Raver1 links Ripk1 RNA splicing to caspase-8-mediated pyroptotic cell death, inflammation, and pathogen resistance

Abstract

Multiple cell death and inflammatory signaling pathways converge on two critical factors: receptor-interacting serine/threonine kinase 1 (RIPK1) and caspase-8. Careful regulation of these molecules is critical to control apoptosis, pyroptosis, and inflammation. Here, we found a pivotal role of Raver1 as an essential regulator of Ripk1 pre-mRNA splicing, expression, and functionality and the subsequent caspase-8-dependent inflammatory cell death. We show that Raver1 influences mRNA diversity primarily by repressing alternative exon inclusion. Macrophages from Raver1-deficient mice exhibit altered splicing of Ripk1. As a result, Raver1-deficient primary macrophages display diminished cell death and decreased interleukin-18 and interleukin-1ß production, when infected with Yersinia bacteria, or by restraining TGF-ß-activated kinase 1 or IKKβ in the presence of lipopolysaccharide, tumor necrosis factor family members, or interferon-γ. These responses are accompanied by reduced activation of caspase-8, Gasdermin D and E, and caspase-1 in the absence of Raver1. Consequently, Raver1-deficient mice showed heightened susceptibility to Yersinia infection. Raver1 and RIPK1 also controlled the expression and function of the C-type lectin receptor Mincle. Our study underscores the critical regulatory role of Raver1 in modulating innate immune responses and highlights its significance in directing in vivo and in vitro inflammatory processes.

Keywords: IL-1ß; RIPK1; caspase-8; gasdermin; pyroptosis.

Fig. 1.

Raver1 regulates RIPK1 RNA splicing and signaling. (A) A CRISPR screen was performed in Cas9-expressing Ripk3^−/− BMDMs infected with Y. pestis. Log₂-fold change and MAGeCK Robust Rank Aggregation enrichment scores are shown to indicate the essentiality of the genes. (B and C) Cas9-expressing BMDMs transduced with Raver1-targeting sgRNAs or a nontargeting control (NTC) (B) or Raver1^−/− BMDMs (C) were challenged with Y. pestis, Y. pestis △YopJ (△YopJ) or LPS and TAK1-i, or Salmonella. Cell death was measured by lactate dehydrogenase (LDH) release after 4 h. (D–F) Differential splicing analysis on Raver1^−/− and WT BMDMs using RNA-seq data (three biological replicates each). (D) Frequencies of significant alternative splicing events in Raver1^−/− vs. WT BMDMs. A3SS, alternative 3′ splice sites; A5SS, alternative 5′ splice sites; MXE, mutually exclusive exons; RI, retained intron; SE, skipped exons (E) Skipped exon event differences between Raver1^−/− vs. WT were depicted, with percent spliced-in (ΔPSI) and −log₁₀(P-value) shown. Ripk1 is highlighted as the most significant SE change upon Raver1 deletion. (F) PSI of Ripk1 alternative exon 4 inclusion. (G) Model for canonical and alternative splicing of Ripk1 in the presence or absence of Raver1. Kinase domain (KD), intermediate domain (ID), RIP homotypic interaction motif (RHIM), death domain (DD). (H and I) qPCR of mRNA levels of Ripk1 exons in indicated resting BMDMs. (J) Protein levels of RIPK1 in WT or Raver1^−/− BMDMs from five mice; the right-hand side shows quantification from the Western blot. (K) qPCR of Raver1 mRNA levels in C57BL/6 BMDMs treated with LPS for the indicated times. Data are presented as the mean ± SD (B, C, and K) or ± SEM (H–J) for triplicates from three or more independent experiments. (H–K) Data normalized to β-Actin. For comparison between two groups, Student’s t test was used; for more than two groups, ANOVA with the Bonferroni post hoc test was used. ns, not significant; *P ≤ 0.05; **P ≤ 0.01; ***P ≤ 0.001. Abbreviations: n.d., not detected; NT, nontreated; Alt exon, alternative exon.

Fig. 2.

Raver1-deficient macrophages and neutrophils are resistant to pyroptotic cell death and activation of RIPK1, caspase-8, and GSDMD/E after Yersinia challenge. (A–E) C57BL/6 or Raver1^−/− BMDMs (A–D) or neutrophils (E) were challenged with Yersinia spp., Salmonella, TAK1-i or IKK-i plus LPS, TNF, or LT-α. Cell death was measured by LDH release after 4 h (A–C) or SYTOX membrane permeability assay (D and E). (F–I) Cell lysates from BMDMs (F, G, and I) or neutrophils (H) from indicated mice were analyzed by immunoblotting for GSDMD/E, caspase-8, caspase-3, and RIPK1 after 2 to 3 h (F–H) or 1 h (I). (J) Caspase-8 immunoprecipitation (IP: Casp8) or cell lysates of C57BL/6 or Raver1^−/− BMDMs treated with LPS+TAK1-i for indicated timepoints were analyzed by immunoblotting on FADD, RIPK1, and caspase-8. (K and L) Indicated BMDMs were stimulated with TNF, TAK1-i, or a SMAC mimetic ± pan-caspase inhibitor zVAD-fmk pretreatment and then analyzed for LDH release after 4 h (K) or SYTOX assay (L). Data are presented as the mean ± SD for triplicate wells from three or more independent experiments. Immunoblots are representative of ≥3 performed. For comparison between two groups, Student’s t test was used; for more than two groups, ANOVA with Bonferroni post hoc test was used. n.s., not significant; *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001. Abbreviations: FL, full-length; +, Raver1^+/+; −, Raver1^−/− (F, G, and I).

Fig. 3.

Raver1 controls TAK1-restrained IL-1β/IL-18 release and IFNγ-driven inflammatory responses. (A–C) C57BL/6 or Raver1^−/− BMDMs were treated with Yersinia spp., Salmonella, or LPS with TAK1-i or IKK-i. IL-1β release into the supernatant as measured by the Enzyme-linked immunosorbent assay (ELISA) after 5 h or at the indicated time points. (D and E) Cell lysates plus supernatants were analyzed by immunoblotting for caspase-1, -8, or IL-1β cleavage 3 h after the indicated stimulation. (F) Oligomerization of ASC in the inflammasome-enriched and crosslinked lysates as detected by immunoblotting. (G–I) BMDMs of indicated genotypes were treated with IFNγ plus TAK1-i. Cell death was measured by SYTOX assay (G and H). GSDMD and caspase-8 processing at 3 h were detected by immunoblotting (I). (J) IL-18 levels as measured by the ELISA on supernatants from C57BL/6 or Raver1^−/− BMDMs challenged with indicated bacteria or TAK1-i for 5 h. (K and L) C57BL/6 or Tnfr1^−/− BMDMs were treated with TAK1-i plus TNF or IFNγ, or Raptinal alone. Cell death was measured by LDH release after 4.5 h (K) or SYTOX assay (L). Data are presented as the mean ± SD for triplicates from three or more independent experiments. Immunoblots are representative of three immunoblots performed. For comparison between two groups, Student’s t test was used; for more than two groups, ANOVA with Bonferroni post hoc test was used. n.s., not significant; *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001. Abbreviations: CHX, cycloheximide.

Fig. 4.

Raver1 dysfunction results in a RIPK1 splicing isoform, Splice I, that negatively regulates cell death. (A and B) WEHI164 clone 13 fibrosarcoma cells (A) or human U2OS osteosarcoma cells (B) transduced with indicated sgRNAs were treated with TNF and TAK1-i. Cell death was measured by SYTOX assay (A) or LDH release after 10 h (B). (C and D) HEK293T cells were cotransfected with vectors encoding RIPK1-HA and RIPK1 variants (Splice I-Flag or RIPK1-kinase domain [KD]-Flag) or empty pcDNA3 for 24 h and analyzed by immunoblotting (C), or immunoprecipitation with RIPK1 antibodies followed by Flag or HA immunoblotting (D). (E) Doxycycline (Dox)-induced RIPK1 or RIPK1-Splice I was expressed in HEK293T cells. Cell death was measured by reduced ATP intensity after Dox treatment for the indicated times. (F) HEK293T cells expressing Dox-induced RIPK1-Splice I were transfected with RIPK1 or empty vector (EV) 12 h post-Dox addition, and cell death was measured by reduced ATP intensity after another 12 h. (G) WEHI164 clone 13 cells expressing Dox-induced RIPK1-Splice I were stimulated with TNF (50 ng/mL) 12 h post-Dox addition, and cell death was assessed by LDH release after 10 h. Data are presented as the mean ± SD for triplicates from three or more independent experiments. Immunoblots are representative of three immunoblots performed. For comparison between two groups, Student’s t test was used; for more than two groups, ANOVA with Bonferroni post hoc test was used. ns, not significant; *P ≤ 0.05, ***P ≤ 0.001. Abbreviations: NTC, nontargeting control; hTNF, human TNF.

Fig. 5.

Raver1 cooperates with Ptbp1 to regulate RIPK1 splicing and cell death. (A) Cas9-expressing BMDMs transduced with sgRNA targeting Raver1 or Ptbp1 were treated with LPS plus TAK1-i and cell death was measured by SYTOX assay. (B–E) WT or Raver1^−/− immortalized macrophages were transfected with Ptbp1-targeting siRNA or NTC siRNA. Cell death was measured by LDH release 4 h after LPS+TAK1-i or TNF+TAK1-i treatment (B). qRT-PCR of mRNA levels of Ripk1 exons normalized to Actin (C and D). Protein levels of RIPK1 were analyzed by immunoblotting (E). Data are presented as the mean ± SD for triplicate wells from three or more independent experiments. Immunoblots are representative of three blots performed. For comparison, ANOVA was used with Bonferroni’s post hoc test. *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001. Abbreviations: Alt exon, alternative exon.

Fig. 6.

Raver1 modulates Mincle-dependent TNF release via RIPK1. (A) WT or Raver1^−/− BMDMs were unstimulated (NT) or stimulated for 1 h with TNF or TNF plus TAK1-i and analyzed by RNAseq. Genes with significant differences in expression between WT and Raver1^−/− are enumerated. (B and C) WT or Raver1^−/− BMDMs were stimulated with TNF or synthetic mycobacterial cord factor, trehalose-6,6′-dibehenate (TDB). (B) Relative mRNA expression of Mincle (Clec4e) was quantified by qRT-PCR after 1-h TNF or 6-h TDB treatment, normalized to Actin. (C) Protein lysates were analyzed for Mincle level by immunoblotting at the indicated times. (D–G) BMDMs transduced with Clec4e sgRNA (D) or from indicated genotypes (E–G) were stimulated with TDB, LPS, Pam₃CSK₄, or cell-death-related β-glucosylceramide (βGlcCer) for 24 h. TNF-α or IL-1β release was assessed by the ELISA. (H) WT, Raver1^−/−, and Ripk1^K45A BMDMs were treated with TDB for 18 h and analyzed by immunoblotting. Data are presented as the mean ± SD for triplicate wells from three or more independent experiments. Immunoblots are representative of three blots performed. For comparison between two groups, Student’s t test was used; for more than two groups, ANOVA was used with Bonferroni’s post hoc test. n.s., not significant; *P ≤ 0.05; **P ≤ 0.01; ***P ≤ 0.001.

Fig. 7.

Raver1 is essential for host resistance against Y. pseudotuberculosis infection in vivo. (A) Survival of C57BL/6, Raver1^−/−, or heterozygous littermates were monitored following oral challenge with 0.5 to 1 × 10⁸ CFU of Y. pseudotuberculosis. (B and C) Tissues from C57BL/6 and Raver1^−/− mice were collected 5 d postinfection to quantify the bacterial load in spleens and livers (B) and analyze subpopulations of CD11b⁺ myeloid cells in spleens for cell death with live/dead stain by flow cytometry (C). (D–G) Liver sections from C57BL/6 and Raver1^−/− mice were stained with hematoxylin and eosin and subjected to microscopy. (D) Inflammatory foci (green arrows) and (F) bacterial microcolonies (yellow asterisks) were quantified (E and G). Magnification: 40× (4× objective, as indicated), 100× (10× objective), or 600× (60× objective). (H) IL-18 in C57BL/6 and Raver1^−/− mice spleen homogenates 3 d postinfection were measured by the ELISA. (I) Survival of C57BL/6 and Il18^−/− mice challenged orally with 1 × 10⁸ CFU of Y. pseudotuberculosis. (J) Bacterial loads in indicated spleens and (K) IL-18 in indicated spleen homogenates were measured by the ELISA 3 d postinfection. Data are presented as medians with interquartile range (B, C, H, J, and K) or mean with SEM (E and G), either pooled from two independent experiments (A–C) or representative of two or more independent experiments (D–K). Images are representative of two individual experiments. For comparisons, datasets were analyzed by the log-rank test (A and I), Mann–Whitney U test (B, C, E, G, H, and K), or Kruskal–Wallis with Dunn’s post hoc test (J). *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001. (L) Proposed model for Raver1-mediated caspase-8-dependent pyroptotic cell death, inflammation, and pathogen resistance.