mTORC1 as a cell-intrinsic rheostat that shapes development, preimmune repertoire, and function of B lymphocytes

Abstract

Ample evidence indicates that nutrient concentrations in extracellular milieux affect signaling mediated by environmental sensor proteins. For instance, the mechanistic target of rapamycin (mTOR) is reduced during protein malnutrition and is known to be modulated by concentrations of several amino acids when in a multiprotein signaling complex that contains regulatory-associated protein of mTOR. We hypothesized that a partial decrease in mTOR complex 1 (mTORC1) activity intrinsic to B-lineage cells would perturb lymphocyte development or function, or both. We show that a cell-intrinsic decrease in mTORC1 activity impacted developmental progression, antigen receptor repertoire, and function along the B lineage. Thus, preimmune repertoires of B-lineage cells were altered in the marrow and periphery in a genetic model of regulatory-associated protein of mTOR haplo-insufficiency. An additional role for mTORC1 was revealed when a B-cell antigen receptor transgene was found to circumvent the abnormal B-cell development: haploinsufficient B cells were profoundly impaired in responses to antigen in vivo. Collectively, our findings indicate that mTORC1 serves as a rheostat that shapes differentiation along the B lineage, the preimmune repertoire, and antigen-driven selection of mature B cells. The findings also reveal a range in the impact of this nutrient sensor on activity-response relationships for distinct endpoints.—Raybuck, A. L., Lee, K., Cho, S. H., Li, J., Thomas, J. W., Boothby, M. R. mTORC1 as a cell-intrinsic rheostat that shapes development, preimmune repertoire, and function of B lymphocytes.

Human and rodent malnutrition increases susceptibility to infection, indicating that immunity is impaired in settings of inadequate nutrient levels (see refs. 1–3; reviewed in refs. 4 and 5). A variety of categories that correlated with lower immunity or resistance to infection have been delineated for these deficiencies. Some are specifically linked to the effect of a deficient micronutrient to the affected cell types. For instance, lack of vitamin A impairs the epithelial cell capacity to deal with measles virus (6) and retinoids derived from this vitamin influence CD4⁺ T cells via nuclear receptors (7). In contrast to this direct action, indirect mechanisms can link inadequacies of food intake to those of immunity as specific hormones respond to the dietary shortcomings. Leptin exemplifies this mechanism. This feeding-responsive hormone acts on CD4⁺ T lymphocytes in a manner that alters their functional characteristics after immune stimulation (8, 9). However, much remains unclear about nutrient and specific immune mechanisms.

One form of malnutrition primarily consists of inadequate intake of protein or specific amino acids (e.g., lysine or leucine) while caloric and vitamin sufficiency are maintained (2, 5). Evidence from in vitro and in vivo experiments indicates that the characteristics or function of lymphocytes can be influenced by the gene expression programs for proteins involved in intermediary metabolism (10–13). Moreover, protein-deprived humans are at increased risk of infection (4, 5, 14). Studies that used controlled diets in laboratory mammals have shown that protein deficiency led to alterations in inflammatory responses, T-cell–mediated immunity, and depressed humoral immunity (1, 2, 15). Protein restriction in experimental rodents led to decreases in T and B cells in some analyses (3, 5, 15). However, much of this work preceded molecular definition of nutrient sensors as well as the cellular subsets that are crucial for specific aspects of immunity. Thus, the mechanisms to account for these experimental observations on protein malnutrition have been relatively unexplored. Immunity might be compromised if the development or function of a crucial hematopoietic cell type such as B or T cells was substantially reduced in protein-starved mammals. To mount antibody responses, B cells require activation by antigen or TLR stimulation, leading to proliferation and potentially culminating in differentiation into antibody-secreting plasma cells (16–18). This process can either be dependent on help from CD4⁺ T cells or T independent, so that antibody production can result from activated B cells that remain largely extrafollicular (19–21) or, alternatively, after recruitment into germinal center (GC) reactions after sustained cognate interactions with CD4⁺ T cells (22, 23).

The GC fosters selection of higher affinities in the antibody repertoire as well as longer persistence of protective antibody concentrations because of more efficient generation of long-lived plasma cells (22–26). The serine-threonine kinase mechanistic target of rapamycin (mTOR) serves as a major sensor of the sufficiency of amino acid supplies when it is in mTOR complex 1 (mTORC1), a complex in which the protein regulatory-associated protein of mTOR (Raptor) is an essential component (27). Intracellular proteins that modulate mTORC1 activity sense juxta-lysosomal concentrations of a few amino acids—especially leucine, lysine, arginine, and glutamine (28–30). Accordingly, one clue as to why protein malnutrition might impair humoral immunity is provided by data showing that restricted protein intake caused not only weaker antitumor immunity but also a partial reduction in phosphorylation of a protein downstream from mTORC1 (31). However, whether these findings are connected has not been explored. Other work found that even 1 wk of dietary protein restriction led to ∼50% reductions in circulating concentrations of leucine and lysine in mice (32), along with evidence that hepatic mTORC1 activity was reduced by the experimental malnutrition. In parallel, however, the dietary restriction also led to lower blood glucose, increased insulin receptor sensitivity, and profound decreases in circulating insulin and IGF-1, each of which could influence immunity. Thus, data from various studies suggest the hypothesis that reduced mTORC1 resulting from protein malnutrition might directly influence immunity. However, the previous work also underscores that hormonal responses to protein or other nutritional deprivation create complexities of systemic alterations.

Accordingly, we took a genetic approach to test the cell-intrinsic effect of reduced mTORC1 on B-lineage cells as the progenitor cell type central to all production of antibody-secreting cells. The prior studies highlight the need for analyses of partial decreases in sensor pathways (33) and not just complete loss of function to test if the mTORC1 levels observed in protein malnutrition might alter humoral immunity. Thus, we hypothesized that haploinsufficiency of mTORC1 would cause immune alterations. The first key issue was whether or not initial B-cell development is susceptible to partially decreased mTOR. We show here that there is a cell-autonomous impact of haploinsufficiency on antigen receptor repertoire and developmental progression. That being the case, we tested if a partial failure of mTORC1 would impair either development of GCs or antibody responses that arise from haploinsufficient B cells in a model in which mTORC1^low B cells all have the same antigen receptor and developmental profile.

MATERIALS AND METHODS

Reagents

mAbs (purified, biotinylated, or fluorophore-conjugated) were procured from Tonbo Biosciences (San Diego, CA, USA) or BD Biosciences (San Jose, CA, USA) unless otherwise specified; clone, catalog, and supplier information is provided in the Supplemental Data. Recombinant B-cell activating factor (BAFF) was purchased from Adipogen (San Diego, CA, USA); recombinant IL-4, bromodeoxyuridine (BrdU), and CellTrace Violet (CTV) for cell cycling and proliferation assays as well as Trizol reagent for nucleic acid isolation were from Thermo Fisher Scientific (Waltham, MA, USA); LPS, tamoxifen, and 4-hydroxy-tamoxifen were from MilliporeSigma (Burlington, MA, USA); hen egg lysozyme (HEL) peptides HEL_46–61 and HEL-OVA peptides (conjugated to ovalbumin aa 323–339) for immunizations, enzyme-linked immune absorbent spot (ELISpot), and ELISAs were synthesized by Lifetein (Somerset, NJ, USA); phosphorylcholine (PC) conjugated to bovine serum albumin was from Biosearch Technologies (Novato, CA, USA). Reagents including double stranded DNA (dsDNA) and histones were purchased from MilliporeSigma. Complete amino acid–sufficient Roswell Park Memorial Institute (RPMI) 1640 medium was purchased from Thermo Fisher Scientific (11875135), and amino acid–deficient RPMI was purchased from U.S. Biologic Life Sciences (R89999-04A; Salem, MA, USA).

Mice and B-cell transfer models

All mice [regulatory-associated protein of mTOR (Rptor)^+/+, Rptor^f/+, Rptor^f/f (34) introgressed to Rosa26–Cre-ER^T2+ (35) or mb1-Cre⁺ (36), MD4 (37); OTII T-cell receptor transgenic (38); recombination activating gene 2 (Rag2)^−/−] were on a C57BL/6 background, housed in sterile ventilated microisolators under specified pathogen-free conditions in a Vanderbilt mouse facility, and used starting at 6–8 wk of age (except as noted for autoantibody analyses) following approved mouse protocols. For assessment of the impact of induced loss of Rptor using Rosa26–Cre-ER^T2+ mice, inactivating recombination was induced with 3 sequential injections of tamoxifen in safflower oil (3 mg/mouse every other day; 0.2 ml of 15 mg/ml dose) followed by harvests 7 d after the third injection (39). For autoantibody analyses, sera were collected from mice aged to 32 wk. Healthy age-matched littermates were used for all experiments. For adoptive transfer experiments to measure primary immune responses, donor (Rptor^+/+, mb1-Cre⁺, MD4⁺; Rptor^f/f, mb1-Cre⁺, MD4⁺) spleen and lymph node cell suspensions were depleted of T cells using biotinylated anti–Thy-1.2 antibodies and streptavidin-conjugated microbeads for the BD Biosciences iMag system. Polyclonal and OTII transgenic CD4 T cells were purified using anti-L3T4 (CD4) microbeads (Miltenyi Biotec, Bergisch Gladbach, Germany). Age-matched recipient mice then received intravenous transfers of 5 × 10⁶ B cells with 2 × 10⁶ polyclonal CD4 T cells and 10⁶ OTII transgenic CD4 T cells in sterile saline. Bone marrow (BM) chimeras were generated after myelo-ablative irradiation of C57BL/6-CD45.1 recipient mice with 2 divided doses of 4 Gy each using a [¹³⁷Cs] irradiator as described in Lee et al. (40). One day after the second exposure, single-cell suspensions of a 1:1 mixture of wild-type (WT) CD45.1 with CD45.2 marrow (mb1-Cre, Rptor^+/+ or mb1-Cre, Rptor^Δ/Δ) were infused (10⁷ cells intravenous/recipient), followed by harvest 6 wk later.

Immunizations, ELISA, and ELISpot

To measure donor-specific humoral responses stemming from transferred B cells into immunodeficient recipients, 5 × 10⁶ purified donor B cells mixed with 2 × 10⁶ polyclonal CD4 T cells and 1 × 10⁶ OTII transgenic CD4 T cells were transferred intravenously into age-matched Rag2^−/− recipient mice followed by intraperitoneal immunization 24 h later with 10 µg HEL-OVA peptide (41) in alum, followed by a booster immunization of HEL-OVA peptide in alum 3 wk later. Mice were then harvested 7 d after boost, and spleen and BM were analyzed for GC responses and antibody-secreting cell analysis by ELISpot, and terminal sera were used for antibody ELISA (39, 42).

Measurement of humoral responses

Terminal sera were collected at harvest and analyzed by ELISA, performed as described in refs. 39 and 42, to determine relative concentrations of antibodies binding PC, histone, dsDNA, and immunogen. Anti-HEL antibodies were measured by ELISA capturing all anti-HEL antibodies with plate-immobilized HEL after immunization or MD4 B-cell activation with HEL peptide spanning aa 46–61 (Lifetein). Antibodies of defined class were then detected using horseradish peroxidase–conjugated isotype-specific antibodies (Southern Biotechnology Associates, Birmingham, AL, USA). Antibody-secreting cells analyzed by ELISpot were quantitated using the ImmunoSpot Analyzer (Cellular Technology, Shaker Heights, OH, USA), after using HEL (46–61) peptide to capture secreted anti-HEL antibodies.

B-cell cultures

Splenocytes from age-matched littermates were enriched for B cells by Thy-1.2 depletion using biotinylated anti–Thy-1.2 mAbs and the BD Biosciences iMag system. For class-switch cultures, cells were cultured at a density of 0.5 × 10⁶ cells/ml in 2-ml cultures on 12-well plates with BAFF; BAFF, anti-CD40, and HEL (46–61) peptide; or BAFF, anti-CD40, HEL (46–61), and IL-4. Unless otherwise specified, cells were cultured in Iscove’s modified Dulbecco’s medium supplemented with 10% fetal bovine serum (FBS). Cells were cultured 2–5 d, followed by flow cytometry and ELISA measurements of relative antibody concentrations in culture supernatants. To assess the impact of amino acid restriction on Ig class switching and proliferation, complete RPMI 1640, supplemented with 10% FBS, 2 mM glutamine, 0.1 mM 2-ME, was mixed with amino acid–free complete RPMI, supplemented with 10% FBS and 0.1 mM 2-ME, at ratios to produce 50, 30, 10, and 5% amino acid–sufficient RPMI. B cells were then cultured for 5 d with BAFF or BAFF, LPS, and IL-4. Supernatants were collected at the end of culture for measurement of relative antibody concentrations by ELISA. To measure proliferative capacity of splenic B cells, cells were first loaded with 10 µM CTV and then cultured for 4 d under conditions described above. B cells from BM of age-matched littermates were purified by biotinylated anti-B220 mAbs and the BD Biosciences iMag system. To assess BM B-cell proliferation in response to IL-7, B220⁺ BM B cells were cocultured with OP9 stromal cells (10⁶ marrow cells/ml, at a ratio of 4:1 with OP9) for 2 and 4 d. To analyze proliferation, BM B220⁺ cells were labeled with carboxyfluorescein diacetate succinimidyl ester as described in Lee et al. (39) and analyzed by flow cytometry after 4 d in culture. To measure IL-7–stimulated cell cycling, cultures were pulsed (4 h) with BrdU after 2 d of coculture and analyzed by flow cytometry after staining for BrdU incorporation into DNA as described in Lee et al. (40).

Flow cytometry

Flow cytometric phenotyping of B-cell development were conducted on age-matched littermates at 6–8 wk of age. For B-cell maturation phenotyping, BM, spleens, and intraperitoneal wash cell suspensions were stained at 10⁶ cells/cocktail. For B-cell development in the BM, 10⁶ BM cells were stained with anti-CD43, - B220, -IgM, -IgD, -CD23, and the viability marker 7-aminoactinomycin D (7-AAD). To measure B-cell maturation in the periphery, 10⁶ splenocytes were stained with 7-AAD along with anti-IgM, -B220, -CD19, -CD93 (AA4.1), and -CD23. B1 B-cell stains were conducted on single-cell suspensions of spleens and peritoneal washes using anti-B220, -IgM, -CD11b, -CD5, -CD23, and viability stain 7-AAD. Flow cytometric phenotyping of in vivo immune responses were conducted 7 d after booster immunization as indicated. For detection of GC responses, 2 × 10⁶ spleen cells were stained with GL7, anti-CD95, -CD38, -B220, -IgM, and a dump cocktail containing viability stain 7-AAD and monoclonal antibodies for IgD, CD11b, CD11c, F_4/80, and Gr1. For these and other flow cytometric analyses, fluorescence emission data on cell suspensions were collected on BD Biosciences LSR or FACSCanto flow cytometers driven by BD Biosciences FACS Diva software, then processed using FlowJo software (FlowJo, Ashland, OR, USA).

Sequencing and quantitation of mRNA by quantitative PCR

mRNA and DNA were isolated from flow-sorted Pro and Pre B cells (B220⁺ IgM^neg CD43⁺ or CD43^neg, respectively, all negative for 7-AAD, IgD, and CD11b) using Trizol reagent and then purified following the standard manufacturer’s protocol. RNA concentrations were measured using a NanoDrop spectrophotometer and then used to synthesize cDNA using the Promega AMV Reverse Transcriptase kit. All mRNA quantifications were normalized to hypoxanthine guanine phosphoribosyl transferase. DNA was redissolved in Tris-Cl buffer per Adaptive Biotechnologies instructions, and concentrations were measured using a NanoDrop spectrophotometer. Amplification and sequencing of IgH complementarity-determining region 3 were performed using the immunoSEQ Platform (Adaptive Biotechnologies, Seattle, WA, USA). The immunoSEQ platform combines multiplex PCR with high-throughput sequencing and a bioinformatics pipeline for IgH complementarity-determining region 3 analyses (43, 44).

Immunoblots

Whole cell lysates were prepared (39, 40) from BM B-lineage cells purified using anti-B220 microbeads (Miltenyi Biotec) and LS columns (Miltenyi Biotec). Relative protein or phosphopeptide abundance in these lysates was then determined by immunoblotting as described in refs. 39, 40. Primary antibodies were anti-actin (Santa Cruz Biotechnology, Dallas, TX, USA), anti-phosphopeptide antisera for measurement of S6 phosphorylation directed to the S235 and S236 phosphorylated epitopes (Cell Signaling Technologies, Danvers, MA, USA); primary anti-S6 (5G10), anti–phosphorylated S6 kinase(T389), and anti–S6 kinase antibodies were from the same supplier. In brief, reduced, denatured proteins separated by SDS-PAGE with a stacking-resolving gel system were transferred to PVDF membranes by electrophoresis. Membranes were probed with primary antisera at supplier-recommended dilutions, after blocking with 5% milk, and rinsed with 0.1% Tween in Tris-buffered saline. After rinsing, membranes were incubated with goat anti-rabbit–specific, or for anti-mouse monoclonal, secondary anti-IgG antibody, conjugated to Alexa 680 nm fluorophore (Thermo Fisher Scientific) and rinsed, and bands were visualized and quantified by scanning with an Odyssey Imaging System (Li-Cor Biosciences, Lincoln, NE, USA).

Statistical methods and tests of significance

The primary analyses were conducted on pooled data points from independent samples and replicates (minimum 3, biologically and temporally independent replicate experiments for all data, with multiple independent samples per experiment), using an unpaired, 2-tailed Student’s t test with posttest validation of its suitability. Two-way ANOVA with Bonferroni correction for multiple comparisons was used for statistical analysis across ELISA titration curves to compare WT with each mutant sample set. When this determination indicated rejection of the null hypothesis (i.e., P < 0.05), WT vs. mutant samples were compared at single dilution values by 2-tailed Student’s t tests. Data are displayed as means ± sem. Results were considered statistically significant when the value for the null hypothesis of a comparison was P < 0.05. All resulted from statistical tests of null hypotheses with post hoc testing to ensure appropriateness. Because the extent or direction of difference between samples was unknown and regulations mandate reducing the number of animals used to the lowest feasible level, no statistical methods were used to determine prespecified sample sizes and mice of both genders were selected in balanced numbers without bias. The experiments were not randomized, and the investigators were not blinded during the experiments.

RESULTS

Reduced mTORC1 activity in B lymphocytes at reduced extracellular amino acid concentrations

Because of the relationships between nutrition and lymphoid cells, and complexities of endocrine responses to whole-body manipulation of nutrition, we explored the effect of amino acids in the extracellular milieu on mTORC1 activity in B cells. Results from in vitro analyses indicated that phosphorylation downstream from this kinase complex was reduced in the absence of exogenous amino acids in the medium (Fig. 1A). Moreover, addition of leucine, which is sensed by mTORC1, to amino acid–free medium only partially restored activity. To model the effects of lower circulating amino acids across worsening stages of protein malnutrition (2, 32, 45), we used a series of media with progressively lower concentrations of amino acids (1-, 0.5-, 0.3-, 0.1-, and 0.05-fold those of RPMI 1640) (Fig. 1B). Immunoblot analyses of mTORC1 activity in activated B lymphoblasts confirmed that mTORC1 activity dropped when amino acid concentration fell to 0.1× normal, a level akin to or greater than circulating amino acid levels in patients that are malnourished (45). It is possible that Raptor protein ordinarily is synthesized in excess, in which case inactivation of only 1 Rptor allele might not affect mTORC1 activity. When we tested this possibility by comparing haplodeficient B cells to Rptor^+/+ controls, however, immunoblot data showed substantial decreases in mTORC1 activity (Fig. 1C), to a degree comparable to that achieved by amino acid restriction.

Figure 1.

Reduced mTORC1 signaling activity at lower amino acid or Raptor levels. A) Immunoblots were performed using the indicated antibodies to analyze extracts of B cells (Rptor^+/+) after activation with anti-IgM, growth (2 d), and division into portions for cultures for 20 h in new medium [amino acid–replete (+a.a), amino acid–free (−a.a), or amino acid–free with leucine added back (Leu), as indicated], followed by reactivation (20 min with anti-IgM) where indicated (+). Analyses from 1 of 4 biologic replicate experiments are shown. B) Results of immunoblots performed using the indicated antibodies to probe membranes after extracts of B lymphoblasts were activated and cultured at the indicated percentages of control amino acid–sufficient medium for 2 d with anti-IgM, rested for 20 h, and then reactivated for 20 min in the same amino acid condition medium, then resolved as described in refs. 39 and 40. Analyses from 1 of 3 independent replicate experiments are shown. C) Immunoblots were performed using the indicated antibodies to analyze extracts of mb1-Cre⁺ splenic B cells that were either Rptor^+/+ or Rptor^Δ/+ after activation with anti-IgM, 2 d of growth, rest for 20 h in fresh medium, and then reactivation for 20 min with anti-IgM where indicated (+). One analysis is shown of 2 biologically independent replicate experiments. Akt, protein kinase B; P-, phosphorylated; S6K, p70 S6 kinase.

Haploinsufficiency of Rptor as the sensor kinase complex mTORC1 promotes selection of distal heavy-chain variable regions and pro– to pre–B-cell development

Based on the preceding data, we hypothesized that inactivation of only 1 Rptor allele could be used as a genetic approach to model effects on B-lineage cells of lower amino acid concentrations akin to those of malnourished states. We first explored the impact of mTORC1 on lymphoid ontogeny in the marrow. Initially, samples from Rosa26–Cre-ER^T2 mice that were Rptor^+/+, Rptor⁺, and Rptor^f/f were analyzed in the short (∼10 d) period while mice remained healthy after tamoxifen injections caused deletion of the Rptor conditional allele in multiple cell types (46). Even in this limited time, the B lineage showed an increase in pro-B cells for Rptor^+/Δ samples that was not accompanied by an increase in their pre-B progeny (Fig. 2A, B). This finding suggested the imposition of a developmental block that preceded any overt change in the naive repertoire of mature B cells in the periphery. In addition to the shift after short-term induction of haplo-insufficiency, the impedance to development appeared to be amplified in homozygous Rptor^Δ/Δ samples. Of note, a modest decrease in marrow cellularity that did not achieve a value of P < 0.05 for Rptor^+/Δ mice became pronounced for the Rptor^Δ/Δ marrow (Fig. 2A). These findings prompted us to test the hypothesis that a partial reduction in mTORC1 activity can affect B-cell development using a means of gene inactivation in which Cre-mediated deletion of conditional Rptor alleles was restricted to the B lineage and acted in situ early in B lineage ontogeny. To do so, we introgressed homo- or heterozygous Rptor fl alleles onto an mb1-Cre background (36). First, we used Western blotting of marrow-derived B-lineage cells to test if mTORC1 activity in marrow cells was affected by inactivation of only 1 allele. This analysis showed a substantial reduction (0.5× control adjusted for tubulin) (Fig. 2C). In some settings, mTORC1 loss relieves a feedback inhibition that restrains PI3K activity, so that mTORC2 would increase. At most, a modest but inconsistent increase was observed at these levels of mTORC1 inhibition (Figs. 1B, C and 2C). To further test if the defects were B lineage-intrinsic or purely lineage-extrinsic, we used mixed BM chimeras in which Rptor-deleted marrow was used to reconstitute the hematopoietic compartments of irradiated recipients. After repopulation of the recipient mice, phenotypes of the resultant B-cell populations were compared with controls (Supplemental Fig. S1).

Figure 2.

Rptor haploinsufficiency reveals a necessary threshold of mTORC1 activity needed for normal B-cell development. A) BM cell numbers 7 d after tamoxifen-induced Rosa26–Cre-ER^T2+ deletion of floxed Rptor alleles. B) Phenotypic analysis of B-cell development and maturation in the BM compartment of Rptor^+/+, Rptor^Δ/+, and Rptor^Δ/Δ after tamoxifen-induced Rosa26–Cre-ER^T2+-mediated deletion (n = 6 of each genotype from 3 independent replicate experiments). C–E) B-cell autonomous loss of Rptor confirms lineage-specific threshold of requirement for mTORC1 in normal B-cell development. C) Immunoblots of whole cell extracts from B220⁺ BM cells from mb1-Cre⁺ Rptor^+/+ and Rptor^Δ/+ mice were probed using the indicated primary antibodies. One set of analyses is shown of 2 biologically independent replicates. AKT, protein kinase B; P-, phosphorylated. D) Representative flow cytometry with gating strategy for analyses of B-cell development in the BM of mb1-Cre⁺ mice that were Rptor^+/+, Rptor^Δ/+, and Rptor^Δ/Δ. E) Quantitation of B-cell developmental stages in mice with B-lineage–restricted reduction in or elimination of mTORC1 (n = 12 Rptor^+/+, 9 Rptor^Δ/+, 9 Rptor^Δ/Δ). Subsets and stages are identified here as IgM^neg B220⁺ (Pre/Pro), IgM^mid B220^lo [immature (Imm)], IgM⁺ B220^hi [mature (Mat) or recirculating], IgM^hi B220⁺ [transitional (Trans)], IgM^neg B220⁺ CD43⁺ (Pro), and IgM^neg B220⁺ CD43^neg (Pre). F) Relative IL-7Rα RNA expression in flow-sorted (7-AAD^neg, IgM^neg, B220⁺, CD43⁺) pro-B cells from Rptor^+/+, Rptor^Δ/+, and Rptor^Δ/Δ mb1-Cre⁺ BM (n = 4 of each genotype between 2 independent replicate experiments). G) Analysis of percent BrdU positive from 2-dimensional cultures in the presence or absence of recombinant IL-7 of B220⁺ purified BM B cells and OP9 stromal cell coculture as a percent of the B220⁺ IgM^lo population. H) BM pro-B cells were purified by flow sorting 7-AAD^neg 220⁺ IgM^neg CD43⁺ from individual mb1-Cre⁺ mice (5 of each Rptor genotype, +/+, Δ/+, and Δ/Δ). A multivariate heat map of sequencing results of Ig heavy-chain variable (V_H) regions, using data generated by Adaptive Biotechnologies and displayed from top to bottom as most Cµ-distal to -proximal V regions, is shown. Examples of V_H regions enriched in heterozygote pro-B cells compared with WT controls are named to the left of the heat map. Quantitation of these differences and variance among samples is in Supplemental Fig. S2C. N.S., not significant.

The observed decrease in mTORC1 activitiy in BM-derived B cells and the defect in B-cell repopulation of irradiated recipients led us to assess B-lineage development in mb1-Cre–expressing Rptor homo- or heterozygous floxed mice. Profiles of B-lineage cells, marked by the appearance of B220 (CD45R) antigen, were determined with marrow harvested from Rptor^+/+, Rptor^f/+, and Rptor^f/f mice that all bore the mb1-Cre transgene. Virtually no B-lineage cells could be detected past the pre-B-cell stage when Rptor was homozygous for the mutated allele (Fig. 2D, E and Supplemental Fig. S2A). Moreover, almost none of the pro- or prestage cells were CD43⁺, which suggests that lack of mTORC1 blocked progression to the pre-B-cell stage (Fig. 2D, E). Further experiments showed a progressive reduction in IgM^hi (immature) B cells as Raptor decreased from hemizygosity to complete loss (Fig. 2D) and a reduced frequency of pro- or pre-B cells in the marrows of Rptor^+/Δ mice (Fig. 2E). An increase in the pro-B cells was followed by a reduced representation of their pre-B descendants (Fig. 2E, right panel). These results indicated that Raptor in marrow-repopulating stem cells promotes the development of B cells and identified an effect at the stage of the pro– to pre–B-cell transition (Fig. 2D, E). The distortion of B lineage-autonomous development resulting from haploinsufficiency was even more evident in the periphery (Supplemental Fig. S1D, E). Importantly, parallel analyses of littermates heterozygous for the Rptor allele revealed a distorted progression of cellular differentiation, such that pre-B descendants decreased while pro-B precursor frequencies increased (Fig. 2D, E). We conclude that the efficiency of developmental progression from pro- to pre-B state is reduced when mTORC1 activity is lowered in developing B-lineage cells.

The pro- to pre-B transition is due to generation of a functional Ig heavy chain and assembly of a pre–B-cell receptor (BCR). This step depends on RAG proteins (RAG1 and 2) and IgH locus accessibility. Others have reported that amounts of Rag1 and Rag2 mRNA in precisely staged fractions of developing B cells were unaffected by complete loss of function for Rptor (47, 48). Expression of IL-7 receptor α (IL-7Rα) chain in progenitors is a crucial step in promoting B-cell development in the marrow, because it mediates survival signaling by stroma-derived IL-7, as well as initiation of RAG-mediated Ig gene rearrangements to generate BCR heavy and then light chains (47, 48). Because of the scarcity of the relevant cells and developmental stage-specific differences in gene expression, we used flow cytometric sorting to purify B220⁺ CD43⁺ cells from the CD11b^neg, IgD^neg, 7-AAD^neg population of the marrow of mb1-Cre mice that were Rptor^+/+, Rptor^f/+, and Rptor^f/f. Quantitative RT-PCR analyses based on the RNA of these cells showed that Il-7ra mRNA was lower in the haploinsufficient B cells than in controls (Fig. 2F), whereas Rptor^Δ/Δ cells had a profound reduction. To test if the reduced Il-7ra mRNA was functionally significant, we measured proliferative responses of BM cells to IL-7 and found these were reduced for B-lineage cells heterozygous for Rptor disruption (Fig. 2G). As expected, the defect of IL-7–induced proliferation was profound for BM precursors homozygous for loss of Raptor. We conclude that mTORC1 haploinsufficiency in B-cell precursors reduced both Il-7ra mRNA and the ability of these cells to respond fully to IL-7.

Along with other factors, IL-7 signaling promotes the efficiency of RAG function in selecting V regions for recombinational assembly with D-J or J exons. One feature of the process is that within the Ig loci, more promoter-proximal V regions tend to be selected first, with lower representation of more distal regions as RAG function or time are restricted (49–51). The observed decrease in proliferative response to IL-7 prompted us to analyze the repertoire of Ig heavy-chain variable (V_H) regions selected by V-DJ recombination (Fig. 2H). Display of the mean percentages of analyzed V_H region representation within the total pro-B-cell population as a heat map of relative abundance across the locus of individual mice revealed striking differences between Rptor^Δ/Δ samples and the WT controls. However, multiple alterations of V region repertoire also were reproducibly observed even in the haploinsufficient samples (Fig. 2H and Supplemental Fig. S2B, C). All together, these findings indicate that developmental progression along the B lineage and the preimmune BCR repertoire were altered as a consequence of a partial reduction in mTORC1 activity.

Full mTORC1 activity is required for normal emergence of transitional B cells in the periphery

Mice whose B lineage developed with reduced or absent Raptor (Rptor^Δ/+ or Rptor^Δ/Δ, respectively) revealed that B-cell numbers were reduced in the haploinsufficient state and that B lymphocytes were almost undetectable in the absence mTORC1 (Fig. 3A). B-lineage cells that express an antigen receptor in the marrow continue to their developmental progression in the periphery as they seed secondary lymphoid organs such as the spleen (52, 53). These B-lineage cells undergo further maturation after passage through transitional stages at which further BCR selection modulates inherent risks of autoreactivity (54, 55). Strikingly, the frequencies of immature (CD93-expressing, i.e., AA4.1⁺) and transitional stage 1 (T1) (IgM⁺CD23^neg) B cells, whose transition to T2 stage (IgM⁺CD23⁺) is an important point of vetting (55–60) were dramatically lower in mb1-Cre mice heterozygous for the mutated Rptor allele (Fig. 3B–D). As a fraction of CD93-positive cells, the T2 subset was normal; however, because of the >7-fold difference between frequencies of AA4.1⁺ B cells in haploinsufficient mice and WT controls (Fig. 3B), this subset was diminished in the spleens of Rptor^f/+ mice. Finally, a T3 (IgM^negCD23⁺) subset is considered largely to represent anergic B cells that result from engagement of autoreactive BCRs with self antigens (61, 62). The fraction of T3 (An1) phenotype cells in the haploinsufficient mice was more than double that in the control AA4.1⁺ population (Fig. 3C, D). Nonetheless, because of the low frequency of AA4.1⁺ cells overall, in absolute terms, this anergic subset was reduced even with only 1 Rptor allele deleted (Fig. 3B). Ultimately, almost no fully mature B2 B cells of either follicular or marginal zone (MZ) phenotype were present in the Rptor^Δ/Δ mice, whereas these end products (follicular B cells and MZ B cells) were at modestly lower numbers, and normal frequencies, in the haploinsufficient state (Fig. 3A, E, F). Together, the data indicate that full mTORC1 activity is required to achieve the normal flux of precursors through stages of maturation in the periphery and for regulation of the autoreactive anergic population.

Figure 3.

mTORC1 is required for normal B-cell maturation in the periphery. A–F) Phenotypic analyses of splenic B-cell maturation of mb1-Cre⁺ mice (Rptor^+/+, Rptor^Δ/+, and Rptor^Δ/Δ, as indicated) at 6–8 wk of age. A) Total numbers of splenocytes and splenic B cells (n = 12 Rptor^+/+, 9 Rptor^Δ/+, 9 Rptor^Δ/Δ). B) Representative flow cytometric plots of mature vs. transitional B cells. C, D) Transitional subsets from within the transitional B-cell compartment, shown as percentage of events within gate for a representative sample set in the B220⁺ CD93⁺ gate (C); quantitative results for each of the mice (D). Horizontal bars represent the means ± sem; significance values are shown for statistical analyses of the null hypothesis with a Student’s t test as described in Materials and Methods. E, F) Flow cytometric analysis of MZ B cells (MZB) (IgM⁺ CD23^neg) and follicular B cells (FOB) (IgM⁺ CD23⁺) from within the mature B-cell (B220⁺ CD93^neg) compartment, shown as a representative sample set (E) and graph (F) quantifying each sample, with means (horizontal lines) and results of statistical analyses. G, H) Altered natural antibodies and accumulation of anti-dsDNA IgM in mice with a B lineage haplo-insufficient for Rptor. G) Sera were collected from 6- to 8-wk-old mice, and relative anti-PC IgM and IgG3 antibody concentrations were measured using capture by plate-immobilized PC conjugated to bovine serum albumin with serial 2-fold dilutions of sera starting at 1:50, followed by detection with horseradish peroxidase–conjugated anti-IgM or -IgG3 (n = 5 of each genotype collected from 2 independent cohorts of littermates). Error bars represent the sem. Statistical analysis was conducted by 2-way ANOVA across the curves; then unpaired Student’s t tests were calculated for each dilution with assumption of unequal distribution using the Bonferroni correction. H) IgM antibodies against dsDNA and histones were measured in serial 4-fold dilutions of sera starting at 1:50 from 32-wk-old age-matched mice (5 mb1-Cre⁺ Rptor^+/+ and mb1-Cre⁺ 6 Rptor^Δ/+). Statistical analysis was conducted by 2-way ANOVA across the curves; then unpaired Student’s t tests were calculated for each dilution with assumption of unequal distribution using the Bonferroni correction. OD_450nm, optical density at 450 nm.

An early branch point in B-cell development leads to a divergence between B1 and B2 lineages (63). Of these, B1 B cells are the predominant source of a steady-state pool of what have been defined as natural antibodies. These are thought to contribute to autoimmunity (64) as well as priming initial defenses against certain pyogenic bacteria because of a repertoire biased toward recognition of microbial coats (65). Consistent with prior work (47, 48), homozygous inactivation of Rptor eliminated B1 B cells in the peritoneum and spleen (Supplemental Fig. S2D). In contrast to the B2 lineage, in which haploinsufficiency yielded a much weaker phenotype than complete loss of Raptor, both B1a and B1b cells were drastically reduced in peritoneal washings of the Rptor loss-of-function mice (Supplemental Fig. S2D). Surprisingly, however, levels of PC-binding antibody circulating in the mice whose B cells were haploinsufficient for Raptor showed that the partial reduction in mTORC1 led to an increase in these natural IgM antibodies (Fig. 3G). Moreover, modest breaches of tolerance in the IgM repertoire were detected for anti-dsDNA and anti-histone antibodies (Fig. 3H). Thus, a partial reduction in mTORC1 activity throughout B-cell ontogeny and life in the periphery promoted an increase in natural antibodies and partially breached censorship of autoreactivity even as complete loss of function eliminated B-cell development and natural antibodies.

Differentiation normalized by a rearranged BCR transgene

Our data suggested that the dose-response curve for mTORC1 function differed for distinct cell populations. However, they also pointed both to altered BCR repertoires and likely alterations in selection during ontogeny, and suggested that efficiencies of RAG-mediated VDJ recombination were part of the mechanism for these findings. To fix the B-cell population with a single specificity and bypass the need for Ig rearrangement during development, we introgressed the HEL-specific MD4 BCR transgene into the mb1-Cre⁺ lines that were Rptor^f/f (not shown), Rptor^+/+, and Rptor^+/f. Although the rearranged BCR could increase numbers of Rptor^Δ/Δ B-lineage cells to a variable extent, they remained only a modest fraction (1–4%) of mb1-Cre, MD4⁺ controls (Supplemental Fig. S2E). In contrast, the anti-HEL BCR appeared to override the impact of haploinsufficiency on development. Thus, the MD4⁺ Rptor^Δ/+ marrow was no different from MD4⁺ (Fig. 4A, B); the deficit in B-cell production and numbers was reduced by a statistically nonsignificant magnitude (Fig. 4C), and abnormalities of the developmental profiles in spleen of MD4⁺ mb1-Cre, Rptor^Δ/+ mice were eliminated (Fig. 4D–G and Supplemental Fig. S2F). Consistent with the designation of T3-like cells as anergic mature B cells, this population was eliminated by fixing the B-cell repertoire with a nonautoreactive specificity (Fig. 4E) (62). In summary, qualitative and quantitative changes in B-cell development to the stage of a fully mature naive population in secondary lymphoid organs that were caused by a partial reduction in mTORC1 were reversed once the haploinsufficient precursors had a functional MD4 BCR.

Figure 4.

Introgression of a rearranged IgH chain eliminates the developmental phenotype in B-lineage Rptor haploinsufficiency. A, B) Effect of MD4 introgression on BM B-cell development. B-cell development in BM was analyzed by flow cytometry using cells eluted from age-matched littermate mice (9 Rptor^+/+ mb1-Cre⁺ MD4^neg; 7 Rptor^Δ/+ mb1-Cre⁺ MD4^neg; 10 Rptor+/+ mb1-Cre⁺ MD4⁺; 10 Rptor^Δ/+ mb1-Cre⁺ MD4⁺). A) Representative flow cytometric phenotyping of BM B-cell developmental stages among the 4 genotypes. B) Quantitation of the pre– and pro–B-cell frequencies in the pre– and pro—B-cell gate (B220⁺ IgM^neg) for each of the mice of indicated genotype. C) Spleen cell numbers in WT and haploinsufficient mice, expressing the MD4 BCR transgene as indicated. D–G) Impact of MD4 introgression on peripheral B-cell maturation in the spleen was assessed by flow cytometry. D, E) Representative flow cytometric phenotyping of total splenic (D) and transitional (E) B-cell compartments, with the transitional [CD93⁺ (AA4.1⁺)] compartment divided into T1, T2, and T3 (An1) subgates. F, G) Graphical compilation across all independent replicates for distribution among T1, T2, and T3 (An1) subpopulations (F) and prevalence of MZ B cells (MZB) and follicular B cells (FOB) within the mature B-cell gate (B220⁺ CD93^neg) (G). Age-matched littermate mice (n = 9 Rptor^+/+ mb1-Cre⁺ MD4^neg; 7 Rptor^Δ/+ mb1-Cre⁺ MD4^neg; 10 Rptor^+/+ mb1-Cre⁺ MD4⁺; and 10 Rptor^Δ/+ mb1-Cre⁺ MD4⁺) distributed across 4 independent replicate experiments. N.S., not significant. Error bars represent the sem. Statistical analysis was conducted using an unpaired Student’s t test with assumption of unequal distribution using the Bonferroni correction.

Decreased antibody production from B cells at lower amino acid levels or haploinsufficient for Raptor

The findings presented above prompted us to test if amino acid reductions to concentrations in the sera of patients who are malnourished affect the capacity of B cells to yield antibodies and the relationship with haploinsufficiency for mTORC1. To control for ambient conditions, we first tested B cells in vitro. When purified MD4⁺ B cells were activated by costimulation (anti-CD40) along with antigen (HEL peptide), the haploinsufficient cells divided less vigorously than MD4⁺ control B cells with normal mTORC1 (Fig. 5A). As with the population of B cells with polyclonal BCRs (Figs. 1C and 2C), B cells from the MD4 model, in which development and repertoire appeared normal, exhibited decreased phosphorylation downstream from mTORC1 in Rptor^+/Δ B cells (Supplemental Fig. S3A). The Rptor^+/Δ B-cell population with a diverse, polyclonal BCR repertoire proliferated less than Rptor^+/+ controls (Fig. 5B). Strikingly, a graded decline in proliferative capacity of WT B cells was observed as amino acid concentrations were reduced while holding all other factors constant (Fig. 5C). Activated B cells that have divided several times can give rise to secreted antibody. IgM concentrations in the culture supernatants after activation of the MD4⁺ Rptor^+/Δ B cells tended to be higher than those from parallel controls (Fig. 5D). In contrast, nontransgenic IgG1 production by B cells from MD4⁺ mice was almost completely eliminated by this reduction in mTORC1 (Fig. 5E). Haploinsufficient B cells without the MD4 BCR yielded results consistent with these findings (Fig. 5F, G). Consistent with this finding, we observed progressively reduced antibody production as amino acid concentrations were reduced in culture medium and, of note, the class-switched IgG1 isotype was more affected than IgM (Fig. 5H, I) (note the results at 0.3× standard amino acid concentrations). Although a polyclonal repertoire of haploinsufficient B cells could divide multiple times and yield IgM-secreting progeny, the frequency of antibody class switching to IgG1 was substantially reduced (Fig. 5J and Supplemental Fig. S3B–J). In summary, antibody production and efficient Ig heavy chain switching to IgG can require full mTORC1 activity, all of which are attenuated by limiting amino acid concentrations.

Figure 5.

Impaired proliferation and function of Rptor haploinsufficient B cells in vitro. A) CTV partitioning of Rptor^+/+ and Rptor^Δ/+ mb1-Cre⁺ MD4⁺ B cells that were unstimulated (No Stim) or cultured (4 d) in the presence of HEL peptide and anti-CD40. Top: representative flow plots of the CTV partitioning (1 experiment’s results, representative of 3 independent replications; 1 mouse of each genotype per experiment). Below: proliferation index derived from CTV partitioning, calculated using FlowJo. Each dot represents an individual biologic sample across 3 biologically independent replicate experiments. Bars represent means ± sem. B) CTV partitioning of polyclonal (MD4^neg) B cells from Rptor^+/+ and Rptor^Δ/+ mb1-Cre⁺ mice analyzed 4 d after activation (LPS). Display of upper and lower plots is as in A. C) CTV partitioning of WT B cells, activated and cultured as in B except that cultures were in fractional concentrations of amino acid–sufficient RPMI medium as indicated. Display of upper and lower plots is as in A. D–I) Results of ELISA measuring IgM (D, F, H) and IgG1 (E, G, I) antibodies in culture supernatants at d 5 after activation of B cells of the indicated genotypes [Rptor^+/+ or Rptor^+/Δ (D, E), MD4⁺ (F, G), MD4^neg] or, for WT B cells, at the indicated fractional concentrations of amino acids in the medium (H, I). D–G) Each titration curve represents analysis of an independent sample from the replicate experiments. H, I) Each curve plots means ± sem from the replicate experiments. Results of the main statistical comparisons in linear portions of the ELISA data are tabulated in Supplemental Table S1. J) Quantification of in vitro class switching from IgM to IgG1 was assessed by culturing B cells purified from mb1-Cre⁺ mice (Rptor^+/+ and Rptor^Δ/+) with BAFF, LPS, and IL-4 for 5 d in indicated fractions of amino acid–sufficient RPMI. Each dot represents an individual mouse B-cell sample across 3 biologically independent replicate experiments with bars representing the means ± sem. *P < 0.05 for the null hypothesis. OD_450nm, optical density at 450 nm; N.S., not significant.

For in vivo analyses of functional capabilities of the B-cell population when mTORC1 was decreased, we used an adoptive transfer model and the MD4 BCR transgenic system because it normalized developmental issues and because its function in vitro was similar to that of a polyclonal population. A CD4 T-cell population enriched for recognition of an ovalbumin-derived peptide was mixed with equal numbers of B cells that were mb1-Cre, MD4⁺ and of each genotype for Rptor, and then transferred into Rag2^−/− mice with no lymphocytes (Fig. 6A). After transfer, recipient mice were immunized with a covalent conjugate peptide, HEL-OVA, to link the B- and T-cell epitopes, and primary immunity was analyzed after a single boost. Although overall splenic cellularity was similar in the RAG-deficient recipients of the transferred cells (Fig. 6B), the heterozygous (Rptor^+/Δ) B cells were at a disadvantage compared with controls (Fig. 6C, D). In contrast, the difference in overall B lymphocyte numbers (<2-fold, and failing to exclude the null hypothesis) (Fig. 6D) was striking when B cells with a GC phenotype [IgD^neg, CD95⁺, GL7 mAb, activated B cell marker GL7⁺)] were enumerated (Fig. 6E, F). Consistent with a relatively modest effect on B-cell numbers, haploinsufficiency of Raptor had little effect on the frequencies of cells in the spleen (Fig. 6G) and marrow (Fig. 6H) that secreted IgM anti-HEL. The MD4 transgene, which lacks IgH regions essential for antibody class switching, allows emergence of some B cells expressing a functional endogenous receptor that can switch. However, the relative steady-state IgM anti-HEL concentrations in sera of mice with haplo-insufficient B cells increased to a modest extent (Fig. 6I). Measurements of switched cells found very few antibody-secreting cells that produced the IgG1 isotype (Fig. 6G, H) and substantially less IgG1 anti-HEL (Fig. 6I) when B cells had reduced mTORC1. We conclude that there were substantial effects of haploinsufficiency on the generation of GC and antibody-secreting cells, despite the capacity of the MD4 BCR to normalize differentiation and achieve normal B-cell numbers in steady-state conditions.

Figure 6.

Rptor haploinsufficient B cells exhibit defective activation and altered humoral responses. A) Schematic of experiments for in vivo testing of MD4 B cells with reduced mTORC1. Thy-1.2–depleted B cells from mb1-Cre⁺ mice (Rptor^+/+ and Rptor^Δ/+) were transferred into Rag2^−/− recipients along with CD4⁺ T cells [nontransgenic and from OTII T-cell receptortransgenics as described in Lee et al. (39)]. Recipients immunized with HEL-OVA peptide received booster immunization 3 wk thereafter, 1 wk prior to harvest. B) Mean ± sem numbers of splenocytes recovered from immunized recipient mice. C, D) Representative flow plot for B-cell gating strategy in splenocyte cell suspensions (C, left), along with quantitation of the individual and mean ± sem frequencies (C, right) and numbers (D) of recovered B cells in recipient mice. E, F) Reduced GC responses of Rptor haploinsufficient MD4 B cells after immunization. A representative flow cytometric result in the viable B-cell gate (E, left) and quantitative graph (E, right) of GC B-cell (GCB) responses in Rptor^+/+ and Rptor^Δ/+ mb1-Cre⁺ recipient mice as well as calculated total number of splenic GCBs (F). G–I) Intact mTORC1 signaling is required for normal humoral response to immunization. Splenocytes (G) and BM cells (H) were used to quantify frequencies of HEL-specific antibody secreting cells by ELISpot. I) Terminal sera were collected at time of harvest, and circulating anti-HEL IgM and IgG1 antibodies were measured by ELISA using plate immobilized HEL (46–61) peptide. For these analyses, recipients (11 Rptor^+/+ mb1-Cre⁺ and 10 Rptor^Δ/+ mb1-Cre⁺) were distributed across 3 independent replicate experiments. Statistical analysis for flow phenotypic analyses was conducted with an unpaired Student’s t test with assumption of unequal distribution using the Bonferroni correction. For ELISA titration curve data, a 2-way ANOVA was used to establish statistical significance across the curve, and then an unpaired Student’s t test was applied at each dilution for assessment of significance with assumption of unequal variance applied by the Bonferroni correction. ASC, antibody secreting cell; GL7, GL7 mAb, activated/GC B cell marker; N.S., not significant; OD_450nm, optical density at 450 nm.

DISCUSSION

Regulation of most biologic processes operates in ranges that yield dose-response curves as physiologic homeostasis typically prevents extremes. The advent of gene targeting and its extension to systems that allow conditional disruption or loss-of-function alleles continues as a rich vein mined for insights into requirements for signal transduction proteins and transcription factors. However, physiologic perturbations in general and malnutrition in particular cause only partial decreases in the function of signaling complexes such as mTORC1 (31, 32). Nutritional deficiencies arise in humans not only from poverty or geographic isolation but also because of postsurgical states and eating disorders (1–5). Altered production of B cells and weaker immune function in human populations have been linked to these complex deficiencies (3, 14). More readily controlled experimental models in rodents have linked protein-deficient diet to partial reductions in mTORC1 activity and in circulating amino acids that modulate activity of this signal transducing kinase complex (15, 31, 32). However, even these important findings are associated with the potential for indirect effects such as changes in endocrine factors. At a nonphysiologic extreme, complete loss of Raptor has been analyzed separately for development of B cells and their establishment of potentially functional population in the periphery (47, 48) and for the function of B cells in which inactivation of mTORC1 was delayed until after production of an apparently normal B-cell compartment (66–68). Here, we showed that a partial reduction in Raptor and hence mTORC1, limited to the B lineage, led to cell-autonomous alterations in B-cell development and function. These findings indicate that the previously documented partial decrease in mTORC1 activity could be the key mechanistic basis for weaker humoral immunity. This finding is consistent with a model in which insufficient nutrition may directly affect B cells because of lower amino acid concentrations in their extracellular milieu. Intriguingly, the developmental stages impacted by cell-intrinsic haplodeficiency were those most dependent on BCR signaling (57–59).

Previous work showed that complete loss of Raptor (homozygous deletion) prior to lineage specification caused an absolute lack of B-cell maturation (47, 48, 69). However, activation, differentiation into antibody-secreting cells, and production of IgM antibodies could occur despite lack of mTORC1 activity in mature B cells that had developed normally (66–68). In the analyses presented here, we confirmed that homozygous inactivation of the Rptor gene from the earliest point of B-cell identity eliminated both B1 and B2 lymphocyte lineages. Strikingly, haploinsufficiency had a more substantial impact not only on the steady-state prevalence of B1a and B1b B cell in the peritoneal cavity than was observed for follicular B2 populations but also on MZ-phenotype B2 B cells in the spleen. An apparent paradox is that these B-cell subsets are the main source of natural and steady-state antibodies in the serum (70), and the sera of the haploinsufficient mice had normal or increased IgM. This may be caused by one or more of several mechanisms at this level of mTOR signaling: an increase signaling by PI3K when the negative feedback from mTORC1 is lower, a greater amount of stimulation from altered microbiota or self-antigens, development of T2 B cells into antibody secretors, specifics of the preimmune BCR repertoire (i.e., better retention of V_H that gives rise to the anti-PC or polyreactive antibodies), or an acceleration of plasma cell development from B1 cells or their rate of antibody secretion. In any case, the finding is akin to observations transgenic models that lack B1 or MZ B cells, or both, and nonetheless have natural antibodies (e.g., MD4). Analytic tools are lacking for assessing some of these possibilities, but in any case, the findings provide evidence that the effect of a partial reduction in mTORC1 activity is substantially different for distinct biologic endpoints in B-lineage development and function.

Early in ontogeny, the partial reduction in mTORC1 led to reduced Il-7ra gene expression and IL-7 responsiveness, along with altered repertoire of V_H gene segments, followed by defects in subset distribution that were reversed by provision of rearranged heavy and light chains that yield a high-affinity BCR. These findings indicate that even a partial decrease in mTORC1 reduced Il-7ra gene expression, indicating that the receptor level is titrated between almost nil when Raptor is absent (47, 48) to intermediate range when mTORC1 is reduced by haploinsufficiency. Although IL-7Rα is essential for lymphocyte development because of its functions in receptors for IL-7 and the related cytokine, thymic stromal lymphopoietin analyses of mice with inactivated Il-7ra alleles used heterozygotes as WT controls (71, 72), suggesting that the reduction observed with a Rptor^Δ/+ B lineage would be unlikely to account for the reduced proliferation we observed. Reversion experiments using normal and mutated IL-7Rα cytoplasmic tails concluded that IL-7–induced PI3K activation was essential for proliferation of early progenitors (71, 73). In contrast, even with complete loss of Raptor, IL-7–induced phosphorylation of signal transducer and activator of transcription 5 (STAT5) at the activating tyrosine was normal (48). STAT5 inhibits expression of the RAG proteins, which is thought likely to delay the onset of light chain gene rearrangement (74). Accordingly, lower receptor induction of STAT5 would not account for the altered repertoire of V_H gene segments observed in the Rptor haploinsufficient B lineage as less STAT5 induction would favor greater gene rearrangement. These findings indicate that neither the reduced proliferation to IL-7 nor the spectrum of recombination were caused by the lower receptor level. Instead, the data provide evidence that even such a partial reduction in mTORC1 suffices to alter accessibility and proliferation. However, the drop we observed in pre-B cells—especially in the competitive repopulation experiments—might, in part, involve a subtle decrease in STAT5 nuclear induction. Furthermore, the capacity of MD4 to suppress a deficiency of T1-stage transitional B cells indicates that mTORC1 activity level required for progression depends on BCR-complex signaling during peripheral maturation.

At the later stage of mature B cells and their capacity to yield antigen-specific antibodies, the combination of our in vitro and adoptive transfer data provide evidence of a B lymphocyte-intrinsic requirement for mTORC1 activity greater than what can be achieved in the haploinsufficient state. Though there may be some reduced proliferation of the total population of Rptor^+/Δ B cells in situ, it is noteworthy that for the high-affinity IgM BCR MD4 there was no reduction in the frequencies of Ag-specific antibody secreting cells (ASCs) or the concentrations of serum antibodies deriving from the B cells with reduced Raptor. Accordingly, we consider it unlikely that a global defect of cell cycling in vivo is the main basis for the observed defect of generating a GC B-cell population. An implication in the setting of chronic protein deprivation with hypo-amino-acidemia is that over time, the high-affinity and high-memory B-cell populations might be reduced.

A caveat relating to immunity in the setting of malnutrition is posed by findings pertaining to mTORC1 in the CD4 T-cell subset specialized to provide help within the follicles and their GCs. Lower circulating amino acids or even short-term dietary restriction decreases mTORC1 in T cells, in part because of low leptin levels (75). Full mTORC1 activity elicited by IL-2 at a late time postactivation in an anti-viral response inhibited the peak of follicular help (76), suggesting that partially decreased mTORC1 might increase a GC reaction. Alternatively, however, complete lack of Raptor in naive T cells, or its loss soon after initial T-cell activation, led to defects in follicular help, GC B cells, and antibody responses, as did restriction of the uptake of glucose (77, 78). Thus, the collective work suggests an impact of metabolism on T cell help to B cells in regulation of humoral responses, but CD4 T-cell type–specific partial reductions of Raptor and analyses of antibodies after immunization will be required to determine how the mTORC1 reduction in malnutrition will affect the T-cell component of immunity. Notwithstanding the need for such further analyses, our results with the B lineage have revealed very different dose-response curves for mTORC1 depending on the biologic end point.

Glossary

7-AAD

7-aminoactinomycin D

BAFF

B-cell activating factor

BCR

B-cell receptor

BM

bone marrow

BrdU

bromodeoxyuridine

CTV

CellTrace Violet

dsDNA

double stranded DNA

ELISpot

enzyme-linked immune absorbent spot

FBS

fetal bovine serum

GC

germinal center

HEL

hen egg lysozyme

HEL-OVA

HEL conjugated to ovalbumin aa 323–339

IL-7Rα

IL-7 receptor α

mTOR

mechanistic target of rapamycin

mTORC1

mTOR complex 1

MZ

marginal zone

PC

phosphorylcholine

Rag2

recombination activating gene 2

Raptor

regulatory-associated protein of mTOR

RPMI

Roswell Park Memorial Institute

Rptor

regulatory-associated protein of mTOR

STAT5

signal transducer and activator of transcription 5

V_H

heavy-chain variable

WT

wild type

AUTHOR CONTRIBUTIONS

A. L. Raybuck, K. Lee, S. H. Cho, J. Li, J. W. Thomas, and M. R. Boothby conceived of, designed, and interpreted experiments and their data; A. L. Raybuck, K. Lee, S. H. Cho, and J. Li performed experiments; and A. L. Raybuck, J. W. Thomas, and M. R. Boothby wrote the manuscript.

Supplementary Material

This article includes supplemental data. Please visit http://www.fasebj.org to obtain this information.