Specific pathogen-free female C57BL/6 wild type (WT) mice (12–16 weeks of age) were bred in-house using breeders from The Jackson Laboratory (Bar Harbor, ME). RIPK3^-/-, which originated from Vishva Dixit (Genentech, San Francisco, CA) (29) were generously provided by Dr. Maziar Divangahi (McGill University) (18). RIPK3^K51A/K51A mice initially described by Mandal et al. (30) were kindly supplied by GlaxoSmithKline Inc. (GSK; Collegeville, PA). MLKL^-/- was generated by Dr Jiahuai Han (Xiamen University (31) and breeders were generously provided by Dr. Maya Saleh (McGill University). All deficient strains were on a C57BL/6 background and bred in-house. Mice were maintained and bred according to Canadian Council on Animal Care (CCAC) guidelines (consistent with the National Research Council Guide for the Care and Use of Laboratory Animals [8th Edition, 2011]), and maintained on food and water ad libitum. Animal experiments were approved by the McGill University Animal Care Committee.

Unless stated otherwise, all reagents were obtained commercially from the following sources and used without further purification: human β2-glycoprotein I (β2GPI) (Crystal Chem, Downers Grove, IL); and bovine heart cardiolipin (CL) (Avanti Polar Lipids, Elk Grove Village, IL). Lipopolysaccharide (LPS; Escherichia coli-derived, serotype O111:B4) was obtained from List Biological Laboratories, Campbell, CA or Sigma-Aldrich Canada Co., Oakville, ON. Other reagents are indicated under the protocol in which they were used.

Mice (12–16 weeks) were immunized subcutaneousy with a mixture of β2GPI (20μg) and LPS (10μg) every 14 days for a total of 3 immunizations, unless otherwise indicated. This immunization protocol with β2GPI and LPS results in the spread of the immune response to IgG autoAbs typically seen in human SLE (including to Ro[SSA], La[SSB], dsDNA, nRNP, and Sm), compared immunization with buffer or LPS alone (27). Mice immunized with LPS alone were included as a control group in each experiment. Serum was collected from saphenous vein bleeds 10 days following each immunization to assess serum autoAb levels.

Development of SLE was assessed by performing in-house enzyme-linked immunosorbent assays (ELISAs) to detect serum autoAbs against β2GPI, CL, and hallmark SLE antigens (dsDNA, Ro [SSA], La [SSB], Sm, and Sm/RNP), as previously described (26, 27).

The spleen, lymph nodes (LN; axial, brachial and inguinal), and thymus were harvested from WT, RIPK3^-/-, RIPK3^K51A/K51A and MLKL^-/- mice. The spleens and LN were treated with collagenase V (1mg/mL; MilliporeSigma) for 15 minutes at 37^°C and filtered through a 70 μm nylon cell strainer to obtain single cell suspensions. Splenocytes were depleted of red blood cells (RBCs) by washing with ammonium-chloride-potassium (ACK) lysis buffer. 1x10⁶ cells/well were stained with fluorescently conjugated antibodies for surface expression of lineage-specific markers and acquired by flow cytometry on a BD LSRFortessa™ Cell Analyzer (BD Biosciences).

Spleens were harvested from WT, RIPK3^-/-, RIPK3^K51A/K51A, and MLKL^-/- mice under sterile conditions and processed into single cell suspensions. T cells were isolated from the splenocyte suspensions by negative selection (Stemcell) and stained with CellTrace Violet™ (CTV). 5x10⁵cells/well were plated in 24-well plates coated with anti-CD3 (clone 145-2C11) and anti-CD28 (clone 37.51), and incubated for 72 h at 37°C (5% CO₂). Cells were harvested and stained for surface expression of CD4 and CD8 and acquired by flow cytometry on a BD LSRFortessa™ Cell Analyzer (BD Biosciences). IFN-γ concentration in the supernatant was measured by ELISA, according to the manufacturer’s instructions (BD Biosciences).

C57BL/6 WT mice expressing the congenic marker CD45.1 were treated with bacitracin in drinking water and irradiated with a dose of 9.5Gy. Bone marrow cells were isolated from the femurs of WT and RIPK3^-/- CD45.2⁺ donor mice, and T cells were depleted using negative selection. Within 24h of irradiation, 4x10⁶ bone marrow cells from either WT or RIPK3^-/- mice were transferred intravenously to WT CD45.1⁺ mice. Immune reconstitution was assessed on peripheral blood at 11 weeks following bone marrow transfer. The murine lupus model was induced in the bone marrow chimeras 12 weeks post-transfer.

Bone marrow cell suspensions were isolated from the femurs of mice and RBCs were depleted using ACK lysis buffer. Bone marrow-derived dendritic cells (BMDCs) were generated from bone marrow cells following a 7-day culture in the presence of GM-CSF (20 ng/mL; PeproTech). BMDCs were plated at 1x10⁶cells/well in 12-well plates and stimulated with LPS (10 ng/mL) for 16 h. Cells were harvested by scraping, stained for MHC II and co-stimulatory molecule expression (CD80, CD86), and acquired by flow cytometry on a BD LSRFortessa™ Cell Analyzer (BD Biosciences). BMDCs were defined as CD11c⁺ MHC-II^hi cells. IL-6 and TNFα in the supernatants were measured by ELISA (BD Biosciences).

RIPK3-dependent cell death and inflammation occurs through both kinase-dependent and -independent (scaffolding) functions. To investigate which of these mechanisms are involved in the generation of SLE-associated autoAbs, we induced our model of murine lupus in wild-type (WT), RIPK3-deficient (RIPK3^-/-), and RIPK3 kinase-dead (RIPK3^K51A/K51A) mice. MLKL-deficient (MLKL^-/-) mice were also included to investigate the contribution of RIPK3-dependent necroptosis. RIPK3^-/- mice developed significantly lower levels of autoAbs to the immunizing antigen β2GPI and the closely related antigen CL, compared to WT mice ( Figure 1A ). This was also the case when anti-β2GPI and anti-CL antibodies were analyzed as a combined antibody subset (2-way ANOVA) ( Figure 1A ). RIPK3^-/- mice also had significantly reduced levels of autoAbs to hallmark SLE antigens when the antibodies were compared as a combined subset, but only anti-Sm/RNP antibodies were significantly reduced when individual autoAbs were compared to those of WT mice ( Figure 1B ). Like RIPK3^-/- mice, RIPK3^K51A/K51A mice had reduced levels of the combined subset of hallmark SLE autoAbs, but only reduced levels of anti-Sm/RNP when individual autoAbs were compared to those of WT mice. The levels of anti-β2GPI and anti-CL autoAb in RIPK3^K51A/K51A mice were comparable to those of WT mice ( Figures 1A, B ). In contrast to the RIPK3-deficient strains, MLKL^-/- did not differ from WT mice in their production of SLE antibodies. MLKL deficiency did not impact autoAb production, suggesting that autoAb production in our murine lupus model results from necroptosis-independent mechanisms of RIPK3 ( Figures 1A, B ).

Deficient strains were compared to C57BL/6 WT mice, their background strain. However, we compared C57BL/6 WT mice to RIPK3^K51A/K51A WT littermates (RIPK3^WT/WT) to ensure that their antibody response was similar. Both strains showed equivalent anti-β2GPI, anti-CL, and anti-dsDNA antibody levels after 2 immunizations with β2GPI and LPS ( Supplementary Figure 1 ). A shortened immunization protocol was used in this experiment as the purpose was solely the comparison of autoAb induction between strains. We conclude that the response of RIPK3^WT/WT mice is comparable to that of C57BL/6 WT mice for induction of the murine lupus model, and that C57BL/6 WT mice are therefore a suitable control in this context.

To confirm that the impact of RIPK3 on autoAb production was not due to differences in immune cell composition among strains, we characterized the immune cell proportions in WT, RIPK3^-/-, RIPK3^K51A/K51A, and MLKL^-/- mice prior to induction of the murine lupus model. We focused on cells involved in antigen presentation and antibody production. No significant differences were observed in the cell numbers or proportions of conventional dendritic cells (cDCs), CD4⁺ T cells, or B cells in the lymph nodes (LNs) and spleens among WT, RIPK3^-/-, RIPK3^K51A/K51A, and MLKL^-/- mice ( Supplementary Figure 2 ). A complete summary of the immune cell proportions in the LNs, spleen, and thymus is included in Supplementary Table 1 .

To investigate whether RIPK3 reduced autoAb generation in our model of murine lupus through an effect on immune cell function, we created bone marrow chimeras in which the hematopoietic cells of WT mice were replaced by either WT or RIPK3^-/- immune cells ( Figures 2A, B ). Mice received 4 immunizations to mitigate any potentially immunosuppressive effect that generation of the chimeric state may have had on the immune response to injected antigens. RIPK3 deficiency in the hematopoietic cell compartment resulted in significantly lower levels of autoAbs to the immunizing antigen (β2GPI), but not to CL, compared to WT mice ( Figure 2C ). However, no differences in the levels of autoAbs to hallmark SLE autoantigens were observed between WT and RIPK3^-/- bone marrow chimera mice ( Figure 2D ).

These results suggest that RIPK3 signaling in immune cells may be dispensable for epitope spread to hallmark SLE autoantibodies in our model, but is important for generating antibodies to the immunizing antigen β2GPI. Since we know that autoAb production in our model is associated with the generation of a β2GPI-specific CD4⁺ T cell response (28), we next focused our investigations on immune cell-dependent mechanisms through which RIPK3 may influence the generation of antibodies to β2GPI.

We first confirmed the importance of a T cell response in the generation of anti-β2GPI antibodies in our model. Two days following the 4^th immunization with β2GPI and LPS, we observed a specific expansion of T follicular helper (T_FH) cells in the draining LN compared to PBS-immunized mice ( Supplementary Figure 3 ). T_FH cells play a crucial role in the establishment of a germinal center (GC) where T cell-dependent activation of B cells takes place (32, 33). Accordingly, we found an expansion of GC B cells and IgG⁺ class-switched B cells in the draining LN 8 days following the 3^rd immunization with β2GPI and LPS ( Supplementary Figure 4 ). These findings suggest that B cell maturation and autoAb production in our model of murine lupus is driven by a classical T cell-dependent GC reaction (33). Together, these results reinforce the importance of the T cell response for autoAb production in our model, and further motivate a focus on the role of RIPK3 deficiency in the T cell-dependent generation of β2GPI-specific antibodies.

The establishment of a robust T cell response requires the complex integration of three major signals provided by the APC: [1] antigen processing and peptide presentation on the major histocompatibility complex class II (MHC II), [2] co-stimulatory molecule expression by the APC, and [3] specific cytokine production by the APC (32, 34, 35).

We first investigated if deficiency in RIPK3-dependent pathways impacted the expression of MHC II (signal 1) and/or co-stimulatory molecules (signal 2) by bone marrow-derived DCs (BMDCs), following their activation with LPS. We found that expression of MHC-II, CD80, and CD86 was comparable between WT and RIPK3-dependent pathway-deficient BMDCs following LPS stimulation ( Figures 3A–F ). We next determined if the production of pro-inflammatory cytokines (signal 3) was impaired in RIPK3-deficient BMDCs. We found a slight but non-significant decrease in the production of TNF-α and IL-6 by RIPK3^-/- and RIPK3^K51A/K51A BMDCs compared to WT and MLKL^-/- BMDCs ( Figures 3G, H ). These data indicate that RIPK3 deficiency does not significantly impact any of the three major APC signals necessary for T cell activation.

To determine whether RIPK3-deficiency in APCs impacts T cell activation (following antigen presentation), we performed an in vitro antigen presentation assay using antigen-specific OT-II T cells that are activated by presentation of the ovalbumin (OVA) peptide 323-339 (OVA_323-339) on MHC II. In these experiments, we investigated multiple aspects of the overall T cell response.

First, we found that a deficiency in RIPK3-dependent pathway in CD4-depleted splenocytes did not impact the proliferation of OT-II T cells following antigen-specific activation ( Figures 4A–C ). Next, we examined cytokine production by activated OT-II T cells. We have previously shown that development of murine lupus in our model is associated with the emergence of IFN-γ-producing β2GPI-specific T cells (28). Similar to proliferation, IFN-γ production by OT-II T cells was comparable following antigen presentation by either WT or RIPK3-dependent pathway-deficient splenocytes ( Figures 4D, E ).

To this point, we had only used the processed peptide, OVA_323-339, in our antigen presentation assays. To determine whether RIPK3 might impact antigen presentation by altering intracellular antigen processing, we co-incubated WT and RIPK3-dependent pathway-deficient LPS-stimulated BMDCs with intact OVA protein, and then used these BMDCs to assess the impact of RIPK3 deficiency on T cell activation and IFN-γ production. As seen with processed OVA peptide, antigen presentation with the additional requirement for processing of the intact protein was unaffected by RIPK3 deficiency. T cell proliferation ( Figures 4F–H ) and IFN-γ production ( Figures 4 I, J ) by OT-II T cells were comparable whether they were co-incubated with WT or RIPK3-dependent pathway-deficient BMDCs.

Finally, we explored the possibility that RIPK3 deficiency may affect the concentration of antigen required for effective antigen presentation. OT-II T cells were co-incubated with LPS-activated BMDCs from WT or RIPK3-dependent pathway-deficient mice in the presence of decreasing concentrations of OVA_323-339. The dose-response curves for proliferation of OT-II T cells were virtually indistinguishable following antigen presentation by RIPK3^-/-, RIPK3^K51A/K51A, or WT BMDCs ( Figures 4K, L ). However, a significant increase in the proliferation of OT-II T cells was observed following co-incubation with MLKL^-/- BMDCs compared to WT BMDCs ( Figures 4K, L ). The significance of this finding is uncertain, although it suggests that MLKL deficiency in APCs may impact T cell proliferation through RIPK3-independent mechanisms ( Figures 4K, L ).

Overall, RIPK3 deficiency in APCs did not affect the in vitro activation of T cells, as assessed by either proliferation or IFN-γ production. These findings suggest that factors extrinsic to the APC may influence antigen presentation and T cell activation in RIPK3-deficient mice.

The outcome of activation of RIPK3-dependent pathways is dependent on a combination of inflammatory triggers and the cell-specific environment in which they take place (11, 12). Given the limited complexity of an in vitro antigen presentation assay compared to that of T cell activation in vivo, we next sought to determine the impact of deficiency in RIPK3-dependent pathways on early T cell activation ex vivo.

Deficiency in RIPK3-dependent pathways did not impair the in vitro activation or IFN-γ production by freshly isolated splenic T cells ( Supplementary Figure 5 ). We therefore conclude that any effects of RIPK3-dependent pathways on T cell activation may be a consequence of extrinsic effects during their activation.

Using our induced model of murine lupus, we evaluated the impact of RIPK3 on APC and T cell activation ex vivo ( Figure 5A ). Here, mice received an additional immunization following the induction of our model in order to evaluate the immune response at a timepoint suitable for T cell activation. The total number of cells in the draining LN 48 h after a 4^th immunization with β2GPI and LPS was comparable between WT and RIPK3-dependent pathway-deficient mice ( Figure 5B ). However, a decreased proportion of activated DCs (MHC II^hi CD86^hi) was found in the draining LN of RIPK3^K51A/K51A and MLKL^-/- mice, compared to WT mice ( Figures 5 C, D ). No difference in the proportion of activated DCs was observed between RIPK3^-/- and WT mice ( Figures 5 C, D ). A change in the proportion of activated DCs could impact T cell phenotype or proportion in the draining LN. We next investigated the proportion of CD4⁺ IFN-γ-producing T cells, as they have previously been associated with the β2GPI-specific T cell response in our model (28). IL-4 was included to demonstrate the selectivity of cytokine expression towards a Th1 profile. Despite significant differences in the proportion of activated DCs in RIPK3^K51A/K51A and MLKL^-/- mice as compared to WT mice, no differences in the proportions of IFN-γ-producing CD4⁺ T cells or T_FH cells were observed between WT and RIPK3-dependent pathway-deficient mice ( Figure 5E–H ). The slight increase in the proportion of T_FH observed in the draining LN of RIPK3^-/- mice appeared to be independent of the response to immunization with β2GPI and LPS, as a greater increase in the proportion of T_FH cells was observed in the non-draining LN of RIPK3^-/- mice, compared to WT mice ( Figures 5G–I ). For Panels C-L in Figure 5 , absolute cell counts were performed and did not differ from cell proportions (data not shown).

B cells can also act as APCs and indeed were found to upregulate their expression of MHC II following induction of our model of murine lupus ( Supplementary Figures 3E, F ). We therefore evaluated the expression of MHC II by B cells in WT and RIPK3-dependent pathway-deficient mice. Although we observed an increased expression of MHC II by B cells in the draining LN of RIPK3^-/- and MLKL^-/- mice compared to WT mice ( Figures 5 J, K ), elevated MHC II expression by RIPK3^-/- B cells was also observed in the non-draining LN. This finding suggests that RIPK3 deficiency may enhance MHC II expression by B cells independently of induction of our model of murine lupus ( Figure 5L ).

Nonetheless, these ex vivo differences in MHC II expression by RIPK3-deficient B cells warranted a closer examination of the impact of RIPK3 deficiency on B cell activation in vitro. No significant differences in the expression of MHC II, CD80, or CD86 were identified between unstimulated WT and RIPK3-dependent pathway-deficient splenic B cells ( Supplementary Figures 6A, B ). In contrast, following LPS stimulation, a significant increase in the level of MHC II expression was observed on B cells from RIPK3^-/- mice compared to WT mice ( Supplementary Figures 6C, D ). CD80 and CD86 expression between RIPK3^-/- and WT B cells was comparable ( Supplementary Figures 6C, D ). All differences were limited to RIPK3^-/- B cells. No differences in MHC II or co-stimulatory marker expression were observed between WT and RIPK3^K51A/K51A or MLKL^-/- B cells, either at baseline or following LPS stimulation ( Supplementary Figures 6C, D ).

While we were unable to observe an effect of RIPK3 deficiency on antigen presentation, its effect on the proportion of T_FH cells and on the level of MHC II expression by B cells following innate immune stimulation raises the possibility that deficiency in RIPK3-dependent pathways may impact the T cell-dependent maturation of B cells. We therefore investigated the proportions of GC and IgG⁺ class-switched B cells in the draining LN 8 days following immunization with β2GPI and LPS ( Figure 6A ). Total cell numbers in the draining LN were comparable between WT and RIPK3-dependent pathway-deficient mice following induction of our model of murine lupus ( Figure 6B ). No differences in the proportions of GC and IgG⁺ class-switched B cells were observed between WT and RIPK3^-/- or RIPK3^K51A/K51A mice ( Figures 6C–F ). However, an increase in the proportion of GC and IgG⁺ class-switched B cells was observed in MLKL^-/- mice compared to WT mice ( Figures 6C–F ). For Panels C-F in Figure 6 , absolute cell counts were performed and did not differ from cell proportions (data not shown).

Although the proportions of GC and IgG⁺ class-switched B cells were comparable between WT and RIPK3-deficient mice, RIPK3 deficiency may specifically impact the number of β2GPI-specific IgG-producing cells. To address this possibility, we quantified total IgG and β2GPI-specific IgG production ex vivo using ELISpot. We found no difference in the proportion of IgG-producing cells, or in the amount of IgG produced per cell, in the draining LN of RIPK3-dependent pathway-deficient mice compared to WT mice ( Figures 6G–I ). The numbers of β2GPI-specific IgG-producing cells, as well as the per-cell production of β2GPI-specific IgG, was also comparable between WT and RIPK3-dependent pathway-deficient mice ( Figures 6J–L ).

Overall, we were unable to identify any major defects in B cell maturation or β2GPI-specific IgG-producing cells in RIPK3-deficient mice. These findings suggest that the increased proportion of T_FH cells and increased MHC II expression on RIPK3^-/- B cells did not significantly impact T cell dependent B cell maturation following induction of our model of murine lupus.