mTOR Activation Underlies Enhanced B Cell Proliferation and Autoimmunity in Prkcd^G510S/G510S Mice Free

Systemic lupus erythematosus (SLE) is a devastating autoimmune disease whose symptoms may include joint pain, fever, skin rash, and organ damage. SLE is currently incurable, and the therapeutic arsenal to treat SLE remains very limited, without specific treatment for the different clinical forms of the disease. SLE usually occurs in adults and less frequently in children (juvenile-onset SLE [jSLE]), but in the latter case it is usually more severe. In jSLE, genetics is thought to be a major contributing factor to the onset of the disease, and a subset of jSLE is driven by inborn errors of a single gene, defining the concept of monogenic SLE (1). Using next generation sequencing, we recently found that at least 7% cases of jSLE are in fact monogenic (1). Gene mutations may alter different immunological pathways, such as the clearance of apoptotic bodies (i.e., efferocytosis), the production of type I IFN (IFN-I), or B cell tolerance (2).

Mechanistically, SLE is caused by an aberrant immune response resulting in the production of inflammatory cytokines, such as IFN-I, and autoantibodies directed against components of apoptotic bodies (nucleosomes, chromatin, histones, etc.), as well as the formation and deposition of immune complexes responsible for tissue inflammation and damage (3). Autoantibodies identified in SLE are generally high-affinity, somatically mutated IgG, which suggests they have arisen in germinal centers (GCs), where T cells provide help to B cells for class switching and somatic mutation (3). They are produced by mature B cells, differentiated into plasma cells, secondary to a break in B cell tolerance. B cell tolerance ensures that B cells expressing BCRs with the ability to recognize self-antigens are either modified (BCR editing) or eliminated to avoid the development of autoimmunity. In particular, strong BCR signaling on engagement by self-antigens induces apoptosis of developing B cells (4). Consequently, mutations in genes affecting the strength or the quality of BCR signal transduction can cause B cell tolerance breakdown and autoimmunity, as previously reviewed (4, 5).

We and others previously identified PRKCD mutations in jSLE patients (6–12). PRKCD encodes for protein kinase C-δ (PKC-δ), a cytosolic kinase of the novel diacylglycerol-dependent and calcium-independent serine/threonine PKC group. In our own study, we described a homozygous mutation (c.1528G>A, p.G510S) of the PRKCD gene in three jSLE patients from the same family (6). The causality of these mutations in SLE was supported by genetic analyses and by previous observations of a lupus phenotype in Prkcd^−/− mice (13, 14). Moreover, multiple other mutations in PRKCD were associated with jSLE since the original observations (8–12). The G510S missense mutation affects a highly conserved amino acid of the protein activation loop and is associated with a low, but not abrogated, PKC-δ expression (6). We also demonstrated that the mutated protein displays a reduced activity, and that it was refractory to stimulation by PMA. PKC-δ is broadly expressed and plays multiple roles in cells of the innate and adaptive immune systems. For example, PKC-δ regulates neutrophil migration and oxidative burst (15, 16) and also influences macrophage effector functions (17). T cell proliferation and cytokine production are slightly increased in Prkcd^−/− mice after TCR triggering (13). In B cells, PKC-δ is phosphorylated after BCR engagement (18, 19) and is involved in BCR signaling. A previous study proposed that PKC-δ amplified the Ca²⁺-dependent Ras/Erk pathway, thereby inducing apoptosis of autoreactive transitional B cells (20). Another study showed that PKC-δ translocation into the nucleus was a proapoptotic event that could be prevented by BAFF signaling (21). PKC-δ nuclear translocation was recently shown to occur in GC B cells, downstream of sphingomyelin synthase 2 (22). Overall, several lines of evidence point to PKC-δ as an important regulator of B cell tolerance, and the proposed mechanism involved the induction of apoptosis in immature cells. Yet, most studies have used Prkcd^−/− mice as a disease model, which may not adequately model SLE pathogenesis in patients with PRKCD mutations. To address this issue, we generated a knock-in mouse strain bearing the Prkcd G510S mutation. Our results showed that Prkcd^G510S/G510S mice developed a strong autoimmune phenotype recapitulating the disease observed in patients. We hypothesized that B cells were the major drivers of the disease. We performed a detailed analysis of their phenotype and of their response to BCR engagement to tackle this question and to find a specific treatment for the disease.

Prkcd^G510S/G510S mice were generated using CRISPR-Cas9 gene editing. guide RNA and single-stranded oligodeoxynucleotides were designed using CRISPOR (Tefor) and Mainworkbench softwares (Qiagen) and were synthesized by Eurogentec. Mouse zygotes were electroporated with ribonucleoprotein complex in our animal facility (Plateau de Biologie Expérimentale de la Souris [PBES]), as previously described (23). Littermate mice were backcrossed three times to C57BL/6 mice, purchased from Charles River Laboratories (L’Arbresle). Mice were bred and maintained under specific pathogen-free conditions in the PBES. Age-matched (8–20 wk old) and sex-matched littermate mice were used as controls. This study was carried out in accordance with the French recommendations in the Guide for the Ethical Evaluation of Experiments Using Laboratory Animals and the European guidelines 86/609/CEE. The local bioethics committee (Comité d’Éthique en Expérimentation Animale de la Région Rhône-Alpes) approved all experimental studies.

For the generation of mixed bone marrow chimeras, donor-derived BM cells of the indicated genotypes (CD45.2⁺) were mixed with BM cells from Ly5a (CD45.1) mice and transferred i.v. into sublethally (10 Gy) irradiated Ly5axC57BL/6 (CD45.1, CD45.2) recipients. Mice were analyzed 8–12 wk after BM reconstitution. For the generation of μMT: Prkcd^+/+ and μMT: Prkcd^G510S/G510S mixed BM chimeras, a 4:1 ratio of donor-derived BM cells from μMT mice and Prkcd^+/+ or Prkcd^G510S/G510S mice was transferred i.v. into sublethally (10 Gy) irradiated Ly5axC57BL/6 (CD45.1, CD45.2) recipients.

Mice were treated with either vehicle: PBS 1× (Life Technologies) or Rapamune (rapamycin; Pfizer) (2 mg/kg/d) by gavage during 28 d. Blood sampling from the retro-orbital sinus was performed before and after treatment to analyze antinuclear autoantibodies and IFN-stimulated gene (ISG) expression.

Antinuclear Abs (ANAs) reactivity were detected by indirect immunofluorescence staining of HEp-2 cell slides (Bio-Rad) using sera from indicated mice at a 1:50 dilution and probed with FITC-conjugated goat anti-murine IgG F(ab′)₂ (ImmunoJackson Research). All images were obtained with a Nikon Eclipse T&2R (Nikon) microscope and were analyzed using ImageJ (National Institutes of Health). For each experiment, the fluorescence intensity of 10 nuclei in each micrograph was quantified and normalized to the value obtained using an ANA-positive serum.

Total murine IgG or IgM serum levels were measured as follows. In brief, 96-microwell plates (Nunc Maxisorp; ThermoFisher) were coated with purified goat anti-mouse IgG or IgM (Jackson ImmunoResearch) in carbonate buffer (Na2CO3 0.1 M, NaHCO3 0.1 M, pH 9.5) for 1 h at 37°C, and total IgG or IgM were detected with HRP-conjugated goat anti-mouse γ or µ (Jackson ImmunoResearch). Levels of Ig were determined by comparison with a standard curve using purified IgG or IgM (Jackson ImmunoResearch). Serum anti-dsDNA and rheumatoid factor (RF) activities were analyzed as follows. In brief, 96-microwell plates (Nunc Maxisorp; ThermoFisher) were coated with purified Calf-Thymus DNA at 100 ng/ml (Sigma-Aldrich) in Tris-HCl buffer (Tris HCl 50 mM, EDTA 10 mM, pH 7.5) for 1 h at 37°C. Then, ssDNA was digested with S1 nuclease at 0.001 U/μl (Promega) in acetate buffer (50 mM Na acetate, 50 mM NaCl, 1 mM ZnSO4, 5% glycerol, pH 4.5) for 1 h at 37°C. For RF detection, 96 microwell plates (Nunc Maxisorp; ThermoFisher) were coated with purified mouse IgG at 50 μg/ml (Jackson Immunoresearch) in carbonate buffer (Na2CO3 0.1 M, NaHCO3 0.1 M, pH 9.5) for 1 h at 37°C. Serum samples were incubated at several dilution overnight at room temperature. Plates were then incubated with HRP-conjugated goat anti-mouse γ or µ (Jackson ImmunoResearch) for 1 h at 37°C. For all ELISA tests, the wells were developed with Fast Ortho-Phenylene Diamine substrate (Sigma-Aldrich). The reaction was stopped with addition of H2SO4 solution (1 M) after 20 min at room temperature, and absorbance was measured at 490 nm.

RNA was extracted from whole blood with Direct-zol RNA MicroPrep (Ozyme), and cDNA was generated using the high-capacity RNA-to-cDNA kit (Applied Biosystems). PCR was carried out with a FastStart Universal SYBR Green Master (Roche) on a StepOne plus instrument (Applied Biosystems). Primers were designed using the Roche software. The following primers were used for quantitative PCR: MM-ISG15-F, 5′-AAGCAGCCAGAAGCAGACTC-3′; MM-ISG15-R, 5′-CACGGACACCAGGAAATCGT-3′; MM-IFIT1-F, 5′-AAGGCACTGAACAACAAGACC-3′; MM-IFIT1-R, 5′-TGACTGCTTATGAAACTGTAGGAAA-3′; MM-RSAD2-F, 5′-GTGGACGAAGACATGAATGAAC-3′; MM-RSAD2-R, 5′-CTCAATTAGGAGGCACTGGAA-3′; MM-GAPDH-F, 5′-GCATGGCCTTCCGTGTTC-3′; MM-GAPDH-R, 5′-TGTCATCATACTTGGCAGGTTTCT-3′. The expression of ISGs was normalized to the level of GAPDH, and the IFN score was calculated as the median of the normalized count of the three ISGs.

Kidneys were fixed in 10% neutral buffered formalin and embedded in paraffin. Kidney sections were cut at 5 μm thickness for H&E staining (Sigma-Aldrich). All images were obtained with an Axioimager Z1 microscope (Zeiss) and Coolsnap camera using MetaMorph software (Molecular Devices).

Urine was collected from mice and deposited on an Albustix strip (Siemens). Proteinuria score was determined as follows: 0 = negative; 1 = trace; 2 = 30 mg/dl; 3 = 100 mg/dl; 4 = 300 mg/dl; and 5 = >2000 mg/dl.

Mouse kidneys were embedded in OCT and snap frozen over liquid nitrogen. Six-micrometer thin sections were cut on a cryostat, mounted on SuperFrost Plus Gold slides (Thermo Fisher Scientific), and fixed in cold acetone for 20 min. Kidney sections were saturated with BSA 1% and donkey serum 5% blocking solution during 30 min, then with BSA 1% blocking solution containing anti-CD16/CD32 (BD Biosciences) during 30 min and were stained using FITC-anti-IgG (Jackson Immunoresearch) or FITC-anti-IgM (Jackson Immunoresearch). The images of stained kidney sections were captured using Axioimager Z1 microscope (Zeiss) and Coolsnap camera using MetaMorph software (Molecular Devices) and were analyzed using ImageJ (National Institutes of Health).

Cells were lysed in radioimmunoprecipitation lysis buffer (25 mM Tris, HCl, pH 7.4, 150 mM NaCl, 1% Nonidet P-40, 1% sodium deoxycholate, and 0.1% SDS containing protease and phosphatase inhibitors; Sigma-Aldrich) for 30 min at 4°C. Supernatant was collected after centrifugation at 16,000 × g, 4°C, and protein content was quantified using the μBCA quantification kit (Thermo Fisher Scientific). PKC-δ protein expression was detected using the automated Jess Simple Western system (ProteinSimple). In brief, cell lysates were diluted to 0.5 µg/µl in 0.1× sample buffer and Fluorescent 5× Master mix containing 400 mM DTT (ProteinSimple). Samples were denatured at 95°C for 5 min, and proteins were separated in capillaries using a 12- to 230-kDa Jess separation module and following the manufacturer’s standard method (ProteinSimple). Anti–PKC-δ (Cell Signaling Technology) was used for protein detection and revealed using HRP-coupled secondary Abs (ProteinSimple). Digital image of chemiluminescence of the capillary was analyzed using Compass Simple Western software (version 4.1.0; Protein Simple) that calculated automatically height (chemiluminescence intensity), area, and signal/noise ratio. PKC-δ level was normalized on total protein expression, using the JESS protein Normalization assay and according to manufacturer’s instructions.

Single-cell suspensions of BM, lymph node (LN), and spleens were stained with Ab cocktails. The following mAbs from eBioscience, BD Biosciences, or BioLegend were used for flow cytometry: anti-CD19 (6D5), anti-B220 (RA3-6B2), anti-CD3 (145-2C11, 500A2), anti-CD8a (53-6.7), anti-CD4 (GK1.5), anti-NK1.1 (PK136), anti-CD11b (M1/70), anti-CD11c (N418), anti-IgM (R6-60.2), anti-IgD (11-26c.2a), anti-IgG1 (A85-1), anti-CD21 (7E9), anti-CD23 (B3B4), anti-CD24 (M1/69), anti-CD27 (LG.7F9), anti-CD45.1 (A20), anti-CD45.2 (104), anti-CD64 (X54-5/7.1), anti-CD69 (H1.2F3), anti-CD86 (GL1), anti-CD95 (Jo2), anti-CD138 (281-2), anti-CD317 (927), anti-TACI (8F10), anti-MHC class II (MHCII; M5/114.15.2), anti-GL7 (GL7), anti-KI67 (solA15), anti-Ly6C (HK1.4), anti-Ly6G (RB6-8C5), anti–Siglec F (E50-2440), anti–Siglec H (440c), anti-F4/80 (T45-2342), and anti-CD107a (1D4B). Intracellular stainings and stainings for phosphorylated proteins were performed using the Foxp3 fixation/permeabilization or Lyse/Fix and Perm-III kits (BD Bioscience). The following Abs from eBioscience, BD Biosciences, BioLegend, and Tube-Bio were used for intracellular staining and phospho-specific flow cytometry: anti–P-Syk (pY348) (I120-722), anti–P-Btk (pY223) (N35-86), anti–p-Akt (pS473) (M89-61), anti–p-Erk (T202/pY204) (201), anti–p-S6 (p240/244) (D57.2.2E), anti–P-4E-BP1 (Thr37/46) (236B4), anti–T-bet (4B10), and anti–phosphatidylinositol (3,4,5)-trisphosphate (PIP3) (Z-P345). Cell viability was measured by staining with the fixable viability dye eFluor 506 or Near IR (Thermo Fisher Scientific). Flow cytometry was carried out on a FACS Fortessa (Becton-Dickinson). Data were analyzed using FlowJo (Treestar).

A total of 2 × 10⁶ splenocytes were cultured in complete medium (RPMI-1640 supplemented with 10% heat-inactivated FCS, 10 mM Hepes, 1 mM sodium pyruvate, and 50 μM 2-ME). After 48 or 72 h, B cells viability was assessed by flow cytometry.

Splenocytes were labeled with 5 μM CellTrace Violet (CTV) dye (Thermo Fisher Scientific) in PBS for 20 min at 37°C. After washing with complete medium (RPMI-1640 supplemented with 10% heat-inactivated FCS, 10 mM Hepes, 1 mM sodium pyruvate, and 50 μM 2-ME), cells were seeded at a density of 2 × 10⁶ cells in 1 ml of stimulation medium containing 10 µg/ml AffiniPure F(ab′)₂ Fragment Goat Anti-Mouse IgM, µ chain specific (Jackson Immunoresearch) and 5 µg/ml CD40 (mouse) mAb (FGK45). At day 3, proliferation of stimulated B cells was assessed by flow cytometry.

For the measurement of phosphorylated proteins, different CD45 Abs or CTV dye (Thermo Fisher Scientific) dilutions were used to barcode splenocytes from the indicated genotypes after 1 h at 4°C. Next, barcoded cells of each genotype were pooled and incubated for 15 min at 4°C with 10 µg/ml AffiniPure F(ab′)₂ Fragment Goat Anti-Mouse IgM, µ chain specific (Jackson Immunoresearch), then with 10 µg/ml Rabbit anti-Goat IgG (H+L), Superclonal Recombinant Secondary Ab (Thermo Fisher Scientific) for 15 min. In some experiments, some inhibitors were used at the following concentrations: cerdulatinib (20 nM; Sigma-Aldrich), PRT062607 (11.11 nM; Sigma-Aldrich), wortmannin (37 nM; Sigma-Aldrich), and rapamycin (1.85 nM; Merck). Splenocytes were incubated at 37°C for the stimulation and fixed by addition of 10 vol of Lyse/Fix at the indicated time point.

Spleen cell suspensions were incubated with biotinylated mAbs against CD3 (14–2C11), TCRβ (H57-597), TCRγδ (GL3), CD19 (ebio1D3), and TER-119 (ter119) (eBioscience), followed by incubation with anti-biotin microbeads (Miltenyi) and NK cell enrichment by magnetic separation using the AutoMACS device (Miltenyi). Enriched NK cells were cocultured for 4 h with YAC-1 cells at 1:1 ratio calculated based on the cell number and the percentage of NK cells after purification. The expression of surface CD107a was then measured on NK cells by flow cytometry.

Follicular B cells CD3⁻CD19⁺CD45R⁺CD24^intCD21^intCD23⁺ were sorted using a FACSAria Cell Sorter (Becton-Dickinson). Purity of sorted cell populations was >98% as measured by flow cytometry. FO B cells were stimulated with 10 µg/ml AffiniPure F(ab′)₂ Fragment Goat Anti-Mouse IgM, µ chain specific (Jackson Immunoresearch) during 4 h. RNA libraries were prepared as follows. In brief, total RNA was purified using the Direct-Zol RNA microprep kit (Zymo Research) according to the manufacturer’s instructions and was quantified using QuantiFluor RNA system (Promega). The preparation of library was performed using the SENSE mRNA-Seq Library Prep Kit (Lexogen) according to the manufacturer’s instructions. Tagged library quality was checked on D1000 screen tape and analyzed on Tape station 4200 (Agilent). Sequencing was performed by the GenomEast platform, a member of the “France Génomique” consortium (ANR-10-INBS-0009), on an Illumina HiSeq 4000 sequencing machine (read length 1 × 50 nt).

Statistical analysis was performed with GraphPad Prism 6 software. We used unpaired Student t test or Mann–Whitney test to calculate statistical significance when comparing a single parameter between two experimental groups. If more than one parameter (sex or cell subsets) was compared, a false discovery rate adjustment was applied. One-way ANOVA was used when more than two experimental groups were compared, and the statistical significance was determined using Tukey’s post hoc test. Two-way ANOVA with Bonferroni multiple comparison correction was used when comparing two parameters between more than two experimental groups.

PRKCD mutations result in a highly variable clinical phenotype, which may result from differences in the level of expression or residual activity of the protein (6–12, 16, 24–26) (summarized in Supplemental Table I). To investigate the immunological mechanism underlying autoimmunity in patients with the PRKCD G510S mutation (6), we generated a knock-in mouse strain bearing the Prkcd G510S mutation using CRISPR-Cas9 genome editing (Fig. 1A). Capillary electrophoresis analysis showed that Prkcd^G510S/G510S mice had decreased expression of PKC-δ at the protein level in various lymphoid organs, such as spleen, thymus, and BM, compared with control mice (Fig. 1B), which suggests a similar impact of the PRKCD mutation on protein expression in mouse and human (6). Prkcd^G510S/G510S mice had a reduced lifespan compared with Prkcd^+/G510S and Prkcd^+/+ control mice (Fig. 1C). Indeed, all Prkcd^G510S/G510S mice died before the age of 48 wk, whereas ∼80% of control mice (Prkcd^+/G510S and Prkcd^+/+) were still alive at this age. Notably, Prkcd^G510S/G510S males died slightly earlier than Prkcd^G510S/G510S females. To address the causes of this premature death, we then carried out a clinical, biological, and histological examination of knock-in mice in comparison with controls. Spleens and LNs were significantly enlarged in Prkcd^G510S/G510S mice, especially in males (Fig. 1D, Supplemental Fig. 1A). Kidneys from Prkcd^G510S/G510S mice and not controls exhibited extensive infiltration in the glomeruli, which correlated with the presence of IgG deposit and proteinuria (Fig. 1E, 1F, Supplemental Fig. 1B). ANA, a hallmark of SLE, were already detected at 2 mo of age in Prkcd^G510S/G510S mice (Fig. 1G). Moreover, ELISA assays revealed the presence of anti-dsDNA IgG, anti-dsDNA IgM, and RF in the serum of Prkcd^G510S/G510S mice, and their levels increased with time except for anti-dsDNA IgM (Fig. 1H, Supplemental Fig. 1C). In addition, total IgG and IgM concentrations were strongly increased (Fig. 1I). We also monitored the expression of several ISGs in blood cells, because the production of IFN-I is often elevated in SLE patients. We calculated an IFN score, following a method previously used for human cells (27). This score was significantly higher in Prkcd^G510S/G510S mice compared with controls (Fig. 1J). In addition, we found no difference between males and females in features related to autoimmunity i.e., autoantibodies, Ig levels, and IFN score (Supplemental Fig. 2A, 2B).

Overall, Prkcd^G510S/G510S mice develop an early-onset and severe autoimmune phenotype combining lymphoproliferation, kidney failure, various autoantibodies, and positive IFN score, leading to premature death. This phenotype recapitulates clinical features observed in SLE patients bearing the PRKCD G510S mutation, validating our translational approach.

To gain insight into the immunological mechanism leading to SLE in Prkcd^G510S/G510S mice, we performed a large-scale flow cytometry analysis of immune subsets in these mice in comparison with controls. Because Prkcd expression is particularly high in myeloid cells (Supplemental Fig. 3), we first focused our analysis on innate immune cells. Macrophages, neutrophils, eosinophils, plasmacytoid dendritic cells (DCs) and conventional DCs were observed at similar frequencies in the spleen of Prkcd^G510S/G510S mice and control mice, and no further investigations were performed for these subsets (Fig. 2A). By contrast, the frequency of NK cells was reduced 3-fold in the spleen of Prkcd^G510S/G510S mice compared with controls. The absolute number of NK cells was similar in both mouse genotypes because of the splenomegaly in Prkcd^G510S/G510S mice (Supplemental Fig. 4A). Nevertheless, we observed that Prkcd^G510S/G510S NK cells had a bigger size and higher Ki67 expression than control NK cells (Supplemental Fig. 4B, 4C). Furthermore, Prkcd^G510S/G510S NK cells displayed defective maturation with an increase of immature CD11b⁻CD27⁺ cells and a decrease of mature CD11b⁺CD27⁻ cells (Supplemental Fig. 4D). We stimulated spleen cells with YAC-1 cells that are classical NK cell targets. We observed higher cell surface expression of the cytotoxic marker CD107a cell surface expression after stimulation in Prkcd^G510S/G510S NK cells than in control NK cells (Supplemental Fig. 4E). To test whether this NK cell phenotype was due to a cell-intrinsic impact of the Prkcd mutation or to the inflammatory environment, we generated mixed BM chimeric mice by reconstituting lethally irradiated Ly5a × C57BL/6 (CD45.1 × CD45.2) with a 1:1 mixture of BM from Prkcd^G510S/G510S or Prkcd^+/+ (CD45.2) mice and Ly5a mice (CD45.1), taking advantage of CD45.1- and CD45.2-specific Abs to identify cells of the different genotypes. In wild-type (WT)/Prkcd^G510S/G510S BM chimeric mice, a lower percentage of NK cells was found in spleen cells from the Prkcd^G510S/G510S genotype compared with the WT genotype (Supplemental Fig. 4F), and their maturation was also impaired (Supplemental Fig. 4G). However, the degranulation capacity of Prkcd^G510S/G510S NK cells was similar to that of WT counterparts in mixed BM chimeras (Supplemental Fig. 4H). Altogether, these results show a cell-intrinsic impact of the Prkcd^G510S/G510S mutation on NK cell development and homeostasis, whereas the impact on degranulation capacities is probably cell extrinsic.

Next, we examined adaptive lymphocyte features in Prkcd^G510S/G510S mice in comparison with controls. As shown in Fig. 2A, we observed a decreased frequency of CD8 T cells and a normal frequency of CD4 T cells in the spleen of Prkcd^G510S/G510S mice compared with controls. The proportion of B cells was similar in the BM and spleen of Prkcd^G510S/G510S and control mice, but this proportion was higher in the LNs of Prkcd^G510S/G510S mice than in controls. B cell size was higher in Prkcd^G510S/G510S mice than in controls, irrespective of the organ analyzed (Fig. 2B), suggesting ongoing activation. In agreement with this hypothesis, peripheral Prkcd^G510S/G510S B cells expressed significantly higher levels of activation markers CD69, MHCII, and CD86 and proliferation marker Ki67 than control B cells (Fig. 2C, Supplemental Fig. 5A). We then performed a detailed analysis of immature and mature B cell subsets. The frequencies of precursor, immature, transitional, and mature B cells in the BM were similar in Prkcd^G510S/G510S and control mice (Fig. 2D), and so were the proportions of transitional and follicular B cells in the spleen of both strains (Fig. 2E, 2F). However, marked differences were observed for other B cell subsets. In particular, the proportion of marginal zone (MZ) B cells was strongly reduced, whereas the proportions of GC and plasma cells (PC) were elevated in the spleen of Prkcd^G510S/G510S compared with control mice (Fig. 2F, 2G). We then inspected aging-associated B cells (ABCs) defined as CD3⁻CD19⁺B220⁺IgM⁺CD11b⁺CD11c⁺T-bet⁺. This population is known to expand with age and in SLE patients (28). We observed an elevated proportion of these cells in the spleen of Prkcd^G510S/G510S mice (Fig. 2G). Notably, PC and MZ B cells had an increased Ki67 expression in Prkcd^G510S/G510S mice compared with controls (Supplemental Fig. 5B). Overall, the analysis of activation markers and peripheral B cell subsets in Prkcd^G510S/G510S mice is evocative of B cell–driven autoimmunity.

To test whether this phenotype was intrinsically caused by the Prkcd mutation in B cells, we analyzed the B cell compartment in the spleen of BM chimeras described earlier. In WT/Prkcd^G510S/G510S chimeric mice, Prkcd^G510S/G510S splenic B cells had a clear competitive advantage over control B cells (Fig. 2H), whereas no difference in B cell proportion was observed between CD45.1 and CD45.2 genotypes in WT/Prkcd^+/+ chimeric mice. Moreover, Prkcd^G510S/G510S B cells had an activated phenotype in chimeric mice, as shown by a high expression level of activation markers (Fig. 2I), and among Prkcd^G510S/G510S B cells, PCs and ABCs were overrepresented compared with controls (Fig. 2J). These results indicate a cell-intrinsic role of the Prkcd^G510S/G510S mutation in B cells, leading to autoimmunity. Notably, MZ B cells were also increased in the Prkcd^G510S/G510S compartment of BM chimeric mice compared with controls (Fig. 2J), which did not correspond to the phenotype of Prkcd^G510S/G510S mice. This suggests that the reduction of MZ B cells in Prkcd^G510S/G510S mice was not due to a cell-intrinsic role of PKC-δ in these cells.

Analyses of immune subsets in Prkcd^G510S/G510S mice point toward an important role of B cells in driving the autoimmune phenotype. To test whether this role was B cell autonomous, we generated mixed BM chimeric mice using a 4:1 ratio of BM from the B cell–deficient mice μMT and Prkcd^+/+ or Prkcd^G510S/G510S mice. In those chimeric mice, most peripheral immune cells are derived from the μMT donor, except B cells that are derived only from the Prkcd^+/+ or the Prkcd^G510S/G510S donors. Autoimmune features and immune subsets were analyzed 3 mo after BM injection. We first measured PKC-δ expression in B cells. PKC-δ expression was much lower in B cells from μMT: Prkcd^G510S/G510S than in those from μMT: Prkcd^+/+ chimeric mice, whereas its expression was comparable in other immune cells from both mice validating the models (Supplemental Fig. 5C). Moreover, μMT: Prkcd^G510S/G510S and not μMT: Prkcd^+/+ mice developed an autoimmune phenotype very similar to that of Prkcd^G510S/G510S mice in terms of splenomegaly, glomerulonephritis, IgG deposits, IFN score, and serum antinuclear and anti-dsDNA IgG (Fig. 3A–E). Notably, some autoimmune features such as the RF, the anti-dsDNA IgM, and total Igs were more moderate in these chimeric mice compared with Prkcd^G510S/G510S mice, suggesting that other cells than B cells may influence the phenotype. Yet, for the most part, the autoimmune phenotype was recapitulated in those mice, demonstrating a B cell–autonomous role of the Prkcd^G510S/G510S mutation in driving B cell tolerance break and autoimmunity. We also examined B cell phenotype and activation status. B cells from μMT: Prkcd^G510S/G510S displayed significantly higher levels of activation markers CD69, MHCII, and CD86 than B cells from μMT: Prkcd^+/+ chimeric mice (Fig. 3F). In addition, the proportion of ABCs was significantly increased in μMT: Prkcd^G510S/G510S chimeric mice than was observed in Prkcd^G510S/G510S mice (Fig. 3G). Thus, these results reinforce our conclusion on a cell-intrinsic role of the Prkcd^G510S/G510S mutation in B cells and highlight that B cells with the Prkcd^G510S/G510S mutation are sufficient to drive autoimmunity.

We then dissected the mechanism driving autoimmune B cell activation in Prkcd^G510S/G510S mice. Considering the splenomegaly and the B cell overproliferation in these mice, and the previous literature on Prkcd^−/− mice, we hypothesized that Prkcd^G510S/G510S B cells responded more than control B cells to BCR engagement. To test this point, we labeled splenic B cells with CTV and stimulated them with anti-IgM^+/− anti-CD40 crosslinking Abs for 3 d. Prkcd^G510S/G510S B cells proliferated more than control B cells (Fig. 4A, Supplemental Fig. 5D). Specifically, Prkcd^G510S/G510S B cells performed more than four division cycles, whereas control B cells divided only twice on average on stimulation by anti-IgM and anti-CD40 Abs. By contrast, Prkcd^G510S/G510S and control B cells responded equally well to LPS stimulation (Supplemental Fig. 5E), indicating a specific hyperreactivity of Prkcd^G510S/G510S B cells to BCR-mediated stimulation. This phenotype was also associated to a better survival of Prkcd^G510S/G510S B cells than control B cells to growth factor deprivation (Fig. 4B).

In a first attempt to understand how Prkcd^G510S/G510S dysregulated B cell activation, we performed RNA sequencing on sorted Prkcd^+/+ and Prkcd^G510S/G510S follicular B cells stimulated or not through their BCR. Although the transcriptional profile of resting Prkcd^+/+ and Prkcd^G510S/G510S follicular B cells was relatively similar, there were greater differences between cells of both genotypes after BCR-mediated activation, with 143 differentially expressed genes (DEGs) (Fig. 4C). A functional annotation of these DEGs using Metascape (Fig. 4D) retrieved pathways related to cellular activation (leukocyte activation, mononuclear cell proliferation, etc.) confirming the impact of the Prkcd^G510S/G510S mutation on B cell activation. The mTOR-signaling pathway was also differentially regulated between both cell types, pointing to a link between mTOR and PKC-δ. To substantiate this link, we analyzed signal transduction downstream of the BCR in Prkcd^G510S/G510S versus Prkcd^+/+ B cells using phospho-specific Abs and flow cytometry, focusing on the major players of this pathway. Syk and Btk were phosphorylated at similar levels and following similar kinetics in B cells of both genotypes after BCR stimulation (Fig. 4E). However, baseline phosphorylation levels tended to be higher in Prkcd^G510S/G510S B cells. ERK phosphorylation was also higher in Prkcd^G510S/G510S B cells than in control B cells at baseline but also after BCR stimulation, especially at the latest time points. These results contrasted with a previous study where Prkcd^−/− B cells displayed lower ERK phosphorylation than controls on thapsigargin stimulation (20, 29). To test whether the discrepancy was explained by the difference in the stimulation mode, we used thapsigargin to stimulate B cells, but a higher ERK phosphorylation was also observed in Prkcd^G510S/G510S B cells in these conditions (Supplemental Fig. 5F). We then assessed activation of the mTOR pathway after BCR engagement. Basal phosphorylation levels of classical mTOR targets Akt (p-Akt), S6 (pS6), and 4EBP1 were higher in Prkcd^G510S/G510S B cells compared with controls and remained significantly higher on BCR stimulation, indicating higher mTOR activity in Prkcd^G510S/G510S B cells than in control B cells (Fig. 4E, 4F). Because the mTOR kinase is activated by PI3K on conversion of phosphatidylinositol (4,5)-bisphosphate into PIP3, we then measured PI3K activation on BCR engagement using a PIP3-specific Ab. PIP3 levels increased earlier than mTOR activity on BCR engagement, and their levels were higher in Prkcd^G510S/G510S B cells than in control B cells (Fig. 4F). We also evaluated mTOR activation after BCR engagement in B cells from Ly5a:Prkcd^G510S/G510S BM chimeric mice by measuring S6 phosphorylation. PS6 levels were higher in CD45.2⁺Prkcd^G510S/G510S B cells compared with CD45.1⁺ WT B cells (Fig. 4G), confirming the B cell–intrinsic role of the Prkcd^G510S/G510S mutation in regulating the PI3K/mTOR pathway. Collectively, our results indicate that BCR engagement induces higher PI3K/mTOR activation and proliferation in Prkcd^G510S/G510S B cells than in control B cells.

Based on the earlier characterization of BCR signaling in Prkcd^G510S/G510S B cells, we treated Prkcd^G510S/G510S and control mice with the mTORC1 inhibitor rapamycin or vehicle for 28 d (Fig. 5A). Rapamycin controlled lymphoproliferation in Prkcd^G510S/G510S mice, reducing the spleen size to nearly WT levels (Fig. 5B). IgG and IgM deposits in kidneys also decreased on rapamycin treatment (Fig. 5C). Specifically, the level of IgG deposits was almost halved in rapamycin-treated mice. The IFN score increased over time after treatment, and there was no difference between vehicle and rapamycin groups (Supplemental Fig. 6). We also monitored serum ANA levels (Fig. 5D). We observed a global increase of ANA level in vehicle-treated mice, whereas the rapamycin treatment prevented it. Next, we examined the effect of rapamycin treatment on B cell phenotype and activation status. Prkcd^G510S/G510S B cells displayed lower CD69 and CD86 activation markers in the rapamycin group compared with the vehicle group, and their levels were close to those of WT B cells (Fig. 5E). In addition, the proportions of plasma cells and GC B cells were normalized by rapamycin treatment. Aging B cells were not significantly increased in the young Prkcd^G510S/G510S mice used in this experiment (Fig. 5F). Rapamycin treatment also decreased mTOR activity in Prkcd^G510S/G510S B cells, close to the control levels, both at basal state and after BCR engagement (Fig. 5G). Altogether, these data demonstrate the effectiveness of rapamycin as a specific treatment of autoimmunity linked to Prkcd mutations.

Since 2013, 17 patients with autosomal recessive PKC-δ deficiency have been described (6–12, 16, 24–26). In many cases, these patients suffered from typical SLE often associated with lupus nephritis and lymphoproliferation or other autoimmune manifestations. In the familial form associated with PRKCD G510S mutations that we previously described (6), all three patients had lupus nephritis and various autoantibodies, while two had skin involvement and other autoimmune manifestations, including arthritis, hepatomegaly, and neurological involvement for one patient and autoimmune anemia, thrombocytopenia, antiphospholipid syndrome, and lymphoproliferative disease for the other, most severe patient. To our knowledge, the novel mouse model bearing homozygous Prkcd G510S mutations that we present in this article essentially recapitulates the SLE features, including glomerulonephritis associated with IgG deposits, splenomegaly, and serum autoantibodies. Moreover, mice had a reduced lifespan, which correlates with the premature death of one of the patients with PRKCD G510S mutation. Prkcd^G510S/G510S mice are a near phenocopy of previously described Prkcd^−/− mice (13, 14). However, there is residual expression of PKC-δ protein in Prkcd^G510S/G510S mice, and one cannot exclude a contribution of this mutated protein to the autoimmune phenotype. Moreover, knockout constructs can sometimes generate artifactual results by modifying expression patterns of adjacent genes or the sequence of noncoding structural elements in the genome. Thus, the point mutant mouse model we describe in this article represents a more reliable model of the human disease than the knockout mouse.

PKC-δ–deficient patients also display variable levels of immune deficiency, associated with susceptibility to bacterial, fungal, or viral infections. Chronic EBV or CMV infection has been reported (7). One of our patients also presented at the age of 30 y with severe papillomatosis caused by human papillomavirus infection (2). A previous study reported a decrease in NK cells and in their cytotoxic activity in one patient with PKC-δ deficiency. However, a decrease in peripheral NK cell numbers is often observed in SLE patients, especially when the disease is active (30). We report in this article that Prkcd^G510S/G510S mice have a specific decrease in mature NK cells, and that this deficiency is caused by a role of PKC-δ in NK cell development and maturation. Although further studies will be needed to understand the precise role of PKC-δ in NK cells, a deficiency in mature NK cells can contribute to the susceptibility to viral infections in patients, as shown in mouse models (31).

We show that the autoimmune phenotype in Prkcd^G510S/G510S mice is due to a B cell–autonomous impact of the Prkcd G510S mutation, confirming previous studies that used PKC-δ knockout mice (13) and pointing to PKC-δ as an important player in B cell central tolerance. Our data using the mouse model do not exclude that T cells may contribute to lupus pathology in patients as previously proposed (32, 33), but they demonstrate that the mutation in T cells is not essential for the development of the disease.

Multiple B cell subsets were altered in Prkcd^G510S/G510S mice, with the most dramatic effect on MZ, GC, and plasma cells, the latter two expanding in the mice, which is evocative of ongoing autoimmune GC reactions. B cell tolerance mechanisms ensure that B cells with high avidity for self-antigens are eliminated in the BM. Central tolerance of B cells operates via two distinct mechanisms, which are BCR editing and clonal deletion, with BCR editing being the predominant route. BCR editing is induced in immature B lymphocytes, which express an autoreactive BCR in the BM. This mechanism operates only if the tonic signal emanating from the BCR is weak, because the recombination machinery and in particular the RAG1/2 enzymes are repressed by a strong tonic signal. Moreover, strong tonic signaling also dampens the BM retention mechanism in immature B cells. In transgenic mouse models, expression of the protein N-Ras-D12 that constitutively activates Erk1/2 phosphorylation, or expression of a constitutively active form of PI3Kδ suppresses BCR editing in autoreactive B cells, which leads to a failure of central tolerance (5). The results we present here support a model by which PKC-δ is a negative regulator of BCR signaling, especially important to limit ERK phosphorylation and PI3K/mTOR activation. In particular, Prkcd^G510S/G510S B cells display elevated basal levels of p-Akt, pS6, and pERK, evocative of strong tonic BCR signaling. This higher tonic signaling may therefore prevent BCR editing in autoreactive clones, which would result in B cell tolerance break and autoimmunity. Notably, a previous article reported a decrease in Erk phosphorylation on thapsigargin stimulation in PKC-δ–deficient B cells (20–22, 24–29). We also used thapsigargin in our study but obtained opposite results, i.e., Prkcd^G510S/G510S B cells displayed stronger ERK phosphorylation after stimulation with thapsigargin. Technical differences may explain this discrepancy, but our results were supported by a comprehensive analysis of other BCR signaling branches that matched with the p-Erk data. Previous reports also showed that PKC-δ was essential for B cell tolerance by shuttling to the nucleus and inducing apoptosis on BCR engagement to induce clonal deletion (21). Our observation that Prkcd^G510S/G510S B cells survive better than control B cells during in vitro cultures are in agreement with this possibility. PKC-δ could therefore have several roles in B cell tolerance by promoting both clonal deletion and BCR editing in immature cells. Further studies will be required to determine the conditions that allow each of these PKC-δ functions and how they are coordinated during B cell development.

Mature B cells in the periphery of Prkcd^G510S/G510S mice displayed higher proliferation than controls on BCR engagement, a phenotype also observed in Prkcd^−/− mice (13) and in PKC-δ–deficient patients (7). Mechanistically, we linked this hyperreactivity to unrestrained BCR signaling, in particular at the level of the PI3K/mTOR pathway. This pathway is known to have multiple roles in mature B cells, promoting survival, proliferation, and differentiation of activated B cells (34). In particular, the reduced survival of mature BCR-deficient B cells can be rescued by activation of PI3K signaling (35). In agreement with our findings, gain-of-function mutations in the PIK3CD gene encoding the catalytic subunit p110d of PI3K cause activated-PI3K-d syndrome that combines massive lymphoproliferation with immunodeficiency and/or autoimmune features. Although this phenotype resembles that of PKC-δ deficiency, a recent article that reviewed genetic defects associated with an autoimmune lymphoproliferative phenotype found highest correspondence in terms of clinical features and immunophenotyping between PKC-δ deficiency and Ras-associated lymphoproliferative diseases (RALDs) (36). RALD is caused by somatic activating mutations in KRAS or NRAS that appear in hematopoietic progenitor or stem cells and are subsequently detectable in lymphocytes and monocytes (37). RALD presents with B cell lymphocytosis and may present with laboratory and clinical features not seen in patients with autoimmune lymphoproliferative syndrome because of FAS mutations such as SLE-like autoimmunity. Ras mediates the cellular response to growth factors, which leads to modulation of gene expression necessary for cell growth and proliferation. Ras operates through multiple downstream effectors, including PI3K and mTORC2 (38), and PKC-δ has been proposed to be an effector of oncogenic Ras in lung tumors (39). In B cells, Ras is activated downstream of BCR engagement through the guanine nucleotide exchange factors RasGRP1 and RasGRP3 (29). Further studies will be needed to clarify the link between PKC-δ and the activation of Ras in B cells, and more broadly to the 22 proteins whose defect have been associated with autoimmune lymphoproliferative syndrome-like disorders (36). Moreover, how PKC-δ negatively regulates mTOR activity in BCR-stimulated B cells remains to be determined considering that in other cellular pathways, PKC-δ has been shown to activate mTOR, in particular through TSC2 phosphorylation (40).

Despite the knowledge gained on the role of PKC-δ in immune regulation, no specific treatment was identified for PRKCD-mutated patients thus far. Such treatments may benefit a broader group of patients because PKC-δ expression appears to be silenced in PBMCs from adult-onset lupus patients (41). Based on the in vitro evidence of strong PI3K/mTOR activation in Prkcd^G510S/G510S B cells, we selected rapamycin as a specific treatment of autoimmune features in Prkcd^G510S/G510S mice. This strategy proved to be relevant as a 28-d daily treatment; rapamycin was sufficient to restore spleen homeostasis and alleviate autoimmune features in mice. This effectiveness correlated with a normalization of PI3K/mTOR activation in B cells after treatment. Interestingly, two patients with PRKCD deficiency received sirolimus, which corresponds to rapamycin, and showed excellent clinical response to treatment (7, 42). In addition, stopping treatment for 6 mo in one patient resulted in a significant relapse, highlighting the effectiveness of this treatment in the patient (42). Thus, and even though the usual clinical practice is to use rapamycin to treat lymphoproliferative disorders, our translational study provides a strong rationale to do it when the cause of the disease is proven to be PKC-δ deficiency.

The authors have no financial conflicts of interest.

We acknowledge the contribution of the SFR Biosciences (UAR3444/CNRS, US8/INSERM, ENS de Lyon, UCBL) facilities: ANIRA-PBES, ANIRA-Cytométrie, PSF, and ANIRA-AGC.