RIPK1 kinase inactive mice are viable and protected from TNF-induced necroptosis in vivo

Abstract

The serine/threonine kinase RIPK1 is recruited to the TNF receptor 1 to mediate pro-inflammatory signalling and to regulate TNF-induced cell death. A RIPK1 deficiency results in perinatal lethality, impaired NFκB and MAPK signalling and sensitivity to TNF-induced apoptosis. Chemical inhibitor and in vitro reconstitution studies suggested RIPK1 displays distinct kinase activity dependent and independent functions. To determine the contribution of RIPK1 kinase to inflammation in vivo, we generated knock-in mice endogenously expressing catalytically inactive RIPK1 D138N. Unlike Ripk1^−/− mice, which die shortly after birth, RIPK1^D138N/D138N mice are viable. Cells expressing RIPK1 D138N are resistant to TNF- and poly (I:C)-induced necroptosis in vitro and RIPK1^D138N/D138N mice are protected from TNF-induced shock in vivo. Moreover, RIPK1^D138N/D138N mice fail to control Vaccinia virus replication in vivo. This study provides genetic evidence that the kinase activity of RIPK1 is not required for survival but is essential for TNF-, TRIF- and viral-initiated necroptosis.

Introduction

RIPK1 is the founding member of a serine-threonine kinase family that transduces inflammatory and cell death signals following death receptor ligation, activation of pattern recognition receptors and DNA damage. RIPK1 is the core component of TNF-induced signalling complexes mediating NF-κB and MAP kinase activation, apoptosis and an alternative form of caspase independent cell death called necroptosis. TNF binding to the TNF receptor 1 (TNFR1) leads to the formation of an intracellular complex that includes RIPK1, TRADD, TRAF2, cIAP1/2 and the components of the linear ubiquitin chain assembly complex (LUBAC) (complex I) (1). RIPK1 is modified by the addition of K63-linked and linear ubiquitin chains, resulting in the recruitment of NEMO/IKKγ and TAK1/TAB2/TAB3 complex to mediate IKK activation (1). Deubiquitination of RIPK1 by CYLD or A20 results in formation of a cytosolic complex, containing RIPK1, FADD, cFLIP and caspase 8 (complex IIb)(2). Caspase 8 then cleaves and inactivates RIPK1, RIPK3 and CYLD and stimulates apoptosis (3-5). In the absence of caspase 8 or when apoptosis is inhibited, RIPK1 interacts with RIPK3 to induce necroptosis via the recruitment of MLKL (2).

Necroptosis is known to contribute to several types of pathological injury and can be triggered by TNF family members (TNF, Fas or TRAIL), Toll-like receptors (TLR3 and TLR4) or the DNA sensor (DAI) (2). Necroptosis is thought to depend on auto- and transphosphorylation of RIPK1 and RIPK3, resulting in the RIPK3-mediated recruitment of MLKL and other substrates (6). Although there is ample genetic evidence linking RIPK3 to necroptosis that occurs during development, inflammation and viral infection (6) genetic evidence implicating RIPK1 to necroptosis in vivo is limited (7).

Materials and Methods

Mice

Ripk1^D138N/D138N knock-in mice were generated by mutating the conserved aspartate (D) at position 138 to asparagine (D138N). The Ripk1^D138N construct was introduced into Bruce 4 embryonic stem (ES) cells derived from C57BL/6 mice. Mice were maintained at the SPF animal facilities of the University of Massachusetts Medical School and the Institute for Genetics at the University of Cologne. All animal procedures were conducted in accordance with national and institutional guidelines. Sex and age-matched mice were used in all experiments.

In vitro kinase assay and immunoblotting

MEF were left untreated or treated with mTNFα (50 ng/ml) and RIPK1 was immunoprecipitated and the kinase assay performed in the presence or absence of Necrostatin-1 (30μM) in kinase buffer containing 10 μCi γ-³²P-ATP. The samples were separated by SDS-PAGE and visualized by autoradiography. Wild type or Ripk1^D138N/D138N MEF or BMDM were left untreated or treated with 10ng/ml mTNF, 25μg/ml Poly (I:C) or 100 ng/ml LPS for the indicated time periods and signaling examined as described previously (8).

Necroptosis Assays

Wild type, Ripk1^D138N/D138N, Ripk1^−/− and Ripk3^−/− MEF were pretreated with 10μg/ml cycloheximide, 20μM zVAD-fmk, 10μM SMAC mimetic or 30μm Nec-1 then treated with 10ng/ml mTNF. Wild type or Ripk1^D138N/D138N BMDM were pretreated with 20μM zVAD-fmk or 30μM Nec-1 then treated with 50μg/ml poly(I:C) and cell survival determined by neutral red assay.

TNF-induced shock experiments

Age- and sex-matched wild type, Ripk1^D138N/D138N and Ripk3^−/− mice were injected with 9μg mTNF only or with zVAD-fmk (16.7mg/kg). Body temperature and survival was determined.

Vaccinia Virus Infections

Ten-week old wild type, Ripk1^D138N/D138N and Ripk3^−/− mice were infected with 2 × 10⁶ pfu of Western Reserve or B13R-deficient strain of Vaccinia virus via the intraperitoneal route. Tissue extracts were harvested and titers were determined by plaque assays on Vero cells.

Results and Discussion

To reveal the in vivo physiologic role(s) of the serine/threonine kinase activity of RIPK1, we generated knock-in mice expressing a kinase inactive mutant RIPK1 from the endogenous Ripk1 locus. The mutation introduced resulted in the replacement of the conserved aspartate (D) at position 138 within the activation loop of the RIPK1 kinase domain with asparagine (Ripk1^D138N)(Figure 1A). To determine whether the D138N mutation alters the kinase activity of RIPK1, we immunoprecipitated RIPK1 from RIPK1^D138N/D138N and wild type MEF that were left untreated or stimulated with TNF and performed an in vitro kinase assay in the presence of γ³²P-ATP. RIPK1 autophosphorylation was detected in both unstimulated and TNF-treated wild type cells, which was inhibited by the RIPK1 inhibitor Necrostatin-1 (Nec-1)(Figure 1B). Although an equivalent amount of the RIPK1 D138N protein was immunoprecipitated, no RIPK1 autophosphorylation was observed even in the presence of TNF (Figure 1B). Moreover, the D138N mutation appeared to have no detectable effects on RIPK1 expression levels (Figure 1C).

Figure 1. Generation of kinase inactive Ripk1 allele.

A. Schematic diagram of the mouse Ripk1 locus and the Ripk1 kinase inactive allele. B. The D138N mutation impairs RIPK1 kinase activity. Wild type and Ripk1^D138N/D138N MEFs were left untreated or stimulated with mTNF and an in vitro kinase assay performed in the presence or absence of Necrostatin-1 (Nec-1). The amount of RIPK1 in each immunoprecipitation was determined by immunoblotting with a RIPK1 Ab. C. RIPK1 protein levels in Ripk1^D138N/D138N mice. RIPK1 protein levels were examined by immunoblotting lysates from MEFs and macrophages isolated from wild type and Ripk1^D138N/D138N mice. D. The kinase activity of RIPK1 is not required for TNF- induced NF-κB, JNK, p38 MAPK or ERK activation. Wild type and Ripk1^D138N/D138N macrophages were left untreated or treated with TNF for the time periods indicated and protein lysates were probed with phospho-specific antibodies for JNK, p38 MAPK and ERK. Lysates were also probed with total IκBα, JNK1/2, p38α, ERK and β-actin.

RIPK1 is recruited to the TNFR1 to mediate the activation of NF-κB, p38 MAPK, JNK and ERK. RIPK1 is also recruited to the Toll like receptor (TLR)3/4 adapter TRIF and contributes to TRIF-mediated NF-κB activation and necroptosis (2, 9). A role for the kinase activity of RIPK1 in TNF-induced ERK activation has been suggested (10). Therefore, we examined TNF- and TRIF-dependent signalling in MEF and primary bone marrow derived macrophages (BMDM). Cells expressing RIPK1 D138N showed normal activation of these signalling pathways in response to TNF, poly (I:C) and LPS (Figure 1D and Supplemental Figure 2). These data are consistent with studies using Necrostatin-1 to inhibit RIPK1 (11, 12) and with our prior studies using Ripk1^−/− MEF reconstituted with a kinase inactive form of RIPK1 (8).

The kinase activity of RIPK1 is not essential for survival or for TNF- or TRIF-dependent activation of NF-κB or MAP kinase signaling

Unlike mice with a complete RIPK1 deficiency, which die at/around time of birth (13), RIPK1^D138N/D138N mice are born at the expected Mendelian ratios and show no gross or histological abnormalities (Supplemental Figure 1). A complete RIPK1 deficiency results in perinatal death with evidence of cell death in the thymus, lymph nodes and subcutaneous tissue, accompanied by an inflammatory response characterized by granulocyte and macrophage infiltration (13). To determine whether loss of RIPK1 kinase activity causes similar pathology in vivo, we examined tissues from Ripk1^D138N/D138N mice for evidence of cell death and inflammation. Histopathological examination and cleaved caspase 3 staining of tissues from Ripk1^D138N/D138N mice revealed no evidence of cell death or inflammation (Supplemental Figure 1). These results demonstrate that the kinase activity of RIPK1 is not required for mouse survival.

RIPK1^D138N/D138N MEF and macrophages are resistant to TNF- and poly (I:C)-induced necroptosis

Necroptosis can be initiated by TNF or pattern recognition receptors (PRR) (2). To investigate necroptotic death in response to TNF, wild type, Ripk1^D138N/D138N, Ripk1^−/− or Ripk3^−/− MEF were pretreated with cycloheximide, the pan caspase inhibitor zVAD-fmk and/or Necrostatin-1 prior to the addition of mTNF. Ripk1^D138N/D138N MEFs were protected from TNF/zVAD-induced necroptosis but exhibit similar levels of TNF-induced apoptosis as wild type cells (Figure 2A). As expected, a RIPK1-deficiency sensitized cells to TNF-induced apoptosis and a RIPK1- or RIPK3-deficiency rendered cells resistant to TNF-induced necroptosis (Figure 2A). The kinase activity has also been implicated in TNF-induced apoptosis induced under conditions of cIAP depletion (14). Consistent with these studies, we found RIPK1 and it’s kinase activity required for apoptosis induced by TNF and Smac mimetic treatment (Figure 2B).

Figure 2. Ripk1^D138N/D138N MEF and BMDM are protected from TNF- and poly (I:C)-induced necroptosis.

A. Wild type, Ripk1^D138N/D138N, Ripk1^−/− and Ripk3^−/− MEFs were pretreated with zVAD-fmk, Nec-1, or cycloheximide prior to treatment with mTNF. B. Wild type, Ripk1^D138N/D138N, and Ripk1^−/− MEFs were treated with TNF and /or SMAC and cell viability determined. C. Wild type or Ripk1^D138N/D138N BMDM were pretreated with zVAD-fmk and/or Nec-1 prior to treatment with poly (I:C). Cell viability was analyzed by neutral red assay and percent survival is shown. ***p<0.0001.

The TLR3/4 specific adapter TRIF recruits RIPK1 via the Rip Homotypic Interaction Motif (RHIM) to mediate NF-κB activation and when apoptosis is inhibited is thought to recruit RIPK1 and RIPK3 to mediate necroptosis (6, 9). However, TRIF-dependent necroptosis can occur in the absence of RIPK1 and may be mediated by the direct recruitment of RIPK3 to TRIF and TLR3/4 (15). To determine whether the kinase activity of RIPK1 mediates TRIF-dependent necroptosis, BMDMs from wild type or Ripk1^D138N/D138N mice were left untreated or pretreated with zVAD-fmk and/or Nec-1 for 1 hour prior to stimulation with poly (I:C). Poly (I:C) and zVAD-fmk treatment of wild type BMDM induced cell death that was prevented by Nec-1 pretreatment (Figure 2C). We find RIPK1^D138N/D138N BMDM protected from poly (I:C)-induced necroptosis (Figure 2C). Thus, an absence of RIPK1 kinase activity had no detectable effect on TNF-induced apoptosis (TNF/Cx) but proved essential for TNF- and poly (I:C)-induced necroptosis in vitro.

Ripk1^D138N/D138N mice are protected from TNF-induced hypothermia

TNF administration to mice mimics the histopathological and pathophysiologic changes associated with septic shock. We challenged Ripk1^D138N/D138N mice along with wild type and Ripk3^−/− mice with 9μg mTNF, a dose determined previously to induce death in 100% of treated wild type mice. In contrast to wild type controls, the Ripk1^D138N/D138N and Ripk3^−/− mice were protected from hypothermia and all the animals survived the TNF challenge (Figure 3). Thus, expression of a kinase inactive form of RIPK1 was sufficient to provide protection from TNF-induced hypothermia and mortality in the presence or absence of caspase inhibition (Figure 3). Collectively these genetic studies reveal that the mortality induced by TNF in this shock model reflects RIPK1- and RIPK3-mediated necroptotic death in vivo.

Figure 3. Ripk1^D138N/D138N mice are protected from TNF-induced shock.

Body temperatures and survival of wild type, Ripk1^D138N/D138N and Ripk3^−/− mice injected with mTNF (A, B, p<0.0001) or mTNF and zVAD-fmk (C, D p<0.0001).

Necroptosis has emerged as an important host response to viral infection. Vaccinia virus (VV) infected cells become susceptible to necroptosis due to the fact the virus encodes the caspase inhibitor B13R, which inhibits caspase 8 and consequently stimulates RIPK3-dependent necroptosis (16). We found Ripk1^D138N/D138N mice unable to control VV replication, resulting in significant increases in viral titers in the spleen and liver of infected mice (Figure 4A,B). Consistent with the viral loads in these animals, and as observed in Ripk3^−/− mice (16), Ripk1^D138N/D138N mice exhibit reduced liver inflammation when compared to infected wild type mice (Figure 4D,E). In contrast infection of wild type and Ripk1^D138N/D138N mice with the B13R-deficient virus, which lacks the caspase 8 inhibitor yielded similar viral titers in both strains (Figure 4C). These data demonstrate that the host defense to VV infection depends on necroptosis mediated by RIPK3 and the kinase activity of RIPK1.

Figure 4. The kinase activity of RIPK1 is required for protection against Vaccinia Virus infection in vivo.

Vaccinia viral titers in the spleens (A) and livers (B) of wild type, Ripk1^D138N/D138N and Ripk3^−/− mice. C. B13R-deficient viral titers of wild type and Ripk1^D138N/D138N mice (p=0.24). D. Histological examination of liver inflammation. E. Number of focal inflammatory regions for infected wild type, Ripk1^D138N/D138N and Ripk3^−/− mice *p<0.05, **p<0.01, ***p<0.0001.

Our in vivo results show that the kinase activity of RIPK1 is not required for mouse survival but has an essential role in TNF-induced necroptosis. In contrast to in vitro studies where necroptosis is observed in the absence of RIPK1 (11, 15, 17), we find MEF and primary BMDM that express RIPK1 D138N resistant to TNF- and TRIF-induced necroptosis. Consistent with published studies (8, 12), the kinase activity of RIPK1 does not contribute to pro-inflammatory signaling induced by TNF or TLR3/4 ligands (Supplemental Figure 2). Apoptosis induced by TNF and cycloheximide treatment was unaffected in Ripk1^D138N/D138N MEFs whereas the kinase activity was required for apoptosis induced by TNF and Smac mimetic treatment (Figure 2). Thus, the pro-inflammatory and pro-survival functions of RIPK1 are kinase independent whereas its pro-death functions remain kinase dependent.

In conclusion, our in vivo studies in Ripk1^D138N/D138N mice demonstrate that the kinase is not responsible for the pro-survival function(s) of RIPK1 but is indispensable for TNF-induced necroptosis in vivo. Our data also provide genetic evidence to implicate RIPK1 kinase in viral-initiated necroptosis. Our findings support the development of more stable, potent and selective RIPK1 inhibitors to control the systemic inflammation associated with chronic infection, sepsis or other types of tissue injury.