The tuberous sclerosis complex (TSC), composed of TSC1/TSC2 heterodimers, is inhibitory to the mammalian target of rapamycin (mTOR). Deletion of either TSC1 or TSC2 renders mTOR constitutively active. To directly explore the impact of mTOR activation on B-cell development, we conditionally deleted TSC1 in murine B cells. This led to impairment in B-cell maturation. Unexpectedly, and in contrast to Akt activation, marginal zone (MZ) B cells were significantly reduced. Administration of rapamycin partially corrected the MZ defect, indicating a direct role for mTOR in controlling MZ development. When challenged with a T-cell-dependent antigen, TSC1 KO mice responded less efficiently. Consistent with the MZ defects, TSC1 KO mice did not respond at all to T-independent antigens. Because activation of Akt upstream of TSC and mTOR yields the reverse phenotype with respect to MZ development, we conclude that, physiologically, Akt simultaneously emits two opposing signals that counterbalance each other in the control of B-cell differentiation.
The mammalian target of rapamycin (mTOR) is an evolutionary conserved serine/threonine kinase, originally identified in Saccharomyces cerevisiae based on its sensitivity to rapamycin 1. mTOR exists in at least two multi-subunit complexes, referred to as mTOR complex 1 (mTORC1) and mTORC2. mTORC1 is sensitive to rapamycin, while inhibition of mTORC2 requires higher concentrations and long duration of incubation with rapamycin 2. mTORC1 funnels multiple signaling pathways from inside and outside the cell. When activated, mTORC1 promotes anabolic processes and enhances protein synthesis and cell growth 3. When inhibited either physiologically or by rapamycin, mTORC1 induces macroautophagy 4. In the recent years mTOR, primarily in the form of mTORC1, has been demonstrated to play major roles in cancer and in immune functions 5, 6.
Much of the knowledge on the role of mTOR in immune regulation has been obtained from loss of function experiments using rapamycin or analogs thereof. mTOR plays various important roles in the T-cell arm of the adaptive immune response. mTOR is required for IL-2-dependent proliferation of CD4+ cells 7. Genetic removal of mTOR or rapamycin treatment promotes the generation of regulatory T cells 8. Interestingly, in CD8+ T cells mTOR antagonizes the generation of memory cells at the priming phase 9.
In contrast to T cells, the role of mTOR in B-cell function is less clear. Most evidence exists for a role of mTOR in controlling IL-7 signaling at the pro-B stage. The addition of rapamycin to B-precursor acute lymphoblastic leukemia lines induced apoptotic cell death in a manner reversible by IL-7 10. Deletion of SIN1, a key element of mTORC2, enhanced IL-7R expression and increased pro-B cell survival 11. At the mature state of B-cell development, mTOR is activated in response to toll-like receptor and B-cell receptor (BCR) ligation downstream to the PI3K/Akt signaling pathway. Akt activates mTORC1 indirectly by reversing the inhibition of mTOR by the tuberous sclerosis complex (TSC), a heterodimer of TSC1 and TSC2 12. TSC2 is a GTPase-activating protein (GAP), which upon association with TSC1 inactivates the small G protein Rheb. Because TSC2 phosphorylation by Akt represses its GAP activity, it allows the accumulation of Rheb in a GTP-bound state. When this occurs, mTORC1 is potently activated 13. Recently, Akt was described to utilize a second mechanism for mTOR activation, by phosphorylation and neutralization of PRAS40, an inhibitor of the interactions between mTORC1 and its substrates 14.
Often, the signaling pathway controlling mTOR is referred to as the PI3K/Akt/mTOR. However, as well as having multiple substrates apart from TSC2, Akt participates in other cellular functions, such as apoptosis, cell cycle, cell migration and cell differentiation. Therefore, Akt activation is not restricted to the mTOR pathway. One of the mTOR-independent functions of Akt in BCR signaling is the inactivation of the forkhead transcription factor FoxO1. Phosphorylation by Akt causes cytoplasmic sequestration and proteasomal degradation of FoxO1 15.
When Akt is activated in mature B cells by genetic ablation of the PI3 phosphatase PTEN, marginal zone (MZ) B cells prosper 16. Recently, it has been shown that FoxO1, rather than mTOR, predominantly transmits the regulation of MZ B cells downstream of Akt, as deletion of FoxO1 yields a similar phenotype as PTEN knockouts 17. Furthermore, deletion of FoxO1 and PTEN are equally potent in rescuing the survival of BCR− mature B cells, reinforcing the idea that the PI3K pathway primarily regulates mature B-cell functions by eliminating FoxO1 18. However, there is no information as to what extent mTOR per se regulates B-cell development.
Here, we investigated the role of mTOR activation in a manner independent of Akt. Mice expressing a floxed allele of TSC1 were crossed with CD19-Cre to yield B cells that constitutively overactivate mTORC1 (referred to as TSC1 KO). TSC1 KO B cells exhibit a partial block in maturation. Interestingly, and in contrast to the phenotype of Akt activation, MZ B cells were greatly reduced in the TSC1 KO. Administration of rapamycin caused increase in MZ B-cell numbers in TSC1 KO as well as WT mice. The TSC1 KO mice failed to respond to T-cell-independent antigens. We observed an atrophied germinal center response in TSC1 KO mice following T-cell-dependent vaccination. Our data suggest that the PI3K pathway emits two opposing signals with respect to MZ B-cell homeostasis, of which the FoxO1 is probably the dominant one.
Results and discussion
The absence of TSC1 confers impairment in B-cell maturation and decrease in MZ B cells
Mice deleted for TSC1 in their B-cell lineage were generated by crossing TSC1f/f to CD19-Cre mice (termed TSC1 KO). B cells were extracted from spleens of TSC1 KO and WT littermates. At first we performed two PCR reactions according to Kwiatkowski et al. 19, one for the recombined allele and another for the non-recombined allele. The analysis demonstrates the appearance of recombination and almost complete loss of the non-recombined TSC1 allele (Fig. 1A). Total cell extracts were subjected to immunoblotting with anti-TSC1 and p97 as a loading control. Consistent with the PCR analysis, we observed efficient reduction in TSC1 protein level (Fig. 1B). To verify that deletion of TSC1 results in activation of the mTOR pathway, we monitored the phosphorylation of ribosomal S6 protein in the course of LPS activation. S6 is phosphorylated by S6K1, whose activation is directly under mTORC1 control. By immunoblotting, we observed induction of phosphorylated S6 (P-S6) expression (Fig. 1B). Intracellular staining showed for TSC1 KO cells heterogeneous expression of P-S6. One population expressed normal level of P-S6 and a second population expressed higher levels. Following LPS stimulation for 3 days, almost all TSC1 KO cells increased P-S6 expression to a higher level than WT cells (Fig. 1C). We conclude that deletion of TSC1 by CD19-Cre yields the expected mTOR hyperactivation phenotype in the vast majority of cells.
Although CD19 is expressed early in B-cell development, gene deletion with CD19-Cre usually occurs at the transitional immature state, when B cells have already emigrated from the bone marrow to the periphery. We monitored the maturation and transition states of B cells in spleens of TSC1 KO and WT mice. Total number of splenocytes and the relative proportion of B cells were similar between the WT and KO (Fig. S1). However, relative to WT, B cells of TSC1 KO cells were overrepresented at the T1 and T2 transition states indicating a partial block in maturation (Fig. 1D). Accordingly, TSC1 KO spleens contained less mature B cells. Further analysis of the mature B cells showed a substantial reduction in the MZ population, based on CD21/CD23 expression (Fig. 1E). This result was particularly surprising as the activation of the PI3K pathway upstream of mTOR by PTEN deletion yields the opposite phenotype, i.e. expansion of MZ B cells 16.
mTOR affects the synthesis of many cellular proteins including those of the secretory pathway. Therefore, in light of the discrepancy with the PTEN KO, and to verify that this result was not due to loss of the CD21 expression, we stained splenic B cells with CD1d, an independent additional marker of MZ B cells. We observed similar phenotype as for CD21 expression i.e. loss of MZ B cells in the TSC1 KO spleens (Fig. 1E). Finally, we analyzed frozen sections of spleens by immunofluorescence. For unclear reasons, in most TSC1 KO spleen sections, we did not detect clearly the MZ macrophages by MOMA-1 staining, while the MOMA-1 positive macrophage ring was evident in the WT controls. When discerned, TSC1 KO B cells were scarce in the outer periphery of the MZ macrophages, confirming the defects in MZ B cells seen by flow cytometry (Fig. 2B, left panel). Collectively, our data indicate that direct overactivation of mTOR impairs B-cell maturation and predominantly leads to the reduced numbers of MZ B cells.
Rapamycin treatment enhances the generation of MZ B cells
To address whether deletion TSC1 perturbs MZ homeostasis by mechanisms that directly involve the hyperactivation of mTOR, we subjected mice to rapamycin treatment. Since MZ B cells are long lived 20, we decided to administer rapamycin for a prolonged time. Following 17 days of treatment, we observed an increase in the percentage of TSC1 KO MZ B cells. Curiously, we also recorded an increase in MZ B-cell proportion in the WT animals following rapamycin treatment (Fig. 2A). The expansion of MZ B cells by rapamycin treatment could also be demonstrated by immunofluorescence of spleen sections (Fig. 2B). Because rapamycin is highly efficient in blocking mTORC1 in vivo, our findings indicate that mTORC1 should be efficiently suppressed for optimal MZ development.
TSC1 KO mice respond poorly to immunization and display defective germinal centers
To determine whether the aberrations in B-cell differentiation also confer functional defects, we challenged mice with T-independent type II (NP-ficoll) and T-dependent (NP-CGG) antigens, which predominantly activate the MZ and follicular subsets, respectively. As expected from the reduced numbers of MZ B cells, TSC1 KO mice did not mount an antigen-specific immune reaction to NP-ficoll. We deliberately show the raw data of a typical ELISA analysis of 2 WT and 2 TSC1 KO mice. TSC1 KO mice did not respond at all to NP-ficoll challenge as seen by lack of specific anti-NP IgM, IgG1 and IgG3 (Fig. 3A). We also observed reduced IgM secretion when naïve splenocytes were stimulated with LPS in vitro, indirectly inferring reduction in MZ activity (Fig. S2). In contrast, TSC1 KO mice did respond to NP-CGG inoculation with antigen-specific antibodies, although with high level of variability. Some mice responded normally, while other did not, in response to two consecutive inoculations (Fig. 3B). Overall, we observed reduced titers for serum isotypes. These results suggested that TSC1 KO have a functional immune response to T-dependent antigens at least to some extent. To further characterize the immune reaction in TSC1 KO mice we analyzed germinal center (GC) B cells 17 days following NP-CGG vaccination. Immunofluorescence analysis for peanut-agglutinin (PNA) showed dispersed staining for KO cells. Furthermore, KO Peyer's patches contained B cells in normal numbers, but staining to PNA was at the background level (Fig. 3C). The defect in GC reaction was confirmed by flow cytometry to two GC markers GL7 and PNA. In both cases, less TSC1 KO cells displayed the GC markers following vaccination (Fig. 3D and 3E). Our data demonstrate that activation of mTOR in B cells yields a defect in immune response to T-dependent and primarily to T-independent antigens.
The PI3K/Akt/mTOR pathway plays major roles in the adaptive and innate arms of the immune response. Somewhat opposing roles for this pathway were demonstrated for different cell types. While in T cells rapamycin blocks proliferation, in macrophages rapamycin promotes the secretion of proinflammatory cytokines 21. Interestingly, in mature B cells, the BCR usurps the vast majority of PI3K activity in a ligand-independent manner. When the BCR is removed, B cells succumb to apoptosis due to lack of PI3K activity 18. However, because rapamycin does not induce apoptosis of B cells, it is unlikely that the pro-survival PI3K activity is utilizing mTOR. Rather, evidence implicates FoxO1 as the main Akt effector in B cells. Moreover, inhibitors of PI3K do not fully block the activity of mTOR, further reinforcing the dissociation of mTOR from the PI3K pathway in B cells 22. Our data, for the first time, demonstrate that in the B-cell lineage mTOR and PI3K may have opposing roles with respect to MZ B-cell differentiation.
Why should MZ B cells be sensitive to mTOR activation? The exact mechanism remains to be determined. mTOR has pleotropic functions that are hard to predict. To account for its predominant effect on MZ B cells, the unique features of this cell type should be taken into account. mTOR hyperactivation may affect the signaling pathways that control generation of MZ and/or elicit metabolic aberrations that are not tolerated by MZ cells. In contrast to FO B cells, MZ B cells do not recirculate, are long-lived, self-renew and exert minimal mobility. Thus, their metabolic demand at the naïve state is presumably low. As such, imbalanced activation of mTOR may confer metabolic aberrations, among which is the induction of endoplasmic reticulum (ER) stress, and consequently the compromise of MZ B-cell viability 23.
Materials and methods
TSC1f/f mice were purchased from the Jackson laboratories. TSC1 KO B cells were generated by crossing to the CD19-Cre strain. All animal studies were conducted in accordance with the Principles of Laboratory Animal Care (NIH publication ♯85-23, revised 1985). The joint ethics committee (IACUC) of the Hebrew University and Hadassah Medical Center approved the study protocol for animal welfare. The Hebrew University Animal Facility is an AAALAC international accredited institute (1285).
Mature B cells were purified from mouse splenocytes by magnetic depletion with anti-CD43 (Miltenyi Biotec, Germany). Cells were plated at 1.5×106 cells/mL in complete medium containing RPMI 1640 (Invitrogen-Gibco, CA, USA) supplemented with 10% FBS (Biological Industries, Israel), 2 mM glutamine, 50 U/mL penicillin, 50 μg/mL streptomycin, 50 μM β-ME, 25 mM 1× nonessential amino acids, 1 mM sodium pyruvate (Biological Industries, Israel) and E. coli LPS 20 μg/mL (Sigma, L3755).
Analysis of TSC1 deletion by PCR
Mature B cells were purified from mouse splenocytes as above. DNA was purified and subjected to PCR analysis (annealing temp 58°C, 40 cycles) with the following primers: F4536: AGG AGG CCT CTT CTG CTA CC and R6548: TGG GTC CTG ACC TAT CTC CTA to amplify the recombined TSC1 allele and R4830: CAG CTC CGA CCA TGA AGT G and COR4771: AGC CGG CTA ACG TTA ACA AC to amplify the floxed allele prior to recombination 19. Primers for G6PDH were used as a loading control.
For flow cytometry we used the following antibodies: rat anti-mouse Syndecan-1-FITC and isotype control rat IgG1-FITC were purchased from R&D Systems (Minneapolis, MN, USA). Rat anti-mouse CD93-PE, rat anti-mouse T- and B-cell activation-Alexa Fluor 647 (GL7), rat anti-mouse CD21/CD35-PE, rat anti-mouse IgM-PE; isotype controls rat IgG2b-PE and rat IgG2a-PE were purchased from e-bioscience. Rat anti-mouse CD45R/B220-PerCP, mouse anti-mouse CD19-FITC, rat anti-mouse CD23-FITC, rat anti-mouse CD1d-PE, rat anti-mouse IgD-FITC, isotype controls rat IgG2a-PerCP, rat IgG2a-FITCand rat IgM-Alexa Fluor 647 were purchased from BioLegend (San Diego, CA, USA). Rabbit anti-human phospho-S6 ribosomal protein and TSC1 were purchased from Cell Signaling Technology (Danvers, MA, USA). Goat anti-rabbit Alexa Fluor 647 was purchased from Molecular Probes (Eugene, OR, USA). Biotin-IgA, isotype and isotype control mouse IgA-FITC was purchased from Southern Biotech (Birmingham, AL, USA). Isotype Botin-SP-conjugated ChromoPure goat IgG was purchased from Jackson ImmunoResearch (West Grove, PA, USA). FITC-PNA was purchase from Vector Laboratories (Burlingame, CA, USA).
Intracellular staining was performed with BD Cytofix/Cytoperm (BD Biosciences, San Jose, CA, USA) according to the manufacturer's instructions. Flow cytometry was performed on BD LSRII flow cytometer (Becton Dickinson, Franklin Lakes, NJ, USA). Data were analyzed using FCS Express V3 analysis software (De Novo, CA, USA).
For immunofluorescence the following antibodies were used: Purified rat anti-mouse CD45R/B220 was purchased from BioLegend). Biotin anti-mouse macrophages (MOMA-1) was purchased from Cedarlane Laboratories (Ontaria, Canada). Biotinylated PNA was purchase from Vector Laboratories (Burlingame, CA, USA). DyLigth 488-conjugated streptavidin and donkey anti-rat Cy3 were purchased from Jackson ImmunoResearch (West Grove, PA, USA).
Spleen and Peyer's Patches were embedded in O.C.T compound, frozen and sectioned at 10 μm. Tissue sections were fixed in ice acetone (10 min, room temperature), washed in Optimax (BioGenex Laboratories, San Ramon, CA, USA) and blocked in Cas Block (Invitrogen, Carmarillo, CA, USA) (10 min, room temperature). Then, sections were labeled with rat anti-mouse CD45R/B220, biotin anti-mouse macrophages (MOMA-1) and biotinylated PNA (30 min, room temperature) and were detected with DyLigth 488-conjugated streptavidin and donkey anti-rat Cy3 (30 min, room temperature). Sections were imaged by confocal microscopy (X200, zoom 60×5, Olympus IX70, Center Valley, PA, USA).
Measurements of serum immunoglobulin titers
The level of IgM, IgG1, IgG2a, and IgG3 was determined by enzyme-linked immunosorbent assay (ELISA) using SBA Clonotyping System/HRP kits (Southern Biotechnology Associates, Birmingham, AL, USA). To evaluate NP-specific antibody plates (Nalge Nunc International, Rochester NY, USA) were coated overnight with 5 μg/ml NP26-BSA (Biosearch Technologies, Novato, CA, USA). A series of dilutions were performed and the titers relative to WT were determined.
Cohorts of five mice were immunized i.p with 30 μg NP-Ficoll (Biosearch Technologies) in PBS or with 100 μg NP-CGG (Biosearch Technologies) in Imject Alum (Pierce, Rockford, IL, USA). Serum titers of anti-NP antibodies following NP-ficoll vaccination were determined 6 days after vaccination. Serum titers of anti-NP antibodies following NP-CGG vaccination were determined at day 17 following vaccinations at days 0 and 10.
In total, 75 μg rapamycin were dissolved in 200 μL of 2% Tween 20 in PBS. Rapamycin was injected i.p every other day for 16 days. On day 17 mice were sacrificed.
Research was funded by grants from David R. Bloom center for pharmacy, the Rosetrees Trust and the Israel Science Foundation (grant 78/09) to B. T. The anonymous and Lady Davis funds to S. B.
Conflict of interests: The authors declare no financial of commercial conflict of interest.