Comprehensive analysis of prognosis and immune function of CD70-CD27 signaling axis in pan-cancer

The immune checkpoint molecule CD70 and its receptor CD27 constitute the signal transduction axis, which is abnormally expressed in many solid tumors and is crucial for T cell co-stimulation and immune escape. Tumor cells regulate CD27 expression in the tumor microenvironment by expressing CD70, which promotes immune escape. Although current research evidence suggests a link between CD70 and tumors, no pan-cancer analysis is available. Using the Cancer Genome Atlas, Gene Expression Omnibus datasets, and online databases, we first explored the potential carcinogenic role of the CD70-CD27 signaling axis in human malignancies. Furthermore, qRT-PCR, Western blot, immunohistochemistry, and a T cell-mediated tumor cell killing assay were used to assess the biological function of the CD70-CD27 signaling axis. CD70 expression is upregulated in most cancers and has an obvious correlation with the prognosis of tumor patients. The expression of CD70 and CD27 is associated with the level of regulatory T cell (Treg) infiltration. In addition, T cell receptor signaling pathways, PI3K-AKT, NF-κB, and TNF signaling pathways are also involved in CD70-mediated immune escape. CD70 mainly regulates tumor immune escape by regulating T cell-mediated tumor killing, with Tregs possibly being its primary T cell subset. Our first pan-cancer study provides a relatively comprehensive understanding of the carcinogenic role of the CD70-CD27 signaling axis in different tumors.


Introduction
The mechanism of tumor genesis and development is extremely complex (Kong et al. 2020a). Thus, extensive target gene/oncogene analysis is important for studying the relationship between target gene/oncogene and clinical prognosis, the potential of targeted therapy, and the underlying molecular mechanism. Currently, analyzing gene transcriptome data sets from publicly accessible tumor databases such as Cancer Genome Atlas (TCGA) and Gene Expression Omnibus (GEO) is an important method of performing single-gene or multi-gene extensive cancer analysis (Blum et al. 2018, Clough and Barrett 2016, Tomczak et al. 2015. Over the past few decades, advances in tumor immunology knowledge have led to the identification of immune checkpoints and the reinvigoration of cancer immunotherapy (Jacobs et al. 2015, Kong et al. 2021. Immune checkpoint inhibitors (ICIs) with strong immunotherapeutic potential in solid tumors have been shown to kill tumor cells by targeting immune checkpoints such as anti-cytotoxic T lymphocyte-associated protein 4 (CTLA-4), antiprogrammed death-1 receptor (PD-1), and anti-programmed death ligand-1 (PDL-1) (Lee et al. 2019). Although ICIs have shown promising results in targeted tumor therapy, drug resistance remains one of their major challenges , Kong et al. 2020b). Because of the high 48 Page 2 of 18 response of immunosuppressive cells, including myeloidderived suppressor cells, tumor-associated macrophages, and regulatory T cells (Tregs), treatment is not guaranteed in the microenvironment of solid tumors (Huang et al. 2022). Concurrently, immunosuppressive components of the tumor microenvironment (TME) may cause T cell failure and immune checkpoint protein dysfunction, leading to increased resistance to ICIs (Lin et al. 2020). research shows that tumor resistance to ICIs is caused by failure of antigen presentation in the TME, an impaired immune microenvironment, changes in immune checkpoint molecules, and immunosuppressive cells (Sharma et al. 2017).
Given this, developing new immunotherapy targets may be a more effective and promising therapeutic strategy. Thus, the CD70-CD27 axis, which belongs to the tumor necrosis factor (TNF) superfamily, has emerged as an appealing target for tumor development (Flieswasser et al. 2019, Flieswasser et al. 2022. CD70 and CD27 are TNF ligands, and receptor family ligand-receptor pairs are critical for T cell co-stimulation (Wajant 2016). CD27 is a costimulatory molecule on T cells that induces intracellular signaling-mediated cell activation, proliferation, effector function, and cell survival by binding to the ligand CD70 (Ansell et al. 2020). CD70, the only ligand of CD27, is a tightly regulated transmembrane glycoprotein and a member of the TNF superfamily. CD70 is expressed in B lymphocytes, T lymphocytes, and antigen-presenting cells (APCs) but not in non-lymphocyte tissues (van Nieuwenhuijze and Liston 2015). CD70 is abnormally expressed in malignant tumor cells, promoting tumor progression and immune escape via TME (Flieswasser, Camara-Clayette, Danu, Bosq, Ribrag, Zabrocki, Van Rompaey, de Haard, Zwaenepoel, Smits, Pauwels andJacobs 2019, Jacobs, Deschoolmeester, Zwaenepoel, Rolfo, Silence, Rottey, Lardon, Smits andPauwels 2015). Varlilumab, a CD70-CD27-mediated monoclonal antibody, has recently undergone extensive phase I/II clinical trials in vivo and in vitro for a variety of blood and solid tumor types, including Hodgkin's lymphoma, non-Hodgkin's lymphoma, glioblastoma, melanoma, renal cell carcinoma, prostatic adenocarcinoma, colorectal cancer, and ovarian cancer (Ansell, Flinn, Taylor, Sikic, Brody, Nemunaitis, Feldman, Hawthorne, Rawls, Keler andYellin 2020, Lutfi et al. 2021). However, clinical data show no link between the CD70-CD27 axis and various tumor types.
Our study is the first to analyze the CD70-CD27 axis using TCGA, GEO, and other databases. We also looked at gene expression, survival status, immune infiltration, and related cellular pathways to explore the potential molecular mechanisms of the CD70-CD27 axis in the pathogenesis, clinical prognosis, and immune escape of different cancers.

Gene expression analysis
The Human Protein Atlas (HPA) database (http:// www. prote inatl as. org/ human prote ome/ patho logy) was used to obtain the expression of CD70-CD27 in different cells and tissues under physiological conditions. The detailed information about the low specificity of CD70-CD27 was stated as "Normalized expression (NX) ≥ 1 in at least one tissue/ region/cell type but not elevated in any tissue/region/cell type," which can be found at http:// prote inatl as. org// search/ CD70 and http:// prote inatl as. org// search/ CD27.
We entered CD70 and CD27 into the "Gene_DE" module of the tumor immune estimation resource, version 2 (TIMER2) web (http:// timer. cistr ome. org/) and observed the expression difference of CD70 and CD27 between tumor and adjacent normal tissues for the different tumors or specific tumor subtypes of the TCGA project.
GEPIA (http:// gepia. cancer-pku. cn/): GEPIA is a web server for profiling and analyzing cancer and normal gene expression (Tang et al. 2017). In this study, the "Multiple Gene Comparison" modules were used to compare the expression of CD70 and CD27. The relationship between CD70 and CD27 expression and tumor stages was analyzed using the "Single Gene Analysis" and "Pathological Stage Plot" modules. We obtained violin plots of CD70 and CD27 expression in all TCGA tumors at different pathological stages (stage I, stage II, stage III, and stage IV). The log2 (Transcripts per million (TPM) +1) transformed expression data were used for the box or violin plots. P-values were calculated using the Student's t-test, with a cut-off of 0.05. F-value indicates the statistical value of the F test; Pr (> F) indicates the p-value. A P-value of < 0.05 was considered statistically significant. UALCAN (http:// ualcan. path. uab. edu/): UALCAN is a clinical database that analyzes gene expression in tumor subgroups, survival, and cancer genome-based mapping (TCGA) data from level 3 RNA-sequencing data across 31 cancer types (Chandrashekar et al. 2017). In this study, UALCAN was used to examine the distinct expression levels between tumor and healthy tissues. The TCGA dataset was downloaded and used for analyzing the expression of CD70 in pan-cancer patients. P-values were calculated using the Student's t-test, with a cut-off of 0.05.

Survival prognosis analysis
We used the "Survival Map" module of GEPIA2 (Tang et al. 2019) to obtain the OS and disease-free survival (DFS) significance map data for CD70 and CD27 across all TCGA tumors. Cut-off-high (50%) and cut-off-low (50%) values were used as expression thresholds to split the highand low-expression cohorts. The log-rank test was used in the hypothesis test, and the survival plots were obtained using the "Survival Analysis" module of GEPIA2.
The Kaplan-Meier (K-M) plotter (https:// kmplot. com/ analy sis/) can be used to assess the effects of approximately 54,000 genes on survival across 21 cancer types (Nagy et al. 2018). In this study, K-M curves were used to analyze the relationship between gene mutations in CD70 and overall survival (OS) in patients with pan-cancer. The logrank test was used to evaluate significant differences in the survival curves. A P-value of < 0.05 was considered statistically significant. Patients were divided into highand low-expression groups based on the median values of mRNA expression levels, and the results were validated using K-M survival curves based on the hazard ratio (HR) with 95% confidence intervals (CIs) and log-rank P-values. Patients were divided into groups to determine the best cutoff values, and survival analysis was performed. A P-value of < 0.05 was considered statistically significant.

Immune infiltration analysis
We used the "Immune-Gene" module of the TIMER2 web server to investigate the association between CD70-CD27 expression and immune infiltrates in all TCGA tumors. Treg cells were selected as immune cells. Immune infiltration was estimated using the TIMER, CIBERSORT, CIB-ERSORT-ABS, QUANTISEQ, and XCELL algorithms. The purity-adjusted Spearman's rank correlation test was used to calculate the P-values and partial correlation (cor) values obtained via the purity-adjusted Spearman's rank correlation test. A heatmap and a scatter plot represented the data.

CD70-related gene enrichment analysis
We first searched the STRING website (https:// string-db. org/) with a single protein name ("CD70") and organism ("Homo sapiens"). Following that, we set the following key parameters: the minimum required interaction score ["Low confidence (0.150)"], the meaning of network edges ("evidence"), the maximum number of interactors to show ("no more than 50 interactors" in the first shell), and active interaction sources ("experiments"). Finally, the experimentally determined 50 CD70-binding proteins were obtained.
We used the "Similar Gene Detection" module of GEPIA2 to identify the top 100 CD70-correlated targeting genes using datasets from all TCGA tumors and normal tissues. We also used the "correlation analysis" module of GEPIA2 to conduct a pairwise gene Pearson correlation analysis of CD70 and selected genes. For the dot plot, the log2 TPM was used. The P-value and the correlation coefficient (R) were shown. Moreover, we used the TIMER2 "Gene_Corr" module to provide the heatmap data for the selected genes, which included the cor and P-value in the purity-adjusted Spearman's rank correlation test.
We used jvenn, an interactive Venn diagram viewer (Bardou et al. 2014), to conduct an intersection analysis to compare the CD70-binding and interacting genes. We also combined the two sets of data to perform the Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis. In brief, we uploaded the gene-listed Database for Annotation Visualization and Integrated Discovery (DAVID) with the selected identifier ("OFFI-CIAL_GENE_SYMBOL") and species ("Homo sapiens") settings and obtained the functional annotation chart data. Finally, the enriched pathways were visualized using the R packages "tidyr" and "ggplot2." Furthermore, we used the R package "clusterProfiler" to conduct Gene Ontology (GO) enrichment analysis. The data for biological process (BP), cellular component (CC), and molecular function (MF) were visualized as cnetplots using the cnetplot function (circular = F, colorEdge = T, and node_label = T). The R language software (R-3.6.3) (https:// www.r-proje ct. org/) was used in this analysis. A twotailed P-value < 0.05 was considered statistically significant.

Cell lines and culture
The HCC-LM3 cell line (Institute of Biochemistry and Cell Biology, CAS) was cultured in Dulbecco's Modified Eagle Medium (DMEM), and the renal cell carcinoma A498 cell line (Institute of Biochemistry and Cell Biology, CAS) was cultured in complete RPMI-1640 medium supplemented with 10% fetal bovine serum and 1% penicillin/ streptomycin (Invitrogen Co., Carlsbad, CA, USA). Cells were maintained at 37 °C in a water-saturated atmosphere with 5% CO 2 /95% air.

qRT-PCR assay
RNA extraction and qRT-PCR: The tumor and paracancerous tissues were treated with the TRIzol reagent (Invitrogen, USA) for 10 min before centrifugation at 12,000 g at 4 °C for 15 min. Following that, the suppressed RNA was collected and mixed with isopropanol for isolation. After obtaining RNA, its purity and concentration were determined using a NanoDrop 1000 spectrophotometer (Thermo Fisher, USA). Genomic DNA was removed from the sample using the FastKing RT Kit (with gDNase) (TIANGEN, China). A high-capacity cDNA reverse transcription kit (Life Tec, USA) was used to synthesize the cDNA. On a LightCycler 96 system (Roche, USA), qRT-PCR was performed using a 2× Universal SYBR Green Fast qPCR mix (ABclonal, China). We also set up a no-RT negative control to confirm no genomic DNA contamination. All experiments were performed in triplicate.

Western blot
The proteins separated on gels after electrophoresis were transferred to poly (vinylidene fluoride) membranes and subsequently incubated with primary and secondary antibodies. The following primary antibodies were incubated for 12 h at 4 °C: GAPDH (1:10,000 dilution, ABclonal), CD70 (1:1,000 dilution, ABclonal, China, Catalog No.: A2032). The secondary antibody (1:5,000 dilution, ABclonal) was incubated for 2 h at room temperature. Immunoreactive bands were developed using an enhanced chemiluminescence detection kit (Genview Scientific Inc., USA). All experiments were performed in triplicate.

Immunohistochemistry
Paracancer and tumor tissue samples were collected from Zhongnan Hospital of Wuhan University and treated with formalin embedding. Each tissue was cut into 4-μm thick slices and laid flat on a slide. After decolorization, the slices were incubated with 3% hydrogen peroxide for 15 min. Heat-induced epitopes were searched for in a sodium citrate buffer (10 mM sodium citrate, 0.05% Tink 20, pH 6.0) at 96 °C for 30 min. After immersing in phosphatebuffered saline (PBS) 3 times (5 min each), rabbit antihuman CD70 antibody (1:100 dilution, ABclonal, Catalog No.: A2032) was incubated for 2 h. Sections were incubated with a few drops of A solution (ChemMateTMEn-Vision+/HRP) for 30 min, followed by diamino-benzidine staining and hematoxylin restaining. The slices were then dehydrated, soaked in xylene, and fixed with a neutral resin.

Cell transfection assay
The siRNA sequences used for targeted silencing of CD70 (NCBI refseq: NM_001252) were designed and selected according to the previously described method (Elbashir et al. 2002). Three CD70 siRNAs and scramble siRNAs (Supplementary Table 1) were synthesized and annealed by the manufacturer (Genechem, Shanghai, China). HCC-LM3 and A498 cells were seeded into 6-well plates (3 × 10 5 cells/ well) and 96-well plates (1 × 10 4 cells/well). The siRNA was mixed in Opti-MEN (Gibco), while Lipofectamine 3000™ (Invitrogen, Thermo Fisher Scientific Inc., UK) was incubated in Opti-MEN at room temperature for 5 min. Then the diluted siRNA and Lipofectamine 3000 were incubated for another 20 min at room temperature for complex formation. The complexes were added to wells, and the final siRNA concentration was 400 nmol/l. This was followed by a 48-h transfection.

T cell-mediated tumor cell killing assay
We co-cultured primary T cells with cancer cells to assess their effect on cancer cells. Human peripheral blood mononuclear cells (PBMCs) were purchased from TPCS (Milestone Biotechnologies, Shanghai, China). A total of 3 × 10 6 /mL T cells were treated with 1 μg/mL of CD3 monoclonal antibody (OKT3) (Cat# 16-0037-81, Thermo, Shanghai, China) and 50 ng/mL of IL-2 (Cat# PHC0026, Thermo, Shanghai, China) to activate T cells. After 48 h, activated T cells were maintained at a density of 1 million cells per milliliter of culture medium in the previously described culture medium, and the medium was changed every 2-3 days. HCC-LM3 and A498 cells were co-cultured with activated CD45+CD3+ T cells isolated from PBMCs. The effector-to-target ratio was 1:1. After 4 days of co-culture of tumor cells and T cells on 6-well plates or 96-well plates, T cells were washed twice with PBS to remove T cells, and the survival tumor cells were fixed and stained with crystal violet solution or added with the cell counting kit-8 (CCK-8) reagent to measure tumor cell viability.

CCK-8 assay
After the T cell-mediated tumor cell killing experiment, cells were plated in 96-well plates with 3 replicates per experimental group. Cells were cultured for 24 h at 37 °C in a humidified incubator with 5% CO 2 . A microplate spectrophotometer (Thermo Fisher, USA) was used to determine the optical density values at 450 nm using a CCK-8 kit (GeneView, USA). All data were normalized to control wells that contained no cells and are presented as the mean ± SD.

Clone formation assay
Following the T cell-mediated tumor cell killing experiment, cells were washed with PBS, fixed with 4% paraformaldehyde, and stained with crystal violet staining solution (Beyotime, China) before being used in a 6-well plate experiment. A digital camera was used to photograph the colonies. Fig. 1 Expression levels of CD70 and CD27 genes in different cells and tissues under normal physiological conditions and different tumor and pathological stages. A The expression of the CD70 gene in different tissues is analyzed using the consensus dataset of HPA, GTEx, and FANTOM5. B The expression of CD70 and CD27 genes in different blood cells is analyzed using the HPA, Monaco, and Schmiedel consensus datasets. C TIMER2 is used to analyze the expression of the CD70 gene in different cancers or specific cancer subtypes. *P < 0.05; **P < 0.01; ***P < 0 .001. D The expression of the CD70 gene in different cancers is analyzed using the TCGA database in UAL-CAN (Pan-cancer view module ◂

Ethics approval and consent to participate
This study was approved by the Medical Ethics Committee of Zhongnan Hospital at Wuhan University. All experiments were conducted following the study protocol.

Statistical analysis
Statistical analyses included one-way ANOVA calculations and the unpaired t-test. SPSS 24.0 software (SPSS, Inc., Chicago, IL, USA) was used for statistical analysis. A P-value of < 0.05 was considered statistically significant.

CD70-CD27 signal axis gene expression analysis data
In this study, we aimed to investigate the role and mechanism of human CD70 (mRNA: NM_001252.5, protein: NP_001243.1) in carcinogenesis and immune escape. Phylogenetic tree data (Supplementary Figure 1) show the evolutionary relationship of the CD70 protein in humans.
We first analyzed the expression patterns of CD70 and CD27 in different cells and non-tumor tissues. Using data from the HPA, GTEx, and functional annotation of mammalian genome 5 (FANTOM5) datasets, we found that CD70 was most expressed in the appendix, spleen, and tonsil, followed by the lymph node and thymus (Fig. 1A). Meanwhile, CD27 expression was highest in the appendix, spleen, tonsil, lymph node, and thymus (Supplementary Figure 2A). CD70 and CD27 were expressed in all tissues studied, indicating low RNA tissue specificity. When we analyzed the expression of CD70 and CD27 in different blood cells from the HPA/Monaco/Schmiedel dataset, we discovered immune cell type specificity and enhanced expression of CD70 and CD27 in Treg immune cells (Fig. 1B). This suggests that the target cells of the CD70-CD27 axis could be Tregs.
We also observed the correlation between the expression of CD70 and CD27 and the pathological staging of cancer using the "pathological staging map" module of GEPIA2. As shown in Fig. 2, the pathological staging expression of CD70 was statistically different in testicular germ cell tumors (TGCT), KIRC, KIRP, KICH, LUAD, and THCA. Concurrently, the pathological staging of CD27 was significantly different in LUAD, KIRC, STAD, and skin cutaneous melanoma (SKCM). In addition, we further analyzed the expression of CD70 in different tumor types and pathological stages using the TCGA database. As shown in Fig. 3 and Supplementary Figure 3, CD70 expression was upregulated to varying degrees in the diffuse tissues, and the expression varied across clinical stages.

Survival analysis data of CD70-CD27 signal axis
Cancer cases were divided into high and low expression groups based on CD70 and CD27 expression levels, and the correlation between CD70 and CD27 expression and the prognosis of different tumor patients was studied using the "Survival Map" module of GEPIA2. As shown in Fig. 4, in the TCGA database, upregulation of CD70 expression was associated with poor OS outcomes in low-grade glioma (LGG) (P = 3 e-05), mesothelioma (P = 0.0076), and uveal melanoma (UM) (P = 0.013). It was associated with a poor prognosis for DFS in COAD (P = 0.036) and LGG (P = 0.0076). In addition, upregulated CD27 expression was associated with poor prognoses of OS for LGG (P = 0.012) and UVM (P = 0.0067) and with a poor prognosis of DFS for LGG (P = 0.035). In addition, we used K-M analysis to analyze CD70 survival data and found that high CD70 expression was associated with poor OS in LUAD (P = 0.0082), lung squamous cell carcinoma (LUSC) (P = 0.0088), STAD (P = 0.023), and THYM (P = 0.0071) (Supplementary Figure 4). Low CD70 expression in UCEC was associated with poor OS (P = 0.00012) (Supplementary Figure 4).

Immune infiltration analysis data of CD70-CD27 signal axis
Tumor-infiltrating immune cells are crucial in TME and are linked to tumor immune escape (Steven and Seliger 2018). Tumor tissue and its microenvironment change the host immune system in many ways, promoting tumor growth. One approach is to increase the number of tumor-infiltrating Tregs to achieve immune escape. Tregs are a potential therapeutic barrier and a cause of immunotherapy failure due to their persistence in TME and immunosuppression (Sarkar et al. 2021). We used the CIBERSORT, CIBERSOR-ABS, QUANTISEQ, and XCELL algorithms in the TIMER2 web. We investigated the potential relationship between Treg invasion level and CD70-CD27 expression in different tumor tissues. The results showed that Treg infiltration was positively correlated with CD70 gene expression in BRCA, LIHC, LUAD, LUSC, OV, PRAD, and THCA (Fig. 5). Treg infiltration levels were also positively correlated with  Figure 5). These findings suggest that CD70 may activate Tregs and enhance their immunosuppressive function through co-stimulation with CD27 in Tregs, thus promoting immune escape, tumor occurrence, and tumor development.

Interactions and correlations of predicted proteins with the CD70-CD27 signal axis
We attempted to screen out genes targeting CD70-binding proteins and related CD70 expression for a series of pathway enrichment analyses to further study the molecular mechanism of the CD70 gene in tumorigenesis. A total of 50 CD70binding proteins were obtained using the STRING tool and supported by experimental evidence (Supplementary Table 2). Figure 6A shows the network of interactions between these proteins. In addition, we used GEPIA2 to combine all of the TCGA expression data to identify the top 100 genes related to CD70 expression (Supplementary Table 2). Through the intersection analysis of the protein data of the above two groups, we found a common protein, tumor necrosis factor superfamily member 9 (TNFSF9) (Fig. 6B). TNFRSF9 was first reported as a T cell inducer in 1989 (Kwon and Weissman 1989). TNFRSF9 is cross-linked to activated T cell receptors and can deliver costimulatory signals to T cells, resulting in T cell proliferation, survival, memory formation, enhanced cytotoxicity, and cytokine production (Wortzman et al. 2013). Recent studies have found that activating TNFSF9 reverse signaling mediates tumor immune escape by forming a tumor suppressor microenvironment. TNFSF9 reverse signaling inhibition may have an anti-tumor effect by remodeling the TME (Wu and Wang 2022). Thus, this preferential proliferation of T cells makes TNFRSF9 a particularly attractive target for anticancer therapy. Currently, antibodies targeting TNFRSF9 are being used in preclinical studies, including urelumab (BMS-663513), which blocks TNFRSF9 interaction with TNFSF9 (Etxeberria et al. 2020). Therefore, CD70 and TNFSF9 are expected to become new immunotherapy targets.
We combined these two data sets for GO and KEGG enrichment analyses. Among them, GO enrichment analysis showed that the tumor cell BP was primarily involved in immune response, T cell stimulation, and T cell activation pathways (Supplementary Table 3), suggesting that the CD70-CD27 axis was primarily a T cell-mediated immune response process (Fig. 7A). CC is primarily made up of the plasma membrane, integral membrane components, and extracellular exosomes (Fig. 7B, Supplementary  Table 4). Furthermore, cytokine activity, tumor necrosis factor receptor binding, and tumor necrosis factor-activated receptor activity comprise MF (Fig 7C, Supplementary  Table 5). KEGG enrichment analysis showed that the pi3K-AKT, NF-κB, and TNF signaling pathways were all activated. In addition, the cytokine receptor interaction pathway was involved in the activation of the CD70-CD27 signaling axis (Fig. 7D, Supplementary Table 6).

CD70 expression was upregulated in tumor tissues
According to an analysis of the UALCAN and GEPIA databases, CD70 expression was upregulated in tumor tissues. To further validate the expression of CD70 in tumor tissues, we collected surgical specimens (paracomcinoma and tumor tissues) from patients with LIHC, KIRP, KIRC, ESCA, and cervical cancer (CESC). The clinical characteristics of the patients are shown in Supplementary Table 7. Immunohistochemical results showed that CD70 expression was upregulated in LIHC, KIRP, KIRC, CESC, and ESCA (Fig. 8). This is consistent with the findings of our analyses in the UALCAN and GEPIA databases. These findings support the abnormal expression of CD70 in tumor tissues, implying that CD70 was involved in the development and immune escape of middle tumors.

CD70-CD27 signaling axis promotes tumor immune escape through T cells
According to the findings of the immune infiltration study, CD70-related partners, GO, and KEGG enrichment analysis, Tregs may be the primary T cell subgroup targeted by the CD70-CD27 signal axis, which primarily achieves immunological escape via the T cell pathway. To verify the results of this analysis, we constructed CD70 knockdown and overexpression plasmids and transfected them into the HCC-LM3 and A498 cell lines using a cell transfection experiment. Figure 9A and B show that CD70 expression was downregulated after CD70 knockdown while CD70 expression was upregulated after CD70 overexpression. Meanwhile, the T cell-mediated tumor killing experiment revealed that under CD70 knockdown, the tumor cell viability was significantly lower than that of the control group, whereas the cell viability of the overexpressed CD70 group was significantly higher than that of the control group (Fig. 9C). Interestingly, in the absence of T cell co-culture, there was no significant difference in tumor cell viability in either the knockdown or overexpression CD70 groups (Fig. 9C). The Fig. 5 TIMER2 database is used to analyze the correlation between immune infiltration of regulatory T cells (Tregs) and CD70 expression. We used different algorithms to analyze the potential correlation between CD70 gene expression level and Treg invasion level in all types of cancer in the TCGA ◂ Fig. 6 Enrichment analysis of CD70-related genes. A We first obtain the available experimentally determined CD70-binding proteins using the STRING tool. B An intersection analysis of the CD70-binding and correlated genes is conducted. C Using the GEPIA2 approach, we also obtain the top 100 CD70-correlated genes in TCGA projects and analyzed the expression correlation between CD70 and selected targeting genes, including IL-2, TNFSF9, CD86, CCR4, CD27, FoxP3, and GAL3ST1. D The corresponding heatmap data in the detailed cancer types are displayed findings were further verified by cloning experiments (Fig. 9D). Therefore, the findings suggest that the immune escape of tumor cells via the CD70-CD27 signaling axis is primarily mediated by T cells. This is also consistent with the above findings of the previous analysis.

Discussion
The importance of CD70 in solid tumors has become increasingly clear in recent decades, and abnormal expression of CD70 on tumor cells has been reported in various solid tumors with different levels of expression (Inaguma et al. 2020, Jacobs, Deschoolmeester, Zwaenepoel, Rolfo, Silence, Rottey, Lardon, Smits and Pauwels 2015, Kong, Ye, Miao, Liu, Huang, Xiong, Wen and Zhang 2021. Therefore, treatments targeting CD70 have great potential for treating both early and advanced cancers. However, some studies have shown that CD70 is not upregulated or is not significantly upregulated in tumor cells from Kaposi's sarcoma, prostate cancer, Langerhans cell histiocytosis, and colorectal cancer (Flieswasser, Van den Eynde, Van Audenaerde, De Waele, Lardon, Riether, de Haard, Smits, Pauwels andJacobs 2022, Ryan et al. 2010). Although CD70 is prevalent in many cancers, its expression pattern varies in terms of spatial distribution, expression intensity, and percentage of positive cells. In addition, recent studies have reported that CD70 is involved in developing malignant tumors, particularly in regulating tumor immune escape. However, whether the CD70-CD27 signaling axis can play a role in immune escape in different tumors via some common molecular mechanisms remains to be further studied. Through a literature search, we were unable to find any literature on a generalized analysis of the CD70-CD27 signaling axis from a global tumor perspective. Therefore, we performed a comprehensive analysis of the CD70-CD27 signaling axis in 33 different tumors using data from the TCGA, CPTAC, and GEO databases, as well as gene expression, survival prognosis, and immune infiltration analysis.
A new understanding of the CD70/CD27 axis has recently revealed its unique characteristics in TME. We found that CD70 was upregulated in tumor tissues by analyzing the abnormal expression of CD70 on tumor cells and the continuous signal transduction in TME. Furthermore, the CD70-CD27 axis mediates immune escape by regulating T cell apoptosis, T cell failure, and Treg survival. In addition, blocking this pathway can prevent the proliferation of malignant tumors. In our study of the immune infiltrating module,  The CD70-CD27 signaling axis promotes tumor immune escape through T cells. A Western blot experiment shows that CD70 expression is downregulated after knockdown and upregulated after overexpression. B qRT-PCR shows that CD70 expression is downregulated after knockdown and upregulated after overexpression. C T cell-mediated tumor killing experiment shows that the tumor cell viability is significantly lower than that of the control group under CD70 knockdown, while the cell viability of the CD70 overexpres-sion group is significantly stronger than that of the control group. D The clonal formation assay shows that the proliferation of tumor cells is significantly lower than that of the control group after CD70 knockdown and significantly higher than that of the control group after CD70 overexpression. Data are shown as the mean ± SD of three independent experiments (*P < 0.05, **P < 0.01 vs. control. ns: not significant) we found that CD70 and CD27 were positively correlated with Treg infiltration intensity. This implies that the CD70-CD27 axis is crucial in regulating Tregs. Treg is critical for immune tolerance and modulating immune responses to pathogens and tumors (Yang 2022). Tregs are an immunosuppressive subpopulation of CD4+ T cells characterized by the regulation of the transcription factor FoxP3, which regulates Treg differentiation and development (Takeuchi and Nishikawa 2016). Tregs are vital in maintaining tolerance and immune homeostasis and promote tumor progression by impeding effective anti-tumor immunity (Tian et al. 2022). However, it remains unknown how Tregs function in the TME. Thus, studying the molecular mechanism of Tregs' function is conducive to developing new immunotherapy targets. Several studies have confirmed that the CD70-CD27 signaling axis mediates immune escape via its effect on Tregs. For example, in a mouse solid tumor model, CD70-CD27 signaling can directly reduce Treg apoptosis, whereas in vitro experiments revealed that an increase in IL-2 secreted by non-Treg CD4+ T cells indirectly increased the apoptosis of Tregs (Claus et al. 2012).
We screened for CD70-binding proteins and genes related to CD70 expression for a series of pathway enrichment analyses to investigate the potential molecular mechanism of the CD70-CD27 signaling axis in regulating Tregs. Using the STRING tool, 50 CD70-binding proteins were identified and supported by experimental evidence. In addition, the GEPIA2 tool was combined with all TCGA tumor expression data to obtain the top 100 genes related to CD70 expression. Interestingly, we found that CD70 expression was linked to the expression of several genes associated with Treg function. FoxP3, CCR4, CD86, and TGNFSF9 genes are a few examples. FoxP3, as a Treg-specific marker molecule, is crucial for maintaining Treg function and increasing Treg numbers (Saleh and Elkord 2020). According to studies, tumor-infiltrating Tregs express higher levels of the chemokine receptor CCR4 than peripheral Tregs in BRCA patients. At the same time, FoxP3, as a CCR4 transcription activator, can increase CCR4 expression in Tregs and promote BRCA immune escape (Sarkar et al. 2022). Furthermore, Tregs can reduce APC CD80 and CD86 expression via CTLA-4-dependent autophagy, thereby inhibiting APC T cell-stimulating activity. Reduced CD80/ CD86 expression on APCs can double-inhibit T cell immune responses by limiting the co-stimulation of naive T cells by CD80/CD86 and increasing the free PDL-1 expression on effector T cells used to inhibit PD-1 expression. Therefore, combined blocking of CTLA-4 and PD-1/PDL-1 may synergistically block Tregs-mediated immunosuppression, effectively enhancing immune responses, including tumor immunity (Tekguc et al. 2021). In GO and KEGG enrichment analyses, we found that GO enrichment analysis showed that the CD70-CD27 pathway is important in immune response, T cell stimulation, and T cell activation pathways. KEGG enrichment analysis showed that the PI3K-AKT signaling pathway, the NF-κB signaling pathway, and the TNF signaling pathway were all activated. In addition, the cytokine receptor interaction pathway is also involved in activating the CD70-CD27 signaling axis. Many studies over the last decade have found that the PI3K/AKT pathway is vital in the development, function, and stability of Tregs (Pompura and Dominguez-Villar 2018). The PI3K/AKT pathway regulates the function of mature Tregs and natural and induced Treg differentiation by integrating signals from multiple cell surface receptors. The NF-κB signaling pathway is essential for Treg differentiation and maturation, and despite significant progress in the last decade, many aspects of NF-κB function in Tregs remain unstudied (Hövelmeyer et al. 2022). Determining the NF-κB target gene, differentiating the common and unique roles of NF-κB in individuals, and elucidating the potential molecular mechanism of NF-κB protein regulation of Tregs-specific genes and tumor immune escape are all important future research tasks. To this end, we used a T cell-mediated tumor destruction assay to confirm the immune escape of the CD70-CD27 signaling axis. Our findings showed that when tumor cells and T cells were co-cultured, the tumor cell viability after CD70 knockdown was significantly lower than that of the control group, whereas the cell viability of the overexpressed CD70 group was significantly higher than that of the control group. In addition, in the absence of T cells, there was no statistically significant difference in cell viability between the treated and control groups, whether CD70 was knocked down or overexpressed. This means that the CD70 gene does not promote tumor cell proliferation, but rather facilitates immune escape by regulating T cells' function in the TME via the CD70-CD27 signaling axis.
Finally, our first pan-cancer analysis of the CD70-CD27 signaling axis revealed that CD70 expression was statistically correlated with clinical prognosis, immune cell infiltration, and tumor cell immune escape in multiple tumors, which aids in understanding the role of the CD70-CD27 signaling axis in tumorigenesis from the perspective of clinical tumor samples. Currently, antibodydrug conjugate compounds and chimeric antigen receptor T cell therapy targeting CD70 have been developed and are under clinical evaluation in blood and solid cancers (Massard et al. 2019, Owonikoko et al. 2016, Pal et al. 2019, Phillips et al. 2019. Although this approach has been successful in hematological malignancies, challenges remain in solid tumors due to multiple barriers, such as metastasis to the tumor site, infiltration into the tumor, and survival maintenance (Donnadieu et al. 2020). Thus, the potential molecular mechanism of the CD70-CD27 signaling axis in regulating T cell-mediated immune escape requires further investigation.

Conclusions
Anti-CD70-targeted therapy combinations have shown promise in preclinical and clinical settings. To date, several promising CD70 treatments are being investigated, both in preclinical and clinical settings, to improve treatment outcomes for cancer patients. Therefore, blocking the CD70-CD27 pathway is a hotspot of clinical research and has great potential as a single drug therapy. In addition, anti-CD70 therapy offers many opportunities for rational combination strategies with conventional treatment and immunotherapy. Therefore, strategies that inhibit the CD70-CD27 axis signaling pathway may be promising new therapeutic options in the future, in addition to existing methods of targeting CD70. Further preclinical studies and clinical evaluation of CD70 targeting strategies will provide new insights into the mechanisms and actions of CD70 and potentially pave the way for new treatment options in the oncology field.