Exosomal circGSE1 promotes immune escape of hepatocellular carcinoma by inducing the expansion of regulatory T cells

Abstract Studies have shown exosomal circRNAs can regulate the immune escape of tumors by carrying cancer‐derived molecules. Regulatory T cells (Tregs) participate in the process of tumor immune escape. However, the mechanism by which exosomal circRNAs regulate Tregs to create a microenvironment for tumor immune escape is unclear. The effect of exosomes on the proliferation, migration, and invasion of tumor cells was evaluated by CCK‐8, transwell, and wound‐healing assays. The expression of circGSE1 was evaluated by real‐time quantitative PCR, and the function of exosomal circGSE1 was explored by Western blot and RNA pull‐down assays. In vivo animal metastasis models and bioluminescence imaging were used to verify the effect of exosomal circGSE1 on tumor progression. Collectively, we revealed that exosomal circGSE1 derived from hepatocellular carcinoma (HCC) cells promotes the progression of HCC by inducing Tregs expansion via regulating the miR‐324‐5p/TGFBR1/Smad3 axis. Therefore, in the future, exosomal circGSE1 can be used as a promising biomarker for immunotherapy of HCC.


| INTRODUC TI ON
Hepatocellular carcinoma (HCC) is reported to be the fifth most common cancer in the world. 1 In the past few decades, the widespread application of immunotherapy in cancer has greatly changed the prognosis of cancer patients. 2 An increasing number of clinical trials are being carried out to explore the efficacy of PD1 inhibitors in advanced liver cancer. However, the phase 3 clinical trials that have been carried out so far have not achieved the expected efficacy. 3 Therefore, it is necessary to find new targets for immunotherapy of liver cancer. Antigen presentation, antigen recognition, T cell activation and migration, and the killing of tumor cells by effector T cells constitute the immune response of cancer. 4,5 The results of single-cell sequencing in liver cancer tissues showed that CD8 + T cells in liver cancer tissues were exhausted and the expansion of regulatory T cells (Tregs) was promoted. 6 Tregs secrete immunosuppressive factors to inhibit the function of CD8 + T cells, causing tumor immune escape. 7 An increasing number of studies have shown exosomes regulate the immune escape of tumors by carrying cancer-derived molecules. 8,9 However, the relationship between HCC-derived exosomes and Treg is unclear.
Circular RNA (circRNA) has been identified as a noncoding RNA with a closed-loop structure and no 5' and 3' ends, which mainly acts as miRNA sponge. [10][11][12][13] A recent study also found that cancer cell-derived exosomal circUHRF1 can participate in the tumor immune escape process by inducing natural killer cell depletion. 14 However, the specific mechanism by which cancer-derived exosomal circRNA regulates Tregs to participate in tumor immune escape is still unclear.
Notably, the TGFβ signaling pathway not only inhibits the production of CD8 + effector T cells, but also participates in promoting the expansion of Tregs. 15,16 In this study, we first demonstrated that HCC cell-derived exosomal circGSE1 promotes the progression of HCC by inducing Tregs expansion via regulating the miR-324-5p/ TGFBR1/Smad3 axis. Therefore, our study reveals the carcinogenic effect of exosomal circGSE1 in the progression of HCC cells and provides a new theoretical basis for the study of exosomal circRNAs in HCC immunotherapy.
The temperature of the incubator was controlled at 37°C, and the CO 2 concentration was 5%.
The cells were then transfected with the Lipofectamine™ 3000 kit (ThermoFisher Scientific) according to the manufacturer's instructions.

| Transmission electron microscopy (TEM)
Exosomes were extracted by ultraspeed ultracentrifugation as described above. 17 The purified exosomes were resuspended in 100 μl 1 × PBS. A total of 5 μl exosome suspension was added to a formvarcarbon sample copper net and fixed with 1% glutaraldehyde solution for 20 minutes. Then, 50 μl of uranyl oxalate was dropped on the copper mesh at pH 7 for 5 minutes and 50 µl of methylcellulose-UA onto the copper mesh for 10 minutes. The copper mesh was dried in the air for 10 minutes and finally photographed at 80 kV with FEI TecnaiG2 spirit transmission electron microscope.

| Nanoparticle-tracking analysis (NTA)
As mentioned above, the purified exosomes were resuspended in PBS and analyzed and identified using the 301 NanoSight NS300 system. 18

| Exosome labeling and tracking
PKH26 red fluorescent membrane linker dye (Sigma) was used to stained the purified exosomes. The labeled exosomes were cocultured with T cells. Fluorescence microscopy was used to observe the uptake process of T cells in exosomes.

| Quantitative reverse-transcription polymerase chain reaction (qRT-PCR)
TRIzol reagent (Invitrogen) was used to extract total RNA of cells or tissues. Then, 1 μg total RNA was added to a 10μl final volume of mixed reagent to reverse transcription. Quantitative RT-PCR was performed with SYBR Green PCR Master Mix (Vazyme) using an ABI Prism 7900 sequence detection system (Applied Biosystems). U6 (Rnu6-1) was used as an endogenous control for miRNA, GAPDH was used as an internal control, and the results of each sample were normalized to U6 or GAPDH expression.

| Luciferase reporter assays
CircGSE1-wt/mut or TGFBR1-wt/mut were cotransfected with miR-324-5p/miR-744-5p mimics or control into T cells with Lipofectamine™ 3000 (Invitrogen). Cell lysates were collected 48 hours after transfection, and then the firefly and Renilla luciferase activities were measured using a dual-luciferase reporter gene detection kit (Promega) according to the manufacturer's protocol.

| Immunohistochemistry (IHC)
Immunohistochemistry staining was performed using the streptavidin-biotin-peroxidase complex method. In short, liver cancer tissue samples underwent several processes, including fixation, paraffin embedding, dewaxing, rehydration, and antigen retrieval.

| RNA fluorescence in situ hybridization (FISH)
T cells were first made into paraffin sections. After permeabilized in PBS with 0.5% Triton X-100, the cells were hybridized in hybridization buffer with Cy3-labeled circGSE1 probes at 37°C overnight.
The hybridization buffer was then gradually washed off with 4× SSC (including 0.1% Tween-20), 2× SSC, and 1× SSC at 42°C. Then, the nucleus was stained with DAPI for 10 minutes, and finally the results were observed with a TCS SP2 AOBS confocal laser microscope.

| Flow cytometry (FCM)
The antibodies used in FCM analysis included the following:

| RNA pull-down
The biotin-labeled circGSE1 probe/negative control probe was After about 4 weeks, the representative bioluminescence imaging of metastases was measured by a Xenogen IVIS 2000 luminal imager.

| Statistical analysis
All statistical analyses were performed in SPSS 25.0 software (SPSS Inc.). GraphPad Prism 8.0 (GraphPad Software) was used to generate graphs. Statistical significance for comparing two or more groups was determined by Student's t test or one-way ANOVA. When P value <0.05, the difference is considered statistically significant.

| HCC-derived exosomes enhance HCC proliferation, migration, and invasion
It is reported that exosomes derived from HCC are related to tumor progression. 19 In order to verify the mechanisms, we isolated HCC cell (Huh7 and HepG2)-derived exosomes from the supernatant.
Nanoparticle-tracking analysis showed that these exosomes had a double-layer membrane structure with a size of about 80-100 nm, which was in line with the size and morphology of common exosomes ( Figure 1A and B). The expression level of CD63 (exosomal marker) detected by IHC in liver cancer tissues was significantly higher than that in nontumor tissues ( Figure 1C). Cell-counting kit-8 showed exosomes significantly increased the proliferation of HCC cells ( Figure 1D). Additionally, wound-healing and transwell analysis showed that HCC-derived exosomes markedly enhanced the migration and invasion capabilities of HCC cells ( Figure 1E and F).
Western blot results showed that HCC-derived exosomes reduced the expression level of E-cadherin, an epithelial cell marker, while increased the expression level of N-cadherin and vimentin, which served as mesenchymal cell markers ( Figure 1G). To study the role of HCC-derived exosomes in HCC metastasis, mice were injected with control/exosome-treated HCC cells via tail vein. Then, using a live animal bioluminescence imaging system, it was found that exosomes significantly increased the metastasis of Huh7 ( Figure 1H). These results demonstrate that exosomes derived from HCC cells increase the ability of HCC to proliferate, migrate, and invade.

| HCC-derived exosomes enhance HCC proliferation, migration, and invasion via inducing Tregs expansion
Tregs in tumor-infiltrating lymphocytes regulate the immune es-  (Figure 2A). At first, IHC results showed that the expression of Tregs marker protein FOXP3 in liver cancer tissues was significantly higher than that in nontumor tissues ( Figure 2B).

| CircGSE1 is significantly upregulated in T cells with tumor-derived exosome addition
The data in the microarray GSE10 0206and GSE10 0207were used to explore several differentially expressed circRNAs in blood exosomes of HCC patients and normal people. On the basis of log2 (fold change) ≥ 1 and p < 0.05, the volcano plot shows the difference in circRNA expression ( Figure 3A). The heat map showed circGSE1 is the most upregulated circRNA in the exosomes from HCC patients' serum ( Figure 3B). CircGSE1 (hsa_circ_0000722, chr16:85667519-85667738) was derived from the protein-coding locus GSE1, which is generated by backsplicing of the second exon of the GSE1 gene with several Alu elements in introns on both sides ( Figure 3C and D).
Immediately afterward, qRT-PCR found that the expression level of circGSE1 in HCC tissues was significantly higher than that of adjacent nontumor tissues ( Figure 3E). Additionally, Kaplan-Meier survival analysis showed that the overall survival (OS) of patients with F I G U R E 4 circGSE1 acts as a sponge for miR-324-5p. A, Distribution of circGSE1 in T cells after treatment with exosomes detected by FISH. B, Quantitative RT-PCR analyses of miR-324-5p, miR-744-5p, and miR-138-5p in T cells transfected with control vector or circGSE1 plasmid. C, Schematic diagram of circGSE1 with the predicted binding site for miR-324-5p and miR-744-5p. D, Luciferase reporter gene assay was carried out to check the binding ability between miR-324-5p/miR-744-5p and circGSE1. Reporter constructs containing either circGSE1wt or circGSE1mut at the predicted miR-324-5p/miR-744-5p target sequences were cotransfected into T cells, along with miR-324-5p/miR-744-5p mimics or miR-NC, (Negative Control). E, Lysates prepared from circGSE1-overexpressing T cells were incubated with biotinylated probes for circGSE1, and then RNA pull-down assay was performed. The expression of circGSE1 and miR-324-5p was detected by qRT-PCR. F, Quantitative RT-PCR analyses of miR-324-5p expression in T cells after exosomes treatment (Exo-Huh7 and Exo-HepG2). G, Quantitative RT-PCR analyses of the expression of miR-324-5p in T cells treated with exosomes and circGSE1-knockdown exosomes. All data are means ± SD; n = 3, *p < 0.05, **p < 0.01, ***p < 0.001 higher circGSE1 expression levels was significantly shorter than that of patients with lower circGSE1 expression ( Figure 3F). Then, qRT-PCR results also showed circGSE1 was significantly upregulated in Huh7 and HepG2 cells compared with other cell lines ( Figure 3G). For further analysis, we explored the effect of exosomal circGSE1 derived from the serum of HCC patients and HCC cells on circGSE1 in T cells. Quantitative RT-PCR showed that after T cells were added with exosomes derived from serum, Huh7, and HepG2 cells, circGSE1 in T cells was markedly higher than that in the control group ( Figure 3H and I). To further verify that HCC-derived exosomes can carry circGSE1 into T cells, we transfected circGSE1 siRNA into T cells, and qRT-PCR assays validated that knockdown of circGSE1 significantly reduced the expression of circGSE1. However, after T cells were cocultured with Huh7-and HepG2-derived exosomes, the expression of circGSE1 in T cells increased significantly ( Figure 3J). In general, circGSE1 was significantly upregulated in T cells treated with tumorderived exosomes.

| circGSE1 acts as a sponge for miR-324-5p
Studies have found that circRNA is mainly present in the cytoplasm and produces a sponge effect by binding to miRNA. 20 FISH results showed that circGSE1 transcription signals were abundant in the cytoplasm of T cells, while there was little hybridization signal in the nucleus ( Figure 4A). To find potential miRNAs that can bind to circGSE1, StarBase 2.0 (http://starb ase.sysu.edu.cn/) and TargetScan (https://www.targe tscan.org/vert_72/) were used to predict the potential complementary miRNAs. MiR-324-5p, miR-744-5p, and miR-138-5p were predicted to bind to circGSE1.
Then, we found that the exosomes of HCC knocking down circGSE1 did not increase miR-324-5p expressions in T cells ( Figure 4G).
These results demonstrated that circGSE1 acts as a sponge for miR-324-5p.
The prediction results indicated that TGFBR1 is one of the most suitable target genes that can bind to miR-324-5p. Studies have shown that TGFBR1, as a receptor for TGFβ1, participates in signal transduction in immune cells, including T cells. 22 In order to explore the relationship between miR-324-5p and TGFBR1, a luciferase reporter gene was constructed and used ( Figure 5A). Dual-luciferase reporter assay results demonstrated miR-324-5p mimics significantly decreased the luciferase activity of wild-type TGFBR1 (TGFBR1-wt) rather than the luciferase activity of mutation TGFBR1 (TGFBR1-mt) ( Figure 5B). Western blot analysis also found that miR-324-5p inhibitor markedly induced the protein expression of TGFBR1 ( Figure 5C and D). Studies have shown that SMD3 can be used as an important substrate of TGFβ and participate in the transcriptional regulation of downstream genes. 15 Therefore, we first explored the protein ex- F I G U R E 5 miR-324-5p activates the TGFBR1/smad3 signaling pathway. Schematic diagram of TGFBR1 with the predicted binding site for miR-324-5p. B, Luciferase reporter analysis was carried out to examine the binding ability between miR-324-5p and TGFBR1. Reporter constructs containing either TGFBR1wt or TGFBR1mut at the predicted binding site for miR-324-5p were cotransfected into T cells with miR-324-5p or miR-NC mimic. C, D, Western blots and grey value of TGFBR1 expression in T cells after being transfected with miR-324-5p inhibitor or miR-NC. E, F, Western blots and grey value of TGFBR1, Smad3, and FOXP3 expression in T cells after being transfected with TGFBR1 siRNA. G, H, Western blots and grey value of TGFBR1, Smad3, and FOXP3 expression in T cells after treatment with Huh7-derived exosomes (Exo-Huh7), miR-324-5p mimics, or miR-NC. I, J, Western blots and grey value of TGFBR1, Smad3, and FOXP3 expression in T cells after treatment with HepG2-derived exosomes (Exo-HepG2), miR-324-5p mimics, or miR-NC. All data are means ± SD; n = 3, *p < 0.05, **p < 0.01, ***p < 0.001 F I G U R E 6 Hepatocellular carcinoma (HCC)-derived exosomal circGSE1 induces the expansion of Tregs via the miR-324-5p/TGFBR1/ Smad3 axis. A-C, Flow cytometry analysis of the ratio of Foxp3+ T cells in CD4 + T cells and the Foxp3 + CD4 + /CD8 + T cells ratio after indicated treatment. D, ELISA analysis of the IFNγ secretion after indicated treatment. All data are means ± SD; n = 3, *p < 0.05, **p < 0.01, ***p < 0.001 axis. The ratio of Tregs to CD8 + T cells is related to the prognosis and clinical efficacy of cancer patients. 24 Therefore, we further found that the trend of Treg/CD8 + T cells ratio is basically consistent with the change in the proportion of Tregs cells in CD4 + T cells ( Figure 6C).
Tregs shape the suppressive immune microenvironment by suppressing effector CD8 + T cells. 25 In order to understand whether Tregs affect the activity of CD8 + T cells, we used ELISA to detect the expression of IFNγ (cytokine secreted by CD8 + T cells) in the supernatant of T cells cocultured with HCC-derived exosomes. ELISA analysis results showed that the IFNγ secretion in the Tregs-amplified group was significantly lower than that in the Tregs-unamplified group ( Figure 6D), indicating that Tregs may exert tumor immunosuppressive function by suppressing the function of effector T cells.

| HCC-derived exosomal circGSE1 facilitates HCC progression in vitro and in vivo via inducing the expansion of Tregs
To

| DISCUSS ION
Studies have reported that circRNAs are abnormally expressed in various cancers, revealing that circRNA is an important molecule involved in tumor progression. 26,27 It is reported that cell-to-cell communication has the effect of promoting cancer immune escape. [28][29][30] The main mechanisms of signal exchange between cells include direct interaction, secreted biologically active molecules, and exosomes. 30,31 Studies have confirmed that exosomal circRNAs are abnormally expressed in the peripheral blood of patients with various cancers. 30,32,33 Notably, circRNA has been shown to be involved in regulating the immune evasion process of cancer. 26 In this study, we identified circGSE1 was highly expressed in HCC serum-derived exosomes. This suggests that circGSE1 may be involved in the occurrence and development of HCC, and it is necessary to further explore its mechanism.
Furthermore, by bioinformatics, luciferase reporter assays, and pull-down, circGSE1 was confirmed to directly bind with miR-324-5p. Moreover, we found that TGFBR1 was a direct common tar- Our results also showed that the trend of the ratio of Treg/CD8 + T cells is basically consistent with the change in the proportion of Tregs in CD4 + T cells, suggesting that Treg/CD8 + T ratio may be an indicator of the prognosis of HCC; further clinical trials are needed to explore this in the future. Additionally, ELISA analysis results showed that Tregs may exert tumor immunosuppressive function by suppressing the function of effector T cells. Of note, recent research also found that PD1 blockade can promote the expansion of Tregs. 34 It has been confirmed that the expression of PD1 on CD8 + T is related to CD8 + T cell apoptosis, which is also the mechanism by that PD1/PDL1 inhibitors work. 35 These results further illustrate that the Tregs/CD8+ T ratio may also predict the antitumor response of PD1/PDL1 inhibitors.

| CON CLUS IONS
In summary, we demonstrated for the first time that circGSE1 leads to immune escape of HCC by promoting the expansion of Tregs via regulating the miR-324-5p/TGFBR1/Smad3 axis. These findings shed light on the mechanism of exosomal circGSE1 in the progression of HCC and provide a promising biomarker for HCC immunotherapy.

ACK N OWLED G EM ENTS
This study was supported by grants from National Natural Science Foundation of China (No.82003040).

D I SCLOS U R E
The authors declare that they have no competing interests.