COP9 Signalosome Suppresses RIPK1-RIPK3 Mediated Cardiomyocyte Necroptosis in Mice

Animal models

Perinatal cardiomyocyte-restricted ablation of the Cops8 gene (Cops8^CKO) was achieved in C57BL/6J inbred mice as we previously reported.¹³ The creation of RIPK3 null mice was previously described.³¹ Mice with germline knockout of the Ppif gene (encoding Cyclophilin D) were provided by Dr. Jeffrey Molkentin of University of Cincinnati.²²

The animal care and use protocols for this study were approved by the Institutional Animal Care and Committee of the University of South Dakota and followed the NIH guide for the care and use of laboratory animals.

Evan’s blue dye (EBD) uptake assay

After absorbed into the circulatory system, EBD is bound by albumin in the blood. Therefore, in this assay, EBD does not enter the cell with intact plasma membrane; EBD only enters a cell whose membrane permeability is abnormally increased, a key feature that distinguishes necrosis from apoptosis.³ Hence, the in vivo EBD uptake assay was performed to detect CM necrosis in mice at 3 weeks of age as we reported.^{13, 14}

Necrostatin-1 (Nec-1) treatment

At 2 weeks of age, 19 Cops8^CKO mice were continuously administered Nec-1 (BML-AP309, Enzo Life Science; 1.56 mg/kg/day) or vehicle (10% DMSO in PBS) by intraperitoneal implantation of osmotic mini-pumps (Alzet Model 1002, designed for continuous drug delivery for 2 weeks). Two cohorts of mice were included. For CM necrosis analysis using the EBD uptake assay as described above, one cohort (3 mice per group) was sacrificed 7 days after implantation of the mini-pump. The other cohort was used for Kaplan-Meier survival analysis.

Dihydroethidium (DHE) staining for reactive oxygen species (ROS)

Three mouse hearts per group were perfused in situ and excised in PBS, embedded in OCT and rapidly frozen. Serial cryosections (10 μm thick) were mounted onto glass slides. The slides were air-dried and incubated with 2.5 μM DHE in PBS at 37°C for 30 min. When oxidized to ethidium bromide by the superoxide anion, DHE produces a red fluorescence.³² The red fluorescence derived from DHE was measured with confocal microscopy and is used as the indicator of ROS content.

Western blot analyses and co-immunoprecipitation assays (Co-IP)

Western blotting and Co-IP were performed as described.^{33, 34}

Protein carbonyl assays

Protein carbonyls, biomarkers of oxidative stress, were assessed as described.³⁵ Briefly, Ventricular protein samples were subjected to 2,4-dinitrophenylhydrazine (DNPH) derivatization. The carbonyl groups in the protein side chains are derivatized to 2,4-dinitrophenylhydrazone (DNP-hydrazone). The DNP-derivatized proteins were then immuno-detected with an anti-DNP antibody.

Caspase 8 activity assays

The activities of caspase 8 in myocardial crude protein extracts were measured using the Caspase-8 Colorimetric Assay Kit (BioVision, Inc., USA).

Statistical analyses

GraphPad Prism software (San Diego, CA) was used. All continuous variables are presented as mean±SEM unless indicated otherwise. All data were examined for normality with the Shaprio Wilk’s test prior to application of parametric statistical tests. Those that failed this test were analyzed with non-parametric methods. Tests are specified in figure legends. In all cases where 1-way ANOVA was used, post hoc Tukey’s test was performed for pairwise comparisons only when ANOVA showed an overall significant effect. A p value <0.05 is considered statistically significant.

Key proteins of the necroptotic pathway are increased in Cops8^CKO mouse hearts

To explore the mechanism governing the CM necrosis in Cops8^CKO mice, we examined the potential involvement of the RIPK1-RIPK3 pathway. Western blot analyses revealed that myocardial protein levels of RIPK1, RIPK3, and MLKL in Cops8^CKO mice were significantly higher than in littermate control mice (Figure 1A, 1B). Co-immunoprecipitation of RIPK1 detected greater association of RIPK3 with RIPK1 in Cops8^CKO hearts compared with littermate controls (Figure 1C, 1D).

Cops8 deficiency increases myocardial oxidative stress

Next, we examined whether oxidative stress was increased in CMs of Cops8^CKO mice. The level of superoxide anion (O₂⁻) in myocardial sections was probed with the DHE staining assay. The red fluorescence intensity of the DHE-probed myocardial sections from homozygous Cops8^CKO mice was greater than that from either heterozygous Cops8^CKO or Cops8^FL/FL control mice (Figure 2A, 2B), indicating that Cops8 deficiency increases myocardial superoxide levels. Myocardial ROS were also assessed via immunoblotting for DNPH-derivatized protein carbonyls. Immuno-probing of DNP in protein dot blots revealed that myocardial protein carbonyls were higher in the homozygous Cops8^CKO mice compared with heterozygous Cops8^CKO, Cops8^FL/FL, or Myh6-Cre^TG mice (Figure 2C, 2D). Western blot analyses further showed that the increased carbonyls were mainly on proteins of a molecular weight ranging from 25 to 37 kDa (Figure 2E). These findings indicate that Cops8 deficiency in CMs increases myocardial oxidative stress.

Impaired caspase 8 activation and upregulated BCL2 expression in Cops8^CKO hearts

Since necroptosis was originally observed in TNFα-treated cells whose caspase 8 is defective or suppressed, we sought to examine myocardial expression and activity of caspase 8 in Cops8^CKO mice. Both the cleaved/activated form of caspase 8 and the activities of caspase 8 were significantly lower but the abundance of the full-length caspase 8 was discernibly greater in Cops8^CKO mice than in littermate controls at 3 weeks of age (Figure 3A ~ 3C). Myocardial protein levels of BCL2, a key inhibitor of the mitochondrial apoptotic pathway, were significantly greater in 3-week-old homozygous Cops8^CKO mice, compared with heterozygous Cops8^CKO and Cops8^FL/FL littermates (2.98±0.10 vs. 1.97±0.20, 1.00±0.16; p=0.010, <0.001, respectively; Figure 3D). Myocardial BCL2 mRNA levels were also greater in homozygous Cops8^CKO mice than littermate controls at both 2 and 3 weeks of age (Figure 3E).

Suppression of CM necrosis and delay of premature death by RIPK1 inhibition in Cops8^CKO mice

To determine whether RIPK1 kinase activity is required for CM necrosis in Cops8^CKO hearts, we tested the impact of necrostatin-1 (Nec-1), a RIPK1 kinase-specific inhibitor.³⁶ Since CM necrosis is detectable in Cops8^CKO mice at 3 weeks, but not at 2 weeks of age, the administration of Nec-1 or vehicle control via intraperitoneal implantation of osmotic mini-pumps was initiated at 2 weeks. CM necrosis was assessed using the in vivo EBD uptake assay in the heart harvested 7 days after mini-pump implantation. EBD positive CMs were not detectable in mice with control genotypes (Myh6-Cre^TG, Cops8^FL/FL, and Cops8^+/+; data not shown) but were readily detectable in homozygous Cops8^CKO mice treated with vehicle control. Strikingly, the CM EBD positivity was strikingly less in the Nec-1 treated Cops8^CKO mice than in the vehicle treated (0.005% vs. 0.20%, p<0.001; Figure 4A, 4B), indicating that RIPK1 kinase activity is required for Cops8 deficiency to induce CM necrosis in mice. Moreover, Kaplan-Meier survival analyses revealed that Nec-1 treatment significantly delayed the premature death observed in Cops8^CKO mice (p=0.007; Figure 4C).

Requirement of RIPK3 for CM necrosis in Cops8^CKO mice

To test the role of RIPK3 in the CM necrosis of Cops8^CKO mice, RIPK3 germline knockout (RIPK3^−/−) mice were cross-bred with Cops8^CKO mice and the resultant Cops8^CKO::RIPK3^+/+ and Cops8^CKO::RIPK3^+/− littermate mice were subjected to EBD CM necrosis assessment at 3 weeks of age as well as Kaplan-Meier survival analysis. The prevalence of EBD-positive CMs in Cops8^CKO::RIPK3^+/− mice was lower than that of littermate Cops8^CKO::RIPK3^+/+ mice (p<0.001; Figure 5A, 5B;); also, the lifespan of the former was longer than that of the latter (p<0.001; Figure 5C). These analyses show that RIPK3 haploinsufficiency is capable of suppressing CM necrosis and delaying premature death in Cops8^CKO mice.

CM necroptosis in Cops8^CKO mice is independent of mitochondrial permeability transition (MPT)

To determine whether MPT-driven necrosis contributes to CM necrosis in Cops8^CKO mice, we cross-bred Cops8^CKO mice with Ppif gene knockout mice to test whether ablation of the Ppif gene would mitigate the CM necrosis and mouse premature death induced by Cops8^CKO. Compared to mice harboring Cops8^CKO alone, mice harboring Cops8^CKO coupled with either heterozygous or homozygous knockout of the Ppif gene did not show a delay in premature death; on the contrary, homozygous Ppif knockout moderately increased CM necrosis (p=0.010) and accelerated mouse premature death caused by Cops8^CKO (p=0.007; Figure 6). These results indicate that MPT is not a mediator for CM necrosis in Cops8^CKO mice.

Cops8 deficient or CSN inhibited CMs die primarily from necroptosis

Massive CM necrosis occurs in Cops8^CKO mice, which is evidenced by increased EBD uptake by CMs in the absence of increased TUNEL positivity, and by the ultrastructural features like CM swelling and a broken plasma membrane.^12–14 Activation of RIPK3 is the centerpiece of necroptotic pathway although RIPK1 is also required in the induction of necroptosis by TNFα at least.³⁷ The detection of necroptosis currently requires a combination of rather sophisticate tests to reveal both the necrotic feature (e.g., loss of plasma membrane integrity) and the dependence on RIPK3 activation, according to recently published guidelines.³⁸ Here we found that CM necrosis in Cops8^CKO mice was associated with increases in RIPK1, RIPK3, MLKL, and RIPK1-bound RIPK3 protein levels (Figure 1) and depended on both RIPK1 kinase activity (Figure 4) and increased expression of RIPK3 (Figure 5). These new findings demonstrate unequivocally that the massive CM necrosis induced by Cops8^CKO in mice belongs to necroptosis and the RIPK1-RIPK3 pathway mediates its occurrence. Notably, in contrast to a recently delineated RIPK3-CamKII-MPT pathway to cardiac necroptosis,¹⁹ MPT does not play a major role in the execution of CM necroptosis in Cops8^CKO mice. This is because Cyclophilin D knockout, which is known to inhibit MPT, did not attenuate but rather exacerbated CM necrosis and premature death in Cops8^CKO mice (Figure 6). Indeed, MPT- and even mitochondria-independent but RIPK3-depdent necroptosis was reported for several other cell types.^39–41 It will be interesting to test whether CaMKII is involved in CM necrosis in Cops8^CKO mice.

How does Cops8 deficiency cause CM necroptosis?

The requirement of both RIPK1 and RIPK3 by the CM necrosis observed here suggests that the induction of CM necroptosis by Cops8^CKO shares the same pathway taken by TNFR1 activation (Figure 7). The ligation of TNFR1 by TNFα can lead to at least 3 possible downstream events: (1) formation of complex 1 where RIPK1 serves as a scaffold in a manner independent of its kinase activity, which provides survival signals via activation of nuclear factor κB (NFκB) and mitogen-activated protein kinases (MAPKs), (2) formation of complex 2a, which induces apoptosis via caspase 8 and downstream cascade, and (3) formation of complex 2b (i.e., the RIPK1-RIPK3-MLKL) and thereby induction of necroptosis when caspase 8 is defective or inhibited.³ The kinase activity of RIPK1 is required for RIPK1 to induce cell death in complex 2. UPS-dependent degradation of IκBα is a key step in the activation of NFκB by TNFα where the ubiquitination of IκBα is driven by Skp1-Cul1-β-TrCP (SCF^β-TrCP),⁴² a member of the CRL1 family E3 ligases whose assembly and disassembly are regulated by the CSN. Hence, the survival signaling from NFκB is likely suppressed by impairment of IκBα ubiquitination due to defective Cullin deneddylation resulting from Cops8 deficiency. Our prior study detected decreases in myocardial F-box protein β-TrCP protein levels in Cops8^CKO mice,¹³ adding a reason to predict a reduction of SCF^β-TrCP ligase activities. Thus, Cops8 deficiency swings TNFR1 signaling towards the cell death direction.

Then, the next question is why necroptosis instead of apoptosis takes place. At least in the case of induction of necroptosis by death receptor activation, two prerequisites must be met. First, the formation of the so-called complex 2 containing RIPK1 and RIPK3 and second, the failure of caspase 8 to activate.³ Indeed, both prerequisites were met in the Cops8^CKO hearts. Not only were RIPK1, RIPK3, and MLKL protein levels markedly increased but also RIPK1-RIPK3 interaction was significantly augmented as indicated by the RIPK1-RIPK3 Co-IP results (Figure 1); and importantly the cleaved form of caspase 8 as well as caspase 8 activity were substantially lower in the homozygous Cops8^CKO hearts compared with CTL hearts (Figure 3). It is very likely that this impairment of caspase 8 activation directly result from the loss of Cullin deneddylation because a prior study has established that Cul3-RBX1 mediated polyubiquitination of caspase 8 is required for further processing and activation of caspase 8 and the signaling of the extrinsic apoptotic pathway.⁹ Both neddylation and deneddylation of Cullins are required for proper functioning of CRLs; hence, the ubiquitination of caspase 8 by Cul3-RBX1 is very likely suppressed by Cops8 deficiency. This argument is strongly supported by a recent report that inhibition of CRLs with the neddylation inhibitor MLN4924 sensitizes monocyte necroptosis in vitro.⁴³ Besides caspase 8 which is essential to the extrinsic pathway of apoptosis, as discussed below, the mitochondrial pathway is likely suppressed by increased BCL2 (Figure 3).¹³

We have previously observed a suppressed autophagic flux in Cops8^CKO mice. This could probably be due to impairment in autophagosome-lysosome fusion that occurs before impairment in the UPS degradation of a surrogate misfolded protein as well as CM necrosis become discernible.¹⁴ We propose dual impairment of both the UPS and the ALP plays an overall causative role in the CM necrosis that now proves to be necroptosis. This proposition now has support from two recent studies that collected evidence from cultured H9c2 cells suggesting a major contribution from impaired autophagy to the induction of necroptosis by TNFα.^{44, 45} According to these reports, the induction of RIPK1-RIPK3 interaction and necroptosis by the combined treatment with TNFα and a broad spectrum caspase inhibitor was associated with suppression of autophagic flux;⁴⁴ improving autophagic flux via mTORC1 inhibition suppressed the necroptosis in an autophagy- and transcription factor EB (TFEB; a master regulator of the ALP)-dependent manner;^{44, 45} and MPT did not play a major role in the execution of necroptosis.⁴⁴ This scenario starkly resembles what we have unveiled in the Cops8^CKO mouse myocardium. Reduced myocardial autophagic flux was found to occur in the Cops8^CKO mice at as young as 1 week of age,¹⁴ which is two weeks earlier than CM necroptosis becomes discernible.¹³ Hence, it will be interesting and important to test whether the autophagic impairment has exacerbated activation of the RIPK1-RIPK3-MLKL necroptotic pathway in Cop8^CKO mice.

Previous reports have shown an important role of increased reactive oxygen species (ROS) in RIPK3-mediated necroptosis in cultured cells.^{46, 47} In TNFα induced necroptosis, the RIPK3-centered necrosome increases ROS production through stimulating aerobic metabolism and RIPK3 does so probably by activating key enzymes of metabolic pathways including glycogen phosphorylase (PYGL), glutamate-ammonia ligase (GLUL), glutamate dehydrogenase 1 (GLUD1),⁴⁶ and more recently pyruvate dehydrogenase (PDH) which is a rate-limiting enzyme linking glycolysis to aerobic respiration.⁴⁸ The increased ROS further promotes necrosome formation and yields cytotoxicity during necroptosis.⁴⁷ As reflected by increased DHE staining and the elevated levels of protein carbonyls in Cops8^CKO hearts (Figure 2), increases in ROS or oxidative stress are indeed associated with CM necroptosis in Cops8^CKO mice. Consistent with increased oxidative stress, our prior transcriptome analysis revealed that the Nrf2 pathway, the master regulator of antioxidant and defensive responses, was markedly upregulated in Cops8^CKO hearts even before CM necrosis becomes discernible.⁴⁹ The exact role of altered redox state in the induction of CM necroptosis by Cops8 deficiency remains to be determined.

Clinical implications

First of all, inadequate cardiac PQC due to UPS malfunction and ALP impairment has been implicated in the progression from a large subset of heart disease to heart failure;^{1, 50} however, the mechanistic link between impaired PQC and heart failure remains obscure. The discoveries of the present study implicate that CM necroptosis could be one of the missing links, because cardiac PQC impairment is obviously the apical defect in Cops8^CKO mice. Accordingly, targeting the necroptotic pathway could potentially help alleviate the adverse outcome of cardiac PQC impairment. Second, a small molecule inhibitor of the CSN (CSN5i) has shown great promise in anti-tumor effects in experimental studies.¹¹ Hence, there is a good possibility for this compound to move into clinical trials for cancer treatment. The findings of the present study caution that cardiac function should be closely monitored should CSN5i or alike be moved into clinical trials.