Caspase-8 auto-cleavage regulates programmed cell death and collaborates with RIPK3/MLKL to prevent lymphopenia

Caspase-8 is an initiator of death receptor-induced apoptosis and an inhibitor of RIPK3-MLKL-dependent necroptosis. In addition, caspase-8 has been implicated in diseases such as lymphoproliferation, immunodeficiency, and autoimmunity in humans. Although auto-cleavage is indispensable for caspase-8 activation, its physiological functions remain poorly understood. Here, we generated a caspase-8 mutant lacking E385 in auto-cleavage site knock-in mouse (Casp8ΔE385/ΔE385). Casp8ΔE385/ΔE385 cells were expectedly resistant to Fas-induced apoptosis, however, Casp8ΔE385/ΔE385 cells could switch TNF-α-induced apoptosis to necroptosis by attenuating RIPK1 cleavage. More importantly, CASP8(ΔE385) sensitized cells to RIPK3-MLKL-dependent necroptosis through promoting complex II formation and RIPK1-RIPK3 activation. Notably, Casp8ΔE385/ΔE385Ripk3−/− mice partially rescued the perinatal death of Ripk1−/− mice by blocking apoptosis and necroptosis. In contrast to the Casp8−/−Ripk3−/− and Casp8−/−Mlkl−/− mice appearing autoimmune lymphoproliferative syndrome (ALPS), both Casp8ΔE385/ΔE385Ripk3−/− and Casp8ΔE385/ΔE385Mlkl−/− mice developed transplantable lymphopenia that could be significantly reversed by RIPK1 heterozygosity, but not by RIPK1 kinase dead mutation. Collectively, these results demonstrate previously unappreciated roles for caspase-8 auto-cleavage in regulating necroptosis and maintaining lymphocytes homeostasis.


RESULTS
Casp8 ΔE385/ΔE385 mice are viable but develop a slight CD8 + T cell lymphopenia in the spleen Previous studies have demonstrated that auto-cleavage of caspase-8 is required for mediating apoptosis but not for inhibiting necroptosis during development because the mice expressing none-cleavable Caspase-8 are viable [7,13,46]. As expected, we observed that caspase-8 cleavage was gradually enhanced when apoptosis was induced by tumor necrosis factor α (TNF-α) plus cycloheximide (CHX) in wild-type mouse dermal fibroblasts (MDFs) (Fig. 1A). Notably, caspase-8 cleavage was also increased in response to necroptotic stimulation with TNF-α plus Smac mimetics (Smac) and the pan-caspase inhibitor Z-VAD-FMK (zVAD). This finding was verified by observing the increased levels of phosphorylated RIPK1, RIPK3, and MLKL necroptotic markers (Fig. 1B). Therefore, in addition to its role in mediating apoptosis, caspase-8 cleavage is also hypothesized to regulate necroptosis.
Previous studies established transgenic mice expressing caspase-8 D387A [7,13,46], which cannot be cleaved between the large and small catalytic subunits. Caspase-8 has a substrate preference for the tetrapeptide (Leu/Val)-Glu-X-Asp [48], which corresponds closely to the caspase-8 auto-processing substrate sequence, L384/E385/V386/D387. We therefore hypothesized that E385 of caspase-8 would also contribute to its auto-cleavage. To explore the contribution of caspase-8 (E385) in its auto-processing in vitro and in vivo, we generated a knock-in mouse that expressed caspase-8 lacking E385 in the auto-cleavage site between the large and small catalytic subunits (Fig. S1A). In contrast to the embryonic lethality observed in caspase-8 deficiency [49] and catalytically inactive caspase-8 mice [6,8], Casp8 ΔE385/ΔE385 mice were viable and matured normally (Fig. S1B), which was consistent with previously reported mouse lines expressing caspase-8(D387A) [7,13,46]. To test whether CASP8 (ΔE385) is indeed unable to auto-process between the large and small catalytic subunits, we treated primary WT and Casp8 ΔE385/ ΔE385 BMDMs with LPS/BV6 to induce apoptosis. Compared with the dramatic caspase-8 cleavage in wild-type BMDMs, caspase-8 cleavage between the large and small catalytic subunits was confirmed to be blocked in Casp8 ΔE385/ΔE385 BMDMs utilizing two different antibodies (Fig. S1C). Besides, it was observed that the expression of CASP8(ΔE385) in multiple tissues including spleen, lung, liver, kidney, colon, heart, ileum, and rectum was normal in Casp8 ΔE385/ΔE385 mice ( Fig. 1C and S1D), suggesting that the cleavage of caspase-8 is dispensable for its expression and stability in vivo. Next, we examined the effect of CASP8(ΔE385) on the pathologies. Histopathological examination demonstrated that the appearance of multiple tissues was indistinguishable in Casp8 ΔE385/ΔE385 mice in comparison with the tissue appearance in WT mice (Fig. S1E). However, we observed that the Casp8 ΔE385/ ΔE385 mice developed slight splenomegaly with a mild decrease in the percentage of the CD8 + T cells in the spleen and bone marrow ( Fig. 1D-F). However, no differences were observed between Casp8 ΔE385/ΔE385 and WT mice with respect to the B cells and the myeloid cell subsets obtained from the spleen, lymph nodes, and bone marrow (Fig. 1F). These results show that the Casp8 ΔE385/ΔE385 mice are viable but develop a slight CD8 + T cell lymphopenia with splenomegaly.
Apoptosis induced by TNF-α was switched to necroptosis by attenuating RIPK1 cleavage in Casp8 ΔE385/ΔE385 cells Previous studies have demonstrated that the auto-cleavage of caspase-8 is essential for the apoptosis induced by the anti-Fas antibody Jo2, in vitro [7,13,46] and in vivo [13,46]. Consistently, we observed that the thymocyte apoptosis induced by anti-Fas from Casp8 ΔE385/ΔE385 mice was compromised compared to that from WT mice (Fig. S2A), and anti-Fas antibody also induced less caspase-3 cleavage in Casp8 ΔE385/ΔE385 thymocytes ( Fig. 2A). To further investigate the role of caspase-8 cleavage in apoptosis, we treated Casp8 ΔE385/ΔE385 MDFs with a RIPK3 kinase inhibitor, GSK'872, to induce apoptosis [50]. We observed that Casp8 ΔE385/ ΔE385 MDFs were strongly resistant to apoptosis induced by GSK'872 ( Fig. 2B). This finding was confirmed by attenuating the cleavage of caspase-3 in Casp8 ΔE385/ΔE385 MDFs (Fig. 2C). To further verify the contribution of caspase-8 cleavage in apoptosis in vivo, we challenged the anti-Fas antibody, Jo2, by intravenous injection in Casp8 ΔE385/ΔE385 and WT mice. In accordance with previous studies [13,46], Casp8 ΔE385/ΔE385 mice were significantly protected from the Jo2-induced lethal effects compared to WT mice (Fig. 2D). Accordingly, Casp8 ΔE385/ΔE385 mice exhibited alleviated liver damage and decreased alanine aminotransferase (ALT)/aspartate aminotransferase (AST) concentrations in the plasma compared to the liver function in WT control mice ( Fig. 2E and S2B). In line with these data, we observed the absence of caspase-8 cleavage and a significant decrease in caspase-3 cleavage in the livers of Casp8 ΔE385/ΔE385 mice (Fig. S2C), suggesting that the lethal effects exerted by the anti-Fas antibody Jo2-induced apoptosis were decreased in Casp8 ΔE385/ΔE385 mice in vivo. These results suggested that blocking cleavage between the large and small catalytic subunits by CASP8(ΔE385) is enough to prevent apoptosis in vitro and in vivo.
and observed a plummeted number of B cells (CD19 + ) and T cells (CD3 + ) in the spleen and bone marrow (Fig. 5B). The macrophages and granulocytes (CD11b + ) were rapidly increased in the spleen but were approximately normal in the bone marrow (Fig. 5B). These results characterized the lymphopenia and myeloid bias disease but excluded the possibility of myeloproliferative disease in these mice. Furthermore, we analyzed the subsets of B cells and T cells in the spleen and bone marrow. Consistent with the percentage results (Fig. S4C), immature and mature B cells (B220 + IgM + /B220 hi CD19 hi ), progenitor B cells (pro-B) and precursor B cells (pre-B) (B220 + IgM − /B220 low CD19 low ), and CD8 + T cells showed a dramatic decrease in the bone marrow of Casp8 ΔE385/ΔE385 Ripk3 −/− and Casp8 ΔE385/

DISCUSSION
Caspase-8 is a key regulator of apoptosis and necroptosis, as well as the inflammatory response through its dimerization and enzymatic activity [1,5,16]. The auto-cleavage activity of Caspase-8 has also been shown to be involved in mediating apoptosis and regulating inflammation [13].
In this study, we demonstrated that CASP8(ΔE385) not only compromised Fas-induced apoptosis and switched TNF-a induced apoptosis to necroptosis but also promoted necroptosis both in vitro and in vivo. However, in contrast to the embryonic lethality observed in caspase-8 deficient [49] or with catalytically inactive caspase-8 mice [6,8], Casp8 ΔE385/ΔE385 mice survived normally, suggesting that primarily caspase-8 catalytic activity rather than caspase-8 cleavage contributes to the suppression of RIPK3-MLKL mediating necroptosis during embryo development.
In this study, we identified the phenotypes of Casp8 ΔE385/ΔE385 mice which resemble those of the Casp8 DA/DA mice from a recent study [13]. Moreover, we also confirmed the enzymatic activity of CASP8(ΔE385) by examining caspase-3 cleavage in thymocytes with FasL treatment [46]. We found that Casp8 ΔE385/ΔE385 thymocytes showed comparable level of caspase-3 cleavage and cell death to that in Casp8 D387A/D387A thymocytes after FasL treatment ( Figs. 2A and S2A), which indicated CASP8(ΔE385) has comparable enzymatic activity as caspase-8(D387A) [46]. However, we still cannot exclude the possibility that deletion of one amino acid in caspase-8 alters other caspase-8-mediated cellular signaling, therefore, whether E385 deletion influences other functions of caspase-8, in addition to its auto-cleavage, needs to be investigated further.

ΔE385
Ripk3 −/− and Casp8 ΔE385/ΔE385 Mlkl −/− mice develop severe lymphopenia that can be prevented by reducing the RIPK1 dosage by half, not by RIPK1 kinase inactive mutant. This indicates that caspase-8 cleavage cooperating RIPK3/MLKL to regulate RIPK1 scaffold-dependent but RIPK1 kinase-independent function contributes to the maintenance of immune cell homeostasis. The exact signaling pathway and mechanism require further investigation.
For flowcytometry analysis, thymocytes were plated in 6-well plates followed by stimulation at a concentration of 1 × 10 6 cells per well, and thymocytes were treated with 2 μg/ml anti-Fas antibody (Jo-2, 554255, BD) for 24 h followed by staining with FITC-Annexin V and PI utilizing apoptosis detection kit (C1062L, Beyotime). After staining, cells were analyzed in cytoflex S flow cytometer (cytoflex S, Beckman Coulter). All analyses were performed using CytExpert software (CytExpert, Beckman Coulter, Inc.).

Anti-Fas induced hepatocellular apoptosis and analysis of the serum and liver damage
The wild-type and Casp8 ΔE385/ΔE385 mice of 8-to 12-week old were injected intravenously with anti-Fas antibody (Jo-2, 554255, BD) in the dose of 0.5 μg/g and their survival time was followed for 20 h. At the indicated times, their livers and peripheral blood were harvested followed by processing for histological analysis, western blot and analyzing the alanine transaminase (ALT) and aspartate transaminase (AST) levels in serum. To analyze the ALT and AST levels in serum, the peripheral blood of the indicated mice were collected in anticoagulation tube followed by centrifugation at 7000 g, 4°C for 30 min. The serum was collected to detect ALT (3040280, Shanghai Shensuo UNF Medical Diagnostic Articles Co.) and AST (3050280, Shanghai Shensuo UNF Medical Diagnostic Articles Co.) level utilizing the kit.

TNF-α induced mice toxicity and analysis of the body temperature
The WT, Casp8 ΔE385/ΔE385 , Casp8 ΔE385/ΔE385 Ripk3 −/− and Casp8 ΔE385/ ΔE385 Ripk1 K45A/K45A mice of 8-to 16-week old were injected intravenously with TNF-α (CRT192C, Cell sciences and obtained from Dr. Yi Zhang at Shanghai Institute of Nutrition and Health, CAS) in the dose of 7 μg each mouse and their body temperature was measured every 2 h until the twelfth hour after injection.

Flow cytometry analyses
Lymphocytes were isolated from the peripheral blood, spleen, bone marrow and lymph nodes of the indicated mice. Total cell numbers were counted using counting slides (SD-100, Nexcelom) in Cellometer Mini Automated Cell Counter (Nexcelom). Surface antigens were stained with indicated conjugated primary antibodies in the staining buffer (1 × PBS, 3% BSA, 1 mM EDTA, 0. Analyses of CD11b + F4/80 + peritoneal macrophages in vivo Wild-type, Casp8 ΔE385/ΔE385 and Ripk1 +/− Ripk3 −/− Casp8 ΔE385/ΔE385 mice were injected intraperitoneally with vehicle or zVAD (20 mg/kg) 1 h before intraperitoneal injection with PBS or LPS (10 mg/kg). Animals were killed at twenty fourth hour after the first injection, resident peritoneal cells were harvested by lavage of the peritoneal cavity with 8 ml PBS. CD11b + F4/80 + peritoneal macrophages were analyzed by flow cytometry.

Bone marrow transplantation assay
All of the recipient mice were wild type with C57BL/6 background, which received 11 Gy of total body irradiation in a split dose (550 rads) with 4-hour rest between doses using a Cesium-137 irradiator. Irradiated recipients were reconstituted by intravenous injection of 2.5 × 10 6 bone marrow cells from femurs and tibias of the 6-week old indicated genotype mice. Recipients were sacrificed at fourth months after reconstitution.

Whole blood count analysis
The whole peripheral blood of the indicated mice was collected in anticoagulation tube followed by diluting in EDTA buffer (0.5 M EDTA pH8.0) at a ratio of 1:1, and then diluted peripheral blood was analyzed on an auto hematology analyzer (BC-2800Vet, Mindray).

Quantification and statistical analysis
Please refer to the figure legends for description of sample size (n) and statistical significance. Data were analyzed with GraphPad Prism 8.0 software using the two-tailed unpaired Student t test or two-sided Log-rank (Mantel-Cox) test. Bars, mean ± standard deviation (mean ± SD). Differences were considered statistically significant when the P < 0.05, where * * * * p < 0.0001, * * * p < 0.001, * * p < 0.01, * p < 0.05, ns, not significant.