Abstract

   Immune checkpoint inhibitors (ICIs) show promise as second‐line
   treatment for advanced bladder cancer (BLCA); however, their
   responsiveness is limited by the immune evasion mechanisms in tumor
   cells. This study conduct a Cox regression analysis to screen
   mRNA‐binding proteins and reveals an association between Ras
   GTPase‐activating protein‐binding protein 1 (G3BP1) and diminished
   effectiveness of ICI therapy in patients with advanced BLCA. Subsequent
   investigation demonstrates that G3BP1 enhances immune evasion in BLCA
   cells by downregulating major histocompatibility complex class I
   (MHC‐I) through phosphoinositide 3‐kinase (PI3K)/Akt signaling
   activation. Mechanistically, G3BP1 interacts with splicing factor
   synergistic lethal with U5 snRNA 7 (SLU7) to form a complex with
   poly(A)‐binding protein cytoplasmic 1 and eukaryotic translation
   initiation factor 4 gamma 1. This complex stabilizes the closed‐loop
   structure of the mRNAs of class IA PI3Ks and consequently facilitates
   their translation and stabilization, thereby activating PI3K/Akt
   signaling to downregulate MHC‐I. Consistently, targeting G3BP1 with
   epigallocatechin gallate (EGCG) impedes immune evasion and sensitizes
   BLCA cells to anti‐programmed cell death (PD)‐1 antibodies in mice.
   Thus, G3BP1 and SLU7 collaboratively contribute to immune evasion in
   BLCA, indicating that EGCG is a precision therapeutic agent to enhance
   the effectiveness of anti‐PD‐1 therapy.

   Keywords: bladder cancer, epigallocatechin gallate, immune checkpoint
   inhibitors, immune evasion, major histocompatibility complex class I,
   Ras GTPase‐activating protein‐binding protein 1
     __________________________________________________________________

   Immune checkpoint inhibitors (ICIs) show promise as second‐line
   treatment for advanced bladder cancer (BLCA), but the immune evasion
   mechanisms in tumor cells hinder their effectiveness. This study links
   G3BP1 to diminished ICI efficacy in advanced BLCA patients. Explore the
   intricate role of G3BP1 in immune evasion, its collaboration with SLU7,
   and the potential of epigallocatechin gallate as a precision
   therapeutic.

   graphic file with name ADVS-11-2305922-g006.jpg

1. Introduction

   Bladder cancer (BLCA) is a prevalent malignancy worldwide, with an
   annual incidence of >500 000 cases and an estimated death toll of
   200 000.^[ [46]^1 ^] Muscle‐invasive BLCA (MIBC) is a malignant type of
   BLCA, with a 5‐year survival rate of only 50%.^[ [47]^2 ^]
   Cisplatin‐based chemotherapy remains the mainstay of treatment for
   unresectable and metastatic MIBCs, but most patients tend to relapse.
   Moreover, the efficacy of traditional second‐line treatment options,
   such as paclitaxel, is not promising. Over the past decade, immune
   checkpoint inhibitors (ICIs), such as monoclonal antibodies targeting
   programmed cell death‐1 (PD‐1) and programmed cell death‐ligand 1
   (PD‐L1), have been approved for patients with advanced BLCA. Although
   ICI represents an important breakthrough, the response rate in BLCA
   remains quite limited at <30%.^[ [48]^3 ^]

   Immunosuppressive tumor microenvironment (TME) is a critical cause of
   tumor immune evasion and immunotherapy insensitivity, as evidenced by
   the exhaustion and hypoinfiltration of cytotoxic tumor‐infiltrating
   lymphocytes (TILs;, e.g., cluster of differentiation 8 [CD8]+ T
   cells).^[ [49]^4 ^] The first step in the recognition of tumor cells by
   cytotoxic CD8+ T cells is the binding of T cell receptor (TCR) to
   antigens presented on major histocompatibility complex class I
   (MHC‐I).^[ [50]^5 ^] However, most tumor cells, including BLCA cells,
   have low MHC‐I expression.^[ [51]^6 ^] Dysregulation of MHC‐I and other
   immune checkpoint signaling has enabled tumor cells to evade immune
   surveillance, thereby creating an immunosuppressive TME.^[ [52]^7 ^]
   Therefore, identifying the mechanisms underlying dysregulated immune
   checkpoint signaling in tumor cells may provide therapeutic targets to
   enhance the efficacy of ICIs for patients with BLCA.

   The physiological functions of somatic cells depend on the homeostasis
   of mRNA turnover.^[ [53]^8 ^] An imbalance in mRNA turnover has been
   linked to several diseases, such as cancer and immune disorders.^[
   [54]^9 ^] RNA‐binding proteins (RBPs) bind to specific RNAs and direct
   their life events, such as maturation, subcellular localization,
   modification, translation, and decay.^[ [55]^10 ^] mRNA‐binding
   proteins (mRBPs) are a major class of RBPs, and different mRBPs can
   bind to and regulate specific mRNAs through their unique RNA‐binding
   domains (RBDs).^[ [56]^10b ^] Recent studies have revealed that mRBP
   dysregulation is closely associated with tumor progression.^[ [57]^11
   ^] However, the mechanism by which mRBPs regulate BLCA immune evasion
   remains unclear. Therefore, we performed a Cox regression analysis to
   screen mRBPs that are potentially associated with immune evasion in
   patients who participated in the IMvigor210 cohort study,^[ [58]^12 ^]
   which investigated the efficacy of anti‐PD‐L1 antibody in patients with
   advanced BLCA. Among the top‐hit mRBPs, Ras GTPase‐activating
   protein‐binding protein 1 (G3BP1), an mRBP known for assembling stress
   granules,^[ [59]^13 ^] attracted our interest because of its unknown
   implication in tumor immune evasion.

   Herein, we revealed that G3BP1 interacts with splicing factor
   synergistic lethal with U5 snRNA 7 (SLU7), an RBP known to regulate
   pre‐mRNA splicing and liver homeostasis,^[ [60]^14 ^] to stabilize the
   closed‐loop structure of the mRNAs of class IA PI3Ks, thereby
   facilitating their translation and stabilization. This leads to the
   activation of phosphoinositide 3‐kinase (PI3K)/Akt signaling, which in
   turn downregulate MHC‐I to promote BLCA immune evasion. Therefore,
   targeting G3BP1 with epigallocatechin gallate (EGCG) impedes immune
   evasion and thus sensitizes BLCA to anti‐PD‐1 antibodies.

2. Results

2.1. G3BP1 Promotes Immune Evasion by Downregulating MHC‐I in BLCA Cells

   To investigate the role of mRBPs in immune evasion, we analyzed the
   effect of mRBPs on the survival of patients from the IMvigor210
   cohort.^[ [61]^12 ^] Several top‐hit mRBPs demonstrating a positive
   correlation with the hazard ratio (Figure [62] 1A) have been identified
   as significant contributors to tumor immune evasion (e.g., SND1,^[
   [63]^15 ^] EIF3B,^[ [64]^16 ^] and RBMS1^[ [65]^17 ^]) as well as BLCA
   progression (e.g., YBX3,^[ [66]^18 ^] EIF4B,^[ [67]^19 ^] and AFF4^[
   [68]^20 ^]). This analysis emphasizes the crucial role of mRBPs in
   facilitating tumor immune evasion. Among other mRBPs, G3BP1 attracted
   our interest due to its unexplored role in tumor immune evasion and
   BLCA progression. Moreover, high G3BP1 expression was associated with
   unfavorable outcomes in patients from the IMvigor210 cohort
   (Figure [69]1B). G3BP1 knockdown markedly promoted the expression of
   HLA‐A, HLA‐B, and HLA‐C, while slightly promoting the expression of
   CD80, CD86, CD276, B7‐H4, PD‐L2, and GITRL in UMUC3 cells (Figure
   [70]S1, Supporting Information). As HLA‐A, HLA‐B, and HLA‐C are major
   components of MHC‐I, we reasoned that G3BP1 may suppress MHC‐I
   expression to promote immune evasion in BLCA cells. This phenomenon was
   confirmed by the finding that G3BP1 knockdown markedly promoted MHC‐I
   expression in both UMUC3 and T24 cells (Figure [71]1C). Likewise, G3bp1
   knockdown increased the cell surface MHC‐I expression in murine BLCA
   cell line MB49, as demonstrated via flow cytometry (Figure [72]1D,E).

Figure 1.

   Figure 1
   [73]Open in a new tab

   G3BP1 knockdown in BLCA cells upregulates MHC‐I and impedes immune
   evasion. A) The effect of mRBPs on the survival of patients with BLCA
   who received anti‐PD‐L1 antibody therapy was determined using
   Kaplan–Meier and Cox regression analyses. The top‐hit mRBPs in terms of
   positive correlation with the hazard ratio are presented. The mRNA
   expression data were derived from the IMvigor210 cohort. B)
   Kaplan–Meier curve constructed using the best cutoff expression of
   G3BP1 in the IMvigor210 cohort. C) Western blotting was used to detect
   the protein expression levels of G3BP1 and MHC‐I in human BLCA cell
   lines UMUC3 and T24 with stable G3BP1 knockdown. GAPDH was used as the
   internal control. D) Western blotting was used to detect the protein
   expression levels of G3bp1 in the murine BLCA cell line MB49 with
   stable G3bp1 knockdown. E) Flow cytometry was used to analyze the
   expression of MHC‐I on the surface of MB49 cells with stable G3bp1
   knockdown. MHC‐I expression was quantified based on the mean
   fluorescence intensity (MFI). F) mCherry‐expressing MB49‐OVA cells with
   G3bp1 knockdown were cocultured with antigen‐activated CTLs for 24 h.
   Cytotoxicity was determined by analyzing the percentage of 7‐AAD+ cells
   among mCherry+ cells. G) Endpoint image of tumors from C57BL/6J mice
   bearing subcutaneous bladder tumors (implanted with 3 × 10^5 MB49 cells
   with stable G3bp1 knockdown). Scale bar, 1 cm. H) Tumor volumes on the
   indicated days after inoculation with MB49 cells. I) Tumors were
   weighed. J) Immunofluorescent staining of CD8 in MB49 tumors, and
   quantification of CD8 signals. Scale bar, 50 µm. Error bars represent
   standard deviation (SD) (n = 3 in E and F; n = 6 in G–J). P‐values are
   presented and were calculated using the log‐rank test (B), one‐way
   analysis of variance (ANOVA) (E, F, I, and J), and two‐way ANOVA (H).

   Tumor‐cell surface MHC‐I is essential for the recognition of tumor
   cells by cytotoxic T lymphocytes (CTLs);^[ [74]^5 ^] therefore, we
   assessed the effect of G3BP1 on immune evasion. The in vitro T cell
   killing assay revealed that G3bp1 knockdown in MB49‐ovalbumin (OVA)
   cells increased the killing capacity of OT1CD8^+ T cells
   (Figure [75]1F). Moreover, in vivo G3bp1 knockdown in MB49 cells
   decreased the tumor volume and weight and enhanced the infiltration of
   CD8+ T cells into tumor (Figure [76]1G–J). These results indicate that
   G3BP1 downregulates MHC‐I in BLCA cells to promote immune evasion.

2.2. G3BP1 Downregulates MHC‐I to Promote Immune Evasion by Activating
PI3K/Akt Signaling in BLCA Cells

   To determine the mechanism by which G3BP1 inhibits MHC‐I expression, we
   performed RNA sequencing (RNA‐seq) in UMUC3 cells with G3BP1 knockdown.
   Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment
   analysis and gene set enrichment analysis (GSEA) demonstrated that
   G3BP1 knockdown inhibited the PI3K/Akt signaling pathway (Figure
   [77]S2, Supporting Information; Figure [78] 2A), which is known to
   repress tumor immunogenicity by downregulating MHC‐I.^[ [79]^21 ^]
   Indeed, G3BP1 knockdown decreased p‐Akt^Thr308 and p‐Akt^Ser473
   expression (PI3K effectors), but not total Akt expression, in UMUC3,
   T24, and MB49 cells (Figure [80]2B–D). Consistently, G3BP1
   overexpression in both human and murine BLCA cells upregulated
   p‐Akt^Thr308 and p‐Akt^Ser473 and downregulated MHC‐I
   (Figure [81]2E–G). Further, G3bp1 overexpression in MB49‐OVA cells
   decreased the killing capacity of OT1CD8^+ T cells (Figure [82]2H). As
   expected, Akt inhibition by MK‐2206 upregulated MHC‐I in both UMUC3 and
   MB49 cells and impeded immune evasion in MB49 cells (Figure [83]S3,
   Supporting Information). Moreover, MK‐2206 restored the effects of
   G3BP1 overexpression on p‐Akt^Thr308, p‐Akt^Ser473, and MHC‐I
   expression as well as the killing capacity of OT1CD8^+ T cells in BLCA
   cells (Figure [84]2I–L). These results indicate that G3BP1
   downregulates MHC‐I to promote immune evasion by activating PI3K/Akt
   signaling in BLCA cells.

Figure 2.

   Figure 2
   [85]Open in a new tab

   G3BP1 downregulates MHC‐I and promotes immune evasion by activating
   PI3K/Akt. A) UMUC3 cells with G3BP1 knockdown were subjected to RNA‐seq
   analysis. GSEA plot showing the correlation between G3BP1 expression
   and the PI3K signaling pathway based on RNA‐seq data. B–D) Western
   blotting was used to detect the protein expression levels of G3BP1,
   total Akt, p‐Akt^Thr308, and p‐Akt^Ser473 in UMUC3, T24, and MB49 cells
   with stable G3BP1 knockdown. GAPDH was used as the internal control. E
   and F) Western blotting was used to detect the protein expression
   levels of G3BP1, total Akt, p‐Akt^Thr308, p‐Akt^Ser473, and/or MHC‐I in
   UMUC3, T24, and MB49 cells with stable G3BP1 overexpression. G) Flow
   cytometry was used to analyze the expression of MHC‐I on the surface of
   MB49 cells in (F). MHC‐I expression was quantified based on the mean
   fluorescence intensity (MFI). H) mCherry‐expressing MB49‐OVA cells with
   G3bp1 overexpression were cocultured with antigen‐activated CTLs for
   24 h. Cytotoxicity was determined by analyzing the percentage of 7‐AAD+
   cells among mCherry+ cells. I,J) G3BP1‐overexpressing UMUC3 and MB49
   cells were treated with or without 1.0 µm MK‐2206 for 24 h, and western
   blotting was used to determine the protein expression levels of G3BP1,
   total Akt, p‐Akt^Thr308, p‐Akt^Ser473, and/or MHC‐I. K) Flow cytometry
   was used to analyze MHC‐I expression on the surface of MB49 cells in
   (J). MHC‐I expression was quantified based on MFI. L)
   mCherry‐expressing MB49‐OVA cells with or without G3bp1 overexpression
   and MK‐2206 treatment were cocultured with antigen‐activated CTLs for
   24 h. Cytotoxicity was determined by analyzing the percentage of 7‐AAD+
   cells among mCherry+ cells. Error bars represent SD (n = 3 in G, H, K,
   and L). P‐values are presented and were calculated using unpaired,
   two‐tailed Student's t‐test (G and H), and one‐way ANOVA (K and L).

2.3. G3BP1 Increases Both mRNA and Protein Levels of Class IA PI3Ks via
Binding to Their mRNAs in BLCA Cells

   To investigate the mechanism by which G3BP1 activates PI3K/Akt
   signaling, we performed RNA immunoprecipitation sequencing (RIP‐seq)
   using UMUC3 cells that stably express HA‐G3BP1. Notably, we revealed
   that G3BP1 binds to the mRNAs of class IA PI3Ks (PIK3CA, PIK3CB,
   PIK3CD, and PIK3R1) through multiple motifs (Figure [86] 3A; Figure
   [87]S4, Supporting Information), which were verified via the RIP assay
   (Figure [88]3B). The colocalization of G3BP1 with the mRNA of PIK3CA,
   PIK3CB, PIK3CD, or PIK3R1 in UMUC3 cells was visualized via
   immunofluorescence in situ hybridization (immuno‐FISH) using an
   antibody against G3BP1 and specific probes for PIK3CA, PIK3CB, PIK3CD
   or PIK3R1 mRNA (Figure [89]3C). Furthermore, we revealed that G3BP1
   knockdown and overexpression decreased and increased both mRNA and
   protein levels of PIK3CA, PIK3CB, PIK3CD, and PIK3R1 in cells,
   respectively (Figure [90]3D–G). These results indicate that G3BP1 binds
   to the mRNAs of class IA PI3Ks to increase their mRNA and protein
   levels.

Figure 3.

   Figure 3
   [91]Open in a new tab

   G3BP1 binds to the mRNAs of class IA PI3Ks and promotes their
   expression. A) Cell lysates from UMUC3 cells with stable expression of
   HA‐G3BP1 were immunoprecipitated with anti‐HA antibody, and the
   purified RNA from the immunoprecipitants was subjected to RNA‐seq
   analysis. Motif spacing analysis was used to identify the significant
   RNA motifs of HA‐G3BP1 on RNAs. B) The RIP assay with anti‐HA
   antibodies was performed using UMUC3 cells with stable expression of
   HA‐G3BP1. qRT–PCR analysis was used to evaluate the mRNA levels of
   class IA PI3Ks in the immunoprecipitants. C) The locations of G3BP1 and
   the mRNAs of class IA PI3Ks in UMUC3 cells were analyzed using
   immuno‐FISH. Scale bar, 10 µm. The fluorescence intensity curves of
   DAPI (blue), G3BP1 (green), and mRNA (purple) that crossed the arrows
   were analyzed. D and E) qRT–PCR analysis was used to evaluate the mRNA
   levels of G3BP1 and the class IA PI3Ks in UMUC3 and T24 cells with
   G3BP1 knockdown or overexpression. F,G) Western blotting was used to
   detect the protein expression levels of class IA PI3Ks in UMUC3 and T24
   cells with G3BP1 knockdown or overexpression. GAPDH was used as the
   internal control. Error bars represent SD (n = 3 in B, D, and E).
   P‐values are presented and were calculated using unpaired two‐tailed
   Student's t‐test (B) and two‐way ANOVA (D and E).

2.4. SLU7 Interacts with G3BP1 and is Required for the G3BP1‐Dependent
Regulation of PI3K/Akt/MHC‐I Signaling

   G3BP1 regulates downstream signaling by interacting with different
   proteins, especially RBPs;^[ [92]^22 ^] therefore, we hypothesized that
   G3BP1 similarly activates PI3K/Akt signaling. Accordingly, we used
   coimmunoprecipitation (co‐IP) and mass spectrometry (MS) to identify
   G3BP1‐interacting proteins in UMUC3 cells (Figure [93]S5A, Supporting
   Information). We identified 213 RBPs that interacted with G3BP1 (Figure
   [94]S5B, Supporting Information). SLU7, one of the top hits among these
   G3BP1‐interacting RBPs (Figure [95]S5C, Supporting Information),
   attracted our attention. This is because the overpression of SLU7, but
   not other top‐hit RBPs, increased p‐Akt^Ser473 expression in UMUC3
   cells (Figure [96]S5D,E, Supporting Information). Similar to G3BP1,
   high SLU7 expression was associated with unfavorable outcomes in
   patients from the IMvigor210 cohort (Figure [97]S6, Supporting
   Information). Indeed, the binding of SLU7 to mRNAs of class IA PI3Ks
   was verified via the RIP assay using UMUC3 cells that express SLU7‐Flag
   (Figure [98]S7, Supporting Information). Further, SLU7 knockdown and
   overexpression decreased and increased both mRNA and protein levels of
   class IA PI3Ks, respectively (Figure [99] 4A–D). These results indicate
   that SLU7 increases both mRNA and protein levels of class IA PI3Ks by
   binding to their mRNAs in BLCA cells.

Figure 4.

   Figure 4
   [100]Open in a new tab

   G3BP1 interacts with SLU7 to promote PI3K/Akt signaling activation,
   downregulating MHC‐I and promoting immune evasion. A,B) qRT–PCR
   analysis was used to evaluate the mRNA levels of SLU7 and the class IA
   PI3Ks in UMUC3 and T24 cells with SLU7 knockdown or overexpression.
   C,D) Western blotting was used to detect the protein expression levels
   of class IA PI3Ks in UMUC3 and T24 cells with SLU7 knockdown or
   overexpression. E) Cell lysates from UMUC3 cells with stable expression
   of SLU7‐Flag or HA‐G3BP1 were immunoprecipitated with anti‐Flag or
   anti‐HA antibodies, and immunoblotting was conducted with the indicated
   antibodies. F) The PLA was used to evaluate the interaction between
   G3BP1 and SLU7 in UMUC3 cells. Scale bar, 10 µm. G) IF was used to
   analyze the locations of G3BP1 and SLU7 in UMUC3 cells. Scale bar,
   10 µm. Fluorescence intensity curves of DAPI (blue), G3BP1 (green), and
   SLU7 (red) that crossed the arrow were analyzed. H,I) The protein
   expression levels of SLU7, total Akt, p‐Akt^Thr308, and p‐Akt^Ser473 in
   UMUC3 and MB49 cells with stable SLU7 knockdown were detected using
   western blotting. J,K) Western blotting was used to detect the protein
   expression levels of G3BP1, SLU7, total Akt, p‐Akt^Thr308,
   p‐Akt^Ser473, and/or MHC‐I in G3BP1‐overexpressing UMUC3 and MB49 cells
   with SLU7 knockdown. L) Flow cytometry was used to analyze the
   expression of MHC‐I on the surface of MB49 cells in (K). MHC‐I
   expression was quantified based on the mean fluorescence intensity
   (MFI). M) mCherry‐expressing MB49‐OVA cells with or without G3bp1
   overexpression and Slu7 knockdown were cocultured with
   antigen‐activated CTLs for 24 h. Cytotoxicity was determined by
   analyzing the percentage of 7‐AAD+ cells among mCherry+ cells. Error
   bars represent SD (n = 3 in A, B, L, and M). P‐values are presented and
   were calculated using two‐way ANOVA (A and B) and one‐way ANOVA (L and
   M).

   Moreover, ectopic SLU7 and G3BP1 formed a complex with endogenous G3BP1
   and SLU7 in cells, respectively (Figure [101]4E; Figure [102]S8A,
   Supporting Information), and the proximity ligation assay (PLA) and
   immunofluorescence (IF) assay using UMUC3 and T24 cells were performed
   to confirm the interaction of endogenous G3BP1 and SLU7
   (Figure [103]4F,G; Figure [104]S8B,C, Supporting Information). SLU7
   knockdown decreased the expression of p‐Akt^Thr308 and p‐Akt^Ser473
   (Figure [105]4H,I) and impaired the effect of G3BP1 overexpression on
   p‐Akt^Thr308, p‐Akt^Ser473, and MHC‐I expression as well as the killing
   capacity of OT1CD8^+ T cells (Figure [106]4J–M). These results revealed
   that SLU7 may interact with G3BP1 and is required for the
   G3BP1‐dependent regulation of the PI3K/Akt/MHC‐I axis in BLCA cells to
   promote immune evasion.

   Interestingly, G3BP1 and SLU7 knockdown substantially reduced the
   binding of SLU7 and G3BP1 to the mRNAs of class IA PI3Ks in cells,
   respectively, as demonstrated via the RIP assay (Figure [107]S9A–D,
   Supporting Information). This was further verified via immuno‐FISH
   analysis, revealing that G3BP1 and SLU7 knockdown disrupted the
   colocalization of SLU7 and G3BP1, respectively, with the mRNAs of class
   IA PI3Ks (Figure [108]S9E, Supporting Information). Taken together,
   these findings indicate that the G3BP1‐SLU7 complex binds to the mRNAs
   of class IA PI3Ks in a G3BP1‐ and SLU7‐dependent manner.

2.5. G3BP1 Interacts with SLU7 to Stabilize the Poly(A) Binding Protein
Cytoplasmic 1 (PABPC1)‐Mediated Closed‐Loop Structure of the mRNAs of Class
IA PI3Ks

   The G3BP1 interactome was established to elucidate the mechanism by
   which the G3BP1‐SLU7 complex affects the mRNAs of class IA PI3Ks
   (Figure [109] 5A). Notably, in addition to SLU7, PABPCs and various
   eukaryotic transcription initiation factors (e.g., eIF3s and eIF4s),
   which are known to constitute the closed‐loop structure of mRNA,^[
   [110]^10 , [111]^23 ^] were listed (Figure [112]5A). We then
   hypothesized the involvement of the G3BP1‐SLU7 complex in stabilizing
   the PABPC‐mediated closed‐loop structure because G3BP1 frequently binds
   to the 3′‐UTR (Figure [113]S4, Supporting Information). Indeed, both
   G3BP1 and SLU7 interacted with PABPC1 and eukaryotic translation
   initiation factor 4 gamma 1 (eIF4G1) (Figure [114]5B,C). These proteins
   formed a complex, as demonstrated by the reimmunoprecipitation (re‐IP)
   assay (Figure [115]5D), and PABPC1‐GFP or eIF4G1‐GFP colocalized with
   both G3BP1 and SLU7 in UMUC3 cells, as demonstrated via the IF assay
   (Figure [116]5E). The interaction of G3BP1‐SLU7 with PABPC1 and eIF4G1
   is dependent on G3BP1, but not on the RBDs of G3BP1 or SLU7, in cells
   (Figures [117]S10 and [118]S11, Supporting Information). Either G3BP1
   or SLU7 knockdown diminishes the binding of PABPC1 to mRNAs of class IA
   PI3Ks, but not to eIF4G1, in cells (Figure [119]5F,G). These results
   indicate that both G3BP1 and SLU7 form a complex with PABPC1 and eIF4G1
   to potentially stabilize the closed‐loop structure of the mRNAs of
   class IA PI3Ks.

Figure 5.

   Figure 5
   [120]Open in a new tab

   G3BP1 and SLU7 form a complex with PABPC1 and eIF4G1 to stabilize the
   closed‐loop structure of the mRNAs of class IA PI3Ks. A) Scatter plot
   of the G3BP1‐interacting proteins identified using MS. B and C) Cell
   lysates from UMUC3 cells with stable expression of SLU7‐Flag or
   HA‐G3BP1 were immunoprecipitated with anti‐Flag or anti‐HA antibodies,
   and immunoblotting was performed with the indicated antibodies. D) Cell
   lysates from UMUC3 cells with simultaneous stable expression of
   SLU7‐Flag and HA‐G3BP1 were primarily immunoprecipitated with anti‐Flag
   antibody and then blocked with Flag peptide. The resultant
   immunoprecipitants were subjected to re‐IP with anti‐HA antibodies and
   immunoblotted with the indicated antibodies. E) IF was used to analyze
   the locations of eIF4G1 or PABPC1 with SLU7 and G3BP1 in UMUC3 cells
   stably expressing eIF4G1‐GFP or PABPC1‐GFP. Scale bar, 10 µm. The
   fluorescence intensity curves of DAPI (blue), eIF4G1‐GFP (green),
   PABPC1‐GFP (green), SLU7 (red), and G3BP1 (purple) that crossed the
   arrows were analyzed. F) Cell lysates from UMUC3 cells with PABPC1‐GFP
   stable expression and SLU7 or G3BP1 knockdown were immunoprecipitated
   with anti‐GFP antibody. Western blotting was used to detect the
   expression levels of PABPC1‐GFP, SLU7, G3BP1, and eIF4G1 in whole cell
   lysates and immunoprecipitants. G) qRT–PCR analysis was used to
   evaluate the mRNA expression levels of class IA PI3Ks in the
   immunoprecipitants. H) Western blotting was used to detect the
   expression levels of G3BP1 and SLU7 in UMUC3 cells with SLU7 knockdown
   or G3BP1 knockdown. I) FISH–FRET was performed in UMUC3 cells with SLU7
   or G3BP1 knockdown using Cy3‐labeled 5′‐UTR probes and Cy5‐labeled
   3′‐UTR probes targeting the mRNAs of PI3Ks. The mean values of FRET
   efficiency are shown. Error bars represent SD (n = 3 in G; n = 5 in I).
   P‐values are presented and were calculated using one‐way ANOVA (G and
   I).

   The direct method to visualize the closed‐loop structure of mRNA in
   situ has not been well established. Therefore, we set up a
   FISH–fluorescence resonance energy transfer (FRET) method by labeling
   the cap and tail of mRNA using Cy3‐conjugated donor primers and
   Cy5‐conjugated receptor primers, respectively, and evaluated the
   closed‐loop structure according to the FRET efficiency between Cy3 and
   Cy5 (Figure [121]S12A, Supporting Information). This FISH–FRET method
   was proven to be effective as knockdown of PABPC1, which is
   indispensable in the closed‐loop structure of mRNAs,^[ [122]^10a ^]
   decreased the FRET efficiency of the paired dyes labeled in the mRNAs
   of class IA PI3Ks (Figure [123]S12B,C, Supporting Information),
   indicating that PABPC1 knockdown impaired the closed‐loop structure of
   the mRNAs of class IA PI3Ks. Importantly, the FRET efficiency on the
   mRNAs of class IA PI3Ks decreased in UMUC3 cells with either G3BP1 or
   SLU7 knockdown (Figure [124]5H,I), indicating that both G3BP1 and SLU7
   are required for stabilizing the closed‐loop structure of the mRNAs of
   class IA PI3Ks.

2.6. G3BP1 and SLU7 Facilitate the Translation and mRNA Stabilization of
Class IA PI3Ks

   The closed‐loop structure of mRNAs is crucial for maintaining
   translation, as evidenced by the fact that mRNAs are usually
   accompanied with enriched polysomes.^[ [125]^10 , [126]^24 ^]
   Therefore, we hypothesized that G3BP1 and SLU7 affect the translational
   activity of class IA PI3Ks (Figure [127] 6A). Polysome profiling
   analysis revealed that either G3BP1 or SLU7 knockdown reduced the
   proportion of class IA PI3K mRNAs distributed to polysomes in cells
   (Figure [128]6B–D), indicating that G3BP1 and SLU7 are critical for the
   translation of class IA PI3Ks.

Figure 6.

   Figure 6
   [129]Open in a new tab

   G3BP1 and SLU7 are essential for the translation and mRNA stabilization
   of class IA PI3Ks. A) A graphical model illustrates the mechanism by
   which G3BP1‐SLU7 stabilizes the closed‐loop structure and ribosome
   binding of the mRNAs of class IA PI3Ks. B) A graphical model
   illustrating the sucrose gradient distribution of subunits of ribosomes
   and polysomes. C) Lysates from UMUC3 cells with or without G3BP1 or
   SLU7 knockdown were separated using 10%–50% sucrose gradients. Polysome
   profiles of the gradients were generated by measuring absorbance at
   254 nm. D) qRT–PCR analysis was used to evaluate the mRNA levels of
   class IA PI3Ks in each sucrose fraction. The mRNA distribution in each
   sucrose fraction was determined by normalizing the mRNA expression to
   that of GAPDH. E,F) UMUC3 cells with G3BP1 or SLU7 knockdown were
   treated with actinomycin D, and total RNA was extracted at the
   indicated time points. qRT–PCR analysis was used to evaluate the mRNA
   levels in each treatment group. The expression of mRNAs remaining at
   different time points was determined by normalizing the mRNA expression
   to that detected at 0 h. Error bars represent SD (n = 3 in E and F).
   P‐values are presented and were calculated using two‐way ANOVA (E and
   F).

   Eukaryotic cells possess well‐established mechanisms of mRNA
   metabolism, and mRNAs with incomplete closed‐loop structures are
   usually susceptible to degradation.^[ [130]^10a ^] Therefore, we
   examined the effect of G3BP1 or SLU7 on the mRNA stability of class IA
   PI3Ks. G3BP1 or SLU7 knockdown markedly decreased the mRNA stability of
   PIK3CA, PIK3CB, PIK3CD, and PIK3R1 (Figure [131]6E,F). These results
   indicate that G3BP1 and SLU7 stabilize the mRNAs of class IA PI3Ks.

2.7. High G3BP1 and SLU7 Expressions Correlate with BLCA Progression and
Unfavorable Outcomes in Patients with BLCA

   Next, we assessed the clinical relevance of both G3BP1 and SLU7 using
   BLCA samples. The protein levels of G3BP1 and SLU7 demonstrated a
   positive correlation, and higher protein levels of G3BP1 or SLU7 were
   associated with higher grades of BLCA, indicative of an unfavorable
   outcome in patients with BLCA (Figure [132] 7A–E). Notably, patients
   with BLCA showing high or low protein levels of both G3BP1 and SLU7
   demonstrated the worst or best outcomes, respectively (Figure [133]7F).
   Altogether, these results indicate that high protein levels of G3BP1
   and SLU7 correlate with BLCA progression and unfavorable outcomes in
   patients with BLCA.

Figure 7.

   Figure 7
   [134]Open in a new tab

   Higher expression levels of G3BP1 and SLU7 correlate with BLCA
   progression. A,B) IHC was used to detect the expression levels of G3BP1
   and SLU7 in BLCA tumors. Images representing the staining of G3BP1 and
   SLU7 from low‐grade and high‐grade tumors. Scale bar, 50 µm. The
   H‐scores of G3BP1 and SLU7 were quantified. Error bars represent SD (n
   = 22 in “Low grade”; n = 122 in “High grade”). P‐values were calculated
   using unpaired two‐tailed Student's t‐test. C) Pearson's correlation
   test was used to analyze correlations between the H‐scores of G3BP1 and
   SLU7 in BLCA tumors; n = 151. D–F) Kaplan–Meier analysis and log‐rank
   test were used to evaluate the impact of G3BP1 and SLU7 on the survival
   of patients with BLCA, with the corresponding P‐values. The G3BP1 and
   SLU7 expression data used for survival analysis comprised the H‐scores
   described in A and B; n = 74 in “G3BP1 low;” n = 77 in “G3BP1 high;” n
   = 74 in “SLU7 low;” n = 77 in “SLU7 high;” n = 44 in “SLU7 low, G3BP1
   low;” n = 30 in “SLU7 low, G3BP1 high;” n = 30 in “SLU7 high, G3BP1
   low;” and n = 47 in “SLU7 high, G3BP1 high.”.

2.8. Targeting G3BP1 with EGCG Impedes Immune Evasion and Increases the
Sensitivity of Anti‐PD‐1 Antibody

   EGCG, a polyphenol isolated from tea leaves, inhibits G3BP1 by binding
   to endogenous G3BP1 in tumor cells.^[ [135]^25 ^] Therefore, we
   investigated the potential use of EGCG to treat patients with BLCA, as
   EGCG has previously been used in patients with other cancer types.^[
   [136]^26 ^] As expected, EGCG downregulated p‐Akt^Thr308 and
   p‐Akt^Ser473 and upregulated MHC‐I in a dose‐dependent manner in both
   UMUC3 and MB49 cells (Figure [137] 8A–C). It also increased the killing
   capacity of OT1CD8^+ T cells in MB49‐OVA cells (Figure [138]8D). More
   importantly, in the MB49 tumor‐bearing immunocompetent C57BL/6J mouse
   model (Figure [139]S13A, Supporting Information), the combination of
   EGCG with anti‐PD‐1 antibody was more efficient in inhibiting tumor
   growth and increasing the infiltration of CD8+ T cells into tumor than
   either EGCG or anti‐PD‐1 antibody alone (Figure [140]8E–H; Figure
   [141]S13B, Supporting Information). Collectively, these results
   indicate that EGCG promotes MHC‐I expression to impede immune evasion
   and increase the sensitivity of anti‐PD‐1 antibodies by inhibiting
   G3BP1‐SLU7/PI3K/Akt signaling.

Figure 8.

   Figure 8
   [142]Open in a new tab

   Targeting G3BP1 with EGCG impedes immune evasion and increases the
   sensitivity of anti‐PD‐1 antibody. A,B) UMUC3 and MB49 cells were
   treated with EGCG with the indicated concentrations for 48 h, and
   western blotting was used to detect the expression levels of G3BP1,
   total Akt, p‐Akt^Thr308, p‐Akt^Ser473, and/or MHC‐I. C) Flow cytometry
   was used to analyze the expression of MHC‐I on the surface of MB49
   cells in (B). MHC‐I expression was quantified based on the mean
   fluorescence intensity (MFI). D) mCherry‐expressing MB49‐OVA cells with
   or without EGCG treatment were cocultured with or without
   antigen‐activated CTLs for 24 h. Cytotoxicity was determined by
   analyzing the percentage of 7‐AAD+ cells among mCherry+ cells. E–H)
   C57BL/6J mice bearing subcutaneous bladder tumors (implanted with 3 ×
   10^5 MB49 cells) received EGCG, vehicle, IgG, or anti‐PD‐1 antibody
   therapy. E) Endpoint image of MB49 tumors from mice that were treated
   as indicated. Scale bar, 1 cm. F) Tumor volumes on the indicated days
   after inoculation with MB49 cells. G) Tumors were weighed. H)
   Immunofluorescent staining of CD8 in MB49 tumors, and quantification of
   CD8 signals. Scale bar, 50 µm. Error bars represent SD (n = 3 in C and
   D; n = 6 in E–H). P‐values are presented and were calculated using
   one‐way ANOVA (C, D, G, and H) and two‐way ANOVA (F).

3. Discussion

   The dysregulated immune checkpoint signaling in tumor cells is a
   critical factor in the low responsiveness of ICIs for advanced BLCA.
   However, the mechanisms underlying immune checkpoint signaling
   dysregulation in tumor cells remain largely unknown. Herein, we
   revealed that G3BP1 downregulates MHC‐I to promote BLCA immune evasion
   by activating PI3K/Akt signaling. G3BP1 interacts with SLU7 to
   stabilize the closed‐loop structure of the mRNAs of class IA PI3Ks,
   thereby facilitating their translation and stabilization. Therefore,
   targeting G3BP1 with EGCG enhances the efficacy of anti‐PD‐1 antibody
   for BLCA (Figure [143] 9 ).

Figure 9.

   Figure 9
   [144]Open in a new tab

   The proposed model in this study. The G3BP1‐SLU7 complex interacts with
   PABPC1 and eIF4G1 to stabilize the closed‐loop structure of the mRNAs
   of class IA PI3Ks, thereby facilitating their translation and
   stabilization and activating the PI3K/Akt axis to downregulate MHC‐I
   and promote immune evasion. In contrast, targeting G3BP1 with EGCG
   disrupts the PABPC1‐mediated closed‐loop structure of the mRNAs of
   class IA PI3Ks and eventually promotes antitumor immunity.

   During the induction of antitumor immunity by cytotoxic CD8+ T cells,
   the first step in immune killing is antigen presentation by MHC‐I on
   tumor cells to the TCR of CD8+ T cells. However, frequent MHC‐I
   deficiency in most tumors,^[ [145]^6 ^] including BLCA, hinders the
   recognition of tumor cells by CTLs, ultimately resulting in tumor
   immune evasion and ICI insensitivity. PI3K/Akt signaling activation has
   recently been shown to cause MHC‐I deficiency in tumor cells,^[
   [146]^21 ^] which was also observed in case of BLCA cells in this
   study. Moreover, G3BP1 downregulates MHC‐I by activating PI3K/Akt
   signaling to promote immune evasion of BLCA, which is a novel function
   of G3BP1, although G3BP1 has previously been shown to enhance antiviral
   immunity^[ [147]^27 ^] and autoimmune response.^[ [148]^25 , [149]^28
   ^]

   Three classes of PI3Ks exist, but only class I PI3Ks have been
   associated with carcinogenesis and BLCA.^[ [150]^29 ^] The class IA
   PI3Ks are heterodimers, including p110 catalytic subunits
   (p110α/PIK3CA, p110β/PIK3CB, and p110δ/PIK3CD) and p85 regulatory
   subunits (p85α/PIK3R1 and p85β/PIK3R2).^[ [151]^29 , [152]^30 ^] We
   revealed that G3BP1 and SLU7 jointly bind to the mRNAs of most class IA
   PI3Ks, stabilize the mRNAs, and enhance their translation. These
   findings are surprising, as previous studies have revealed that G3BP1
   promotes mRNA degradation by binding to UPF1 and that G3BP1 binds to
   the rG4 structure to maintain mRNA stability but inhibit translation.^[
   [153]^22 , [154]^31 ^] This discrepancy may be because G3BP1
   coordinates with SLU7 to form a complex with PABPC1 and eIF4G1 to
   stabilize the closed‐loop structure of the mRNAs of class IA PI3Ks,^[
   [155]^10 , [156]^23 ^] indicating the new role of SLU7 in sustaining
   PI3K/Akt signaling to downregulate MHC‐I in BLCA cells.

   SLU7 is predominantly located in the nucleus,^[ [157]^14c ^] although a
   minor proportion is also detected in the cytoplasm, particularly after
   exposure to ultraviolet‐C stress.^[ [158]^32 ^] However, the role of
   SLU7 in the cytoplasm remains poorly understood. This study identified
   an interaction between SLU7 and G3BP1 in BLCA cells, which is
   consistent with the finding of a previous MS analysis conducted in
   hepatoma cells.^[ [159]^33 ^] The SLU7 interactome analysis revealed
   the presence of cytoplasmic proteins^[ [160]^33 ^] such as G3BP1 as
   well as other proteins responsible for preserving the closed‐loop
   structure of mRNAs (e.g., PABPCs, eIF3s, and eIF4s).^[ [161]^10 ,
   [162]^23 ^] Consequently, Slu7 may enhance Akt activation in the mouse
   liver^[ [163]^14b ^] through our proposed mechanism, although further
   investigation is warranted to validate this hypothesis. Thus, this
   study provides novel insights into the cytoplasmic role of SLU7,
   particularly in co‐sustaining PI3K/Akt signaling with G3BP1 in BLCA
   cells.

   The therapeutic upregulation of MHC‐I has recently been shown to
   enhance the sensitivity of ICIs.^[ [164]^34 ^] However, as MHC‐I is
   also expressed in nontumor cells, an increase in MHC‐I expression in
   vivo may cause autoimmune disorders.^[ [165]^35 ^] Therefore,
   inhibition of G3BP1, a target present at high levels in BLCA cells, may
   be suitable for enhancing MHC‐I expression in BLCA cells. Indeed,
   inhibition of G3BP1 by EGCG notably inactivates PI3K/Akt signaling to
   upregulate MHC‐I in human and murine BLCA cells and suppresses the
   immune evasion of murine BLCA cells in vitro and in vivo. Importantly,
   we revealed that EGCG enhanced the efficacy of anti‐PD‐1 antibodies in
   murine BLCA tumors. High G3BP1 levels are associated with a poor
   prognosis in patients with BLCA treated with anti‐PD‐L1 antibody; thus,
   the use of EGCG may also enhance the efficacy of ICIs in human BLCA. To
   date, >100 clinical trials related to EGCG have been registered
   worldwide ([166]https://clinicaltrials.gov/), including phase 3 and
   phase 4 clinical trials. However, no clinical trials have yet
   investigated the efficacy of EGCG in combination with ICIs for cancer
   (especially advanced BLCA). Therefore, we highly recommend that
   clinical trials should be conducted to investigate whether EGCG
   enhances the efficacy of ICIs (e.g., anti‐PD‐1 antibody) in patients
   with advanced BLCA.

   In conclusion, G3BP1 interacts with SLU7 to form a complex with PABPC1
   and eIF4G1, which stabilizes the closed‐loop structure of class IA PI3K
   mRNAs and facilitates their translation and stabilization, thereby
   activating PI3K/Akt signaling to promote immune evasion by suppressing
   MHC‐I expression. Therefore, targeting G3BP1 with EGCG impedes immune
   evasion and sensitizes BLCA cells to anti‐PD‐1 antibodies in mice.
   Thus, this study indicates the potential of EGCG as a precision
   therapeutic agent to enhance the effectiveness of anti‐PD‐1 therapy for
   BLCA.

4. Experimental Section

Ethics Approval and Consent to Participate

   The use of human tissues in this study was reviewed and approved by the
   Committee for Ethical Review of Research Involving Human Subjects of
   Sun Yat‐sen University Cancer Center (Guangzhou, China; Registration
   ID: B2022‐350‐01). Informed consent was obtained from all patients. All
   animal‐related procedures were approved by the Animal Research
   Committee of Sun Yat‐sen University Cancer Center (Guangzhou, China;
   Registration ID: L102042021040I).

Patients and Tissue Specimens

   One hunded fifty‐one formalin‐fixed and paraffin‐embedded BLCA
   specimens were otained from patients who underwent radical resection
   for BLCA at Sun Yat‐sen University Cancer Center from 2008 to 2016.
   None of the patients had received any antitumor therapy prior to
   biopsy. All samples were pathologically confirmed by two experienced
   pathologists using hematoxylin and eosin staining.

Cell Lines and Cell Culture

   BLCA cell lines, including T24 (human), UMUC3 (human), MB49 (murine),
   and human embryonic kidney 293T (HEK293T) cells, were obtained from the
   American Type Culture Collection (Manassas, VA, USA). T24 and MB49
   cells were cultured in Roswell Park Memorial Institute (RPMI) 1640
   (Invitrogen, Carlsbad, CA, USA), UMUC3 cells were cultured in minimum
   essential medium (containing nonessential amino acids; Invitrogen), and
   HEK293T cells were cultured in Dulbecco's modified Eagle medium
   (Invitrogen). All media were supplemented with 10% fetal bovine serum
   (FBS; Invitrogen) and 1% penicillin/streptomycin. The cells were
   incubated at 37°C under 5% CO[2]. All cell lines used in this study
   were authenticated using short tandem repeat profiling analysis and
   tested negative for mycoplasma contaminations.

Plasmid Construction

   HA‐tagged human or mouse G3BP1, Flag‐tagged human or mouse SLU7, and
   their truncations were cloned into the PSIN‐EF1α‐blasticidin vector.
   Further, Flag‐tagged LARP1, DDX21, DHX9, PHAX, RBM22, EIF2S3, FXR1,
   SERBP1, and LARP4 as well as GFP‐tagged PABPC1 and eIF4G1 were cloned
   into the PSIN‐EF1α‐blasticidin vector. Next, OVA was cloned into the
   CV547‐IRES‐mCherry vector. Then, shRNAs targeting human or mouse G3BP1,
   SLU7, and PABPC1 were cloned into the PLKO.1‐puro vector. The target
   sequences of shRNAs are presented in Table [167]S1 (Supporting
   Information).

Establishment of Stable Expression Cell Lines

   Stable expression cell lines were established for T24, UMUC3, and MB49
   by infecting them with lentiviruses (MOI = 30), followed by selection
   for 1 week using 0.75 µg mL^−1 puromycin and/or 5 µg mL^−1 blasticidin.
   For T cell killing assay, MB49 cells were infected with OVA‐mCherry
   lentivirus (MOI = 30), and mCherry‐positive single‐cell clones were
   selected to establish stable mCherry‐expressing MB49‐OVA cell lines.
   Stable MB49‐OVA cell lines with modifications using shNC, shG3bp1,
   shSlu7, vector, HA‐G3bp1, or their different combinations were
   established via reinfection with the specified lentiviruses, followed
   by selection using the abovementioned concentration of puromycin and/or
   blasticidin.

RNA Extraction and Quantitative Reverse Transcriptase Polymerase Chain
Reaction (qRT–PCR)

   Total RNA was extracted from cells using TRIzol reagent (Thermo Fisher
   Scientific) according to the manufacturer's instructions. cDNA was
   synthesized using HiScript II Q RT SuperMix for qPCR (Vazyme, Nanjing,
   China), according to the manufacturer's instructions. qRT–PCR was
   performed using SYBR Green Master Mix (Vazyme) via a LightCycler 480
   real‐time PCR system (Roche Applied Science). All mRNA levels were
   normalized to those of GAPDH. The sequences of primers used for qRT–PCR
   are listed in Table [168]S2 (Supporting Information).

Western Blotting

   Cells were lysed using RIPA buffer (1% NP‐40, 0.1% sodium
   dodecyl‐sulfate [SDS], 5 mm ethylenediaminetetraacetic acid [EDTA],
   150 mm NaCl, and 50 mm Tris‐HCL; pH 8.0) supplemented with protease and
   phosphatase inhibitors (Sigma‐Aldrich, St. Louis, MO, USA). The lysates
   were cleared via centrifugation at 14 000 ×g and 4 °C for 15 min. The
   study used 8%–12% SDS–polyacrylamide gel electrophoresis (PAGE) to
   separate equal amounts of proteins, after which the separated proteins
   were transferred onto polyvinylidene difluoride membranes (Millipore,
   Bedford, MA, USA) and blocked with QuickBlock solution (Beyotime
   Biotechnology, Shanghai, China) for 10 min at room temperature (RT).
   The membranes were then incubated with primary antibodies, followed by
   incubation with horseradish peroxidase‐conjugated secondary antibodies.
   The signals were visualized using an enhanced chemiluminescence reagent
   (Tanon, Shanghai, China). Primary antibodies were used against G3BP1
   (13057‐2‐AP, ProteinTech; 1:2000 dilution), MHC‐I (15240‐1‐AP,
   ProteinTech; 1:1000 dilution), SLU7 (12050‐1‐AP, ProteinTech; 1:500
   dilution), Flag (14793, Cell Signaling Technology; 1:2000 dilution), HA
   (3724, Cell Signaling Technology; 1:2000 dilution), Akt (4691, Cell
   Signaling Technology; 1:500 dilution), p‐Akt^Thr308 (9275, Cell
   Signaling Technology; 1:1000 dilution), p‐Akt^Ser473 (4060, Cell
   Signaling Technology; 1:500 dilution), PIK3CA (4249, Cell Signaling
   Technology; 1:500 dilution), PIK3CB (3011, Cell Signaling Technology;
   1:500 dilution), PIK3CD (34050, Cell Signaling Technology; 1:500
   dilution), PIK3R1 (60225‐1‐Ig, ProteinTech; 1:500 dilution), GFP (2956,
   Cell Signaling Technology; 1:1000 dilution), eIF4G1 (15704‐1‐AP,
   ProteinTech; 1:500 dilution), PABPC1 (10970‐1‐AP, ProteinTech; 1:500
   dilution), GAPDH (10494‐1‐AP, ProteinTech; 1:20 000 dilution), and
   alpha tubulin (11224‐1‐AP, ProteinTech; 1:5000 dilution).

Flow Cytometry

   The MHC‐I levels on the MB49 cell surface were detected using
   PE‐conjugated MHC class I (H‐2Kd/H‐2Dd) monoclonal antibody
   (#12‐5998‐82, Thermo Fisher Scientific), following the manufacturer's
   instructions. Briefly, MB49 cells were trypsinized and washed twice
   with PBS. Approximately half a million MB49 cells were resuspended in
   100 µL of staining buffer (2% FBS in PBS) supplemented with 1.25 µL of
   MHC‐I antibody solution, and the cell suspension was incubated for
   30 min at 4 °C under dark conditions. The cells were then washed twice
   and resuspended in 200 µL of PBS. CytoFLEX Flow Cytometer (Beckman
   Coulter, USA) were used for sample analysis and FlowJo 10.8.1 software
   for data analysis.

In Vitro T Killing Assay

   OT1CD8^+ T cells were isolated from the spleen of OT1 mice (Cavens
   Biogle, Suzhou, China) using EasySepTM Mouse CD8+ T Cell Isolation Kit
   (#19853, STEMCELL Technologies), following the manufacturer's
   instructions. OT1CD8^+ T cells were maintained in RPMI 1640
   supplemented with 5% FBS, 55 µm β‐mercaptoethanol, 1%
   penicillin/streptomycin, and 10 ng mL^−1 IL‐2. OT1CD8^+ T cells were
   stimulated with 2 µg mL^−1 SIINFEKL and 10 ng mL^−1 IL‐2 for 3 days to
   generate CTLs. CTLs were cocultured with mCherry‐expressing MB49‐OVA
   cells for 24 h at a ratio of 2:1, with or without MK‐2206 (S1078,
   Selleck) or EGCG (S2250, Selleck). Cytotoxicity was assessed by
   calculating the percentage of 7‐AAD+ cells among mCherry+ cells using
   flow cytometry.

Animal Experiments

   Animal care and experiments were strictly performed according to the
   Guide for the Care and Use of Laboratory Animals and the Principles for
   the Utilization and Care of Vertebrate Animals. Five‐week‐old male
   C57BL/6J mice were purchased from Vital River Laboratories (Beijing,
   China) and housed them under specific pathogen‐free conditions (i.e., a
   12/12 h light/dark cycle, 50% relative humidity, and temperature of
   25–27 °C) with ad libitum access to food (standard chow) and water. The
   mice were acclimatized for 5 days before the experiment.

   To determine the role of G3bp1 in tumor growth and CD8^+ T cell
   infiltration, the mice were randomly divided into three groups of six
   mice each. MB49 cells with G3bp1 knockdown were trypsinized and washed
   twice with PBS. The mice were subcutaneously injected with 3 × 10^5
   MB49 cells (shNC or shG3bp1) suspended in 100 µL of PBS into the
   posterior flanks to establish the tumor‐bearing model. Tumor growth was
   monitored every 2 days, and tumor volume was calculated using the
   modified ellipsoidal formula (tumor volume = 1/2 [length × width^2]).
   The mice were sacrificed on day 14, and the tumors were harvested,
   weighed, and imaged. The tumor of each mouse was fixed in 4%
   formaldehyde and embedded in paraffin for histological analysis.

   To determine the effect of EGCG on the efficacy of anti‐PD‐1 antibody,
   24 mice were further used to establish the above‐described
   tumor‐bearing model with wild‐type MB49 cells, and the mice were
   randomly categorized into four groups of six mice each. For treatment
   with EGCG and antibody, EGCG (50 mg kg^−1) or vehicle was injected
   intraperitoneally on days 3, 5, 7, 9, 11, and 13 after inoculation with
   MB49 cells, and 200 µg of anti‐mouse PD‐1 antibody (BE0146, Bio X cell)
   or rat IgG (Bio X cell) was injected intraperitoneally on days 3, 6, 9,
   and 12. The body weight of each mouse was monitored on days 3, 9, and
   13. The subsequent steps were performed as described above.

RNA‐seq Analysis

   RNA‐seq of BLCA cells with G3BP1 knockdown was performed by RiboBio
   (Guangzhou, China) using the Illumina NovaSeq 6000 System. Six
   gigabytes of clean data per sample were collected for RNA‐seq, and the
   clean reads were aligned with the human genome (hg19) using TopHat
   (version 2.0.13) software. The downstream signaling pathways of G3BP1
   were revealed using KEGG pathway enrichment analysis and GSEA based on
   RNA‐seq data.

RIP and RIP‐seq

   RIP was performed using Magna RIP RNA‐Binding Protein
   Immunoprecipitation Kit (17‐700, Sigma‐Aldrich) according to the
   manufacturer's protocol. Briefly, UMUC3 cells with stable expression of
   HA‐G3BP1, SLU7‐Flag, or PABPC1‐GFP were lysed using RIP lysis buffer
   supplemented with a protease inhibitor cocktail and RNase inhibitor.
   Cell lysates were then incubated overnight at 4 °C under gentle
   rotation with antibodies against HA (26183, Thermo Fisher Scientific),
   Flag (8146, Cell Signaling Technology), GFP (66002‐1‐Ig, ProteinTech),
   or mouse IgG (B900620, ProteinTech). Each mixture was then incubated
   with protein A/G magnetic beads for 2 h at 4 °C under gentle rotation,
   and the immunoprecipitants were washed eight times with RIP wash buffer
   supplemented with a protease inhibitor cocktail and RNase inhibitor.
   Finally, total RNA was extracted using TRIzol reagent, after which
   RNA‐seq or qRT–PCR analysis were performed.

   RIP‐seq was performed by RiboBio using the Illumina NovaSeq 6000
   System. Ribosomal RNA sequences were removed. Twelve gigabytes of clean
   data per sample were obtained using RIP‐seq, and the clean reads were
   aligned to the human genome (hg19) using TopHat (version 2.0.13)
   software. Fisher's exact test was used to determine the statistical
   difference between HA‐G3BP1‐immunoprecipitated and input samples.

Co‐IP, re‐IP, and MS Analysis

   For co‐IP, cell lysates were prepared as described in section Western
   Blotting. Then, they were incubated overnight at 4 °C under gentle
   rotation with antibodies against Flag (8146, Cell Signaling
   Technology), HA (26183, Thermo Fisher Scientific), or mouse IgG
   (B900620, ProteinTech). Each mixture was then incubated with protein
   A/G magnetic beads (88803, Thermo Fisher Scientific) for 2 h at 4 °C
   under gentle rotation. Subsequently, the immunoprecipitants were washed
   eight times with RIPA buffer supplemented with protease inhibitors,
   after which they were analyzed using western blotting or MS analysis.

   For re‐IP, UMUC3 cell lysates with simultaneous stable expression of
   SLU7‐Flag and HA‐G3BP1 were primarily immunoprecipitated with anti‐Flag
   antibody, after which they were blocked with 15 µg mL^−1 Flag peptide
   (TP1273, TargetMol, USA) for 2 h at 4°C under gentle rotation. The
   resultant immunoprecipitants were subjected to re‐IP with anti‐HA
   antibody, washed as described previously, and analyzed via western
   blotting.

   For MS analysis, the HA‐G3BP1 complex was obtained using co‐IP with
   anti‐HA antibody from UMUC3 cells stably expressing HA‐G3BP1. The
   purified immunoprecipitants were resolved via gradient SDS–PAGE,
   silver‐stained, and subjected to MS analysis. Briefly, the gels were
   minced and then destained with 50% acetonitrile in 50 mm ammonium
   bicarbonate. The proteins were reduced with 10 mm dithiothreitol at 56
   °C and alkylated with 55 mm iodoacetamide at RT under dark conditions.
   Trypsin digestion was then performed for 20 h at 37 °C, after which the
   peptides were extracted with 1% trifluoroacetic acid in 50%
   acetonitrile. For MS analysis, the samples were vacuum‐dried and
   reconstituted with 0.1% formic acid. The pretreated samples were
   analyzed using nanoscale liquid chromatography coupled with tandem MS.
   Data were acquired using data‐dependent analysis and processed using
   PLGS software (Waters). The resultant peak list was searched against
   the National Center for Biotechnology Information database with the
   Mascot search engine.

PLA

   PLA was performed using Duolink In Situ Red Starter Kit (DUO92101,
   Sigma‐Aldrich) according to the manufacturer's instructions. Cells were
   seeded into confocal dishes and cultured overnight for adherence. They
   were then fixed with 4% paraformaldehyde for 15 min and penetrated with
   0.5% Triton X‐100 for 10 min. The cells were blocked with a blocking
   solution for 1 h at 37 °C and coincubated overnight at 4 °C with
   antibodies against SLU7 (sc‐376985, Santa Cruz; 1:25 dilution) and
   G3BP1 (13057‐2‐AP, ProteinTech; 1:500 dilution). The cells were then
   incubated with PLA probes for 1 h at 37 °C, after which ligation was
   performed for 30 min at 37 °C. Amplification was performed using an
   amplification solution for 100 min at 37 °C. Finally, the samples were
   mounted using Fluoroshield with DAPI. Images were then captured using
   confocal laser scanning microscopy.

Immunohistochemistry (IHC)

   Formalin‐fixed and paraffin‐embedded BLCA tissues were sliced into 5‐µm
   thick sections and heated for 2 h at 60°C. The sections were
   deparaffinized with xylene and rehydrated in gradient ethanol, after
   which antigen retrieval was conducted in a pressure cooker at high
   pressure using EDTA buffer (pH 6.0–9.0). Subsequently, the sections
   were blocked with PBS supplemented with 3% bovine serum albumin (BSA)
   for 1 h at RT. The sections were then incubated with primary antibodies
   overnight at 4 °C, followed by incubation with secondary antibodies for
   1 h at RT. Primary antibodies were used against SLU7 (NBP2‐42921, Novus
   Biologicals; 1:100 dilution) and G3BP1 (13057‐2‐AP, ProteinTech; 1:400
   dilution). Further, the signals were visualized using DAB Visualization
   Kit (Dako Omnis, Santa Clara, CA, USA) according to the manufacturer's
   instructions. Finally, the sections were counterstained with
   hematoxylin. The H‐score was determined using HALO software (Indica
   Labs).

IF Analysis and Immuno‐FISH

   For IF analysis in cells, cell seeding, fixation, and penetration were
   performed similarly as described in section PLA. Subsequently, cells
   were blocked with 10% BSA for 1 h at RT and incubated with primary
   antibodies overnight at 4 °C. Then, the cells were washed and incubated
   with secondary antibodies for 1 h at RT, after which they were mounted
   using Fluoroshield with DAPI (F6057, Sigma‐Aldrich). For IF analysis in
   BLCA tissues, the sections were prepared and pretreated as described in
   section Immunohistochemistry (IHC); then, blocking, antibody
   incubation, and mounting were performed as described previously. Images
   were captured using confocal laser scanning microscopy. The following
   antibodies were used for IF analysis: G3BP1 (13057‐2‐AP, ProteinTech;
   1:500 dilution), SLU7 (sc‐376985, Santa Cruz; 1:25 dilution), CD8
   (GB11068, Servicebio; 1:200 dilution), Alexa Fluor 488‐conjugated
   anti‐rabbit IgG (A‐11034, ThermoFisher; 1:500 dilution), Alexa Fluor
   555‐conjugated anti‐mouse IgG (A‐31570, ThermoFisher; 1:200 dilution),
   and Alexa Fluor 647‐conjugated anti‐rabbit IgG (A‐32733, ThermoFisher;
   1:500 dilution).

   For immuno‐FISH, IF analysis was performed as described previously.
   After incubation with secondary antibody, the cells were washed with
   PBS and subjected to single‐molecule inexpensive FISH, according to a
   previously reported method.^[ [169]^36 ^] Briefly, the cells were
   prehybridized with 15% deionized formamide in 1× saline sodium citrate
   (SSC) for 15 min at RT and incubated overnight at 37 °C with a
   hybridization solution. After hybridization, the samples were washed
   thrice with 15% formamide in 1× SSC at 37 °C for 30 min and mounted
   using Fluoroshield with DAPI. The sequences of the primary probes and
   FLAP probes are presented in Table [170]S3 (Supporting Information).

FISH–FRET

   To identify the closed‐loop structure of a single mRNA, FISH–FRET
   method was established based on a previous report.^[ [171]^37 ^]
   Briefly, the cap and tail of an mRNA were labeled using Cy3 probes
   (donor) and Cy5 probes (receptor), respectively, and then evaluated its
   closed‐loop structure based on the FRET efficiency between the two
   probes. Theoretically, the FRET efficiency between Cy3 and Cy5 probes
   should decrease when the closed‐loop structure is disrupted.

   Cell seeding, fixation, penetration, and blocking were performed
   similarly as described in section PLA. The Cy3 probes and Cy5 probes
   were created by annealing Cy3‐conjugated or Cy5‐conjugated FLAP probes
   with primary probes designed specifically for the cap or tail of an
   mRNA, following a previously reported method.^[ [172]^36 ^] Through
   FISH, the cap and tail of each mRNA were successively labeled with Cy3
   and Cy5 probes, respectively. Briefly, cells were prehybridized with
   15% deionized formamide in 1× SSC for 15 min at RT, after which they
   were incubated with a hybridization solution containing Cy3 probes,
   followed by incubation with Cy5 probes overnight at 37 °C. After
   hybridization, the samples were washed thrice with 15% formamide in 1×
   SSC at 37 °C for 30 min and mounted using Fluoroshield with DAPI. The
   sequences of the primary probes and FLAP probes are presented in Table
   [173]S4 (Supporting Information).

   FRET was performed using an inverted Zeiss LSM880 laser scanning
   confocal microscope with fast AiryScan (Zeiss, Germany) and the
   receptor photobleaching method. The cells used for FRET analysis were
   scanned, selected three random regions in the cytoplasmic region of a
   single cell, and continuously bleached the receptors (Cy5 probes) with
   a 633 laser at 100% intensity. The FRET efficiency in each region was
   calculated using ZEN 2.3 imaging software.

Polysome Profiling Analysis

   Polysome profiling analysis of the mRNAs of class IA PI3Ks was
   performed as previously described.^[ [174]^38 ^] Briefly, cells were
   incubated with 100 µg mL^−1 cycloheximide (CHX) at 37 °C for 10 min.
   The cells were rinsed with ice‐cold PBS solution containing 100 µg
   mL^−1 CHX and then lysed using 1 mL of lysis buffer (10 mM Tris‐HCl [pH
   7.4], 5 mm MgCl[2], 100 mm KCl, and 1% Triton X‐100) supplemented with
   2 mm DTT, 100 µg mL^−1 CHX, 50 U mL^−1 RNase inhibitor, and EDTA‐free
   protease inhibitors, followed by centrifugation at 12 000 ×g and 4 °C
   for 10 min. The lysate was then applied to sucrose gradients of 10%–50%
   and ultracentrifuged at 4 °C and 36 000 rpm for 2 h using a Beckman
   SW41Ti rotor. Polysome profiles were measured via a UV monitor using
   Data Quest software at 254 nm with a piston gradient fractionator. RNA
   extraction and qRT–PCR were performed as previously described. The mRNA
   levels of class IA PI3Ks in each fraction were normalized to those of
   GAPDH. The squences of primers used for qRT–PCR are presented in Table
   [175]S2 (Supporting Information).

RNA Stability Assay

   Cells were treated with 2 µg mL^−1 actinomycin D (sbr00013,
   Sigma‐Aldrich). The cells were then harvested at 0, 4, 8, and 12 h
   after treatment with actinomycin D and performed RNA extraction and
   qRT–PCR analysis. The mRNA levels at different time points were
   determined by normalizing the mRNA expression to that detected at 0 h.

Statistical Analysis

   GraphPad Prism (Version 9.0) was used to construct graphs and perform
   statistical analyses. The mean values obtained for the control and
   experimental groups were analyzed for significant differences. All data
   were expressed as means ± standard deviation. The chi‐square test or
   Fisher's exact test was used to assess differences between variables.
   Statistical significance was determined using two‐tailed Student's
   t‐test, one‐way analysis of variance (ANOVA), or two‐way ANOVA.
   Survival analysis was performed using Kaplan–Meier analysis, Cox
   regression analysis, and log‐rank test. Correlations between variables
   were analyzed using Pearson's correlation test. A P‐value of <0.05 was
   considered to indicate statistical significance.

Conflict of Interest

   The authors declare no conflict of interest.

Author Contributions

   X.Z., J.C., and M.D. contributed equally to this work. X.Z. and J.C.
   designed the study, performed most of the experiments and drafted the
   manuscript; M.D., Zhenhua L., H.T., and Y.L. assisted the experiments
   and analyzed the data; K.N. and Y.P. performed the bioinformatic
   analysis; X.L. and Z.Z. provided clinical samples and relevant clinical
   information; T.K. and Zhuowei L. supervised the study and revised the
   manuscript. All authors have read and approved the final manuscript.

Supporting information

   Supporting Information
   [176]Click here for additional data file.^ (1.6MB, pdf)

Acknowledgements