Pten^fl/fl, Pten^fl/fl:Raptor^fl/fl, and Pten^fl/fl:Rictor^fl/fl mice were crossed with ROSA26^Cre-ERT2 mice. To induce deletion of conditional alleles, mice were treated intraperitoneally with tamoxifen (3 mg per dose) on two consecutive days, as previously described (27). Splenic B cells were isolated five days after the third injection for in vitro experiments or adoptive transfer, allowing sufficient time for Cre-mediated recombination and depletion of pre-existing PTEN Raptor, or Rictor protein. Mice are hereafter referred to as Pten^δ, Pten^δ:Raptor^δ, and Pten^δ:Rictor^δ, respectively. For deletion in the AID-expressing B cell compartment, these strains were crossed with Aicda^Cre mice to generate Aicda^Cre/+:Pten^fl/fl and Aicda^Cre/+:Pten^fl/fl:Rictor^fl/fl, hereafter referred to as Pten^B and Pten^B: Rictor^B, respectively. Control ROSA26^Cre-ERT2 and Aicda^Cre/+ littermates or age-matched controls were used throughout. Pten^fl/fl, Raptor^fl/fl, Rictor^fl/fl and ROSA26^Cre-ERT2 mice were provided by M. Boothby (Vanderbilt University). Aicda^Cre/Cre and Rag2^-/- mice were purchased from Jackson Laboratory. C57BL/6J mice were obtained from DBL Korea. All mice were housed under specific pathogen-free conditions at the Laboratory Animal Center of Hallym University. Influenza A virus (IAV) infection experiments were conducted under biosafety level 2 (BSL-2) at the Medical and Bio-Convergence Research Institute of Hallym University. All animal procedures complied with guidelines from the Korean Ministry of Food and Drug Safety and were approved by the Institutional Animal Care and Use Committee of Hallym University [Hallym 2018-58; Hallym 2020-47; Hallym 2023-55].

Single-cell suspensions from spleens or lymph nodes were prepared as described (28). B cells were isolated by depleting CD43⁺ cells using CD43 (Ly-48) MicroBeads and MACS Separator (Miltenyi Biotec). GC B cells were purified using the Mouse Germinal Center B cell MicroBead Kit (Miltenyi Biotec) according to the manufacturer’s instructions. CD4 T cells were isolated using CD4 (L3T4) MicroBeads (Miltenyi Biotec).

CD40LB feeder cells that express murine CD40L and B cell activation factor (BAFF) (kind gift from S.G. Kang, Kangwon National University) were cultured in DMEM containing 10% FBS (29). Feeder cells were plated in 60-mm dishes, treated with 7 µM mitomycin C (Sigma) for 3 h, and washed several times to remove residual mitomycin C. Naïve B cells were co-cultured with CD40LB feeder cells in complete RPMI-1640 (Welgene) supplemented with 2 ng/ml mIL-4 (R&D Systems) to generate in vitro GC-like B cells (iGCBs) over 4 days. iGCBs were then harvested and cultured with mitomycin C-treated CD40LB cells in the presence of 10 ng/ml mIL-21 (PeproTech) for an additional four days to induce plasmablast (iPB) differentiation. In some experiments, iGCB cultures were treated with 5 nM Rapamycin (Tocris), 25 nM Torin-2 (Tocris), 1 μM MK-2206 (Selleckchem), or 50 μM DAPT (MCE).

Single-cell suspensions were prepared from iGCB and iPBs cultures and spleen and incubated with Fc Block (BD Bioscience). Cells were stained with fluorescence-conjugated antibodies in the presence of 7-AAD (ThermoFisher) for viability exclusion, as previously described (28). The following antibodies were used: CD19 (1D3), B220 (RA3-6B2), GL7 (GL7), Fas (Jo2), CD138 (clone 281-2), CXCR4 (clone 2B11), IgM (II/41), IgD (11-26c.2a), IgG1 (A85-1), and IgE (RME-1). Flow cytometry antibodies were purchased from BD Bioscience, ThermoFisher, or BioLegend. Data were acquired on a BD FACS Canto II instrument using FACSDiva software and analyzed with FlowJo software V10.8.1 (BD Bioscience).

A fraction of 12 x 10⁶ B cells isolated from tamoxifen-treated ROSA26^Cre-ERT2:Pten^fl/fl, ROSA26^Cre-ERT2:Pten^fl/fl:Raptor^fl/fl, ROSA26^Cre-ERT2:Pten^fl/fl:Rictor^fl/fl, or control mice were mixed with 4 x 10⁶ wild-type CD4⁺ T cells and intravenously transferred into Rag2^-/- mice. Mice were immunized intraperitoneally with 50 µg 4-hydroxy-3-nitrophenylacetyl (NP)-conjugated ovalbumin (OVA) emulsified in alum and boosted with NP-OVA in PBS 14 days later to ensure robust, synchronized germinal center responses in the adoptive transfer setting. Sera were collected 5 days after the boost immunization. To examine GC reactions in vivo, Aicda^Cre/+:Pten^fl/fl, Aicda^Cre/+:Pten^fl/fl:Rictor^fl/fl, and control mice were injected intraperitoneally with 0.2 ml of a 100% sRBC suspension (Innovative research). Splenic GC B cells, plasma cells and Ig isotype switching were analyzed by flow cytometry on day 9 post-immunization.

The H1N1 mouse-adapted IAV strain A/WSN/1933 was provided by Hyung-Joo Kwon (Hallym University). Virus stocks were prepared in embryonated chicken eggs under BSL-2 conditions and quantified by plaque assay (30). 8–10 week-old male Aicda^Cre/+:Pten^fl/fl, Aicda^Cre/+:Pten^fl/fl:Rictor^fl/fl, and control mice were immunized intraperitoneally with 1 × 10⁷ pfu WSN virus. After 20 days, mice were intranasally infected with 1 × 10⁶ pfu WSN. For passive immunization, sera were collected 21 days post-immunization, pooled, and 100 μl sera were injected intravenously into C57BL/6J males, followed by intranasal infection WSN infection. Viral titers in lungs were assessed at 6 days post-infection by plaque assay or quantitative real-time PCR detecting the WSN nucleocapsid protein (NP) gene. Lung tissues were harvested after perfusion and fixed with 4% paraformaldehyde for histological analysis with hematoxylin and eosin staining (Merck). Images were acquired using an Eclipse Ni-U upright microscope (Nikon).

Relative levels of serum NP-specific antibodies were measured by capture ELISA using NP₃₀-BSA-coated plates, as described (28). Anti-NP IgM, IgG1, IgG2b, IgG3 and IgA were detected using SBA Clonotyping System-HRP kit (SouthernBiotech). High-affinity and cross-reactive antibodies were assessed using NP₂-BSA or 4-hydroxy-3-iodo-5-nitrophenylacetyl (NIP)-BSA as capture antigens, respectively. Anti-IAV antibodies were detected using plates coated with 1 x 10⁶ pfu WSN virus.

Total RNA was extracted from iGCBs using TRIzol (Invitrogen). Libraries were prepared using the TruSeq Standard mRNA Library Prep Kit (Illumina) according to the manufacturer’s instructions. Sequencing was performed on an Illumina NovaSeq 6000 platform by Macrogen. Sequencing reads were aligned to the Mus musculus reference genome mm10 using HISAT2 v2.1.0. and assembled with StringTie v2.1.3b. Differentially expressed genes (DEGs) were identified using DESeq2 with threshold set a false discovery rate (FDR) < 0.05 and fold change > 1.5. Functional annotation of DEGs was conducted using DAVID Bioinformatics Resources (https://davidbioinformatics.nih.gov), and pathway enrichment was analyzed using the Kyoto Encyclopedia of Gens and Genomes (KEGG) database within DAVID. Heatmaps were generated using the heatmap.2 in the gplots package in RStudio v3.4.4. Gene expression correlations were assessed using FPKM values and Pearson’s correlation coefficients were calculated using GraphPad Prism v7.04 (GraphPad Software).

Total RNA was extracted from iGCBs or splenic GL-7^hi Fas⁺ B cells using TRIzol reagent (ThermoFisher) and reversed-transcribed with the ImProm-II Reverse Transcription system (Promega) according to the manufacturers’ instructions. Quantitative real-time PCR was performed using SYBR qPCR mix (Toyobo) on a CFX96 Real-Time PCR Detection System (Bio-Rad). Primer pairs are listed in Supplementary Table 1. Data were normalized to Actb in each sample and calculated using the comparative 2^-δδCT method (31). For immunoblotting, iGCBs lysates were prepared in RIPA lysis buffer supplemented with protease and phosphatase inhibitor cocktails (GenDEPOT). Protein lysates were separated by SDS-PAGE and transferred to PVDF membranes. Membranes were incubated with the indicated primary antibodies, followed by HRP-conjugated secondary antibodies (Enzo Life Sciences). Monoclonal antibodies against phospho-S6^Ser240/244 (D68F8), S6 ribosomal protein (5G10), phospho-4E-BP1^Thr37/46 (236B4), 4E-BP1 (53H11), phospho-Akt^Ser473 (D9E), PTEN (D4.3), Raptor (24C12), phospho-GSK-3β^S9 (D3A4), GSK-3β (D5C5Z), AID (L7E7), Hes1 (D6P2U), Notch2 (D76A6), α-tubulin (DM1A) and β-actin (8H10D10), as well as polyclonal antibodies against Akt and Rictor, were purchased from Cell Signaling Technology. Protein signals were detected using an Enhanced Chemiluminescence Detection kit (Merk Millipore).

All data are representative of at least two independent experiments with consistent results and are presented as means ± standard deviation (S.D). Statistical analyses were performed using Prism v7.04 (GraphPad Software). Data distributions were assessed for normality using the Shapiro–Wilk test prior to parametric testing. Comparisons between two groups were performed using an unpaired two-tailed Student’s t-test. For experiments involving multiple comparisons, two-way analysis of variance (ANOVA) was used, followed by Bonferroni’s multiple-comparison post hoc. Outliers were identified and excluded using the ROUT method (Q = 5%). Statistically significance was reported using predefined threshold p values (p < 0.05, p < 0.01, p < 0.001), as indicated in the figure legends. Statistical tests used and the numbers of biological replicates (n) are specified in the corresponding figure legends.

GC reactions are orchestrated through dynamic interactions between B cells, follicular helper T cells and stromal cells to generate of high-affinity, isotype-switched antibodies, as well as long-lived plasma cells and memory B cells (1). Although mTOR signaling has been implicated in GC B cell biology, the precise regulation of mTOR activity required for optimal humoral immunity remains incompletely understood. To address this, we utilized an in vitro GC-like B cell (iGCB) culture system in which naïve B cells were cocultured with CD40L- and BAFF-expressing feeder cells to mimic GC-like activation in vivo (29) (Figure 1A). During iGCB activation and differentiation, we observed robust induction of both mTORC1 and mTORC2 activities, as evidenced by phosphorylation of S6 and 4E-BP1 (mTORC1) and Akt (mTORC2), respectively (Figure 1B; Supplementary Figure S1a). Notably, Akt phosphorylation declined during in vitro plasmablast (iPB) differentiation, whereas 4E-BP1 phosphorylation remained sustained.

To assess the functional consequences of mTOR activity, we treated iGCBs and iPBs with low doses of the mTOR inhibitors rapamycin (Rapa) and torin-2 (Torin), which were selected to minimize cytotoxic and cytostatic effects (Supplementary Figure S1b) (19, 32). Torin treatment efficiently suppressed phosphorylation of S6, 4E-BP1, and Akt without impairing iGCB differentiation or expression of the activation markers such as CD40 (Figures 1C, D; Supplementary Figures S1c, d). While Rapa had only a marginal effect on S6 phosphorylation, both inhibitors abrogated IL-4-induced isotype switching to IgE and increased the frequency of unswitched IgM⁺ iGCBs (Supplementary Figure S1c; Figures 1E–G). Intriguingly, Torin significantly enhanced the frequency of IgG1⁺ iGCBs, whereas Rapa had no detectable effect on IgG1 class switching (Figures 1F, G). Analysis of absolute cell numbers revealed that Torin treatment had minimal effects on iGCB expansion, whereas rapamycin modestly but significantly reduced iGCB viability (Supplementary Figure S1b), indicating that the differential effects of Torin on IgG1 switching are not attributable to major differences in cell survival or proliferation. Upon differentiation into iPBs, CD138⁺ B220^lo cells were comparably induced and more than 99% of iPBs had lost surface IgM and IgD expression (Figures 1H, I). Remarkably, both Rapa and Torin treatment promoted the emergence of IgG1⁺ iPBs at the expense of IgE⁺ cells (Figures 1J, K). Together, these results demonstrate that mTOR signaling acts as a negative regulator of IgG1 isotype switching in GC-like activated B cells.

PI3K signaling is essential for B cell activation and differentiation, but its amplitude must be tightly regulated to support efficient CSR and affinity maturation. Loss of PTEN, a negative regulator of PI3K signaling, results in pathologic pathway hyperactivation and defective GC output (10, 13). To dissect the individual contributions of mTOR complexes within this hyperactive signaling context, we generated inducible KO mice lacking PTEN alone or in combination with either Raptor (an essential component of mTORC1) or Rictor (a regulatory subunit of mTORC2) (Figure 2A). Consistent with previous studies (17, 27), deletion of either Raptor or Rictor alone resulted in markedly reduced B cell viability, whereas combined deletion with PTEN did not impair expansion of iGCBs (Supplementary Figure S2a). This genetic framework therefore enabled direct comparison of PTEN-deficient B cells (Pten^δ) with double-deficient cells (Pten^δ:Raptor^δ and Pten^δ:Rictor^δ) to functionally interrogate the specific roles of mTORC1 and mTORC2 downstream of PI3K, while minimizing confounding effects on B cells survival and proliferation (17, 27). Western blotting analysis confirmed that PTEN loss resulted in elevated phosphorylation of both S6 and Akt, indicative of increased mTORC1 and mTORC2 activities, respectively (Figure 2B; Supplementary Figure S2b). Although Raptor depletion reduced S6 phosphorylation, residual S6 phosphorylation remained detectable, consistent with incomplete depletion of mTORC1 activity in this inducible system (Figures 2A, B, Supplementary Figure S2b). Nonetheless, phosphatidylinositol (3,4,5)-triphosphate (PIP3) levels were comparably increased across all PTEN-deficient genotypes, irrespective of Raptor or Rictor deletion (Supplementary Figures S2c, d), confirming persistent upstream PI3K pathway activation. Despite comparable induction of GL7^hi Fas⁺ iGCBs across all genotypes, the surface expression of the costimulatory molecules CD80, CD86, and CD40 was selectively reduced in Pten^δ:Raptor^δ iGCBs (Figure 2C; Supplementary Figure S2e). This phenotype may reflect the functional impact of reduced, though not completely abolished, mTORC1 signaling in this context.

Functionally, PTEN deficiency impaired IL-4-induced isotype switching to both IgG1 and IgE, leading to accumulation of unswitched IgM⁺ iGCBs (Figure 2D). Notably, genetic deletion of Rictor selectively restored IgG1 class switching in PTEN-deficient iGCBs, whereas Raptor deletion further exacerbated the CSR defect (Figures 2D–F). Upon differentiation into iPBs, PTEN deficiency attenuated the generation of B220^lo CD138⁺ iPBs, while combined deletion of PTEN and Raptor enhanced iPB formation (Figure 2G). Consistent with the iGCBs phenotype, both Pten^δ and Pten^δ:Raptor^δ iPBs displayed defective isotype switching, characterized by retention of IgM⁺ cells and reduced IgG1 and IgE expression (Figures 2H–J). In contrast, Pten^δ:Rictor^δ iPBs exhibited markedly increased IgG1 switching relative to control cells (Figures 2H–J). Enhanced IgG1 secretion and productive recombination at the IgG1 heavy chain gene locus were further confirmed in these cells (Supplementary Figures S2f, g). Nonetheless, mTORC2 inactivation did not rescue IgE class switching in PTEN-deficient iGCBs or iPBs (Figures 2F–J). Together with our pharmacologic data, these results demonstrate that mTORC2, but not mTORC1, negatively regulate IgG1 isotype switching in the setting of hyperactivated PI3K signaling. Extending these findings beyond the GC-like culture system, we further observed that mTORC2 inactivation restored IL-4-induced IgG1 switching in PTEN-deficient B cells stimulated with LPS in a T-independent plasmablast culture (Supplementary Figure S2h). Collectively, these data underscore the central role of the PTEN-mTORC2 signaling axis in calibrating optimal IgG1 switching across both T-dependent and T-independent activated B cell activation contexts.

Our in vitro data indicated that the PTEN-mTORC2 signaling module plays a critical role in promoting IgG1 CSR under conditions of PI3K hyperactivation. To evaluate the in vivo relevance of this pathway, we reconstituted immunodeficient Rag2^-/- mice with B cell isolated from tamoxifen-treated inducible KO mice, together with wild type CD4⁺ T cells (Figure 3A). Because adoptive transfer into Rag2^-/- hosts can result in variable or heterogeneous primary GC responses, a booster immunization was included to reinforce robust, synchronized GC-dependent antibody responses at the time of analysis (15). Following immunization with NP-conjugated ovalbumin (NP-OVA), frequencies of GL7^hi Fas⁺ GC B cells and B220^lo CD138⁺ plasma cells were comparable across groups, whereas the total numbers of plasma cells were reduced in mice receiving Pten^δ:Raptor^δ B cells (Figures 3B, C; Supplementary Figure S3). Consistent with our in vitro iPB culture data, PTEN-deficient B cells generated significantly fewer IgG1⁺ plasma cells in vivo. Remarkably, this defect was selectively rescued by concurrent deletion of Rictor, but not Raptor, highlighting a specific requirement for mTORC2 downregulation in IgG1 CSR (Figure 3D).

In line with these results, serum levels of NP-specific class-switched antibodies, including IgG1, IgG2b, IgG3, and IgA, were significantly diminished in mice reconstituted with Pten^δ B cells and were restored by mTORC2 inactivation in the PTEN-deficient background (Figure 3E). Notably, NP-specific IgM levels remained unchanged across all groups, indicating that the PTEN-mTORC2 axis primarily affects isotype switching rather than global antibody production. Although the frequency of IgG1⁺ plasma cells was increased in mice receiving Pten^δ:Rictor^δ B cells compared with controls, NP-specific IgG1 levels in these mice was comparable to controls (Figures 3D, E). Quantification of absolute plasma cell numbers revealed a reduced despite variable total plasma cell pool in this group (Supplementary Figure S3c, d), which likely accounts for the apparent discrepancy between cellular frequencies and serum antibody levels. These findings suggest that reduced plasma cell output may contribute to the limited accumulation of antigen-specific IgG1 in this setting, despite increased frequencies of IgG1⁺ plasma cells. Importantly, high-affinity NP-specific IgG1 responses and cross-reactive binding to the structurally related hapten NIP were significantly enhanced in mice reconstituted with Pten^δ:Rictor^δ B cells (Figures 3F, G). Together, these data demonstrate that mTORC2 inactivation can compensate for PI3K hyperactivation in PTEN-deficient B cells, effectively rescuing both CSR and affinity maturation during the humoral immune responses.

Given the requirement for tightly regulated PI3K signaling in orchestrating effective antibody responses, we next assessed the functional relevance of PTEN-mTORC2 axis during viral infection. To preferentially interrogate B cell-intrinsic effects within the AID-expressing compartment and minimize potential confounding from pre-GC perturbations (17), we crossed Pten^fl/fl and Rictor^fl/fl mice with Aicda^Cre mice, in which Cre recombinase is expressed under the control of the Aicda promoter (33). Using a Cre-reporter system, we confirmed that Cre activity was robustly induced following sheep red blood cells (sRBCs) immunization and was predominantly detected within the GC B cell compartment (Supplementary Figure S4). Deletion of PTEN (Pten^B) or both PTEN and Rictor (Pten^B: Rictor^B) in the Aicda-expressing B cell population did not alter B cell development or steady-state homeostasis (data not shown). Mice were immunized intraperitoneally with influenza A/WSN virus (IAV) and challenged with a lethal dose of the same strain 21 days post-immunization (Figure 4A). All groups exhibited protection and viral clearance following the immunization (Figures 4B C), consistent with prior reports (30). Antibody responses to WSN virus were dominated by class-switched isotypes, including IgG1, IgG2b and IgG3 (Figure 4D). However, Pten^B mice showed significantly reduced WSN-specific IgG1 and IgG3 levels, accompanied by a concomitant increase in IgM levels, indicative of defective CSR. Remarkably, mTORC2 inactivation in the PTEN-deficient background (Pten^B: Rictor^B) restored virus-specific IgG1 and IgG3 levels comparable to control mice (Figure 4D), highlighting a critical role of mTORC2 in constraining isotype switching during IAV infection.

Although antiviral T cell immunity likely contributed to overall protection in all groups, the CSR defect in Pten^B mice could compromise antibody-mediated protection. To directly test this, we performed passive immunization by transferring sera from WSN-immunized mice into naïve recipients prior to viral challenge (Figure 4E). Sera from immunized control mice conferred robust protection, as evidenced by attenuated weight loss and reduced pulmonary viral loads, whereas sera from Pten^B mice failed to provide such protection, resembling outcomes in mice that received unimmunized sera (Figures 4F, G). Notably, sera from Pten^B: Rictor^B mice conferred protection equivalent to that of controls immune sera, confirming functional restoration of protective antibody responses by mTORC2 inactivation (Figures 4F, G). Histopathological analysis revealed extensive pulmonary inflammation, vascular obstruction, and interstitial hemorrhage in mice that received Pten^B or unimmunized sera (Figure 4H). In contrast, lungs from recipients of control or Pten^B: Rictor^B sera displayed preserved tissue architecture and reduced inflammation (Figure 4H). Together, these results establish that the PTEN-mTORC2 axis critically governs protective, isotype-switched antibody responses against to IAV, and that mTORC2 inhibition can functionally rescue defective humoral immunity caused by PTEN loss.

To define how PTEN-mTORC2 signaling influences GC B cell function at the transcriptional level, we performed RNA-seq analysis on iGCBs. PTEN deficiency led to broad transcriptional perturbation, with 1,341 genes significantly up- or downregulated compared to control cells (Figure 5A). Inactivation of mTORC2 in this context resulted in a partial normalization of the transcriptome, with 814 differentially expressed genes (DEGs) identified in Pten^δ:Rictor^δ iGCBs relative to control. Principle component analysis (PCA) demonstrated clear separation between naïve B cells and iGCBs and further revealed that Pten^δ:Rictor^δ iGCBs clustered closer to control iGCBs than did Pten^δ cells (Figure 5B). Direct comparison of DEGs revealed that 392 transcripts dysregulated by PTEN loss were reversed upon Rictor deletion, including 289 genes downregulated and 103 genes upregulated (Figure 5C). These selectively rescued genes encompassed several regulators relevant to GC identity and isotype switching, including surface receptors (e.g., Cxcr4 and Cd86), transcriptional factors (e.g., Myc, Id2, Pou6f1, and Myb), PI3K pathway feedback components (e.g., Inpp4b and Nlrc3), and Aicda, which encodes activation-induced cytidine deaminase (AID). Correlation analyses further showed that expression of Myc, Id2, Pou6f1, and Myb, which were positively or negatively associated with Aicda, was selectively normalized following mTORC2 inactivation in PTEN-deficient iGCBs (Figure 5D). These findings indicate that mTORC2 inhibition restores defined subsets of GC-relevant transcriptional programs disrupted by hyperactive PI3K signaling.

Gene ontology and pathway enrichment analyses of Pten^δ iGCBs revealed altered representation of multiple signaling pathways, including BCR, MAPK, insulin, and calcium signaling (Figure 5E). Importantly, these alterations were partially corrected in Pten^δ:Rictor^δ iGCBs (Figures 5E, F; Supplementary Figures S5a–d), supporting a model in which mTORC2 modulates specific transcriptional outputs downstream of PI3K rather than globally resetting GC-associated pathways. Notably, gene sets linked to immunoglobulin (Ig) production and ER function, such as protein processing in ER, were among those most consistently disrupted by PTEN loss and partially restored by mTORC2 inactivation (Figure 5E; Supplementary Figure S5e), reflecting a potential role of PI3K-mTOR signaling in supporting Ig synthesis and secretory machinery. Consistent with this, expression of Xbp1, a transcription factor essential for secretory capacity and plasma cell differentiation, was reduced in Pten^δ iGCBs and partially restored upon Rictor depletion (Figure 5G). Although global enrichment of canonical CSR and GC B cell signature pathways were not observed, heatmap analysis revealed selective normalization of key GC-associated genes, including Aicda, Myc, Cxcr4, Tnfsf13c, Cd40, and Bach2 in Pten^δ:Rictor^δ iGCBs (Figures 5H, I). FoxO1, a direct target of Akt that is critical GC zonal organization and Aicda regulation (11, 18), was downregulated in PTEN-deficient iGCBs and restored upon mTORC2 inactivation (Supplementary Figure S5f). Concordantly, expression of multiple FoxO pathway genes showed partial normalization following mTORC2 inhibition (Supplementary Figure S5f). Taken together, these data demonstrate that mTORC2 inhibition selectively rebalances discrete transcriptional programs downstream of hyperactive PI3K signaling, rather than globally restoring GC-associated gene expression. This selective transcriptional rewiring likely underlies the functional rescue of IgG1 class switching and antibody output observed upon mTORC2 inactivation in PTEN-deficient B cells.

To validate the transcriptional and functional impact of PTEN-mTORC2 signaling in vivo, we immunized Aicda^CrePten^fl/fl and Pten^fl/fl:Rictor^fl/fl mice with sheep red blood cells (sRBCs) and analyzed splenic B cell populations. Consistent with previous finding (13), deletion of PTEN in Aicda-expressing B cells resulted in increased frequencies of GL7^hi Fas⁺ B cell, a population that includes GC B cells but may also encompass activated extrafollicular Aicda-expressing B cells (Figures 6A, B; Supplementary Figures S6a, b). Notably, Pten^B mice exhibited an altered distribution of GC-associated subsets, characterized by a reduction in CXCR4^hi CD86^lo cells and concomitant accumulation of CXCR4^lo CD86^hi cells, indicative of a skewed DZ-LZ balance and perturbed GC-associated dynamics (Figures 6A, B; Supplementary Figure S6c). This imbalance was normalized in Pten^B: Rictor^B mice, implicating mTORC2 in regulating zonal organization and homeostatic differentiation programs within activated Aicda-expressing B cells. In agreement with these in vivo findings, PTEN-deficient iGCBs displayed reduced Cxcr4 and elevated Cd86 expression, features that were reversed upon Rictor deletion (Figures 5G–I). Although the overall frequency of CD138⁺ B220^lo plasma cells was comparable across groups, the proportion of IgG1⁺ plasma cells was significantly reduced in Pten^B mice and restored to control levels following mTORC2 inactivation (Figures 6C, D). These observations align with our in vitro iGCB and iPB differentiation data and support the conclusion that the PTEN-mTORC2 axis negatively regulates IgG1 class switching and downstream antibody-secreting cell output in activated B cells.

To assess transcriptional alterations in vivo, we purified splenic GL7^hi Fas⁺ B cells from sRBC-immunized mice and analyzed mRNA expression. Transcript levels of Bcl6 and Prdm1, the lineage defining transcriptional factors of GC and plasma cell fate respectively (34), were largely unchanged across genotypes (Figure 6E), although Prdm1 was notably reduced in PTEN-deficient iGCBs (Figure 5I). Consistent with RNA-seq data from iGCBs, PTEN deficiency resulted in significant downregulation of Aicda and Xbp1, both of which are required for CSR and secretory function (35, 36). Importantly, Aicda expression was restored to control levels in Pten^B: Rictor^B cells (Figure 6E). While Pax5 downregulation and Irf4 induction are associated with GC exit and plasma cell commitment (34), neither transcript was significantly altered in Pten^B cells. However, both mRNAs were significantly upregulated following mTORC2 inactivation (Figure 6E), suggesting that suppression of mTORC2 facilitates transcriptional programs linked to productive plasma cell differentiation. Collectively, these findings support a model in which the PTEN–mTORC2 signaling module integrates transcriptional networks to coordinate differentiation, isotype switching, and effector programming in activated B cells during humoral immune responses.

mTORC2 phosphorylates Akt at S473 within its C-terminal hydrophobic motif, a modification required for full Akt activation and optimal phosphorylation of downstream substrates (17, 37). To interrogate the contribution of Akt activity downstream of mTORC2, we used MK-2206, an allosteric Akt inhibitor that stabilizes Akt in an inactive conformation and thereby prevents both Akt phosphorylation and signaling to downstream targets (38). MK-2206 treatment reduced phosphorylation of Akt as well as that of the Akt substrate GSK-3β, confirming effective inhibition of Akt signaling rather than indirect feedback suppression of mTORC2 activity (Figure 7A). Functionally, Akt inhibition restored IgG1 class switching in PTEN-deficient iGCBs (Figure 7B). This was accompanied by increased AID expression, whereas IgE levels were only minimally affected (Figures 7A, B; Supplementary Figure S7a). In contrast, in control iGCBs with intact PTEN, Akt inhibition nearly abolished Akt phosphorylation and resulted in reduced IgG1 switching with a reciprocal increase in IgE expression (Figures 7A, B). These data indicate that Akt activity exerts a context-dependent influence on isotype selection, supporting IgG1 switching under a physiological signaling conditions while suppressing CSR when excessively activated downstream of hyperactive PI3K signaling. Consistent with this model, enforced expression of constitutively active Akt impaired IgG1 switching in both control and Pten^δ:Rictor^δ iGCBs (Supplementary Figure S7b), underscoring the requirement for tightly regulated Akt signaling in effective IgG1 switching. Together, these findings support a model in which mTORC2-dependent Akt activity must be precisely tuned: basal Akt signaling is permissive for IgG1 CSR, whereas excessive Akt activation suppresses AID expression and class switching in activated B cells.

The canonical Notch signaling pathway is essential for marginal zone B cell development and has also been implicated in antibody-secreting plasma cell differentiation (39, 40). During in vitro GC B cell differentiation, cleavage of intracellular Notch2 (ICN2) was transiently induced and declined during subsequent iPB differentiation (Supplementary Figure S5c). Consistent with this observation, transcriptomic analysis revealed enrichment of Notch pathway-associated gene signature in PTEN-deficient iGCBs (Figure 5E, Figure 7C). Although global Notch pathway enrichment was not observed following mTORC2 inactivation, expression of canonical Notch target genes, including Hes1 and Notch2, was significantly reduced in Pten^δ:Rictor^δ iGCBs, as confirmed at the protein levels by immunoblotting (Figures 7C, D). Notably, Hes1 expression inversely correlated with Aicda transcript levels across genotypes (Figure 7E), suggesting that elevated Notch activity might be associated with reduced CSR competence downstream of hyperactive PI3K-mTORC2 signaling.

To explore this relationship functionally, we performed pharmacological perturbation experiments. In control iGCBs, inhibition of Notch cleavage enhanced IgG1 switching and suppressed IgE induction (Figures 7F, G). Conversely, enforced expression of the intracellular Notch1 domain (ICN1) modestly but significantly reduced IgG1 switching in both control and Pten^δ:Rictor^δ iGCBs (Supplementary Figure S7d). However, Notch inhibition alone was insufficient to restore IgG1 switching or AID expression in PTEN-deficient iGCBs (Figures 7F, G; Supplementary Figure S7e), indicating that suppression of Notch cannot overcome the dominant CSR defects imposed by hyperactive PI3K-mTORC2-Akt signaling. Importantly, combined inhibition of Akt and Notch resulted in a significantly greater enhancement of IgG1 switching in PTEN-deficient iGCBs than either perturbation alone (Figures 7H, I; Supplementary Figures S7f, g). These findings support a model in which Notch signaling does not function as a primary determinant of CSR in this context but instead acts as a cooperating modulatory pathway that can influence isotype selection when PI3K-Akt signaling is partially restrained. Together, our data identify Notch signaling as an auxiliary regulatory input downstream of the PTEN-mTORC2-Akt axis that contributes to fine-tuning antibody isotype outcomes.