, the independence of the TIL-12, TIL-16, TIL-19, and TIL-20 clusters was examined by analyzing TCR sharing in multicellular TCR clonotypes, which included 2 or more T cells in each TIL cluster (Fig. 1 E ) . The TIL-12 clonotypes were only shared with the peripheral blood (PB)-12 cluster, indicating that the T cells belonging to PB-12 do not significantly change their gene expression patterns even after infiltration into the TME. TCR sharing with TIL-19 cells was found not only for the PB-19 cluster but also for the TIL-16 cluster. On the other hand, the clonotypes belonging to TIL-16 were mostly found in the PB-16 cluster. Thus, T cells in the PB-19 cluster are likely to change their gene expression patterns to the cluster 16 state after infiltrating the TME, suggesting that clusters 16 and 19 exhibit two gene expression states of one type of differentiated T cell. Accordingly, clusters 16 and 19 were considered a single cluster. TCR clonotypes belonging to the TIL-20 cluster were rarely detected in other clusters. The total cell counts and TCR clone sizes for each cluster are shown in Fig. 1 F. Cluster12 had greatly expanded TCR clones in the peripheral blood but lacked expanded TCR clones in the TME. It is likely that the abundant multicellular clonotypes in the PB-12 cluster reflect clonal expansion in the secondary lymphoid organs and that some of the expanded cells infiltrate the TME. In contrast, cluster 16 exhibited few expanded TCR clonotypes in the peripheral blood but the greatest number of expanded TCR clonotypes in the TIL-16 cluster, suggesting clonal expansion in the TME rather than in the secondary lymphoid organs. The SELL low effector CD4 + T cells in the peripheral blood included canonical Th1-type cells, Th17-type cells and Th7R cells, and chemokine receptor expression patterns revealed that these cells were CXCR3 + CCR4 - CCR6 - Th1, CXCR3 - CCR4 + CCR6 + Th17 cells, and CXCR3 ± CCR4 - CCR6 + Th7R cells, respectively 13 . Next, we examined the signature genes that could discriminate among Th1, Th17, and Th7R cells. First, CXCR3 + CCR4 - CCR6 - , CXCR3 - CCR4 + CCR6 + , and CXCR3 ± CCR4 - CCR6 + T cells sorted from the peripheral blood of CD62L low CD4 + T cells from lung cancer patients (n = 3) other than the 5 patients in this study were analyzed by scRNA-seq. Differential expression analysis identified 64 genes that were differentially expressed among these subsets, including both upregulated and downregulated genes, thereby enabling the characterization of gene expression signatures associated with the Th1, Th7R, and Th17 phenotypes. ( Table S2 ). We refined the 64-gene set to 58 genes by applying support vector machine-based recursive feature elimination (SVM-RFE) for feature selection ( Fig. S2 A-C ). By using the expression data for the 58 selected genes from sorted Th1, Th7R, and Th17 cells, we trained an SVM classifier and applied it to discriminate the Th subtypes of the cell clusters shown in Fig. 1 A. As a result, cluster 12 was predicted to be Th1 cells, cluster 20 was predicted to be Th17 cells, and cluster 16 + 19 was predicted to be Th7R cells (Fig. 1 G, Fig. S2 D ). We evaluated the similarity of the sorted subsets and SVM-predicted clusters using an expression heatmap of 58 genes combined with cosine similarity analysis. This approach revealed strong concordance ( Fig. S2 D ). Furthermore, to investigate whether this gene expression pattern is derived from epigenetic changes, we used the 10x Genomics Multiome platform to perform both single-nucleus RNA-seq (snRNA-seq) and single-cell ATAC-seq (scATAC-seq) of the same nuclei (Fig. 1 H, Fig. S3 A ). Among the clusters obtained on the WNN-UMAP diagram, clusters 4, 6, 11, 14, and 20 were considered SELL low-expression clusters ( Fig. S3 B, C ). On the basis of SVM-based prediction of Th subtypes using the 58 signature genes, clusters 6 and 20 were classified as Th1 cells, cluster 14 was classified as Th17 cells, and clusters 4 and 11 were classified as Th7R cells (Fig. 1 I, Fig. S3 D ). The overall gene expression patterns of the clusters classified as Th1, Th17, and Th7R cells were very similar to those of sorted Th cells (Fig. 1 J ) . The gene activity (GA) scores of 54 of the 58 signature genes obtained from DNA chromatin information also matched the gene expression patterns of each Th subset (Fig. 1 K, Fig. S4 ). These results indicate that Th7R cells have a characteristic gene expression pattern that differs from those of Th1 and Th17 cells and that they have a DNA chromatin state that supports this gene expression profile. Th7R cell gene expression in the TME To perform high-resolution CD4 + T-cell clustering and gene expression analysis of TILs, we combined the results of three additional TIL and PB scRNA-seq analyses. A total of 79,130 CD4 + CD3D + cells (65,572 PB-CD3D + CD4 + cells and 13,558 TIL-CD3D + CD4 + cells) were subjected to unsupervised clustering on the basis of 2,000 genes with highly variable expression, generating 27 clusters (Fig. 2 A, Table S3 , Fig. S5A, B ). In the UMAP plot, clusters 13, 14, 18, 21, and 26 were located in the SELL -low expression area ( Fig. S5C ). On the basis of the highest match rate in the SVM-based prediction using the 58 signature genes, the cells in cluster 13 + 26 were annotated as Th1 cells, the cells in cluster 18 + 21 were annotated as Th7R cells, and the cells in cluster 14 were annotated as Th17 cells (Fig. 2 B). For Th1 and Th17 cells, the peripheral blood cluster (prediction probability of 1.000, 1.000) and the TIL cluster (1.000, 0.842), showed good predictive agreement. In contrast, the Th7R cluster had a greater probability in the peripheral blood (0.993) but a lower probability in TILs (0.557), suggesting altered gene expression in the TME. However, the probability of the TIL Th7R cluster containing Th1 (0.171) and Th17 (0.272) cells was low, suggesting that Th7R cells have a unique gene expression pattern in the TME. Therefore, we compared the gene expression profile of the Th7R cluster between peripheral blood and TILs and found that 115 genes were upregulated more than 2-fold in TILs and 18 genes were downregulated more than 2-fold in TILs (Fig. 2 C, D; Table S4 ; Fig. S5D ). The genes upregulated in the TME included CCL4, CXCL13, CREM, CCL4L2, GZMB, CTLA4, CXCR4, IFNG, ID2, XCL1, CXCR6, CCL3 and ICOS . Next, we investigated the differences in gene expression between Th7R and Th1 or Th17 cells in the TME (Fig. 2 E, Fig. S5E ). The genes whose expression was more than 2-fold greater in Th7R cells than in canonical Th1 cells in the TME included LTB, CXCR6, CTLA4 , and CXCL13 (Fig. 2 F, Table S5 ). The genes whose expression was more than 2-fold greater in Th7R cells than in Th17 cells included CCL4, CCL4L2, CXCL13, CCL5, GZMK, GZMB, IFNG , and CCL3 . Relationship between Th7R cells and HEV formation in the TME CXCL13 was found to be upregulated in Th7R cells in the TME, with a significant difference from the expression observed in Th1 and Th17 cells. Th7R cells also expressed more LTB ; LTβ and CXCL13 are known to be essential for the formation of ectopic HEVs and TLSs 14 . To investigate the relationship between Th7R cells and ectopic HEV formation, immunohistochemical (IHC) analysis and Hyperion™ analysis were performed (n = 56, Table S6 ). Sections were stained with antibodies against the normal vascular marker CD31 and the HEV marker peripheral node addressin (PNAd), and the ratio of PNAd-positive areas to CD31-positive areas (the MECA index) was calculated to assess HEV density and its relationship with the proportion of Th7R cells in the peripheral blood (Fig. 3 A). Consistent with previous reports, significantly better disease-free survival (DFS) was observed in patients with more HEV formation in the TME, and a MECA index of 0.044 was found to predict better DFS (Fig. 3 B). A chi-square test using a Th7R percentage of 3.05, which was used as a threshold for predicting better DFS after surgery for lung cancer in a previous study, and a MECA index of 0.044 showed that HEV density was associated with Th7R abundance (P = 0.0080, Fig. 3 C) 12 . Tpex as a source of Tex in the TME Tpex have self-renewal ability and contribute to a sustained supply of CD8 + T cells to maintain long-term antitumor immunity through stem cell-like asymmetric cell division 5 . To analyze the relationship between Th7R cells and Tpex in the TME, we attempted to identify Tpex in lung cancer tissues. We performed scRNA-seq analysis of 25,728 CD3D + CD8 + cells (17,997 PB-CD3D + CD8 + cells and 7,731 TIL-CD3D + CD8 + cells) derived from 8 surgery patients. Unsupervised clustering was performed on the basis of the expression of 2,000 highly variable genes, generating 16 clusters, as shown in the UMAP plot (Fig. 4 A, Fig. S6 ). We analyzed the gene expression of GZMK and GZMB as candidate markers for discriminating among Tpex, exhausted CD8 + T cells (Tex), and unexhausted cytotoxic T lymphocytes (CTLs) (Fig. 4 B) 2,3,15 . GZMK expression was observed in clusters 1, 2, 3, 4, 8, 10, 11, 13 and 14 adjacent to the SELL high CD45RA + naive clusters (6, 12, and 15) and the SELL high CD45RA - central memory (CM) cluster (7) (Fig. 4 C). GZMB expression was observed in cluster 5, which was farthest from the naive clusters, and clusters 1, 4, 8, 11, and 13. Cluster 9 was CCR6 + and was considered mucosa-associated invariant T (MAIT) cells. Next, we assessed the expression of signature genes thought to characterize Tpex in each cluster (Fig. 4 D) 5 . GZMK + GZMB - clusters 2, 3, 10, and 14 expressed TCF7 , IL-7R , PDCD1 , SELL , and CCR7 and were consistent with Tpex. On the other hand, GZMK - GZMB + cluster 5 was considered a CTL cluster because it expressed no exhaustion genes and highly expressed PRF1, GZMB, GNLY, FGFBP2 , and NKG7 . The GZMK + GZMB + clusters 1, 4, 8, 11, and 13 were considered Tex, as they did not express IL-7R or TCF7 but highly expressed exhaustion genes such as PDCD1 , CTLA-4 , ENTPD1 , and HAVCR2 . These results indicate that Tpex, Tex, and CTLs can be distinguished on the basis of their GZMB and GZMK expression patterns. When the cell density of each pixel in the PB-UMAP plot was subtracted from the cell density of each pixel in the TIL-UMAP plot, clusters 1, 2, 10 and 13 presented a distribution with a TME-dominance (Fig. 4 E, Table S7 ). The direction of the state change for each cluster in the TME and peripheral blood was also examined by velocity analysis (Fig. 4 F). In the peripheral blood, changes to the Tex clusters starting from cluster 5, which is a cluster of CTLs that have not been exhausted, were observed, and it was thought that CTLs that proliferated in the secondary lymphoid organs changed their state in response to repeated antigen stimulation. In contrast, TIL velocity analysis revealed that cluster 2, which was thought to be a Tpex cluster, was the source of the surrounding clusters, including Tex. Considering that clusters 1 and 2 were dominant in the TME, the main progression of T-cell state changes is thought to be from the Tpex cluster 2 to the terminally exhausted cluster 1, which has the highest expression of HAVCR2, CTLA4 , and ENTPD1 . These results indicate that Tex, which have cytotoxic activity, are supplied mainly by Tpex, which replicate in the TME. Spatial relationship between Th7R cells and Tpex in TLSs IHC analysis revealed that CXCR3 + CCR6 + CD4 + T cells were located in TLSs ( Fig. S7 ). Next, we performed a neighborhood analysis of Th7R cells and Tpex on the basis of the molecular expression information obtained using Hyperion™ (Fig. 5 A). We used software that can segment and annotate individual cells on the basis of their molecular expression patterns (Fig. 5 B, Fig. S8A ). GZMB + cells, which constitute a population with cytolytic functions, were found in the cancer cell nest. In contrast, GZMB − GZMK + CD8 + T cells, which are considered Tpex, were found to be abundant within the B-cell aggregates considered TLSs. CXCR3 + CCR6 + CD4 + T cells, which are considered Th7R cells, were found near Tpex within TLSs (Fig. 5 C). Correlations between Th7R and Tpex numbers To determine whether increased Th7R abundance in the TME affects CD8 + T-cell counts, we analyzed TILs from mice subjected to adoptive transfer of Th7R cells. Th7R cells were sorted as CXCR3 ± CCR4 − CCR6 + CD4 + T cells from inguinal lymph nodes draining a methylcholanthrene-induced fibrosarcoma MCA205 skin tumor, and after ex vivo culture, they were adoptively transferred into mice bearing day-11 MCA205 skin tumors. Tumors were harvested 11 days after cell transfer from 3 mice each, and TILs were analyzed. Tumors (310 mg) from untreated mice and tumors (270 mg) from mice subjected to Th7R-transfer were digested, yielding 49.1 million and 43.1 million cells, respectively (Fig. 6 A). TILs obtained by density gradient separation using Percoll were analyzed using a mass cytometer. Adoptive transfer of Th7R cells resulted in an increase in the number of CD8 + T cells in the tumors, with a marked increase in the size of the Tpex and Tex populations. Therefore, Th7R cells likely promote asymmetric cell division of Tpex, resulting in increases in Tpex as abundance and the supply of Tex. Next, we examined whether the percentages of CTLs and Tpex among total CD8 + T cells were correlated with the percentages of Th1 and Th7R cells among total CD4 + T cells in cluster 12 in the draining lymph nodes from lung cancer patients (the hilar lymph nodes) and peripheral blood derived from lung cancer patients (Fig. 6 B, C). In both locations, Th1 cell abundance was positively correlated with CTL abundance, and Th7R abundance was positively correlated with Tpex abundance. To eliminate overlapping relationships and clarify the relationships between cells, a quantitative network analysis of the proportions of clusters in CD4 + T cells and CD8 + T cells relative to CD3 + cells in the peripheral blood was performed ( Fig. S8B ). Notably, the only CD4 + T-cell cluster whose abundance directly correlated with that of Tpex was Th7R cells, with a connection established in 1000 out of 1000 interferences. These results support the hypothesis that Th7R cells are involved in the intratumoral infiltration and proliferation of Tpex. To evaluate whether Th7R cells and Tpex in the TME contribute to antitumor immunity, we investigated their associations with those responses to preoperative neoadjuvant therapy including anti-PD-1 antibody ( Table S8 ) 16 . There was a significantly greater abundance of Tpex in the patients who achieved a RECIST partial response by imaging. In addition, Th7R abundance was significantly greater in patients who achieved a complete pathological response or major pathological response (Fig. 6 D). Thus, Th7R cells and Tpex in the TME likely play important roles as partners in antitumor immunity. Discussion In this study, we showed that Th7R cells, a novel Th1-like CD4 + T-cell lineage that exhibits different gene expression patterns from Th1 and Th17 cells, plays a distinctive role in antitumor immunity as a partner of Tpex. To define Th7R cells, an analysis of the expression of 58 signature genes obtained by machine learning was performed, and IL7R, TCF7, IFNG-AS1 , and NELL2 were selected as those that exhibited Th7R-dominant expression. IL7R and TCF7 , which are known to contribute to the survival and proliferation of T cells, are expressed at extremely low levels in Th1 cells but are expressed to some extent in Th17 cells. Th7R cells, which express RORC , are thought to share the characteristics of high survival and self-replication ability with Th17 cells. On the other hand, Th7R cells express transcription factors, cytokines, and chemokines that are common to Th1 cells, such as TBX21, EOMES, IFNG, CCL4 , and CCL5 , and it is highly likely that many of their effector functions are carried out via the expression of Th1-related genes. On the other hand, Th7R cells express LTB and CXCL13 , which are not found in Th1 cells, in the TME, and Th7R abundance is correlated with the density of HEVs in the TME. Since ectopic HEV/TLS formation is disrupted when IL-7R, RORC, or LTB is knocked out, these three factors are considered essential for the function of LTis 17 . Th7R cells are characterized by high expression of IL-7R and express RORC , LTB , and CXCL13; thus, Th7R cells are considered good candidates for LTis in the TME. Compared with their expression levels in the peripheral blood, Th7R cells exhibited significantly upregulated expression of more than 100 genes in the TME (by more than twofold). The upregulated genes included CCL4, CCL4L2, CXCL13, XCL1 , and XCL2 , and it is thought that they have the ability to recruit CD8 + T cells, CXCR5-positive immune cells, and XCR1-positive dendritic cells to the TME 18,19 . In fact, the number of CD45 + cells, including CD8 + T cells, increased more than threefold in the tumors of mice subjected to adoptive transfer of Th7R cells. Compared with Th1 and Th17 cells, Th7R cells had the strongest TME-dominant distribution and were the Th cells that included the most clonally expanded TCR clonotypes in the TME. This is in contrast to Th1 cells, which include the most clonally expanded TCR clonotypes in the peripheral blood. There are two possible explanations for this distribution pattern: one is that Th7R cells, which have a TCR specific to cancer antigens, may receive a stop signal in the TME and accumulate in large numbers, and the other is that Th7R cells may self-replicate in the TME in response to cancer antigen stimulation 20 . The upregulation of JUNB and FOSB in the TME suggests that Th7R cells self-replicate in the TME 21 . In any case, the fact that Th7R cells contain many multicellular TCR clones in the TME that are not found in the peripheral blood strongly suggests that many TIL-Th7R cells recognize cancer antigens. As evidence that Th7R cells are repeatedly stimulated by antigens in the TME, the expression of exhaustion-related genes such as TOX, PDCD1, TIGIT, LAG3 and CTLA4 was upregulated in TIL-Th7R cells ( Table S4 ). One characteristic of the exhaustion-related expression pattern of Th7R cells is that the expression of CTLA4 was significantly greater than that of Th1 and Th17 cells. These findings suggest that anti-CTLA-4 antibody therapy may be effective in reinvigorating Th7R cells. Consistent with previous reports, the IHC and Hyperion ™ results revealed that Tpex are localized to TLSs, and velocity analysis revealed that Tpex play a central role in providing cytotoxic Tex. On the other hand, surprisingly, most of the Th7R cells in the TME were found to be distributed in TLSs and to be in very close proximity to Tpex. It has been reported that CD8 + T cells are reprogrammed to express IL7R and TCF7 , avoiding terminal exhaustion, by forming a triad of physical contacts with dendritic cells and CD4 + T cells 7 . This study indicates that Th7R cells are important candidate partners for CD4 + T cells that play a role in triad formation, as they are present near Tpex within the TLS and are capable of recruiting dendritic cells due to their upregulated expression of XCL1 and XCL2 . In fact, a marked increase in Tpex and Tex abundance was observed in tumors from mice that had been subjected to adoptive transfer of Th7R. In contrast, the number of CTLs in the tumor did not increase substantially. Thus, Th7R cells likely promoted Tpex proliferation but did not facilitate CTL proliferation in the draining lymph nodes. Consistent with these findings, the results of a correlation analysis of the abundance of CD8 + T-cell subpopulations in the lymph nodes and peripheral blood of lung cancer patients revealed that Th7R numbers were positively correlated with Tpex numbers. On the other hand, Th1 cell abundance was positively correlated with CTL abundance. Taken together, Th7R and Th1 cells share some transcription factors, cytokines, and chemokines but have distinct CD8 + T-cell subpopulations as partners; as a result, only Th7R cells likely play an important role in long-term antitumor immunity. In the early stage of cancer development, CD8 + T cells that clonally proliferate in the draining lymph nodes, as shown in the 7 steps of the cancer immunity cycle, are considered the main players in cancer cell eradication 22 . On the other hand, Tpex, which are responsible for the sustained supply of Tex in the TME, are thought to control the long-term equilibrium phase according to the cancer immunoediting theory and are a critical target for reinvigorating effective antitumor immunity via PD-1 blockade therapy 2,3,23,24 . Consistent with this idea, Tpex and Th7R cells, but not CTLs or Th1 cells, were significantly more abundant in the TME of lung cancer patients who achieved pCR or MPR after neoadjuvant ICI therapy. In summary, Th7R cells, Th1-like CD4 + T cells that are involved in the generation of HEVs and TLSs in the TME, are deeply involved in long-lasting antitumor immunity by partnering with Tpex. Th7R cells, which exhibit TCR sharing between the TME and the peripheral blood, are distributed systemically, so their abundance is thought to be useful as a biomarker that reflects parameters associated with long-lasting antitumor immunity in the TME, such as HEVs/TLSs and Tpex. In addition, cell therapy using Th7R cells in combination with ICI therapy has promise as a treatment to promote Tpex proliferation. Online Methods Patients TIL and peripheral blood samples for single-cell RNA sequencing (scRNA-seq, n = 5), along with tissue samples for IHC analysis (n = 56, Table S6 ) and Hyperion™ analysis, were collected from patients with stage I–II non-small cell lung cancer (NSCLC) who underwent radical surgery between October 2017 and October 2019 at Saitama Medical University International Medical Center (Hidaka city, Saitama pref., Japan). TIL samples for mass cytometry analysis were collected from patients (n = 20, Table S8 ) with stage II-IIIA NSCLC who underwent surgery after 3 courses of neoadjuvant therapy, including nivolumab and chemotherapy, between August 2023 and October 2024 at Saitama Medical University International Medical Center. The response to treatment was assessed using Response Evaluation Criteria in Solid Tumors (RECIST), version 1.1. Pathological response was independently assessed by pathologists. All the samples were collected after written informed consent was obtained from the patients. The Internal Review Board of Saitama Medical University International Medical Center approved the study protocol, in accordance with the Declaration of Helsinki (ethical approval Nos. 17–084, 15–221, and 2023-011). Blood samples Peripheral blood samples were collected using heparinized CPT Vacutainer tubes (Becton Dickinson Vacutainer Systems, Franklin Lakes, NJ, USA), as previously described 25 . The samples were preserved using Cellbanker2 (Nippon Zenyaku Kogyo Co., Koriyama, Japan) in a liquid nitrogen tank. For T-cell subset analyses, the cells were incubated for 32–48 h in RPMI 1640 medium supplemented with 10% fetal calf serum (FCS) in a 5% CO 2 incubator at 37°C. TIL samples To extract TILs, the collected tumors were sectioned into 2–3-mm pieces and incubated with Dulbecco's modified Eagle’s medium supplemented with Dri Tissue & Tumor Dissociation Reagent (BD horizon: 661563; BD Biosciences, Franklin Lakes, NJ, USA) or with RPMI 1640 medium supplemented with Liberase TL Research Grade (05401020001; Roche, Basel, Switzerland) following the manufacturer’s instructions. Erythrocytes were removed using a biotin anti-human CD235ab antibody (306617; BioLegend, San Diego, CA, USA) and streptavidin nanobeads (480016; BioLegend, San Diego, CA, USA) according to the manufacturer’s instructions. Multiplex immunohistochemical staining (OPAL™), and imaging mass cytometry (Hyperion™) image acquisition and data analysis for tissue samples The tumor samples were formalin fixed and paraffin embedded (FFPE), and the three tumors with the largest areas of viable tumor cells were selected. Five-micron-thick sections were deparaffinized and rehydrated with xylene and ethanol for multiplex immunohistochemical staining. All the slides were sequentially treated with 0.3% hydrogen peroxide in methanol for 30 min to block endogenous peroxidase activity. To expose the antigens, the sections were autoclaved in 10 mM sodium citrate buffer (pH 6.0) for 20 min, heated in a microwave at 98°C for 15 min, and then cooled for 30 min. After rinsing in 0.05 M Tris-buffered saline containing 0.1% Tween 20, the sections were incubated with various antibodies: mouse monoclonal CD4 (Leica Biosystems, clone 4B12, 1:300, high pH, retrieval), mouse monoclonal CD8 (DAKO, clone C8/144B, 1:150, high pH, retrieval), mouse monoclonal CD31 (DAKO, clone JC70A, 1:400, pH 6, retrieval), rabbit monoclonal CCR6 (Abcam, clone EPR22259, 1:1000, high pH, retrieval), rat monoclonal peripheral node addressin (Novus, clone MECA79, 1:200, pH 6, retrieval), mouse monoclonal CXCR3 (BioLegend, clone G025H7, 1:300, pH 6, retrieval), CD20 (Akoya, clone L26, 1:200, high pH, retrieval), and mouse monoclonal pan cytokeratin (Abcam, clone AE1/AE3, 1:100, pH 6, retrieval). All the slides were stained using an OPAL™ 7-color IHC kit (Akoya Biosciences NEL811001KT, MA, USA). Immunofluorescence analysis was performed on the Mantra imaging platform (Akoya Biosciences) using Mantra Snap software. Color separation was conducted with inForm® Software v2.5.1 (Akoya Biosciences, MA) to extract image data. Multiplex immunohistochemical staining and data analysis were performed according to previously described procedures 26 . After heat-induced epitope retrieval, the tumor samples were subjected to Hyperion™ analysis. The sections were incubated with conjugated monoclonal antibodies overnight at 4°C ( Table S9 ). Following antibody incubation and washing, selected regions of interest (ROIs) were ablated with a laser at a resolution of 1 µm, and the vaporized particles were analyzed. The acquired data were visualized and analyzed using MCD Viewer and MCD Smart Viewer software (Standard BioTools). Pseudo-colored composite images were generated for each marker to assess tissue morphology and marker expression patterns. All images were reviewed at high resolution, and representative ROIs were selected for figure preparation. Cell sorting for scRNA-seq and scTCR-seq For peripheral blood scRNA-seq analyses, CD3 + T cells were isolated from frozen PBMCs that had been incubated for 36 hrs in CM with 5% CO 2 at 37°C, using Dynabeads Untouched Human T Cells (11344D; Invitrogen). For TIL scRNA-seq analysis, samples were used after erythrocyte removal. For presorted scRNA-seq analysis, CD4 + T cells were isolated as follows. CD62L low CD4 + cells were separated from cultured frozen PBMCs using Dynabeads. Untouched Human CD4 T Cells (11346D; Invitrogen) and Human CD62L MicroBeads (130-091-758; Miltenyi Biotec, Bergisch Gladbach, Germany) were used. CD62L low CD4 + T cells were subsequently sorted with CXCR3 (353715, 353716; BioLegend), CCR4 (353429, 353430; BioLegend), and CCR6 (359410; BioLegend) antibodies using a Cell Sorter SH800Z (Sony Imaging Products & Solutions Inc., Tokyo, Japan) prior to single-cell library construction. Library construction and sequencing for scRNA-seq and scTCR-seq All single-cell RNA-seq and scTCR-seq libraries were constructed using the Chromium controller and Chromium Single-Cell Immune Profiling v2 kit (10x Genomics Inc., Pleasanton, CA, USA) to analyze gene expression differences and T-cell receptor (TCR) repertoires among T-cell subpopulations. Several TotalSeq-C antibody types (BioLegend Inc., San Diego, CA, USA) were used as feature barcodes to separate T-cell subpopulations ( Table S10 ). All the libraries were constructed according to the manufacturers’ standard protocols. The quality and quantity of the libraries were evaluated using a Bioanalyzer High-Sensitivity Kit (Agilent Technologies, CA, USA). All libraries were sequenced on DNBSEQ-G400 sequencers (MGI, Shenzhen, China) in paired-end mode (read 1:28 bp; read 2:90 bp) according to the manufacturer’s standard protocol. The sequencing depth was targeted at 50,000 reads/cell for gene expression libraries and 10,000 reads/cell for feature barcoding and TCR libraries. Details of the libraries and sequence results are summarized in Table S11 . Data preprocessing and quality control for scRNA-seq and scTCR-seq The sequenced reads were processed using Cell Ranger 6 and 7 software (10x Genomics Inc., Pleasanton, CA, USA) with the GRCh38 reference dataset (version 2020-A for gene expression and 5.0.0 (samples from surgery patients) or 7.1.0 (presorted samples) for the TCR repertoire) obtained from 10x Genomics. This processing generated the gene expression/feature barcode (TotalSeq-C) unique molecular identifier (UMI) count matrix and TCR repertoire data. Quality control, statistical analysis, and graphical representation were performed using the Loupe browser (10x Genomics, Inc.) and the Seurat 5.2.1 package in R software (version 4.4.1) ( https://satijalab.org/seurat/ ) 27,28 . The gene expression library and feature barcoding library from the same sample were processed simultaneously using the ‘cellranger count’ command. TCR libraries were processed using the ‘cellranger vdj’ command for each sample. TCR repertoires from different samples (i.e., TILs and PBMCs) from the same patient were aggregated using the ‘cellranger aggr’ command to identify shared clonotypes across different libraries. Quality control of the scRNA-seq data was executed as follows. cells with a valid TCR clonotype within the CD3D + cluster were selected using the Loupe browser. Genes expressed in fewer than three cells were excluded from the analysis. Cells with either too few (≤ 200) or too many (> 2,500–3000) expressed genes were excluded to remove empty or multiple GEMs. Cells with a high percentage of mitochondrial genes (≥ 5–8%) were also excluded to eliminate dead cells, and those with excessively high TotalSeq-C counts were removed to minimize the impact of aggregated antibodies. These cutoff thresholds were set on the basis of the tail of the distribution. Prior to analysis, the UMI gene expression counts were log-normalized using Seurat's ‘LogNormalize’ function (scale factor = 10,000). scRNA-seq and scTCR-seq data integration and analysis of CD4 + and CD8 + T cells derived from surgery patients The filtered CD3D + T cells derived from surgical patients were categorized into CD4 + CD3D + T cells and CD8 + CD3D + T cells on the basis of the RNA and surface protein expression of CD4 and CD8A prior to integration. The integration of these CD4 + and CD8 + T cells was conducted following the standard integration workflow of Seurat version 5, utilizing the reciprocal PCA integration method. Initially, 2,000 highly variable genes (excluding TCR-related genes) were identified using the 'FindVariableFeatures' function with the vst method. The expression of these variable genes was then scaled and centered using the 'ScaleData' function. Principal component analysis was performed using the 'RunPCA' function, followed by batch correction among samples using the reciprocal PCA method with the 'IntegrateLayers' function. Unsupervised clustering of the integrated gene expression matrix was carried out using the 'FindNeighbors' (with the parameter “dims = 1:30”) and 'FindClusters' functions of Seurat on the basis of the shared nearest neighbor modularity optimization clustering algorithm 29 . The resolution parameter of 'FindClusters' was optimized on the basis of the number of expected clusters and the stability of the clustering results. Visualization via a UMAP plot was performed with the 'RunUMAP' function. Differential expression analysis was conducted using the MAST package in R software 30 , implemented in the 'FindMarkers' function of the Seurat package. Genes with an adjusted P value 1 were considered significant and are indicated by red dots on volcano plots drawn with the EnhancedVolcano package in R 31 . The RNA velocity was estimated using the velocyto 0.17.17 package and scVelo 0.2.5 in Python 3.6 32 . The Loom file with UMAP embedding calculated via Seurat analysis was merged with the Loom file containing RNA velocity data from velocyto, and merged data were filtered and normalized using the 'scvelo.pp.filter_and_normalize' function (min_shared_counts = 5, n_top_genes = 2000). First- and second-order moments were computed for each cell across its nearest neighbors using the 'scvelo.pp.moments' function (n_pcs = 30, n_neighbors = 100). Finally, RNA velocities were estimated in the dynamic model using the 'scvelo.tl.recover_dynamics' and 'scvelo.tl.velocity' functions. The estimated velocity was visualized as a stream plot using the 'scvelo.pl.velocity_embedding_stream' function. scRNA-seq analysis of presorted T cells Preparation and quality control of scRNA-seq libraries from presorted CD4 + T cells were performed in a manner similar to that used for PBMCs from surgical specimens, except that feature barcoding technology with TotalSeq was not utilized. For cells that passed quality control, log-normalized count data were merged for each patient, and differential expression gene analysis was conducted among the four cell types, Th1, Th17, CXCR3 + Th7R, and CXCR3 − Th7R, on a round-robin basis using the ‘FindMarkers’ function with the MAST method. This statistical test was applied to genes that were expressed in more than 10% of the cells and exhibited a minimum log 2 -fold change of 0.25. Genes with an adjusted P value < 0.05 were considered differentially expressed. The 64 DEGs common to all three patients were selected as candidate signature genes ( Table S2 ). To identify CD4 signature genes and construct a predictive model for CD4 cell subtypes, we employed support vector machines (SVMs). To mitigate the influence of variability in RNA expression at the single-cell level, we created mini-clusters by averaging the gene expression profiles of 10 randomly sampled cells. For each sorted population—Th1, Th7R, and Th17—we generated 4,000 mini-clusters and used them for downstream analysis. Feature selection was conducted using the 'rfeControl' function in the caret R package in combination with a radial kernel SVM implemented using the e1071 package. On the basis of this feature selection process, 58 genes were selected as relevant markers for classification. By using the expression profiles of these 58 genes, we built an SVM-based classifier to distinguish among the Th1, Th7R, and Th17 CD4 subtypes. Of the 4,000 mini-clusters generated per subtype, 2,000 were used for training, and the remaining 2,000 were used for testing to evaluate the model’s predictive performance. To annotate SELL low-expression clusters, the classifier was trained on all 12,000 mini-clusters (4,000 per subtype) generated from the sorted reference samples. Single-nucleus multiomic profiling of gene expression and chromatin accessibility analysis After a two-day culture of thawed frozen cells, CD4 staining was performed, and CD4 + cells were sorted using a SH800Z cell sorter (Sony). Nuclei were extracted according to the 10x Genomics protocol (CG000365, Rev C). The nuclear density was assessed by microscopy using the BZ-X700 system (Keyence). All multiome libraries were constructed using the Chromium controller and Chromium Single Cell Multiome ATAC + Gene Expression Kit (10x Genomics, Inc.) following the 10x Genomics protocol (CG000338, Rev F). All multiome libraries were sequenced on DNBSEQ-G400 sequencers (MGI) according to the manufacturer’s standard protocol. The sequenced reads were processed using Cell Ranger ARC 2.0.2 software (10x Genomics, Inc.) with the GRCh38 reference dataset (version 2020-A-2.0.0) obtained from 10x Genomics. Quality control, statistical analysis, and graphical representation were performed using the Loupe browser (10x Genomics, Inc.), Seurat 5.2.1, and the Signac1.14 package in R software (version 4.4.1) 27,28,33 . Quality control was executed as follows. Clusters enriched for CD3D expression were identified using the Loupe Browser (10x Genomics) from 'cloupe' files generated by Cell Ranger ARC and subsequently imported into Seurat for further analysis. Cells with either too few (≤ 500) or too many (> 2,500) expressed genes were excluded to remove empty or multiple GEMs. Cells with either too few (≤ 1500) or too many (> 10,000) ATAC counts were also excluded. Cells with a high percentage of mitochondrial genes (≥ 15%) were excluded to eliminate dead cells. Cells with a nucleosome signal that was too high (> 2) or a TSS enrichment score that was too low (≤ 1) were excluded, as these characteristics are indicative of compromised chromatin structure or degraded nuclei. Following quality control and normalization, dimensionality reduction was performed separately for each modality, with principal component analysis (PCA) applied to the RNA data and latent semantic indexing (LSI) to the ATAC data. Integration of the two modalities was achieved using the weighted nearest neighbor (WNN) approach, which enables joint clustering and visualization of cellular states in a unified embedding using the standard Seurat and Signac multiome analysis workflow. Unsupervised clustering of the integrated data was performed using the 'FindMultiModalNeighbors' function (reduction.list = list(“pca”, “lsi”), dims.list = list(1:50, 2:50)) and the 'FindClusters' function (resolution = 1.5). To relate chromatin accessibility to gene expression, gene activity scores were computed using the 'GeneActivity' function (extend.upstream = 2000, extend.downstream = 0) on the basis of gene annotations from Ensembl Hsapiens version 98. To visualize chromatin accessibility signals and peak distributions around specific genes of interest, we generated coverage plots using the 'CoveragePlot' function implemented in Signac. Mass cytometry The monoclonal antibodies used for Helios ™ mass cytometry analysis are listed in Table S12 . Cell preparation and measurements were performed according to the manufacturer’s instructions for Helios (Standard BioTools, San Francisco, CA, USA). Briefly, 5.0 × 10 6 cells were stained with mass cytometric antibodies. For intracellular staining, samples were prepared using Maxpar Nuclear Antigen Staining Buffer working solution prior to staining. After being washed twice with Maxpar Cell Staining Buffer, the samples were fixed with Maxpar Fix and Perm Buffer supplemented with 125 nM iridium nucleic acid intercalator (Standard BioTools). Following fixation, the cells were washed once with Maxpar cell-staining buffer, twice with Maxpar water, and resuspended in Maxpar water. Over 200,000 cells per sample were analyzed using Helios ™ and Cytobank ( https://www.cytobank.org ) software. The gating strategy is detailed in Fig. S9 . Associations between T-cell subpopulations Network analysis was performed to evaluate associations between T-cell subpopulations in the peripheral blood. Network analysis eliminates false (pseudo)correlations due to confounding effects between multiple variables. In addition, the modified path consistency (PC) algorithm, a modification of the original path consistency algorithm, was used to address the redundant data that frequently occur in biological data 34,35 . The PC algorithm requires variable selection when calculating partial correlation coefficients (associations among three or more variables) of the first order or higher. Therefore, the estimation is repeated several times to reduce the dependence on variable selection, and edges with a high frequency of occurrence are selected. In the present analysis, the estimation was repeated 1,000 times, and edges inferred more than 500 times were chosen as thresholds. Statistical analyses GraphPad Prism 10 (GraphPad Software, San Diego, CA, USA) was used for statistical analysis. The data are presented as the means ± standard errors of the means unless otherwise specified. Differences between two groups were evaluated using Student’s t test. Comparisons across multiple groups were conducted using one-way analysis of variance (ANOVA) with Tukey’s post hoc test. Survival curves were generated using the Kaplan–Meier method, with differences and hazard ratios assessed using the log-rank (Mantel‒Cox) test. All P values were two-sided, and values < 0.05 were considered to indicate statistical significance. Animal models C57BL/6N female mice aged 8–11 weeks were subcutaneously inoculated with 2.0 × 10⁶ MCA 205 fibrosarcoma tumor cells in both flanks. On day 11, CXCR3 ± CCR4 − CCR6 + CD4 + T cells were isolated from tumor-draining inguinal lymph nodes as Th7R cells using a cell sorter (SH800Z, Sony Corporation, Tokyo, Japan). The cells were activated with anti-CD3/CD28 microbeads and cultured in vitro in the presence of 10 U/ml IL-2. Subsequently, 2 × 10⁶ Th7R cells were intravenously transferred into mice bearing MCA 205 subcutaneous tumors along the midline of the abdomen for 15 days. On day 7 post-Th7R administration, the tumors were harvested and digested with Liberase TL Research Grade to obtain a single-cell suspension. TILs were recovered by Percoll density gradient centrifugation and analyzed by mass cytometry. Declarations Funding: KAKENHI program of the Japan Society for the Promotion of Science grant 17H04184, Japan Agency for Medical Research and Development grant 19ae0101074h0001. Author contributions : ST, SY, and HK equally contributed to this study. ST conducted animal studies and analyzed the neoadjuvant IO treatment cohort. SY analyzed the scRNA-seq data and wrote the manuscript. OY, AM, AS, YM, KH, HI, KK, YI, HN, HS, and TH recruited the patients. KH analyzed the data and wrote the manuscript. HK designed the study, analyzed the data, and wrote the manuscript. Competing interests: HK is listed as an inventor on a patent application filed by Saitama Medical University that incorporates the discoveries described in this manuscript. HK has received grant support from Boehringer-Ingelheim, Inc. The other authors declare no competing interests. Data and materials availability: The raw scRNA-seq and scTCR-seq, and snRNA-seq + ATAC-seq data from this study are publicly available from the Gene Expression Omnibus (GEO) in GSE267026, GSE267027, GSE295601 and GSE295977. The data from our previous study (GSE215219) were also used. Acknowledgments We thank Mrs. Koko Kodaira, Mrs. Kozue Watanabe, Mrs. Hiroko Noguchi, Mr. Jyoji Shiotani, and Mrs. Chieko Ono for their technical assistance in this study. References Im, S.J., et al. Defining CD8 + T cells that provide the proliferative burst after PD-1 therapy. Nature 537 , 417–421 (2016). Kallies, A., Zehn, D. & Utzschneider, D.T. Precursor exhausted T cells: key to successful immunotherapy? Nat Rev Immunol 20 , 128–136 (2020). Liu, B., et al. Temporal single-cell tracing reveals clonal revival and expansion of precursor exhausted T cells during anti-PD-1 therapy in lung cancer. Nat Cancer 3 , 108–121 (2022). Im, S.J., et al. Characteristics and anatomic location of PD-1(+)TCF1(+) stem-like CD8 T cells in chronic viral infection and cancer. Proc Natl Acad Sci U S A 120 , e2221985120 (2023). Gebhardt, T., Park, S.L. & Parish, I.A. Stem-like exhausted and memory CD8(+) T cells in cancer. Nat Rev Cancer 23 , 780–798 (2023). Ferris, S.T., et al. cDC1 prime and are licensed by CD4(+) T cells to induce anti-tumour immunity. Nature 584 , 624–629 (2020). Espinosa-Carrasco, G., et al. Intratumoral immune triads are required for immunotherapy-mediated elimination of solid tumors. Cancer Cell 42 , 1202–1216 e1208 (2024). Magen, A., et al. Intratumoral dendritic cell-CD4(+) T helper cell niches enable CD8(+) T cell differentiation following PD-1 blockade in hepatocellular carcinoma. Nat Med 29 , 1389–1399 (2023). Asrir, A., et al. Tumor-associated high endothelial venules mediate lymphocyte entry into tumors and predict response to PD-1 plus CTLA-4 combination immunotherapy. Cancer Cell 40 , 318–334 e319 (2022). Sallusto, F. Heterogeneity of Human CD4(+) T Cells Against Microbes. Annu Rev Immunol 34 , 317–334 (2016). Kagamu, H., et al. Single-cell analysis reveals a CD4 + T cell cluster that correlates with PD-1 blockade efficacy. Cancer Res (2022). Yanagihara, A., et al. A Th1-like CD4(+) T-cell Cluster That Predicts Disease-free Survival in Early-stage Lung Cancer. Cancer Res Commun 3 , 1277–1285 (2023). Kagamu, H., et al. Single-Cell Analysis Reveals a CD4 + T-cell Cluster That Correlates with PD-1 Blockade Efficacy. Cancer Res 82 , 4641–4653 (2022). Sautes-Fridman, C., Petitprez, F., Calderaro, J. & Fridman, W.H. Tertiary lymphoid structures in the era of cancer immunotherapy. Nat Rev Cancer 19 , 307–325 (2019). Galletti, G., et al. Two subsets of stem-like CD8(+) memory T cell progenitors with distinct fate commitments in humans. Nat Immunol 21 , 1552–1562 (2020). Forde, P.M., et al. Neoadjuvant Nivolumab plus Chemotherapy in Resectable Lung Cancer. N Engl J Med 386 , 1973–1985 (2022). Meier, D., et al. Ectopic lymphoid-organ development occurs through interleukin 7-mediated enhanced survival of lymphoid-tissue-inducer cells. Immunity 26 , 643–654 (2007). Luther, S.A. & Cyster, J.G. Chemokines as regulators of T cell differentiation. Nat Immunol 2 , 102–107 (2001). Ferry, A., et al. The XCL1-XCR1 axis supports intestinal tissue residency and antitumor immunity. J Exp Med 222 (2025). Wu, T.D., et al. Peripheral T cell expansion predicts tumour infiltration and clinical response. Nature 579 , 274–278 (2020). Hsieh, T., et al. JunB Is Critical for Survival of T Helper Cells. Front Immunol 13 , 901030 (2022). Chen, D.S. & Mellman, I. Oncology meets immunology: the cancer-immunity cycle. Immunity 39 , 1–10 (2013). Dunn, G.P., Bruce, A.T., Ikeda, H., Old, L.J. & Schreiber, R.D. Cancer immunoediting: from immunosurveillance to tumor escape. Nat Immunol 3 , 991–998 (2002). Mellman, I., Chen, D.S., Powles, T. & Turley, S.J. The cancer-immunity cycle: Indication, genotype, and immunotype. Immunity 56 , 2188–2205 (2023). Kagamu, H., et al. CD4(+) T-cell Immunity in the Peripheral Blood Correlates with Response to Anti-PD-1 Therapy. Cancer Immunol Res 8 , 334–344 (2020). Halse, H., et al. Multiplex immunohistochemistry accurately defines the immune context of metastatic melanoma. Sci Rep 8 , 11158 (2018). Stuart, T., et al. Comprehensive Integration of Single-Cell Data. Cell 177 , 1888–1902 e1821 (2019). Butler, A., Hoffman, P., Smibert, P., Papalexi, E. & Satija, R. Integrating single-cell transcriptomic data across different conditions, technologies, and species. Nat Biotechnol 36 , 411–420 (2018). Ludo Waltman, N.J.v.E. A smart local moving algorithm for large-scale modularity-based community detection. The European Physical Journal B 86 (2013). Finak, G., et al. MAST: a flexible statistical framework for assessing transcriptional changes and characterizing heterogeneity in single-cell RNA sequencing data. Genome Biol 16 , 278 (2015). Kevin Blighe, S.R., Myles Lewis. EnhancedVolcano: publication-ready volcano plots with enhanced colouring and labeling. (2021). Bergen, V., Lange, M., Peidli, S., Wolf, F.A. & Theis, F.J. Generalizing RNA velocity to transient cell states through dynamical modeling. Nat Biotechnol 38 , 1408–1414 (2020). Stuart, T., Srivastava, A., Madad, S., Lareau, C.A. & Satija, R. Single-cell chromatin state analysis with Signac. Nat Methods 18 , 1333–1341 (2021). Leday, G.G.R. & Richardson, S. Fast Bayesian inference in large Gaussian graphical models. Biometrics 75 , 1288–1298 (2019). Saito, S., Hirokawa, T. & Horimoto, K. Discovery of chemical compound groups with common structures by a network analysis approach (affinity prediction method). J Chem Inf Model 51 , 61–68 (2011). Edgar, R., Domrachev, M. & Lash, A.E. Gene Expression Omnibus: NCBI gene expression and hybridization array data repository. Nucleic Acids Res 30 , 207–210 (2002). Additional Declarations Yes there is potential Competing Interest. HK is listed as an inventor on a patent application filed by Saitama Medical University that incorporates the discoveries described in this manuscript. HK has received grant support from Boehringer-Ingelheim, Inc. The other authors declare no competing interests. Supplementary Files Supplementaryinformation.docx Supplementary information SupplementaryTables.xlsx Supplementary Tables SupplementaryFigures.pdf Supplementary Figures Cite Share Download PDF Status: Under Review Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-6764582","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":463733968,"identity":"0dcfc039-4b92-4a20-ac09-c5c37a1917a7","order_by":0,"name":"Hiroshi Kagamu","email":"data:image/png;base64,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","orcid":"https://orcid.org/0000-0003-4523-8517","institution":"Saitama Medical University International Medical Center","correspondingAuthor":true,"prefix":"","firstName":"Hiroshi","middleName":"","lastName":"Kagamu","suffix":""},{"id":463733969,"identity":"bb2f9df9-255d-4860-93a8-3c85e4618b43","order_by":1,"name":"Shota Takei","email":"","orcid":"","institution":"Saitama Medical University International Medical Center","correspondingAuthor":false,"prefix":"","firstName":"Shota","middleName":"","lastName":"Takei","suffix":""},{"id":463733970,"identity":"1625120f-6230-4af2-a39e-8c260618c644","order_by":2,"name":"Satoshi Yamasaki","email":"","orcid":"","institution":"Saitama Medical University International Medical Center","correspondingAuthor":false,"prefix":"","firstName":"Satoshi","middleName":"","lastName":"Yamasaki","suffix":""},{"id":463733971,"identity":"1a6b8f97-6537-4808-91f5-56577072cd3d","order_by":3,"name":"Ou Yamaguchi","email":"","orcid":"","institution":"Saitama Medical University International Medical Center, Japan","correspondingAuthor":false,"prefix":"","firstName":"Ou","middleName":"","lastName":"Yamaguchi","suffix":""},{"id":463733972,"identity":"1903d843-2233-47de-b386-8a10918ac8e0","order_by":4,"name":"Atsuto Mouri","email":"","orcid":"","institution":"Saitama Medical University International Medical Center","correspondingAuthor":false,"prefix":"","firstName":"Atsuto","middleName":"","lastName":"Mouri","suffix":""},{"id":463733973,"identity":"d587eaa1-863e-4a19-899e-ab7f58c17f09","order_by":5,"name":"Ayako Shiono","email":"","orcid":"","institution":"Saitama Medical University International Medical Center, Japan","correspondingAuthor":false,"prefix":"","firstName":"Ayako","middleName":"","lastName":"Shiono","suffix":""},{"id":463733974,"identity":"716665e6-fc92-435d-a233-aba9b78ccb88","order_by":6,"name":"Yu Miura","email":"","orcid":"","institution":"Saitama Medical University International Medical Center, Japan","correspondingAuthor":false,"prefix":"","firstName":"Yu","middleName":"","lastName":"Miura","suffix":""},{"id":463733975,"identity":"11cfab0a-e615-47c2-856a-09bf3d6b51c6","order_by":7,"name":"Kosuke Hashimoto","email":"","orcid":"","institution":"Saitama Medical University International Medical Center","correspondingAuthor":false,"prefix":"","firstName":"Kosuke","middleName":"","lastName":"Hashimoto","suffix":""},{"id":463733976,"identity":"fe630e19-919e-45d3-902d-764f0a887472","order_by":8,"name":"Hisao Imai","email":"","orcid":"","institution":"Saitama Medical University International Medical Center","correspondingAuthor":false,"prefix":"","firstName":"Hisao","middleName":"","lastName":"Imai","suffix":""},{"id":463733977,"identity":"838fdd77-0226-4ef7-ace7-a711f250f3c4","order_by":9,"name":"Kyoichi Kaira","email":"","orcid":"","institution":"Saitama Medical University International Medical Center, Japan","correspondingAuthor":false,"prefix":"","firstName":"Kyoichi","middleName":"","lastName":"Kaira","suffix":""},{"id":463733978,"identity":"2f8177dc-0248-479e-b50f-e4b6b31cba5f","order_by":10,"name":"Yoshinobu Ichiki","email":"","orcid":"","institution":"Saitama Medical University International Medical Center","correspondingAuthor":false,"prefix":"","firstName":"Yoshinobu","middleName":"","lastName":"Ichiki","suffix":""},{"id":463733979,"identity":"4a9a036c-9d87-4e63-832a-6660337f146d","order_by":11,"name":"Hiroyuki Nitanda","email":"","orcid":"","institution":"Saitama Medical University International Medical Center","correspondingAuthor":false,"prefix":"","firstName":"Hiroyuki","middleName":"","lastName":"Nitanda","suffix":""},{"id":463733980,"identity":"da50d9b8-d59c-4260-ba62-2d030942a7c9","order_by":12,"name":"Hirozo Sakaguchi","email":"","orcid":"","institution":"Saitama Medical University International Medical Center","correspondingAuthor":false,"prefix":"","firstName":"Hirozo","middleName":"","lastName":"Sakaguchi","suffix":""},{"id":463733981,"identity":"fd78c06d-ffc6-41f1-b384-09588e1db0ed","order_by":13,"name":"Tomoyuki Hishida","email":"","orcid":"","institution":"Saitama Medical University International Medical Center","correspondingAuthor":false,"prefix":"","firstName":"Tomoyuki","middleName":"","lastName":"Hishida","suffix":""},{"id":463733982,"identity":"6d419da3-5481-463b-9320-d1d46146efcf","order_by":14,"name":"Katsuhisa Horimoto","email":"","orcid":"","institution":"","correspondingAuthor":false,"prefix":"","firstName":"Katsuhisa","middleName":"","lastName":"Horimoto","suffix":""}],"badges":[],"createdAt":"2025-05-28 06:05:22","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-6764582/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-6764582/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":83916561,"identity":"f62fa95b-8b77-49a8-a5af-c82494e16d7f","added_by":"auto","created_at":"2025-06-04 12:52:17","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":1071346,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eSimultaneous clustering of CD4\u003c/strong\u003e\u003csup\u003e\u003cstrong\u003e+\u003c/strong\u003e\u003c/sup\u003e\u003cstrong\u003e T cells in the peripheral blood and tumor microenvironment on the basis of single-cell gene expression.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eA.\u0026nbsp;\u0026nbsp;\u0026nbsp; UMAP plots of CD3D\u003csup\u003e+\u003c/sup\u003eCD4\u003csup\u003e+\u003c/sup\u003e cells generated with integrated gene expression data from single-cell RNA sequencing (scRNA-seq) of peripheral blood (PB) and tumor-infiltrating lymphocytes (TILs) derived from 5 stage I-II non-small cell lung cancer patients. All 58,687 CD3D\u003csup\u003e+\u003c/sup\u003eCD4\u003csup\u003e+\u003c/sup\u003e cells (50,608 PB-CD3D\u003csup\u003e+\u003c/sup\u003eCD4\u003csup\u003e+\u003c/sup\u003e cells and 8,079 TIL-CD3D\u003csup\u003e+\u003c/sup\u003eCD4\u003csup\u003e+\u003c/sup\u003e cells) were categorized into 21 clusters, as illustrated in the UMAP plot.\u003c/p\u003e\n\u003cp\u003eB.\u0026nbsp;\u0026nbsp;\u0026nbsp; The expression of \u003cem\u003eSELL\u003c/em\u003e in the UMAP plot is indicated. The area outlined by the red dotted line represents clusters composed of effector T cells with downregulated expression of \u003cem\u003eSELL\u003c/em\u003e.\u003c/p\u003e\n\u003cp\u003eC.\u0026nbsp;\u0026nbsp;\u0026nbsp; On the basis of T-cell receptor (TCR) repertoire analysis using scTCR-seq, cells belonging to TCR clonotypes with two or more cells detected were mapped as pink dots on the UMAP plot. Although \u003cem\u003eFoxp3\u003c/em\u003e\u003csup\u003e+\u003c/sup\u003e Tregs in TILs included multicellular TCR clonotype cells, most effector multicellular TCR clonotype cells belong to clusters in which \u003cem\u003eSELL\u003c/em\u003e expression was downregulated (outlined by a red dotted line).\u003c/p\u003e\n\u003cp\u003eD.\u0026nbsp;\u0026nbsp;\u0026nbsp; Expression of \u003cem\u003eFoxP3\u003c/em\u003e in the UMAP plot is indicated.\u003c/p\u003e\n\u003cp\u003eE.\u0026nbsp;\u0026nbsp;\u0026nbsp; Distribution of cells with the same TCR clonotype found in the TIL-12, TIL-16, TIL-19, and TIL-20 clusters in both the TIL and the PB clusters. The TCR clonotypes found in the TIL-12 cluster were found almost exclusively in cluster 12 in TILs or PB. However, TIL-19 clonotypes were shared with the cluster 16 in TILs.\u003c/p\u003e\n\u003cp\u003eF.\u0026nbsp;\u0026nbsp;\u0026nbsp;\u0026nbsp; The number of cells in each TIL and PB cluster, excluding those belonging to the Treg, naive, and central memory populations, is plotted in a color-coded stacked bar graph divided by five levels of clonal expansion. The PB-16 cluster had very few multicellular TCR clonotypes, whereas the TIL-16 cluster contained substantially expanded TCR clonotypes.\u003c/p\u003e\n\u003cp\u003eG.\u0026nbsp;\u0026nbsp;\u0026nbsp; Expression patterns of the 58 signature genes from sorted Th1, Th7R, and Th17 cells, and clusters 12, 16+19, and 20 from unsupervised clustering of surgical samples are shown as a heatmap.\u003c/p\u003e\n\u003cp\u003eH.\u0026nbsp;\u0026nbsp;\u0026nbsp; UMAP plot based on the WNN method using nuclear mRNA expression (nRNA) and chromatin accessibility (ATAC) profiles. A total of 10,699 nuclei were classified into 27 distinct clusters.\u003c/p\u003e\n\u003cp\u003eI.\u0026nbsp;\u0026nbsp;\u0026nbsp;\u0026nbsp; Nuclear mRNA expression patterns of 58 signature genes. The results of the cosine similarity analysis of the gene expression patterns in each cluster compared with the gene expression of the sorted Th cells are also presented.\u003c/p\u003e\n\u003cp\u003eJ.\u0026nbsp;\u0026nbsp;\u0026nbsp;\u0026nbsp; Annotation of Th1, Th7R, Th17, Treg and other cell type were visualized in a UMAP plot.\u003c/p\u003e\n\u003cp\u003eK.\u0026nbsp;\u0026nbsp;\u0026nbsp; Gene activity of 54 of 58 signature genes were shown as heatmap. The results of the cosine similarity analysis of the gene expression patterns in each cluster compared with those in the sorted Th cells are also presented.\u003c/p\u003e","description":"","filename":"image1.png","url":"https://assets-eu.researchsquare.com/files/rs-6764582/v1/6a804e050648bea4423456fa.png"},{"id":83916032,"identity":"576752de-4276-4954-b0cd-e8b30db28694","added_by":"auto","created_at":"2025-06-04 12:44:17","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":1143374,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eTh7R gene expression in the tumor microenvironment\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eA, B.\u0026nbsp;\u0026nbsp;\u0026nbsp; UMAP plots of CD3D\u003csup\u003e+\u003c/sup\u003eCD4\u003csup\u003e+\u003c/sup\u003e cells derived from the TILs and peripheral blood of 8 lung cancer patients (A). Annotation based on the expression of 58 signature genes with the prediction index is shown in the UMAP plot (B).\u003c/p\u003e\n\u003cp\u003eC, D. The results of differential expression analysis between the TIL-Th7R cluster (18+21) and the PB-Th7R cluster are shown in a volcano plot. Genes with an adjusted \u003cem\u003eP\u003c/em\u003e value \u0026lt; 0.05 and a fold change ≥ 2.0 were considered significant and are represented by red dots (C). Representative gene expression data are shown in violin plots (D).\u003c/p\u003e\n\u003cp\u003eE. F. Differential gene expression analysis between the TIL-Th7R cluster and the TIL-Th1 cluster (13+26) via scRNA-seq analysis is shown in a volcano plot. Genes with an adjusted \u003cem\u003eP\u003c/em\u003e value \u0026lt; 0.05 and a fold change ≥ 2.0 were considered significant and are represented by red dots (E). Representative gene expression data are shown in violin plots (F).\u003c/p\u003e","description":"","filename":"image2.png","url":"https://assets-eu.researchsquare.com/files/rs-6764582/v1/842f52d97550869b8e4a6fe9.png"},{"id":83916033,"identity":"c186588b-d14a-42b3-995b-593ffdbaf268","added_by":"auto","created_at":"2025-06-04 12:44:17","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":1229008,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eAssociation between Th7R cell abundance and HEV formation in the tumor microenvironment\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eA.\u0026nbsp;\u0026nbsp; Representative results of immunohistochemical analysis. Formalin-fixed, paraffin-embedded sections of biopsy samples or resected samples from non-small cell lung cancer patients were stained for MECA79, CD31, CD4, CD8, CXCR3, CCR6, and cytokeratin using OPAL. The stained sections were imaged using the Vectra Automated Imaging System. Double-positive vessels (MECA79\u003csup\u003e+\u003c/sup\u003e and CD31\u003csup\u003e+\u003c/sup\u003e) were considered HEVs.\u003c/p\u003e\n\u003cp\u003eB.\u0026nbsp;\u0026nbsp;\u0026nbsp; We analyzed the percentages of Th7R cells in the peripheral blood and the ratio of the MECA79-positive vascular area to the CD31\u003csup\u003e+\u003c/sup\u003e vessel area in tumors (MECA index) in 56 lung cancer surgery patients. A χ-square test to evaluate the association between the two metrics revealed a significant correlation between peripheral Th7R percentages and intratumoral HEV density (\u003cem\u003eP\u003c/em\u003e = 0.0080).\u003c/p\u003e\n\u003cp\u003eC. \u0026nbsp;\u0026nbsp; The threshold for Th7R percentages was based on the value that best predicted postoperative recurrence-free survival in our previous study\u003csup\u003e12\u003c/sup\u003e, whereas the MECA index threshold of 0.044 was determined to be the value that best discriminated progression-free survival rates by Kaplan‒Meier analysis (\u003cem\u003eP\u003c/em\u003e = 0.0207, hazard ratio (HR) 0.2006 (95%C.I. 0.0647-0.6220)).\u003c/p\u003e","description":"","filename":"image3.png","url":"https://assets-eu.researchsquare.com/files/rs-6764582/v1/5c44d4dbd75fe0cc7b4f33bf.png"},{"id":83916038,"identity":"afeb7217-5f18-4827-a205-05154a546728","added_by":"auto","created_at":"2025-06-04 12:44:17","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":1302592,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003ePrecursor exhausted CD8\u003c/strong\u003e\u003csup\u003e\u003cstrong\u003e+\u003c/strong\u003e\u003c/sup\u003e\u003cstrong\u003e T cells in the peripheral blood and tumor microenvironment\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eA.\u0026nbsp;\u0026nbsp;\u0026nbsp; Combined UMAP plots of CD3D\u003csup\u003e+\u003c/sup\u003eCD8\u003csup\u003e+\u003c/sup\u003e cells derived from peripheral blood (PB) and tumor-infiltrating lymphocytes (TILs) were generated using integrated gene expression data from single-cell RNA sequencing (scRNA-seq) (n = 8). All 25,728 CD3D\u003csup\u003e+\u003c/sup\u003eCD8\u003csup\u003e+\u003c/sup\u003e cells (17,997 PB-CD3D\u003csup\u003e+\u003c/sup\u003eCD8\u003csup\u003e+\u003c/sup\u003e cells and 7,731 TIL-CD3D\u003csup\u003e+\u003c/sup\u003eCD8\u003csup\u003e+\u003c/sup\u003e cells) were divided into 15 clusters via unsupervised clustering.\u003c/p\u003e\n\u003cp\u003eB, C.\u0026nbsp;\u0026nbsp;\u0026nbsp; The expression of the genes \u003cem\u003eGZMB\u003c/em\u003e and \u003cem\u003eGZMK\u003c/em\u003e in the UMAP plot (left: TIL; right: PB) is shown (B). The expression of \u003cem\u003eGZMB\u003c/em\u003e and \u003cem\u003eGZMK\u003c/em\u003e in each cluster is shown in violin plots. The cell-level expression patterns of \u003cem\u003eGZMB\u003c/em\u003e and \u003cem\u003eGZMK\u003c/em\u003e are shown in the UMAP plot (C).\u003c/p\u003e\n\u003cp\u003eD.\u0026nbsp;\u0026nbsp;\u0026nbsp; Heatmap illustrating the expression of signature genes of precursor exhausted CD8\u003csup\u003e+\u003c/sup\u003e T cells (Tpex) in each CD8\u003csup\u003e+\u003c/sup\u003e T-cell cluster, excluding the MAIT, naïve T-cell, and central memory T-cell clusters.\u003c/p\u003e\n\u003cp\u003eE.\u0026nbsp;\u0026nbsp;\u0026nbsp; Subtraction density plot of TILs and PB cells. The density was calculated for each grid with a size of 0.2 on the UMAP plot. The intensity of the red color indicates the TIL-dominant density, whereas the intensity of the blue color indicates the PB-dominant density.\u003c/p\u003e\n\u003cp\u003eF.\u0026nbsp;\u0026nbsp;\u0026nbsp;\u0026nbsp; The results of RNA velocity analysis of TILs and PB are shown as a stream plot.\u003c/p\u003e","description":"","filename":"image4.png","url":"https://assets-eu.researchsquare.com/files/rs-6764582/v1/02f24502f8fa7b751647310f.png"},{"id":83917435,"identity":"4c09e31c-66b6-4d2a-8b69-73ebf2f3e90f","added_by":"auto","created_at":"2025-06-04 13:00:17","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":929621,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eSpatial relationships between Th7R cells and Tpex in TLSs\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eA-C. Single-cell segmentation and annotation were performed via Hyperion\u003csup\u003eTM\u003c/sup\u003e analysis, according to the expression of 18 molecules (A-B). Red circles indicate TLSs (B). Spatial relationship analysis between the annotated cells, including Tpex and Th7R cells, is shown (C).\u003c/p\u003e","description":"","filename":"image5.png","url":"https://assets-eu.researchsquare.com/files/rs-6764582/v1/0b9ade487483914041a76c9a.png"},{"id":83916036,"identity":"e133f561-cf77-45f2-ae7a-1f0e8b60c589","added_by":"auto","created_at":"2025-06-04 12:44:17","extension":"png","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":470255,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eNumerical relationships between Th7R cells and Tpex\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eA.\u0026nbsp;\u0026nbsp;\u0026nbsp; Th7R cells were sorted as CXCR3\u003csup\u003e+\u003c/sup\u003eCCR4\u003csup\u003e-\u003c/sup\u003eCCR6\u003csup\u003e+\u003c/sup\u003e CD4\u003csup\u003e+\u003c/sup\u003e T cells from inguinal lymph nodes draining MCA 205 subcutaneous tumors for 11 days in C57BL/6N mice. After ex vivo expansion via anti-CD3/CD28 microbead activation in the presence of 40 U/ml IL-2, 2 × 10\u003csup\u003e6\u003c/sup\u003e Th7R cells were adoptively transferred i.v. to mice bearing 15-day MCA 205 subcutaneous tumors. Tumors were harvested 7 days after Th7R infusion and analyzed for tumor-infiltrating cells. The percentages of cells among total TILs are shown.\u003c/p\u003e\n\u003cp\u003eB-C. Linear correlations between the percentages of CD8\u003csup\u003e+\u003c/sup\u003e T-cell clusters among total CD8\u003csup\u003e+\u003c/sup\u003e T cells and CD4\u003csup\u003e+\u003c/sup\u003e T-cell clusters among total CD4\u003csup\u003e+\u003c/sup\u003e T cells in the hilar lymph nodes nearest to lung cancer (B) or in the peripheral blood (C) are shown.\u003c/p\u003e\n\u003cp\u003eD.\u0026nbsp;\u0026nbsp;\u0026nbsp; The percentages of CD8\u003csup\u003e+\u003c/sup\u003e T-cell clusters among total CD8\u003csup\u003e+\u003c/sup\u003e T cells and the percentages of CD4\u003csup\u003e+\u003c/sup\u003e T-cell clusters among total CD4\u003csup\u003e+\u003c/sup\u003e T cells in surgically resected lung cancer tissues after neoadjuvant therapy, including anti-PD-1 antibody therapy and chemotherapy, are shown (n = 20). The lung cancer patients (Table S8) who were treated with 3 courses of CBDCA + paclitaxel or pemetrexed + nivolumab underwent surgical resection. Tumor infiltrating lymphocytes were analyzed with CyTOF\u003csup\u003eTM\u003c/sup\u003e.\u003c/p\u003e","description":"","filename":"image6.png","url":"https://assets-eu.researchsquare.com/files/rs-6764582/v1/5cd80cbe8bd8f2f8f316b7e5.png"},{"id":83918001,"identity":"b057d0f9-85ed-4ae5-ace6-48b50356d4a1","added_by":"auto","created_at":"2025-06-04 13:08:21","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":7328412,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-6764582/v1/86a39c3b-ff22-478c-9733-baec1ff2f077.pdf"},{"id":83916030,"identity":"5cce816c-653d-4781-afb4-10c4c067a3a2","added_by":"auto","created_at":"2025-06-04 12:44:17","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":56050,"visible":true,"origin":"","legend":"Supplementary information","description":"","filename":"Supplementaryinformation.docx","url":"https://assets-eu.researchsquare.com/files/rs-6764582/v1/75290eaf9ae9cd0d6b4160be.docx"},{"id":83916031,"identity":"a273ce4d-7fd4-496c-bf29-1decbe7e66ae","added_by":"auto","created_at":"2025-06-04 12:44:17","extension":"xlsx","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":342592,"visible":true,"origin":"","legend":"\u003cp\u003eSupplementary Tables\u003c/p\u003e","description":"","filename":"SupplementaryTables.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-6764582/v1/9468819aa8a9c467bdcabefa.xlsx"},{"id":83916037,"identity":"03b714f7-ebbd-46c6-8592-9a46d9c795dd","added_by":"auto","created_at":"2025-06-04 12:44:17","extension":"pdf","order_by":3,"title":"","display":"","copyAsset":false,"role":"supplement","size":12028008,"visible":true,"origin":"","legend":"\u003cp\u003eSupplementary Figures\u003c/p\u003e","description":"","filename":"SupplementaryFigures.pdf","url":"https://assets-eu.researchsquare.com/files/rs-6764582/v1/eae089c1dffa8038e89ea408.pdf"}],"financialInterests":"\u003cb\u003eYes\u003c/b\u003e there is potential Competing Interest.\nHK is listed as an inventor on a patent application filed by Saitama Medical University that incorporates the discoveries described in this manuscript. HK has received grant support from Boehringer-Ingelheim, Inc. The other authors declare no competing interests.","formattedTitle":"\u003cp\u003eTh7R Cells: CD4\u003csup\u003e⁺\u003c/sup\u003e T-Cell Partners That Drive Tpex-Mediated Antitumor Immunity in the Tumor Microenvironment\u003c/p\u003e","fulltext":[{"header":"Introduction","content":"\u003cp\u003ePrecursor exhausted CD8⁺ T cells (Tpex) were identified by scRNA-seq as a subset of effector T cells with self-renewal capacity that support long-term antitumor immunity in the tumor microenvironment (TME) \u003csup\u003e1\u0026ndash;3\u003c/sup\u003e. Unlike cytotoxic T lymphocytes (CTLs), which lose TCF7 expression and die after tumor killing, Tpex maintain TCF7 and IL-7 receptor expression, infiltrate tumors via high endothelial venules (HEVs), and localize within tertiary lymphoid structures (TLSs) \u003csup\u003e4,5\u003c/sup\u003e. Tpex are thought to be the source of T-cell expansion following PD-1 blockade therapy \u003csup\u003e1\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eCD4⁺ T-cell help is essential for inducing effective CD8⁺ T-cell responses, including Tpex \u003csup\u003e6\u003c/sup\u003e. It has been shown that CD4⁺ T-cell-mediated signals, particularly in the presence of antigen-presenting dendritic cells, prevent deep exhaustion of CD8⁺ T cells and promote TCF7 and IL-7 receptor expression \u003csup\u003e7,8\u003c/sup\u003e. CD4⁺ T cells are also involved in forming ectopic HEVs, crucial for Tpex entry into tumors \u003csup\u003e9\u003c/sup\u003e. However, the specific CD4⁺ T-cell subset responsible for supporting Tpex is not well defined.\u003c/p\u003e \u003cp\u003eUpon priming, CD4⁺ T cells differentiate into functionally distinct subsets \u003csup\u003e10\u003c/sup\u003e. While Th1 cells are traditionally linked with antitumor immunity through IFN-γ and T-bet expression, recent analyses of immune checkpoint inhibitor (ICI) responses have identified Th1-like, non-canonical CD4⁺ T cells in tumors. We previously described a unique CD4⁺ T-cell population, Th7R, which expresses TCF7 and IL-7R and is associated with better outcomes in lung cancer patients treated with ICIs or surgery \u003csup\u003e11, 12\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eBased on their shared molecular features with Tpex, we hypothesized that Th7R cells may function as CD4⁺ T-cell partners of Tpex. In this study, we analyzed CD4⁺ and CD8⁺ T cells from lung cancer patients using scRNA-seq, scTCR-seq, and imaging mass cytometry to define the relationship between Th7R cells and Tpex in antitumor immunity.\u003c/p\u003e"},{"header":"Results","content":"\u003cp\u003e \u003cb\u003eCells with the same TCR clonotypes and gene expression signature as Th7R cells in the peripheral blood are present in the tumor microenvironment.\u003c/b\u003e \u003c/p\u003e \u003cp\u003eTh7R is a novel Th1-like CD4\u003csup\u003e+\u003c/sup\u003e T-cell cluster identified by scRNA-seq analysis of peripheral blood from advanced-stage lung cancer patients who responded to immune checkpoint inhibitor (ICI) therapy \u003csup\u003e11\u003c/sup\u003e. To clarify whether Th7R cells are also present in the tumor microenvironment (TME), we analyzed the scRNA-seq and scTCR-seq data of CD4\u003csup\u003e+\u003c/sup\u003e T cells derived from the peripheral blood and tumor tissues of five lung cancer patients who had undergone surgery and examined whether CD4\u003csup\u003e+\u003c/sup\u003e T-cell clusters with the same gene expression patterns and TCR clonotypes as peripheral blood Th7R cells are present in lung cancer tissues. A total of 58,687 CD3D\u003csup\u003e+\u003c/sup\u003eCD4\u003csup\u003e+\u003c/sup\u003e cells (50,608 PB-CD3D\u003csup\u003e+\u003c/sup\u003eCD4\u003csup\u003e+\u003c/sup\u003e cells and 8,079 TIL-CD3D\u003csup\u003e+\u003c/sup\u003eCD4\u003csup\u003e+\u003c/sup\u003e cells) were analyzed, and unsupervised clustering was performed on the basis of the expression of 2,000 highly variable genes (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eA). Since Th7R cells were discovered in a cluster of effector T cells with low \u003cem\u003eSELL\u003c/em\u003e expression in the peripheral blood, we first focused on the \u003cem\u003eSELL\u003c/em\u003e low-expression clusters 12, 16, 19, and 20 (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eB, \u003cb\u003eFig. \u003cspan refid=\"MOESM1\" class=\"InternalRef\"\u003eS1\u003c/span\u003e\u003c/b\u003e). The \u003cem\u003eSELL\u003c/em\u003e low-expression clusters contained many clonally expanded T cells (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eC, \u003cb\u003eTable \u003cspan refid=\"MOESM1\" class=\"InternalRef\"\u003eS1\u003c/span\u003e\u003c/b\u003e). Cluster 8, which contained many multicellular clonotypes with high \u003cem\u003eSELL\u003c/em\u003e expression in the TME, was thought to include regulatory T cells (Tregs) with high \u003cem\u003eFoxP3\u003c/em\u003e expression (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eD).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eNext, the independence of the TIL-12, TIL-16, TIL-19, and TIL-20 clusters was examined by analyzing TCR sharing in multicellular TCR clonotypes, which included 2 or more T cells in each TIL cluster (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eE\u003cb\u003e)\u003c/b\u003e. The TIL-12 clonotypes were only shared with the peripheral blood (PB)-12 cluster, indicating that the T cells belonging to PB-12 do not significantly change their gene expression patterns even after infiltration into the TME. TCR sharing with TIL-19 cells was found not only for the PB-19 cluster but also for the TIL-16 cluster. On the other hand, the clonotypes belonging to TIL-16 were mostly found in the PB-16 cluster. Thus, T cells in the PB-19 cluster are likely to change their gene expression patterns to the cluster 16 state after infiltrating the TME, suggesting that clusters 16 and 19 exhibit two gene expression states of one type of differentiated T cell. Accordingly, clusters 16 and 19 were considered a single cluster. TCR clonotypes belonging to the TIL-20 cluster were rarely detected in other clusters. The total cell counts and TCR clone sizes for each cluster are shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eF. Cluster12 had greatly expanded TCR clones in the peripheral blood but lacked expanded TCR clones in the TME. It is likely that the abundant multicellular clonotypes in the PB-12 cluster reflect clonal expansion in the secondary lymphoid organs and that some of the expanded cells infiltrate the TME. In contrast, cluster 16 exhibited few expanded TCR clonotypes in the peripheral blood but the greatest number of expanded TCR clonotypes in the TIL-16 cluster, suggesting clonal expansion in the TME rather than in the secondary lymphoid organs.\u003c/p\u003e \u003cp\u003eThe \u003cem\u003eSELL\u003c/em\u003e\u003csup\u003elow\u003c/sup\u003e effector CD4\u003csup\u003e+\u003c/sup\u003e T cells in the peripheral blood included canonical Th1-type cells, Th17-type cells and Th7R cells, and chemokine receptor expression patterns revealed that these cells were CXCR3\u003csup\u003e+\u003c/sup\u003e CCR4\u003csup\u003e-\u003c/sup\u003e CCR6\u003csup\u003e-\u003c/sup\u003e Th1, CXCR3\u003csup\u003e-\u003c/sup\u003e CCR4\u003csup\u003e+\u003c/sup\u003eCCR6\u003csup\u003e+\u003c/sup\u003e Th17 cells, and CXCR3\u003csup\u003e\u0026plusmn;\u003c/sup\u003e CCR4\u003csup\u003e-\u003c/sup\u003e CCR6\u003csup\u003e+\u003c/sup\u003e Th7R cells, respectively \u003csup\u003e13\u003c/sup\u003e. Next, we examined the signature genes that could discriminate among Th1, Th17, and Th7R cells. First, CXCR3\u003csup\u003e+\u003c/sup\u003e CCR4\u003csup\u003e-\u003c/sup\u003e CCR6\u003csup\u003e-\u003c/sup\u003e, CXCR3\u003csup\u003e-\u003c/sup\u003e CCR4\u003csup\u003e+\u003c/sup\u003e CCR6\u003csup\u003e+\u003c/sup\u003e, and CXCR3\u003csup\u003e\u0026plusmn;\u003c/sup\u003e CCR4\u003csup\u003e-\u003c/sup\u003e CCR6\u003csup\u003e+\u003c/sup\u003e T cells sorted from the peripheral blood of CD62L\u003csup\u003elow\u003c/sup\u003e CD4\u003csup\u003e+\u003c/sup\u003e T cells from lung cancer patients (n\u0026thinsp;=\u0026thinsp;3) other than the 5 patients in this study were analyzed by scRNA-seq.\u0026nbsp;Differential expression analysis identified 64 genes that were differentially expressed among these subsets, including both upregulated and downregulated genes, thereby enabling the characterization of gene expression signatures associated with the Th1, Th7R, and Th17 phenotypes. (\u003cb\u003eTable \u003cspan refid=\"MOESM2\" class=\"InternalRef\"\u003eS2\u003c/span\u003e\u003c/b\u003e). We refined the 64-gene set to 58 genes by applying support vector machine-based recursive feature elimination (SVM-RFE) for feature selection (\u003cb\u003eFig. \u003cspan refid=\"MOESM2\" class=\"InternalRef\"\u003eS2\u003c/span\u003eA-C\u003c/b\u003e). By using the expression data for the 58 selected genes from sorted Th1, Th7R, and Th17 cells, we trained an SVM classifier and applied it to discriminate the Th subtypes of the cell clusters shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eA. As a result, cluster 12 was predicted to be Th1 cells, cluster 20 was predicted to be Th17 cells, and cluster 16\u0026thinsp;+\u0026thinsp;19 was predicted to be Th7R cells (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eG, \u003cb\u003eFig. \u003cspan refid=\"MOESM2\" class=\"InternalRef\"\u003eS2\u003c/span\u003eD\u003c/b\u003e). We evaluated the similarity of the sorted subsets and SVM-predicted clusters using an expression heatmap of 58 genes combined with cosine similarity analysis. This approach revealed strong concordance (\u003cb\u003eFig. \u003cspan refid=\"MOESM2\" class=\"InternalRef\"\u003eS2\u003c/span\u003eD\u003c/b\u003e).\u003c/p\u003e \u003cp\u003eFurthermore, to investigate whether this gene expression pattern is derived from epigenetic changes, we used the 10x Genomics Multiome platform to perform both single-nucleus RNA-seq (snRNA-seq) and single-cell ATAC-seq (scATAC-seq) of the same nuclei (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eH, \u003cb\u003eFig. \u003cspan refid=\"MOESM3\" class=\"InternalRef\"\u003eS3\u003c/span\u003eA\u003c/b\u003e). Among the clusters obtained on the WNN-UMAP diagram, clusters 4, 6, 11, 14, and 20 were considered \u003cem\u003eSELL\u003c/em\u003e low-expression clusters (\u003cb\u003eFig. \u003cspan refid=\"MOESM3\" class=\"InternalRef\"\u003eS3\u003c/span\u003eB, C\u003c/b\u003e). On the basis of SVM-based prediction of Th subtypes using the 58 signature genes, clusters 6 and 20 were classified as Th1 cells, cluster 14 was classified as Th17 cells, and clusters 4 and 11 were classified as Th7R cells (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eI, \u003cb\u003eFig. \u003cspan refid=\"MOESM3\" class=\"InternalRef\"\u003eS3\u003c/span\u003eD\u003c/b\u003e). The overall gene expression patterns of the clusters classified as Th1, Th17, and Th7R cells were very similar to those of sorted Th cells (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eJ\u003cb\u003e)\u003c/b\u003e. The gene activity (GA) scores of 54 of the 58 signature genes obtained from DNA chromatin information also matched the gene expression patterns of each Th subset (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eK, \u003cb\u003eFig. \u003cspan refid=\"MOESM4\" class=\"InternalRef\"\u003eS4\u003c/span\u003e\u003c/b\u003e). These results indicate that Th7R cells have a characteristic gene expression pattern that differs from those of Th1 and Th17 cells and that they have a DNA chromatin state that supports this gene expression profile.\u003c/p\u003e \u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003eTh7R cell gene expression in the TME\u003c/h2\u003e \u003cp\u003eTo perform high-resolution CD4\u003csup\u003e+\u003c/sup\u003e T-cell clustering and gene expression analysis of TILs, we combined the results of three additional TIL and PB scRNA-seq analyses. A total of 79,130 CD4\u003csup\u003e+\u003c/sup\u003eCD3D\u003csup\u003e+\u003c/sup\u003e cells (65,572 PB-CD3D\u003csup\u003e+\u003c/sup\u003eCD4\u003csup\u003e+\u003c/sup\u003e cells and 13,558 TIL-CD3D\u003csup\u003e+\u003c/sup\u003eCD4\u003csup\u003e+\u003c/sup\u003e cells) were subjected to unsupervised clustering on the basis of 2,000 genes with highly variable expression, generating 27 clusters (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eA, \u003cb\u003eTable \u003cspan refid=\"MOESM3\" class=\"InternalRef\"\u003eS3\u003c/span\u003e, Fig. S5A, B\u003c/b\u003e). In the UMAP plot, clusters 13, 14, 18, 21, and 26 were located in the \u003cem\u003eSELL\u003c/em\u003e-low expression area (\u003cb\u003eFig. S5C\u003c/b\u003e). On the basis of the highest match rate in the SVM-based prediction using the 58 signature genes, the cells in cluster 13\u0026thinsp;+\u0026thinsp;26 were annotated as Th1 cells, the cells in cluster 18\u0026thinsp;+\u0026thinsp;21 were annotated as Th7R cells, and the cells in cluster 14 were annotated as Th17 cells (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eB). For Th1 and Th17 cells, the peripheral blood cluster (prediction probability of 1.000, 1.000) and the TIL cluster (1.000, 0.842), showed good predictive agreement. In contrast, the Th7R cluster had a greater probability in the peripheral blood (0.993) but a lower probability in TILs (0.557), suggesting altered gene expression in the TME. However, the probability of the TIL Th7R cluster containing Th1 (0.171) and Th17 (0.272) cells was low, suggesting that Th7R cells have a unique gene expression pattern in the TME. Therefore, we compared the gene expression profile of the Th7R cluster between peripheral blood and TILs and found that 115 genes were upregulated more than 2-fold in TILs and 18 genes were downregulated more than 2-fold in TILs (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eC, D; \u003cb\u003eTable \u003cspan refid=\"MOESM4\" class=\"InternalRef\"\u003eS4\u003c/span\u003e; Fig. S5D\u003c/b\u003e). The genes upregulated in the TME included \u003cem\u003eCCL4, CXCL13, CREM, CCL4L2, GZMB, CTLA4, CXCR4, IFNG, ID2, XCL1, CXCR6, CCL3\u003c/em\u003e and \u003cem\u003eICOS\u003c/em\u003e. Next, we investigated the differences in gene expression between Th7R and Th1 or Th17 cells in the TME (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eE, \u003cb\u003eFig. S5E\u003c/b\u003e). The genes whose expression was more than 2-fold greater in Th7R cells than in canonical Th1 cells in the TME included \u003cem\u003eLTB, CXCR6, CTLA4\u003c/em\u003e, and \u003cem\u003eCXCL13\u003c/em\u003e (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eF, \u003cb\u003eTable S5\u003c/b\u003e). The genes whose expression was more than 2-fold greater in Th7R cells than in Th17 cells included \u003cem\u003eCCL4, CCL4L2, CXCL13, CCL5, GZMK, GZMB, IFNG\u003c/em\u003e, and \u003cem\u003eCCL3\u003c/em\u003e.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003eRelationship between Th7R cells and HEV formation in the TME\u003c/h3\u003e\n\u003cp\u003e \u003cem\u003eCXCL13\u003c/em\u003e was found to be upregulated in Th7R cells in the TME, with a significant difference from the expression observed in Th1 and Th17 cells. Th7R cells also expressed more \u003cem\u003eLTB\u003c/em\u003e; LTβ and CXCL13 are known to be essential for the formation of ectopic HEVs and TLSs \u003csup\u003e14\u003c/sup\u003e. To investigate the relationship between Th7R cells and ectopic HEV formation, immunohistochemical (IHC) analysis and Hyperion\u0026trade; analysis were performed (n\u0026thinsp;=\u0026thinsp;56, \u003cb\u003eTable S6\u003c/b\u003e). Sections were stained with antibodies against the normal vascular marker CD31 and the HEV marker peripheral node addressin (PNAd), and the ratio of PNAd-positive areas to CD31-positive areas (the MECA index) was calculated to assess HEV density and its relationship with the proportion of Th7R cells in the peripheral blood (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eA). Consistent with previous reports, significantly better disease-free survival (DFS) was observed in patients with more HEV formation in the TME, and a MECA index of 0.044 was found to predict better DFS (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eB). A chi-square test using a Th7R percentage of 3.05, which was used as a threshold for predicting better DFS after surgery for lung cancer in a previous study, and a MECA index of 0.044 showed that HEV density was associated with Th7R abundance (P\u0026thinsp;=\u0026thinsp;0.0080, Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eC) \u003csup\u003e12\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e\n\u003ch3\u003eTpex as a source of Tex in the TME\u003c/h3\u003e\n\u003cp\u003eTpex have self-renewal ability and contribute to a sustained supply of CD8\u003csup\u003e+\u003c/sup\u003e T cells to maintain long-term antitumor immunity through stem cell-like asymmetric cell division \u003csup\u003e5\u003c/sup\u003e. To analyze the relationship between Th7R cells and Tpex in the TME, we attempted to identify Tpex in lung cancer tissues. We performed scRNA-seq analysis of 25,728 CD3D\u003csup\u003e+\u003c/sup\u003eCD8\u003csup\u003e+\u003c/sup\u003e cells (17,997 PB-CD3D\u003csup\u003e+\u003c/sup\u003eCD8\u003csup\u003e+\u003c/sup\u003e cells and 7,731 TIL-CD3D\u003csup\u003e+\u003c/sup\u003eCD8\u003csup\u003e+\u003c/sup\u003e cells) derived from 8 surgery patients. Unsupervised clustering was performed on the basis of the expression of 2,000 highly variable genes, generating 16 clusters, as shown in the UMAP plot (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eA, \u003cb\u003eFig. S6\u003c/b\u003e). We analyzed the gene expression of GZMK and GZMB as candidate markers for discriminating among Tpex, exhausted CD8\u003csup\u003e+\u003c/sup\u003e T cells (Tex), and unexhausted cytotoxic T lymphocytes (CTLs) (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eB) \u003csup\u003e2,3,15\u003c/sup\u003e. GZMK expression was observed in clusters 1, 2, 3, 4, 8, 10, 11, 13 and 14 adjacent to the \u003cem\u003eSELL\u003c/em\u003e\u003csup\u003ehigh\u003c/sup\u003eCD45RA\u003csup\u003e+\u003c/sup\u003e naive clusters (6, 12, and 15) and the \u003cem\u003eSELL\u003c/em\u003e\u003csup\u003ehigh\u003c/sup\u003eCD45RA\u003csup\u003e-\u003c/sup\u003e central memory (CM) cluster (7) (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eC). GZMB expression was observed in cluster 5, which was farthest from the naive clusters, and clusters 1, 4, 8, 11, and 13. Cluster 9 was CCR6\u003csup\u003e+\u003c/sup\u003e and was considered mucosa-associated invariant T (MAIT) cells. Next, we assessed the expression of signature genes thought to characterize Tpex in each cluster (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eD) \u003csup\u003e5\u003c/sup\u003e. GZMK\u003csup\u003e+\u003c/sup\u003eGZMB\u003csup\u003e-\u003c/sup\u003e clusters 2, 3, 10, and 14 expressed \u003cem\u003eTCF7\u003c/em\u003e, \u003cem\u003eIL-7R\u003c/em\u003e, \u003cem\u003ePDCD1\u003c/em\u003e, \u003cem\u003eSELL\u003c/em\u003e, and \u003cem\u003eCCR7\u003c/em\u003e and were consistent with Tpex. On the other hand, GZMK\u003csup\u003e-\u003c/sup\u003eGZMB\u003csup\u003e+\u003c/sup\u003e cluster 5 was considered a CTL cluster because it expressed no exhaustion genes and highly expressed \u003cem\u003ePRF1, GZMB, GNLY, FGFBP2\u003c/em\u003e, and \u003cem\u003eNKG7\u003c/em\u003e. The GZMK\u003csup\u003e+\u003c/sup\u003eGZMB\u003csup\u003e+\u003c/sup\u003e clusters 1, 4, 8, 11, and 13 were considered Tex, as they did not express \u003cem\u003eIL-7R\u003c/em\u003e or \u003cem\u003eTCF7\u003c/em\u003e but highly expressed exhaustion genes such as \u003cem\u003ePDCD1\u003c/em\u003e, \u003cem\u003eCTLA-4\u003c/em\u003e, \u003cem\u003eENTPD1\u003c/em\u003e, and \u003cem\u003eHAVCR2\u003c/em\u003e. These results indicate that Tpex, Tex, and CTLs can be distinguished on the basis of their GZMB and GZMK expression patterns.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eWhen the cell density of each pixel in the PB-UMAP plot was subtracted from the cell density of each pixel in the TIL-UMAP plot, clusters 1, 2, 10 and 13 presented a distribution with a TME-dominance (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eE, \u003cb\u003eTable S7\u003c/b\u003e). The direction of the state change for each cluster in the TME and peripheral blood was also examined by velocity analysis (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eF). In the peripheral blood, changes to the Tex clusters starting from cluster 5, which is a cluster of CTLs that have not been exhausted, were observed, and it was thought that CTLs that proliferated in the secondary lymphoid organs changed their state in response to repeated antigen stimulation. In contrast, TIL velocity analysis revealed that cluster 2, which was thought to be a Tpex cluster, was the source of the surrounding clusters, including Tex. Considering that clusters 1 and 2 were dominant in the TME, the main progression of T-cell state changes is thought to be from the Tpex cluster 2 to the terminally exhausted cluster 1, which has the highest expression of \u003cem\u003eHAVCR2, CTLA4\u003c/em\u003e, and \u003cem\u003eENTPD1\u003c/em\u003e. These results indicate that Tex, which have cytotoxic activity, are supplied mainly by Tpex, which replicate in the TME.\u003c/p\u003e\n\u003ch3\u003eSpatial relationship between Th7R cells and Tpex in TLSs\u003c/h3\u003e\n\u003cp\u003eIHC analysis revealed that CXCR3\u003csup\u003e+\u003c/sup\u003eCCR6\u003csup\u003e+\u003c/sup\u003e CD4\u003csup\u003e+\u003c/sup\u003e T cells were located in TLSs (\u003cb\u003eFig. S7\u003c/b\u003e). Next, we performed a neighborhood analysis of Th7R cells and Tpex on the basis of the molecular expression information obtained using Hyperion\u0026trade; (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eA). We used software that can segment and annotate individual cells on the basis of their molecular expression patterns (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eB, \u003cb\u003eFig. S8A\u003c/b\u003e). GZMB\u003csup\u003e+\u003c/sup\u003e cells, which constitute a population with cytolytic functions, were found in the cancer cell nest. In contrast, GZMB\u003csup\u003e\u0026minus;\u003c/sup\u003eGZMK\u003csup\u003e+\u003c/sup\u003e CD8\u003csup\u003e+\u003c/sup\u003e T cells, which are considered Tpex, were found to be abundant within the B-cell aggregates considered TLSs. CXCR3\u003csup\u003e+\u003c/sup\u003e CCR6\u003csup\u003e+\u003c/sup\u003e CD4\u003csup\u003e+\u003c/sup\u003e T cells, which are considered Th7R cells, were found near Tpex within TLSs (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eC).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e\n\u003ch3\u003eCorrelations between Th7R and Tpex numbers\u003c/h3\u003e\n\u003cp\u003eTo determine whether increased Th7R abundance in the TME affects CD8\u003csup\u003e+\u003c/sup\u003e T-cell counts, we analyzed TILs from mice subjected to adoptive transfer of Th7R cells. Th7R cells were sorted as CXCR3\u003csup\u003e\u0026plusmn;\u003c/sup\u003eCCR4\u003csup\u003e\u0026minus;\u003c/sup\u003eCCR6\u003csup\u003e+\u003c/sup\u003e CD4\u003csup\u003e+\u003c/sup\u003e T cells from inguinal lymph nodes draining a methylcholanthrene-induced fibrosarcoma MCA205 skin tumor, and after ex vivo culture, they were adoptively transferred into mice bearing day-11 MCA205 skin tumors. Tumors were harvested 11 days after cell transfer from 3 mice each, and TILs were analyzed. Tumors (310 mg) from untreated mice and tumors (270 mg) from mice subjected to Th7R-transfer were digested, yielding 49.1\u0026nbsp;million and 43.1\u0026nbsp;million cells, respectively (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eA). TILs obtained by density gradient separation using Percoll were analyzed using a mass cytometer. Adoptive transfer of Th7R cells resulted in an increase in the number of CD8\u003csup\u003e+\u003c/sup\u003e T cells in the tumors, with a marked increase in the size of the Tpex and Tex populations. Therefore, Th7R cells likely promote asymmetric cell division of Tpex, resulting in increases in Tpex as abundance and the supply of Tex.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eNext, we examined whether the percentages of CTLs and Tpex among total CD8\u003csup\u003e+\u003c/sup\u003e T cells were correlated with the percentages of Th1 and Th7R cells among total CD4\u003csup\u003e+\u003c/sup\u003e T cells in cluster 12 in the draining lymph nodes from lung cancer patients (the hilar lymph nodes) and peripheral blood derived from lung cancer patients (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eB, C). In both locations, Th1 cell abundance was positively correlated with CTL abundance, and Th7R abundance was positively correlated with Tpex abundance. To eliminate overlapping relationships and clarify the relationships between cells, a quantitative network analysis of the proportions of clusters in CD4\u003csup\u003e+\u003c/sup\u003e T cells and CD8\u003csup\u003e+\u003c/sup\u003e T cells relative to CD3\u003csup\u003e+\u003c/sup\u003e cells in the peripheral blood was performed (\u003cb\u003eFig. S8B\u003c/b\u003e). Notably, the only CD4\u003csup\u003e+\u003c/sup\u003e T-cell cluster whose abundance directly correlated with that of Tpex was Th7R cells, with a connection established in 1000 out of 1000 interferences. These results support the hypothesis that Th7R cells are involved in the intratumoral infiltration and proliferation of Tpex.\u003c/p\u003e \u003cp\u003eTo evaluate whether Th7R cells and Tpex in the TME contribute to antitumor immunity, we investigated their associations with those responses to preoperative neoadjuvant therapy including anti-PD-1 antibody (\u003cb\u003eTable S8\u003c/b\u003e) \u003csup\u003e16\u003c/sup\u003e. There was a significantly greater abundance of Tpex in the patients who achieved a RECIST partial response by imaging. In addition, Th7R abundance was significantly greater in patients who achieved a complete pathological response or major pathological response (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eD). Thus, Th7R cells and Tpex in the TME likely play important roles as partners in antitumor immunity.\u003c/p\u003e"},{"header":"Discussion","content":"\u003cp\u003eIn this study, we showed that Th7R cells, a novel Th1-like CD4\u003csup\u003e+\u003c/sup\u003e T-cell lineage that exhibits different gene expression patterns from Th1 and Th17 cells, plays a distinctive role in antitumor immunity as a partner of Tpex. To define Th7R cells, an analysis of the expression of 58 signature genes obtained by machine learning was performed, and \u003cem\u003eIL7R, TCF7, IFNG-AS1\u003c/em\u003e, and \u003cem\u003eNELL2\u003c/em\u003e were selected as those that exhibited Th7R-dominant expression. \u003cem\u003eIL7R\u003c/em\u003e and \u003cem\u003eTCF7\u003c/em\u003e, which are known to contribute to the survival and proliferation of T cells, are expressed at extremely low levels in Th1 cells but are expressed to some extent in Th17 cells. Th7R cells, which express \u003cem\u003eRORC\u003c/em\u003e, are thought to share the characteristics of high survival and self-replication ability with Th17 cells. On the other hand, Th7R cells express transcription factors, cytokines, and chemokines that are common to Th1 cells, such as \u003cem\u003eTBX21, EOMES, IFNG, CCL4\u003c/em\u003e, and \u003cem\u003eCCL5\u003c/em\u003e, and it is highly likely that many of their effector functions are carried out via the expression of Th1-related genes. On the other hand, Th7R cells express \u003cem\u003eLTB\u003c/em\u003e and \u003cem\u003eCXCL13\u003c/em\u003e, which are not found in Th1 cells, in the TME, and Th7R abundance is correlated with the density of HEVs in the TME. Since ectopic HEV/TLS formation is disrupted when IL-7R, RORC, or LTB is knocked out, these three factors are considered essential for the function of LTis \u003csup\u003e17\u003c/sup\u003e. Th7R cells are characterized by high expression of \u003cem\u003eIL-7R\u003c/em\u003e and express \u003cem\u003eRORC\u003c/em\u003e, \u003cem\u003eLTB\u003c/em\u003e, and \u003cem\u003eCXCL13;\u003c/em\u003e thus, Th7R cells are considered good candidates for LTis in the TME.\u003c/p\u003e \u003cp\u003eCompared with their expression levels in the peripheral blood, Th7R cells exhibited significantly upregulated expression of more than 100 genes in the TME (by more than twofold). The upregulated genes included \u003cem\u003eCCL4, CCL4L2, CXCL13, XCL1\u003c/em\u003e, and \u003cem\u003eXCL2\u003c/em\u003e, and it is thought that they have the ability to recruit CD8\u003csup\u003e+\u003c/sup\u003e T cells, CXCR5-positive immune cells, and XCR1-positive dendritic cells to the TME \u003csup\u003e18,19\u003c/sup\u003e. In fact, the number of CD45\u003csup\u003e+\u003c/sup\u003e cells, including CD8\u003csup\u003e+\u003c/sup\u003e T cells, increased more than threefold in the tumors of mice subjected to adoptive transfer of Th7R cells.\u003c/p\u003e \u003cp\u003eCompared with Th1 and Th17 cells, Th7R cells had the strongest TME-dominant distribution and were the Th cells that included the most clonally expanded TCR clonotypes in the TME. This is in contrast to Th1 cells, which include the most clonally expanded TCR clonotypes in the peripheral blood. There are two possible explanations for this distribution pattern: one is that Th7R cells, which have a TCR specific to cancer antigens, may receive a stop signal in the TME and accumulate in large numbers, and the other is that Th7R cells may self-replicate in the TME in response to cancer antigen stimulation \u003csup\u003e20\u003c/sup\u003e. The upregulation of \u003cem\u003eJUNB\u003c/em\u003e and \u003cem\u003eFOSB\u003c/em\u003e in the TME suggests that Th7R cells self-replicate in the TME \u003csup\u003e21\u003c/sup\u003e. In any case, the fact that Th7R cells contain many multicellular TCR clones in the TME that are not found in the peripheral blood strongly suggests that many TIL-Th7R cells recognize cancer antigens. As evidence that Th7R cells are repeatedly stimulated by antigens in the TME, the expression of exhaustion-related genes such as \u003cem\u003eTOX, PDCD1, TIGIT, LAG3\u003c/em\u003e and \u003cem\u003eCTLA4\u003c/em\u003e was upregulated in TIL-Th7R cells (\u003cb\u003eTable \u003cspan refid=\"MOESM4\" class=\"InternalRef\"\u003eS4\u003c/span\u003e\u003c/b\u003e). One characteristic of the exhaustion-related expression pattern of Th7R cells is that the expression of \u003cem\u003eCTLA4\u003c/em\u003e was significantly greater than that of Th1 and Th17 cells. These findings suggest that anti-CTLA-4 antibody therapy may be effective in reinvigorating Th7R cells.\u003c/p\u003e \u003cp\u003eConsistent with previous reports, the IHC and Hyperion\u003csup\u003e\u0026trade;\u003c/sup\u003e results revealed that Tpex are localized to TLSs, and velocity analysis revealed that Tpex play a central role in providing cytotoxic Tex. On the other hand, surprisingly, most of the Th7R cells in the TME were found to be distributed in TLSs and to be in very close proximity to Tpex. It has been reported that CD8\u003csup\u003e+\u003c/sup\u003e T cells are reprogrammed to express \u003cem\u003eIL7R\u003c/em\u003e and \u003cem\u003eTCF7\u003c/em\u003e, avoiding terminal exhaustion, by forming a triad of physical contacts with dendritic cells and CD4\u003csup\u003e+\u003c/sup\u003e T cells \u003csup\u003e7\u003c/sup\u003e. This study indicates that Th7R cells are important candidate partners for CD4\u003csup\u003e+\u003c/sup\u003e T cells that play a role in triad formation, as they are present near Tpex within the TLS and are capable of recruiting dendritic cells due to their upregulated expression of \u003cem\u003eXCL1\u003c/em\u003e and \u003cem\u003eXCL2\u003c/em\u003e. In fact, a marked increase in Tpex and Tex abundance was observed in tumors from mice that had been subjected to adoptive transfer of Th7R. In contrast, the number of CTLs in the tumor did not increase substantially. Thus, Th7R cells likely promoted Tpex proliferation but did not facilitate CTL proliferation in the draining lymph nodes. Consistent with these findings, the results of a correlation analysis of the abundance of CD8\u003csup\u003e+\u003c/sup\u003e T-cell subpopulations in the lymph nodes and peripheral blood of lung cancer patients revealed that Th7R numbers were positively correlated with Tpex numbers. On the other hand, Th1 cell abundance was positively correlated with CTL abundance. Taken together, Th7R and Th1 cells share some transcription factors, cytokines, and chemokines but have distinct CD8\u003csup\u003e+\u003c/sup\u003e T-cell subpopulations as partners; as a result, only Th7R cells likely play an important role in long-term antitumor immunity.\u003c/p\u003e \u003cp\u003eIn the early stage of cancer development, CD8\u003csup\u003e+\u003c/sup\u003e T cells that clonally proliferate in the draining lymph nodes, as shown in the 7 steps of the cancer immunity cycle, are considered the main players in cancer cell eradication \u003csup\u003e22\u003c/sup\u003e. On the other hand, Tpex, which are responsible for the sustained supply of Tex in the TME, are thought to control the long-term equilibrium phase according to the cancer immunoediting theory and are a critical target for reinvigorating effective antitumor immunity via PD-1 blockade therapy \u003csup\u003e2,3,23,24\u003c/sup\u003e. Consistent with this idea, Tpex and Th7R cells, but not CTLs or Th1 cells, were significantly more abundant in the TME of lung cancer patients who achieved pCR or MPR after neoadjuvant ICI therapy.\u003c/p\u003e \u003cp\u003eIn summary, Th7R cells, Th1-like CD4\u003csup\u003e+\u003c/sup\u003e T cells that are involved in the generation of HEVs and TLSs in the TME, are deeply involved in long-lasting antitumor immunity by partnering with Tpex. Th7R cells, which exhibit TCR sharing between the TME and the peripheral blood, are distributed systemically, so their abundance is thought to be useful as a biomarker that reflects parameters associated with long-lasting antitumor immunity in the TME, such as HEVs/TLSs and Tpex. In addition, cell therapy using Th7R cells in combination with ICI therapy has promise as a treatment to promote Tpex proliferation.\u003c/p\u003e"},{"header":"Online Methods","content":"\u003cdiv id=\"Sec10\" class=\"Section2\"\u003e \u003ch2\u003ePatients\u003c/h2\u003e \u003cp\u003eTIL and peripheral blood samples for single-cell RNA sequencing (scRNA-seq, n\u0026thinsp;=\u0026thinsp;5), along with tissue samples for IHC analysis (n\u0026thinsp;=\u0026thinsp;56, \u003cb\u003eTable S6\u003c/b\u003e) and Hyperion\u0026trade; analysis, were collected from patients with stage I\u0026ndash;II non-small cell lung cancer (NSCLC) who underwent radical surgery between October 2017 and October 2019 at Saitama Medical University International Medical Center (Hidaka city, Saitama pref., Japan).\u003c/p\u003e \u003cp\u003eTIL samples for mass cytometry analysis were collected from patients (n\u0026thinsp;=\u0026thinsp;20, \u003cb\u003eTable S8\u003c/b\u003e) with stage II-IIIA NSCLC who underwent surgery after 3 courses of neoadjuvant therapy, including nivolumab and chemotherapy, between August 2023 and October 2024 at Saitama Medical University International Medical Center. The response to treatment was assessed using Response Evaluation Criteria in Solid Tumors (RECIST), version 1.1. Pathological response was independently assessed by pathologists.\u003c/p\u003e \u003cp\u003eAll the samples were collected after written informed consent was obtained from the patients. The Internal Review Board of Saitama Medical University International Medical Center approved the study protocol, in accordance with the Declaration of Helsinki (ethical approval Nos. 17\u0026ndash;084, 15\u0026ndash;221, and 2023-011).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec11\" class=\"Section2\"\u003e \u003ch2\u003eBlood samples\u003c/h2\u003e \u003cp\u003ePeripheral blood samples were collected using heparinized CPT Vacutainer tubes (Becton Dickinson Vacutainer Systems, Franklin Lakes, NJ, USA), as previously described\u003csup\u003e25\u003c/sup\u003e. The samples were preserved using Cellbanker2 (Nippon Zenyaku Kogyo Co., Koriyama, Japan) in a liquid nitrogen tank. For T-cell subset analyses, the cells were incubated for 32\u0026ndash;48 h in RPMI 1640 medium supplemented with 10% fetal calf serum (FCS) in a 5% CO\u003csub\u003e2\u003c/sub\u003e incubator at 37\u0026deg;C.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec12\" class=\"Section2\"\u003e \u003ch2\u003eTIL samples\u003c/h2\u003e \u003cp\u003eTo extract TILs, the collected tumors were sectioned into 2\u0026ndash;3-mm pieces and incubated with Dulbecco's modified Eagle\u0026rsquo;s medium supplemented with Dri Tissue \u0026amp; Tumor Dissociation Reagent (BD horizon: 661563; BD Biosciences, Franklin Lakes, NJ, USA) or with RPMI 1640 medium supplemented with Liberase TL Research Grade (05401020001; Roche, Basel, Switzerland) following the manufacturer\u0026rsquo;s instructions. Erythrocytes were removed using a biotin anti-human CD235ab antibody (306617; BioLegend, San Diego, CA, USA) and streptavidin nanobeads (480016; BioLegend, San Diego, CA, USA) according to the manufacturer\u0026rsquo;s instructions.\u003c/p\u003e \u003cp\u003e \u003cb\u003eMultiplex immunohistochemical staining (OPAL\u0026trade;), and imaging mass cytometry (Hyperion\u0026trade;) image acquisition and data analysis for tissue samples\u003c/b\u003e \u003c/p\u003e \u003cp\u003eThe tumor samples were formalin fixed and paraffin embedded (FFPE), and the three tumors with the largest areas of viable tumor cells were selected. Five-micron-thick sections were deparaffinized and rehydrated with xylene and ethanol for multiplex immunohistochemical staining. All the slides were sequentially treated with 0.3% hydrogen peroxide in methanol for 30 min to block endogenous peroxidase activity. To expose the antigens, the sections were autoclaved in 10 mM sodium citrate buffer (pH 6.0) for 20 min, heated in a microwave at 98\u0026deg;C for 15 min, and then cooled for 30 min.\u003c/p\u003e \u003cp\u003eAfter rinsing in 0.05 M Tris-buffered saline containing 0.1% Tween 20, the sections were incubated with various antibodies: mouse monoclonal CD4 (Leica Biosystems, clone 4B12, 1:300, high pH, retrieval), mouse monoclonal CD8 (DAKO, clone C8/144B, 1:150, high pH, retrieval), mouse monoclonal CD31 (DAKO, clone JC70A, 1:400, pH 6, retrieval), rabbit monoclonal CCR6 (Abcam, clone EPR22259, 1:1000, high pH, retrieval), rat monoclonal peripheral node addressin (Novus, clone MECA79, 1:200, pH 6, retrieval), mouse monoclonal CXCR3 (BioLegend, clone G025H7, 1:300, pH 6, retrieval), CD20 (Akoya, clone L26, 1:200, high pH, retrieval), and mouse monoclonal pan cytokeratin (Abcam, clone AE1/AE3, 1:100, pH 6, retrieval). All the slides were stained using an OPAL\u0026trade; 7-color IHC kit (Akoya Biosciences NEL811001KT, MA, USA). Immunofluorescence analysis was performed on the Mantra imaging platform (Akoya Biosciences) using Mantra Snap software. Color separation was conducted with inForm\u0026reg; Software v2.5.1 (Akoya Biosciences, MA) to extract image data. Multiplex immunohistochemical staining and data analysis were performed according to previously described procedures\u003csup\u003e26\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eAfter heat-induced epitope retrieval, the tumor samples were subjected to Hyperion\u0026trade; analysis. The sections were incubated with conjugated monoclonal antibodies overnight at 4\u0026deg;C (\u003cb\u003eTable S9\u003c/b\u003e). Following antibody incubation and washing, selected regions of interest (ROIs) were ablated with a laser at a resolution of 1 \u0026micro;m, and the vaporized particles were analyzed. The acquired data were visualized and analyzed using MCD Viewer and MCD Smart Viewer software (Standard BioTools). Pseudo-colored composite images were generated for each marker to assess tissue morphology and marker expression patterns. All images were reviewed at high resolution, and representative ROIs were selected for figure preparation.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec13\" class=\"Section2\"\u003e \u003ch2\u003eCell sorting for scRNA-seq and scTCR-seq\u003c/h2\u003e \u003cp\u003eFor peripheral blood scRNA-seq analyses, CD3\u003csup\u003e+\u003c/sup\u003e T cells were isolated from frozen PBMCs that had been incubated for 36 hrs in CM with 5% CO\u003csub\u003e2\u003c/sub\u003e at 37\u0026deg;C, using Dynabeads Untouched Human T Cells (11344D; Invitrogen). For TIL scRNA-seq analysis, samples were used after erythrocyte removal. For presorted scRNA-seq analysis, CD4\u003csup\u003e+\u003c/sup\u003e T cells were isolated as follows. CD62L\u003csup\u003elow\u003c/sup\u003e CD4\u003csup\u003e+\u003c/sup\u003e cells were separated from cultured frozen PBMCs using Dynabeads. Untouched Human CD4 T Cells (11346D; Invitrogen) and Human CD62L MicroBeads (130-091-758; Miltenyi Biotec, Bergisch Gladbach, Germany) were used. CD62L\u003csup\u003elow\u003c/sup\u003eCD4\u003csup\u003e+\u003c/sup\u003e T cells were subsequently sorted with CXCR3 (353715, 353716; BioLegend), CCR4 (353429, 353430; BioLegend), and CCR6 (359410; BioLegend) antibodies using a Cell Sorter SH800Z (Sony Imaging Products \u0026amp; Solutions Inc., Tokyo, Japan) prior to single-cell library construction.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec14\" class=\"Section2\"\u003e \u003ch2\u003eLibrary construction and sequencing for scRNA-seq and scTCR-seq\u003c/h2\u003e \u003cp\u003eAll single-cell RNA-seq and scTCR-seq libraries were constructed using the Chromium controller and Chromium Single-Cell Immune Profiling v2 kit (10x Genomics Inc., Pleasanton, CA, USA) to analyze gene expression differences and T-cell receptor (TCR) repertoires among T-cell subpopulations. Several TotalSeq-C antibody types (BioLegend Inc., San Diego, CA, USA) were used as feature barcodes to separate T-cell subpopulations (\u003cb\u003eTable S10\u003c/b\u003e). All the libraries were constructed according to the manufacturers\u0026rsquo; standard protocols. The quality and quantity of the libraries were evaluated using a Bioanalyzer High-Sensitivity Kit (Agilent Technologies, CA, USA). All libraries were sequenced on DNBSEQ-G400 sequencers (MGI, Shenzhen, China) in paired-end mode (read 1:28 bp; read 2:90 bp) according to the manufacturer\u0026rsquo;s standard protocol. The sequencing depth was targeted at 50,000 reads/cell for gene expression libraries and 10,000 reads/cell for feature barcoding and TCR libraries. Details of the libraries and sequence results are summarized in \u003cb\u003eTable S11\u003c/b\u003e.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec15\" class=\"Section2\"\u003e \u003ch2\u003eData preprocessing and quality control for scRNA-seq and scTCR-seq\u003c/h2\u003e \u003cp\u003eThe sequenced reads were processed using Cell Ranger 6 and 7 software (10x Genomics Inc., Pleasanton, CA, USA) with the GRCh38 reference dataset (version 2020-A for gene expression and 5.0.0 (samples from surgery patients) or 7.1.0 (presorted samples) for the TCR repertoire) obtained from 10x Genomics. This processing generated the gene expression/feature barcode (TotalSeq-C) unique molecular identifier (UMI) count matrix and TCR repertoire data. Quality control, statistical analysis, and graphical representation were performed using the Loupe browser (10x Genomics, Inc.) and the Seurat 5.2.1 package in R software (version 4.4.1) (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://satijalab.org/seurat/\u003c/span\u003e\u003cspan address=\"https://satijalab.org/seurat/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e)\u003csup\u003e27,28\u003c/sup\u003e. The gene expression library and feature barcoding library from the same sample were processed simultaneously using the \u0026lsquo;cellranger count\u0026rsquo; command. TCR libraries were processed using the \u0026lsquo;cellranger vdj\u0026rsquo; command for each sample. TCR repertoires from different samples (i.e., TILs and PBMCs) from the same patient were aggregated using the \u0026lsquo;cellranger aggr\u0026rsquo; command to identify shared clonotypes across different libraries. Quality control of the scRNA-seq data was executed as follows. cells with a valid TCR clonotype within the CD3D\u003csup\u003e+\u003c/sup\u003e cluster were selected using the Loupe browser. Genes expressed in fewer than three cells were excluded from the analysis. Cells with either too few (\u0026le;\u0026thinsp;200) or too many (\u0026gt;\u0026thinsp;2,500\u0026ndash;3000) expressed genes were excluded to remove empty or multiple GEMs. Cells with a high percentage of mitochondrial genes (\u0026ge;\u0026thinsp;5\u0026ndash;8%) were also excluded to eliminate dead cells, and those with excessively high TotalSeq-C counts were removed to minimize the impact of aggregated antibodies. These cutoff thresholds were set on the basis of the tail of the distribution. Prior to analysis, the UMI gene expression counts were log-normalized using Seurat's \u0026lsquo;LogNormalize\u0026rsquo; function (scale factor\u0026thinsp;=\u0026thinsp;10,000).\u003c/p\u003e \u003cp\u003e \u003cb\u003escRNA-seq and scTCR-seq data integration and analysis of CD4\u003c/b\u003e \u003csup\u003e \u003cb\u003e+\u003c/b\u003e \u003c/sup\u003e \u003cb\u003eand CD8\u003c/b\u003e\u003csup\u003e\u003cb\u003e+\u003c/b\u003e\u003c/sup\u003e \u003cb\u003eT cells derived from surgery patients\u003c/b\u003e\u003c/p\u003e \u003cp\u003eThe filtered CD3D\u003csup\u003e+\u003c/sup\u003e T cells derived from surgical patients were categorized into CD4\u003csup\u003e+\u003c/sup\u003eCD3D\u003csup\u003e+\u003c/sup\u003e T cells and CD8\u003csup\u003e+\u003c/sup\u003eCD3D\u003csup\u003e+\u003c/sup\u003e T cells on the basis of the RNA and surface protein expression of CD4 and CD8A prior to integration. The integration of these CD4\u003csup\u003e+\u003c/sup\u003e and CD8\u003csup\u003e+\u003c/sup\u003e T cells was conducted following the standard integration workflow of Seurat version 5, utilizing the reciprocal PCA integration method. Initially, 2,000 highly variable genes (excluding TCR-related genes) were identified using the 'FindVariableFeatures' function with the vst method. The expression of these variable genes was then scaled and centered using the 'ScaleData' function. Principal component analysis was performed using the 'RunPCA' function, followed by batch correction among samples using the reciprocal PCA method with the 'IntegrateLayers' function. Unsupervised clustering of the integrated gene expression matrix was carried out using the 'FindNeighbors' (with the parameter \u0026ldquo;dims\u0026thinsp;=\u0026thinsp;1:30\u0026rdquo;) and 'FindClusters' functions of Seurat on the basis of the shared nearest neighbor modularity optimization clustering algorithm\u003csup\u003e29\u003c/sup\u003e. The resolution parameter of 'FindClusters' was optimized on the basis of the number of expected clusters and the stability of the clustering results. Visualization via a UMAP plot was performed with the 'RunUMAP' function. Differential expression analysis was conducted using the MAST package in R software \u003csup\u003e30\u003c/sup\u003e, implemented in the 'FindMarkers' function of the Seurat package. Genes with an adjusted \u003cem\u003eP\u003c/em\u003e value\u0026thinsp;\u0026lt;\u0026thinsp;0.05 and log\u003csub\u003e2\u003c/sub\u003e-fold change\u0026thinsp;\u0026gt;\u0026thinsp;1 were considered significant and are indicated by red dots on volcano plots drawn with the EnhancedVolcano package in R\u003csup\u003e31\u003c/sup\u003e. The RNA velocity was estimated using the velocyto 0.17.17 package and scVelo 0.2.5 in Python 3.6\u003csup\u003e32\u003c/sup\u003e. The Loom file with UMAP embedding calculated via Seurat analysis was merged with the Loom file containing RNA velocity data from velocyto, and merged data were filtered and normalized using the 'scvelo.pp.filter_and_normalize' function (min_shared_counts\u0026thinsp;=\u0026thinsp;5, n_top_genes\u0026thinsp;=\u0026thinsp;2000). First- and second-order moments were computed for each cell across its nearest neighbors using the 'scvelo.pp.moments' function (n_pcs\u0026thinsp;=\u0026thinsp;30, n_neighbors\u0026thinsp;=\u0026thinsp;100). Finally, RNA velocities were estimated in the dynamic model using the 'scvelo.tl.recover_dynamics' and 'scvelo.tl.velocity' functions. The estimated velocity was visualized as a stream plot using the 'scvelo.pl.velocity_embedding_stream' function.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec16\" class=\"Section2\"\u003e \u003ch2\u003escRNA-seq analysis of presorted T cells\u003c/h2\u003e \u003cp\u003ePreparation and quality control of scRNA-seq libraries from presorted CD4\u003csup\u003e+\u003c/sup\u003e T cells were performed in a manner similar to that used for PBMCs from surgical specimens, except that feature barcoding technology with TotalSeq was not utilized. For cells that passed quality control, log-normalized count data were merged for each patient, and differential expression gene analysis was conducted among the four cell types, Th1, Th17, CXCR3\u003csup\u003e+\u003c/sup\u003e Th7R, and CXCR3\u003csup\u003e\u0026minus;\u003c/sup\u003e Th7R, on a round-robin basis using the \u0026lsquo;FindMarkers\u0026rsquo; function with the MAST method. This statistical test was applied to genes that were expressed in more than 10% of the cells and exhibited a minimum log\u003csub\u003e2\u003c/sub\u003e-fold change of 0.25. Genes with an adjusted \u003cem\u003eP\u003c/em\u003e value\u0026thinsp;\u0026lt;\u0026thinsp;0.05 were considered differentially expressed. The 64 DEGs common to all three patients were selected as candidate signature genes (\u003cb\u003eTable \u003cspan refid=\"MOESM2\" class=\"InternalRef\"\u003eS2\u003c/span\u003e\u003c/b\u003e).\u003c/p\u003e \u003cp\u003eTo identify CD4 signature genes and construct a predictive model for CD4 cell subtypes, we employed support vector machines (SVMs). To mitigate the influence of variability in RNA expression at the single-cell level, we created mini-clusters by averaging the gene expression profiles of 10 randomly sampled cells. For each sorted population\u0026mdash;Th1, Th7R, and Th17\u0026mdash;we generated 4,000 mini-clusters and used them for downstream analysis.\u003c/p\u003e \u003cp\u003eFeature selection was conducted using the 'rfeControl' function in the caret R package in combination with a radial kernel SVM implemented using the e1071 package. On the basis of this feature selection process, 58 genes were selected as relevant markers for classification.\u003c/p\u003e \u003cp\u003eBy using the expression profiles of these 58 genes, we built an SVM-based classifier to distinguish among the Th1, Th7R, and Th17 CD4 subtypes. Of the 4,000 mini-clusters generated per subtype, 2,000 were used for training, and the remaining 2,000 were used for testing to evaluate the model\u0026rsquo;s predictive performance.\u003c/p\u003e \u003cp\u003eTo annotate \u003cem\u003eSELL\u003c/em\u003e low-expression clusters, the classifier was trained on all 12,000 mini-clusters (4,000 per subtype) generated from the sorted reference samples.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec17\" class=\"Section2\"\u003e \u003ch2\u003eSingle-nucleus multiomic profiling of gene expression and chromatin accessibility analysis\u003c/h2\u003e \u003cp\u003eAfter a two-day culture of thawed frozen cells, CD4 staining was performed, and CD4\u003csup\u003e+\u003c/sup\u003e cells were sorted using a SH800Z cell sorter (Sony). Nuclei were extracted according to the 10x Genomics protocol (CG000365, Rev C). The nuclear density was assessed by microscopy using the BZ-X700 system (Keyence). All multiome libraries were constructed using the Chromium controller and Chromium Single Cell Multiome ATAC\u0026thinsp;+\u0026thinsp;Gene Expression Kit (10x Genomics, Inc.) following the 10x Genomics protocol (CG000338, Rev F). All multiome libraries were sequenced on DNBSEQ-G400 sequencers (MGI) according to the manufacturer\u0026rsquo;s standard protocol.\u003c/p\u003e \u003cp\u003eThe sequenced reads were processed using Cell Ranger ARC 2.0.2 software (10x Genomics, Inc.) with the GRCh38 reference dataset (version 2020-A-2.0.0) obtained from 10x Genomics. Quality control, statistical analysis, and graphical representation were performed using the Loupe browser (10x Genomics, Inc.), Seurat 5.2.1, and the Signac1.14 package in R software (version 4.4.1) \u003csup\u003e27,28,33\u003c/sup\u003e. Quality control was executed as follows. Clusters enriched for CD3D expression were identified using the Loupe Browser (10x Genomics) from 'cloupe' files generated by Cell Ranger ARC and subsequently imported into Seurat for further analysis. Cells with either too few (\u0026le;\u0026thinsp;500) or too many (\u0026gt;\u0026thinsp;2,500) expressed genes were excluded to remove empty or multiple GEMs. Cells with either too few (\u0026le;\u0026thinsp;1500) or too many (\u0026gt;\u0026thinsp;10,000) ATAC counts were also excluded. Cells with a high percentage of mitochondrial genes (\u0026ge;\u0026thinsp;15%) were excluded to eliminate dead cells. Cells with a nucleosome signal that was too high (\u0026gt;\u0026thinsp;2) or a TSS enrichment score that was too low (\u0026le;\u0026thinsp;1) were excluded, as these characteristics are indicative of compromised chromatin structure or degraded nuclei. Following quality control and normalization, dimensionality reduction was performed separately for each modality, with principal component analysis (PCA) applied to the RNA data and latent semantic indexing (LSI) to the ATAC data. Integration of the two modalities was achieved using the weighted nearest neighbor (WNN) approach, which enables joint clustering and visualization of cellular states in a unified embedding using the standard Seurat and Signac multiome analysis workflow. Unsupervised clustering of the integrated data was performed using the 'FindMultiModalNeighbors' function (reduction.list\u0026thinsp;=\u0026thinsp;list(\u0026ldquo;pca\u0026rdquo;, \u0026ldquo;lsi\u0026rdquo;), dims.list\u0026thinsp;=\u0026thinsp;list(1:50, 2:50)) and the 'FindClusters' function (resolution\u0026thinsp;=\u0026thinsp;1.5). To relate chromatin accessibility to gene expression, gene activity scores were computed using the 'GeneActivity' function (extend.upstream\u0026thinsp;=\u0026thinsp;2000, extend.downstream\u0026thinsp;=\u0026thinsp;0) on the basis of gene annotations from Ensembl Hsapiens version 98. To visualize chromatin accessibility signals and peak distributions around specific genes of interest, we generated coverage plots using the 'CoveragePlot' function implemented in Signac.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec18\" class=\"Section2\"\u003e \u003ch2\u003eMass cytometry\u003c/h2\u003e \u003cp\u003eThe monoclonal antibodies used for Helios\u003csup\u003e\u0026trade;\u003c/sup\u003e mass cytometry analysis are listed in \u003cb\u003eTable S12\u003c/b\u003e. Cell preparation and measurements were performed according to the manufacturer\u0026rsquo;s instructions for Helios (Standard BioTools, San Francisco, CA, USA). Briefly, 5.0 \u0026times; 10\u003csup\u003e6\u003c/sup\u003e cells were stained with mass cytometric antibodies. For intracellular staining, samples were prepared using Maxpar Nuclear Antigen Staining Buffer working solution prior to staining. After being washed twice with Maxpar Cell Staining Buffer, the samples were fixed with Maxpar Fix and Perm Buffer supplemented with 125 nM iridium nucleic acid intercalator (Standard BioTools). Following fixation, the cells were washed once with Maxpar cell-staining buffer, twice with Maxpar water, and resuspended in Maxpar water. Over 200,000 cells per sample were analyzed using Helios\u003csup\u003e\u0026trade;\u003c/sup\u003e and Cytobank (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.cytobank.org\u003c/span\u003e\u003cspan address=\"https://www.cytobank.org\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e) software. The gating strategy is detailed in \u003cb\u003eFig. S9\u003c/b\u003e.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec19\" class=\"Section2\"\u003e \u003ch2\u003eAssociations between T-cell subpopulations\u003c/h2\u003e \u003cp\u003eNetwork analysis was performed to evaluate associations between T-cell subpopulations in the peripheral blood. Network analysis eliminates false (pseudo)correlations due to confounding effects between multiple variables. In addition, the modified path consistency (PC) algorithm, a modification of the original path consistency algorithm, was used to address the redundant data that frequently occur in biological data \u003csup\u003e34,35\u003c/sup\u003e. The PC algorithm requires variable selection when calculating partial correlation coefficients (associations among three or more variables) of the first order or higher. Therefore, the estimation is repeated several times to reduce the dependence on variable selection, and edges with a high frequency of occurrence are selected. In the present analysis, the estimation was repeated 1,000 times, and edges inferred more than 500 times were chosen as thresholds.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec20\" class=\"Section2\"\u003e \u003ch2\u003eStatistical analyses\u003c/h2\u003e \u003cp\u003eGraphPad Prism 10 (GraphPad Software, San Diego, CA, USA) was used for statistical analysis. The data are presented as the means\u0026thinsp;\u0026plusmn;\u0026thinsp;standard errors of the means unless otherwise specified. Differences between two groups were evaluated using Student\u0026rsquo;s \u003cem\u003et\u003c/em\u003e test. Comparisons across multiple groups were conducted using one-way analysis of variance (ANOVA) with Tukey\u0026rsquo;s \u003cem\u003epost hoc\u003c/em\u003e test. Survival curves were generated using the Kaplan\u0026ndash;Meier method, with differences and hazard ratios assessed using the log-rank (Mantel‒Cox) test. All \u003cem\u003eP\u003c/em\u003e values were two-sided, and values\u0026thinsp;\u0026lt;\u0026thinsp;0.05 were considered to indicate statistical significance.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec21\" class=\"Section2\"\u003e \u003ch2\u003eAnimal models\u003c/h2\u003e \u003cp\u003eC57BL/6N female mice aged 8\u0026ndash;11 weeks were subcutaneously inoculated with 2.0 \u0026times; 10⁶ MCA 205 fibrosarcoma tumor cells in both flanks. On day 11, CXCR3\u003csup\u003e\u0026plusmn;\u003c/sup\u003eCCR4\u003csup\u003e\u0026minus;\u003c/sup\u003eCCR6\u003csup\u003e+\u003c/sup\u003e CD4\u003csup\u003e+\u003c/sup\u003e T cells were isolated from tumor-draining inguinal lymph nodes as Th7R cells using a cell sorter (SH800Z, Sony Corporation, Tokyo, Japan). The cells were activated with anti-CD3/CD28 microbeads and cultured in vitro in the presence of 10 U/ml IL-2. Subsequently, 2 \u0026times; 10⁶ Th7R cells were intravenously transferred into mice bearing MCA 205 subcutaneous tumors along the midline of the abdomen for 15 days. On day 7 post-Th7R administration, the tumors were harvested and digested with Liberase TL Research Grade to obtain a single-cell suspension. TILs were recovered by Percoll density gradient centrifugation and analyzed by mass cytometry.\u003c/p\u003e \u003c/div\u003e "},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eFunding:\u0026nbsp;\u003c/strong\u003eKAKENHI program of the Japan Society for the Promotion of Science grant 17H04184, Japan Agency for Medical Research and Development grant 19ae0101074h0001.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthor contributions\u003c/strong\u003e\u003cstrong\u003e:\u0026nbsp;\u003c/strong\u003eST, SY, and HK equally contributed to this study. ST conducted animal studies and analyzed the neoadjuvant IO treatment cohort. SY analyzed the scRNA-seq data and wrote the manuscript. OY, AM, AS, YM, KH, HI, KK, YI, HN, HS, and TH recruited the patients. KH analyzed the data and wrote the manuscript. HK designed the study, analyzed the data, and wrote the manuscript.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCompeting interests:\u0026nbsp;\u003c/strong\u003eHK is listed as an inventor on a patent application filed by Saitama Medical University that incorporates the discoveries described in this manuscript. HK has received grant support from Boehringer-Ingelheim, Inc. The other authors declare no competing interests.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eData and materials availability:\u0026nbsp;\u003c/strong\u003eThe raw scRNA-seq and scTCR-seq, and snRNA-seq + ATAC-seq data from this study are publicly available from the Gene Expression Omnibus (GEO) in GSE267026, GSE267027, GSE295601 and GSE295977. The data from our previous study (GSE215219) were also used.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAcknowledgments\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe thank Mrs. Koko Kodaira, Mrs. Kozue Watanabe, Mrs. Hiroko Noguchi, Mr. Jyoji Shiotani, and Mrs. Chieko Ono for their technical assistance in this study.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003e Im, S.J., \u003cem\u003eet al.\u003c/em\u003e Defining CD8\u0026thinsp;+\u0026thinsp;T cells that provide the proliferative burst after PD-1 therapy. \u003cem\u003eNature\u003c/em\u003e \u003cb\u003e537\u003c/b\u003e, 417\u0026ndash;421 (2016).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003e Kallies, A., Zehn, D. \u0026amp; Utzschneider, D.T. Precursor exhausted T cells: key to successful immunotherapy? \u003cem\u003eNat Rev Immunol\u003c/em\u003e \u003cb\u003e20\u003c/b\u003e, 128\u0026ndash;136 (2020).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003e Liu, B., \u003cem\u003eet al.\u003c/em\u003e Temporal single-cell tracing reveals clonal revival and expansion of precursor exhausted T cells during anti-PD-1 therapy in lung cancer. \u003cem\u003eNat Cancer\u003c/em\u003e \u003cb\u003e3\u003c/b\u003e, 108\u0026ndash;121 (2022).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003e Im, S.J., \u003cem\u003eet al.\u003c/em\u003e Characteristics and anatomic location of PD-1(+)TCF1(+) stem-like CD8 T cells in chronic viral infection and cancer. \u003cem\u003eProc Natl Acad Sci U S A\u003c/em\u003e \u003cb\u003e120\u003c/b\u003e, e2221985120 (2023).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003e Gebhardt, T., Park, S.L. \u0026amp; Parish, I.A. Stem-like exhausted and memory CD8(+) T cells in cancer. \u003cem\u003eNat Rev Cancer\u003c/em\u003e \u003cb\u003e23\u003c/b\u003e, 780\u0026ndash;798 (2023).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003e Ferris, S.T., \u003cem\u003eet al.\u003c/em\u003e cDC1 prime and are licensed by CD4(+) T cells to induce anti-tumour immunity. \u003cem\u003eNature\u003c/em\u003e \u003cb\u003e584\u003c/b\u003e, 624\u0026ndash;629 (2020).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003e Espinosa-Carrasco, G., \u003cem\u003eet al.\u003c/em\u003e Intratumoral immune triads are required for immunotherapy-mediated elimination of solid tumors. \u003cem\u003eCancer Cell\u003c/em\u003e \u003cb\u003e42\u003c/b\u003e, 1202\u0026ndash;1216 e1208 (2024).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003e Magen, A., \u003cem\u003eet al.\u003c/em\u003e Intratumoral dendritic cell-CD4(+) T helper cell niches enable CD8(+) T cell differentiation following PD-1 blockade in hepatocellular carcinoma. \u003cem\u003eNat Med\u003c/em\u003e \u003cb\u003e29\u003c/b\u003e, 1389\u0026ndash;1399 (2023).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003e Asrir, A., \u003cem\u003eet al.\u003c/em\u003e Tumor-associated high endothelial venules mediate lymphocyte entry into tumors and predict response to PD-1 plus CTLA-4 combination immunotherapy. \u003cem\u003eCancer Cell\u003c/em\u003e \u003cb\u003e40\u003c/b\u003e, 318\u0026ndash;334 e319 (2022).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003e Sallusto, F. Heterogeneity of Human CD4(+) T Cells Against Microbes. \u003cem\u003eAnnu Rev Immunol\u003c/em\u003e \u003cb\u003e34\u003c/b\u003e, 317\u0026ndash;334 (2016).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003e Kagamu, H., \u003cem\u003eet al.\u003c/em\u003e Single-cell analysis reveals a CD4\u0026thinsp;+\u0026thinsp;T cell cluster that correlates with PD-1 blockade efficacy. \u003cem\u003eCancer Res\u003c/em\u003e (2022).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003e Yanagihara, A., \u003cem\u003eet al.\u003c/em\u003e A Th1-like CD4(+) T-cell Cluster That Predicts Disease-free Survival in Early-stage Lung Cancer. \u003cem\u003eCancer Res Commun\u003c/em\u003e \u003cb\u003e3\u003c/b\u003e, 1277\u0026ndash;1285 (2023).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003e Kagamu, H., \u003cem\u003eet al.\u003c/em\u003e Single-Cell Analysis Reveals a CD4\u0026thinsp;+\u0026thinsp;T-cell Cluster That Correlates with PD-1 Blockade Efficacy. \u003cem\u003eCancer Res\u003c/em\u003e \u003cb\u003e82\u003c/b\u003e, 4641\u0026ndash;4653 (2022).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003e Sautes-Fridman, C., Petitprez, F., Calderaro, J. \u0026amp; Fridman, W.H. Tertiary lymphoid structures in the era of cancer immunotherapy. \u003cem\u003eNat Rev Cancer\u003c/em\u003e \u003cb\u003e19\u003c/b\u003e, 307\u0026ndash;325 (2019).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003e Galletti, G., \u003cem\u003eet al.\u003c/em\u003e Two subsets of stem-like CD8(+) memory T cell progenitors with distinct fate commitments in humans. \u003cem\u003eNat Immunol\u003c/em\u003e \u003cb\u003e21\u003c/b\u003e, 1552\u0026ndash;1562 (2020).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003e Forde, P.M., \u003cem\u003eet al.\u003c/em\u003e Neoadjuvant Nivolumab plus Chemotherapy in Resectable Lung Cancer. \u003cem\u003eN Engl J Med\u003c/em\u003e \u003cb\u003e386\u003c/b\u003e, 1973\u0026ndash;1985 (2022).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003e Meier, D., \u003cem\u003eet al.\u003c/em\u003e Ectopic lymphoid-organ development occurs through interleukin 7-mediated enhanced survival of lymphoid-tissue-inducer cells. \u003cem\u003eImmunity\u003c/em\u003e \u003cb\u003e26\u003c/b\u003e, 643\u0026ndash;654 (2007).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003e Luther, S.A. \u0026amp; Cyster, J.G. Chemokines as regulators of T cell differentiation. \u003cem\u003eNat Immunol\u003c/em\u003e \u003cb\u003e2\u003c/b\u003e, 102\u0026ndash;107 (2001).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003e Ferry, A., \u003cem\u003eet al.\u003c/em\u003e The XCL1-XCR1 axis supports intestinal tissue residency and antitumor immunity. \u003cem\u003eJ Exp Med\u003c/em\u003e \u003cb\u003e222\u003c/b\u003e(2025).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003e Wu, T.D., \u003cem\u003eet al.\u003c/em\u003e Peripheral T cell expansion predicts tumour infiltration and clinical response. \u003cem\u003eNature\u003c/em\u003e \u003cb\u003e579\u003c/b\u003e, 274\u0026ndash;278 (2020).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003e Hsieh, T., \u003cem\u003eet al.\u003c/em\u003e JunB Is Critical for Survival of T Helper Cells. \u003cem\u003eFront Immunol\u003c/em\u003e \u003cb\u003e13\u003c/b\u003e, 901030 (2022).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003e Chen, D.S. \u0026amp; Mellman, I. Oncology meets immunology: the cancer-immunity cycle. \u003cem\u003eImmunity\u003c/em\u003e \u003cb\u003e39\u003c/b\u003e, 1\u0026ndash;10 (2013).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003e Dunn, G.P., Bruce, A.T., Ikeda, H., Old, L.J. \u0026amp; Schreiber, R.D. Cancer immunoediting: from immunosurveillance to tumor escape. \u003cem\u003eNat Immunol\u003c/em\u003e \u003cb\u003e3\u003c/b\u003e, 991\u0026ndash;998 (2002).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003e Mellman, I., Chen, D.S., Powles, T. \u0026amp; Turley, S.J. The cancer-immunity cycle: Indication, genotype, and immunotype. \u003cem\u003eImmunity\u003c/em\u003e \u003cb\u003e56\u003c/b\u003e, 2188\u0026ndash;2205 (2023).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003e Kagamu, H., \u003cem\u003eet al.\u003c/em\u003e CD4(+) T-cell Immunity in the Peripheral Blood Correlates with Response to Anti-PD-1 Therapy. \u003cem\u003eCancer Immunol Res\u003c/em\u003e \u003cb\u003e8\u003c/b\u003e, 334\u0026ndash;344 (2020).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003e Halse, H., \u003cem\u003eet al.\u003c/em\u003e Multiplex immunohistochemistry accurately defines the immune context of metastatic melanoma. \u003cem\u003eSci Rep\u003c/em\u003e \u003cb\u003e8\u003c/b\u003e, 11158 (2018).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003e Stuart, T., \u003cem\u003eet al.\u003c/em\u003e Comprehensive Integration of Single-Cell Data. \u003cem\u003eCell\u003c/em\u003e \u003cb\u003e177\u003c/b\u003e, 1888\u0026ndash;1902 e1821 (2019).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003e Butler, A., Hoffman, P., Smibert, P., Papalexi, E. \u0026amp; Satija, R. Integrating single-cell transcriptomic data across different conditions, technologies, and species. \u003cem\u003eNat Biotechnol\u003c/em\u003e \u003cb\u003e36\u003c/b\u003e, 411\u0026ndash;420 (2018).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003e Ludo Waltman, N.J.v.E. A smart local moving algorithm for large-scale modularity-based community detection. \u003cem\u003eThe European Physical Journal B\u003c/em\u003e \u003cb\u003e86\u003c/b\u003e(2013).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003e Finak, G., \u003cem\u003eet al.\u003c/em\u003e MAST: a flexible statistical framework for assessing transcriptional changes and characterizing heterogeneity in single-cell RNA sequencing data. \u003cem\u003eGenome Biol\u003c/em\u003e \u003cb\u003e16\u003c/b\u003e, 278 (2015).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003e Kevin Blighe, S.R., Myles Lewis. EnhancedVolcano: publication-ready volcano plots with enhanced colouring and labeling. (2021).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003e Bergen, V., Lange, M., Peidli, S., Wolf, F.A. \u0026amp; Theis, F.J. Generalizing RNA velocity to transient cell states through dynamical modeling. \u003cem\u003eNat Biotechnol\u003c/em\u003e \u003cb\u003e38\u003c/b\u003e, 1408\u0026ndash;1414 (2020).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003e Stuart, T., Srivastava, A., Madad, S., Lareau, C.A. \u0026amp; Satija, R. Single-cell chromatin state analysis with Signac. \u003cem\u003eNat Methods\u003c/em\u003e \u003cb\u003e18\u003c/b\u003e, 1333\u0026ndash;1341 (2021).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003e Leday, G.G.R. \u0026amp; Richardson, S. Fast Bayesian inference in large Gaussian graphical models. \u003cem\u003eBiometrics\u003c/em\u003e \u003cb\u003e75\u003c/b\u003e, 1288\u0026ndash;1298 (2019).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003e Saito, S., Hirokawa, T. \u0026amp; Horimoto, K. Discovery of chemical compound groups with common structures by a network analysis approach (affinity prediction method). \u003cem\u003eJ Chem Inf Model\u003c/em\u003e \u003cb\u003e51\u003c/b\u003e, 61\u0026ndash;68 (2011).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003e Edgar, R., Domrachev, M. \u0026amp; Lash, A.E. Gene Expression Omnibus: NCBI gene expression and hybridization array data repository. \u003cem\u003eNucleic Acids Res\u003c/em\u003e \u003cb\u003e30\u003c/b\u003e, 207\u0026ndash;210 (2002).\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":true,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"nature-portfolio","isNatureJournal":true,"hasQc":false,"allowDirectSubmit":false,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"","title":"Nature Portfolio","twitterHandle":"","acdcEnabled":false,"dfaEnabled":false,"editorialSystem":"ejp","reportingPortfolio":"","inReviewEnabled":true,"inReviewRevisionsEnabled":false},"keywords":"tumor microenvironment, immune checkpoint inhibitors, T-cell immunity","lastPublishedDoi":"10.21203/rs.3.rs-6764582/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-6764582/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eThe priming, expansion, and function of CD8⁺ T cells depend on CD4⁺ T-cell help via dendritic cells. Precursor exhausted T cells (Tpex) maintain self-renewal and supply cytotoxic CD8⁺ T cells in the tumor microenvironment (TME), but their CD4⁺ T-cell partners were unclear. We performed scRNA-seq, scTCR-seq, and mass cytometry on peripheral blood, tumor, and lymph nodes from lung cancer patients and identified that IL-7R\u003csup\u003ehigh\u003c/sup\u003e CCR6⁺ Th1-like CD4⁺ T cells (Th7R) are numerically and spatially partnered with Tpex. Th7R expressed lymphotoxin-β and CXCL13, correlated with high endothelial venules, and co-localized with Tpex in tertiary lymphoid structures. Th7R abundance correlated with Tpex numbers in TME and lymph nodes, and adoptive transfer of Th7R increased Tpex in a mouse model. In patients, Th7R and Tpex in TME were associated with better response to neoadjuvant PD-1 blockade therapy. These results suggest that Th7R act as partners of Tpex to support sustained antitumor T-cell immunity.\u003c/p\u003e","manuscriptTitle":"Th7R Cells: CD4⁺ T-Cell Partners That Drive Tpex-Mediated Antitumor Immunity in the Tumor Microenvironment","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-06-04 12:44:12","doi":"10.21203/rs.3.rs-6764582/v1","editorialEvents":[],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"nature-communications","isNatureJournal":true,"hasQc":false,"allowDirectSubmit":false,"externalIdentity":"NCOMMS","sideBox":"Learn more about [Nature Communications](http://www.nature.com/ncomms/)","snPcode":"","submissionUrl":"https://mts-ncomms.nature.com/","title":"Nature Communications","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"ejp","reportingPortfolio":"Nature Communications","inReviewEnabled":true,"inReviewRevisionsEnabled":false}}],"origin":"","ownerIdentity":"ff6dd4c2-1840-4a77-9e2d-2fea365a48dd","owner":[],"postedDate":"June 4th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[{"id":49492548,"name":"Biological sciences/Immunology/Tumour immunology"},{"id":49492549,"name":"Biological sciences/Immunology/Translational immunology"}],"tags":[{"value":"featured","date":"2025-06-04 14:00:28"}],"updatedAt":"2025-06-04T12:44:12+00:00","versionOfRecord":[],"versionCreatedAt":"2025-06-04 12:44:12","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-6764582","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-6764582","identity":"rs-6764582","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}