Both authors contributed equally.
RIPK3-Mediated Necroptosis Promotes Donor Kidney Inflammatory Injury and Reduces Allograft Survival
Version of Record online: 18 SEP 2013
© Copyright 2013 The American Society of Transplantation and the American Society of Transplant Surgeons
American Journal of Transplantation
Volume 13, Issue 11, pages 2805–2818, November 2013
How to Cite
Lau, A., Wang, S., Jiang, J., Haig, A., Pavlosky, A., Linkermann, A., Zhang, Z.-X. and Jevnikar, A. M. (2013), RIPK3-Mediated Necroptosis Promotes Donor Kidney Inflammatory Injury and Reduces Allograft Survival. American Journal of Transplantation, 13: 2805–2818. doi: 10.1111/ajt.12447
- Issue online: 28 OCT 2013
- Version of Record online: 18 SEP 2013
- Manuscript Accepted: 2 JUL 2013
- Manuscript Revised: 12 JUN 2013
- Manuscript Received: 6 APR 2013
- Canadian Institutes of Health Research. Grant Numbers: XTW-90932, MOP-111180, MOP-115048
- Kidney Foundation of Canada
- Program of Experimental Medicine (POEM) at LHSC
- kidney allograft survival;
- kidney IRI;
- Top of page
- Materials and Methods
Kidney transplant injury occurs with ischemia and alloimmunity. Members of the receptor interacting protein kinase family (RIPK1,3) are key regulators of “necroptosis,” a newly recognized, regulated form of necrosis. Necroptosis and apoptosis death appear to be counterbalanced as caspase-8 inhibition can divert death from apoptosis to necrosis. Inhibition of necroptosis in donor organs to limit injury has not been studied in transplant models. In this study, necroptosis was triggered in caspase inhibited tubular epithelial cells (TEC) exposed to tumor necrosis factor alpha in vitro, while RIPK1 inhibition with necrostatin-1 or use of RIPK3−/− TEC, prevented necroptosis. In vivo, short hairpin RNA silencing of caspase-8 in donor B6 mouse kidneys increased necroptosis, enhanced high-mobility group box 1 release, reduced renal function and accelerated rejection when transplanted into BALB/c recipients. Using ethidium homodimer perfusion to assess necrosis in vivo, necrosis was abrogated in RIPK3−/− kidneys postischemia. Following transplantation, recipients receiving RIPK3−/− kidneys had longer survival (p = 0.002) and improved renal function (p = 0.03) when compared to controls. In summary, we show for the first time that RIPK3-mediated necroptosis in donor kidneys can promote inflammatory injury, and has a major impact on renal ischemia–reperfusion injury and transplant survival. We suggest inhibition of necroptosis in donor organs may similarly provide a major clinical benefit.
cellular death associated molecular pattern
caspase-independent cell death
hematoxylin and eosin
high-mobility group box 1
human tumor necrosis factor-alpha
total in group
programmed cell death
polymerase chain reaction
receptor interacting protein kinase
standard error of the mean
short hairpin RNA
tubular epithelial cells
tumor necrosis factor-alpha
tumor necrosis factor receptor
- Top of page
- Materials and Methods
Kidney dysfunction following transplantation has multiple etiologies. Tubular epithelial cells (TEC) comprise more than 75% of renal parenchymal cells and are highly susceptible to death from ischemia–reperfusion injury (IRI), reactive oxidative species, nitric oxide, pro-inflammatory cytokines, antibodies and cytotoxic T and natural killer (NK) cells. Their viability, as well as that of other renal parenchymal cells, directs both short- and long-term kidney allograft survival [1-5]. IRI enhances adaptive immune responses and pro-inflammatory cytokine expression that promote rejection [6-8] and the recruitment of T, NK and other cell effectors [9, 10]. Collectively these mechanisms affect organ function and homeostasis by the elimination of parenchymal cells as well as promoting further inflammatory organ injury. It follows that prevention of TEC and inflammatory forms of parenchymal cell death would be expected to reduce delayed graft function and attenuate rejection responses in transplanted kidneys.
While targeting cell death might be useful as a therapeutic strategy in transplantation, a greater understanding of complex intracellular interactions that lead to various forms of cell death will be required for such strategies to be applied effectively. Caspase-dependent apoptosis or “programmed cell death” (PCD) has been regarded as the prototypic form of regulated cell death. While apoptosis induces minimal inflammation and may contribute to immune tolerance [11-13], necrosis and caspase-independent cell death (CICD) are regarded as unregulated forms of cell death induced by severe nonspecific and nonphysiological stress . Necrosis promotes inflammatory injury in kidneys [15, 16] as membrane rupture results in the release of pro-inflammatory endogenous molecules including heat-shock proteins, unprocessed high-mobility group box 1 (HMGB1), uric acid, fibronectin, IL-33 and others [17, 18]. These cellular death associated molecular patterns (CDAMPs) participate in IRI and allograft rejection through interaction with Toll-like receptors (TLRs) and other innate receptors [16, 19-21]. Ligation of surface death receptors (DRs) (TNF receptor 1 [TNFR1], Fas/CD95 and TRAIL-R) recruits adapter proteins, such as Fas-associated death domain, TNFR-associated factor with death domain, receptor interacting protein kinase (RIPK1) and other proteins that allow formation of a complex, which triggers the autocatalytic activation of caspase-8 homodimers and apoptosis. However, recent studies have indicated that a primary function of caspase-8/FLIP-long heterodimers is to prevent a “regulated” form of necrosis (RN) termed necroptosis, which is mediated by RIPK1 and 3 proteins [22, 23]. Necroptosis is morphologically and biochemically indistinguishable from most other forms of necrosis [23-31]. RIPK1 and 3 are serine/threonine kinase family members that interact through RIPK homotypic interaction motifs to permit necroptosis to take place [24, 32], in addition to mixed lineage kinase domain-like protein. As necroptosis is a “failsafe” mechanism to eliminate caspase-8 inhibiting virus infections [32, 33], inhibition of caspase-8 may be detrimental by triggering necroptosis. A recent study showing benefit in renal IRI by blocking necroptosis through RIPK1  suggested that necroptosis may play a role in the pathogenesis of diverse kidney injury including allograft rejection.
Our previous studies have demonstrated that tumor necrosis factor-alpha (TNF-α) can induce apoptosis in renal TEC and that TEC participate in cytokine-enabled, Fas-Fas ligand (FasL)-mediated fratricide . This has been recently confirmed in cisplatin-activated TEC . We and others have also shown that inhibition of caspase-8 , interleukin-2 [36, 37] and indolamine 2,3-dioxygenase [38, 39] can attenuate various forms of TEC death and improve renal injury in short-term IRI models. In the present study, we demonstrate the differential effects of inhibiting caspase-8-mediated apoptosis and necroptosis following IRI or kidney transplantation. While caspase-8 inhibition can improve IRI, in the present study, we show that inhibition can augment necroptosis-mediated kidney allograft injury. Importantly, loss of donor kidney RIPK3 promoted allograft survival in an allogeneic mouse kidney transplant model.
Materials and Methods
- Top of page
- Materials and Methods
B6 (H-2b), Balb/c (H-2d; Jackson Laboratories, Bar Harbor, ME) and B6-RIPK3−/− (H-2b; generously provided by Genentech, Inc., South San Francisco, CA)  were maintained in the animal facility at the University of Western Ontario using approved protocols and procedures. RIPK3−/− mice are phenotypically unremarkable, and have normal kidney function and breeding . All experimental procedures were approved by the University of Western Ontario Animal Care Committee.
Primary cultures were derived from B6 and B6-RIPK3−/− mouse kidney cortex and grown in sterile full media at 37°C in 5% CO2. Primary culture TEC were trypsinized to release from plates and were used for up to two passages. Typical cobblestone appearance of renal epithelial cells was confirmed by visual analysis, and expression of TEC markers (cytokeratin, CD13, CD26 and E-cadherin) was confirmed.
Stable expression and delivery of short hairpin RNA
Generation of short hairpin RNA (shRNA) targeting caspase-8 was as described . The expression vector, pHEX6300, was ligated to the oligonucleotide sequence for caspase-8 mRNA (5′-AAC CTC GGG GAT ACT GTC TGA) to generate caspase-8 shRNA. Empty vector or caspase-8 targeting vector (150 µg of DNA) was delivered to the kidney in donor B6 via inferior vena cava injection as described  48 h prior to kidney transplantation.
A renal clamp was applied to the right kidney pedicle and removed after 45 min at 34°C and the left kidney was removed [4, 9, 39]. Kidneys were collected at 24, 48 and 72 h post-IRI after being flushed with normal saline until clear. Serum was tested for creatinine using an automated CX5 clinic analyzer (Beckman Coulter, Fullerton, CA).
Allogeneic (H-2b to H-2d) kidney transplantation
Male Balb/c recipient mice were bilaterally nephrectomized and transplanted with kidneys from male B6, caspase-8 shRNA-treated B6 or RIPK3−/− mice in a single procedure . Total ischemic time was limited to 35–40 min. Mice with weight loss of 15% or clinical deterioration were euthanized according to animal care protocols. In addition, all recipients terminated prior to 100 days were assessed for rejection by elevation of serum creatinine (>50 µmol/L) and histology. Serum and kidneys for histology were collected at the time of sacrifice for all euthanized mice to establish rejection.
Protein was isolated from tissue and cells using cytoplasmic lysis or nuclear lysis buffer. Blots were incubated with polyclonal rabbit anti-RIPK3 (Abcam, Cambridge, MA), rabbit anti-HMGB1 (Abcam) or mouse anti-β-actin (Sigma, St. Louis, MO) and quantified by densitometry (Alphaview; ProteinSimple, Santa Clara, CA) using β-actin.
RNA isolation and real-time polymerase chain reaction
Total RNA was extracted from tissue and cells by Trizol (Invitrogen, Carlsbad, CA). cDNA was generated using Superscript II (Invitrogen) and quantified by real-time polymerase chain reaction (PCR) MX3005 (Stratagene, Santa Clara, CA) using SybrGreen (Bio-Rad, Hercules, CA). Primers (Invitrogen) used for quantitative PCR (qPCR) include: RIPK3: 5′-GGGACCTCAAGCCCTCTAAC-3′ and 5′-GATCCCTGATCCTGACCCTGA-3′. β-Actin was used as the endogenous control. The normalized delta threshold cycle (Ct) value and relative expression levels (2ΔΔCt) were calculated according to the manufacturer's protocol.
Cell death assays
Primary renal TEC from B6 or RIPK3−/− mice were grown to confluent monolayers and treated with recombinant human TNF-α (hTNF-α; Peprotech, Rocky Hill, NJ), cycloheximide (CHX; Sigma), carbobenzoxy-valyl-alanyl-aspartyl-[O-methyl]-fluoromethylketone (Z-VAD-fmk; BD Bioscience, Franklin Lakes, NJ) and necrostatin-1 (Nec-1; Calbiochem, Darmstadt, Germany) in serum-free media. Cell viability and necrosis death were assessed using propidium iodide (PI; BD Bioscience) labeling and analyzed by flow cytometry (Beckman Coulter). Assessment of apoptosis utilized Annexin-V (BD Bioscience) along with PI by flow cytometry.
Histology and immunohistochemistry
Tissue sections were stained with hematoxylin and eosin (H&E) and scored by a renal pathologist blinded to groups using a semiquantitative method as described . Scoring included tubular cell and glomerular necrosis, mononuclear cell infiltration, tubulitis, fibrosis and vascular injury. To quantify necrosis in sections, ethidium homodimer (EHD; Invitrogen) was perfused into kidneys and areas of necrosis were assessed in frozen tissue sections . Briefly, 5 µM EHD was injected at 1 mL/min. for 10 min into the renal artery via the aorta and then flushed with perfusion buffer at 1 mL/min for 5 min. Total nuclei were labeled by 4′,6-diamidino-2-phenylindole (DAPI) in kidney sections. Sections were quantified using a fluorescent microscope and an automated analysis program (Nikon, Mississauga, Ontario, Canada) that measures the area and fluorescent intensity of five random fields of the outer renal cortex per slide. Immunohistochemistry was performed using polyclonal rabbit anti-RIPK3 (Abcam) and anti-CD3 (Dako, Burlington, Ontario, Canada) and standardized immunoperoxidase methods. Allograft fibrosis was assessed using Masson trichrome staining.
Shapiro–Wilk testing was used to assess data sets for normality. Parametric data were compared using Student's t-test for unpaired values and analysis of variance for multiple groups, while nonparametric data were compared using a Mann–Whitney test. Graft survival was analyzed by log-rank testing Mantel-Cox) using GraphPad Prism software (GraphPad Software, Inc., La Jolla, CA). Data are presented as mean ± SEM using p < 0.05 for significance.
- Top of page
- Materials and Methods
Inhibition of apoptosis by caspase-8 silencing reduced renal allograft survival
Apoptosis can be inhibited by c-FLIP or caspase-8 shRNA in TEC, and treatment of kidneys in vivo by Fas or caspase-8 shRNA can attenuate kidney IRI [3, 4]. However, the potential benefit of caspase-8 inhibition has not been tested in an allogeneic renal transplantation model. Donor kidneys were treated with caspase-8 shRNA or empty vector shRNA via direct inferior vena cava injection 48 h prior to being used for transplantation using a previously described method . While the duration of shRNA effect was not tested, the differential effect of caspase-8 shRNA in treated kidneys was clear compared to controls. Recipients of caspase-8 silenced donor kidneys had reduced survival compared to those that received B6 wild-type donor kidneys (mean of 33.3 ± 8.7 days, n = 8 vs. 68.3 ± 10.9 days, n = 17, p = 0.01; Figure 1A). Recipients treated with vector shRNA control kidneys had similar survival rates as unmanipulated B6 kidneys. One-third of the recipients that received naïve donor kidneys demonstrated acceptance, consistent with previous reports of spontaneous acceptance . In marked contrast, none of the caspase-8 silenced allograft recipients here survived to day 100 posttransplant (p = 0.01). Consistent with this shortened survival, increased mononuclear graft infiltration was evident in caspase-8 silenced grafts (Figure 1B).
Previous studies have demonstrated that TNF-α triggers necroptosis rather than apoptosis when caspase-8 is inhibited and unable to block the RIPK1/3 complex [22, 23, 30, 31, 45-47]. Therefore, we tested the possibility that caspase-8 shRNA inhibition augmented donor kidney necrosis to shorten survival. EHD, which labels necrotic cells with loss of cell membrane integrity, was used to quantitatively assess necrosis in kidneys [43, 48]. There was increased tissue necrosis in caspase-8 silenced kidney allografts on day 4 posttransplant compared to shRNA controls, as indicated by red fluorescence after EHD perfusion. Accordingly, increased release of nonnuclear HMGB1 was detected in caspase-8 shRNA-treated kidneys compared to shRNA controls or naïve donor kidneys (Figure 1B). EHD fluorescence was also increased in day 8 caspase-8 shRNA-treated allografts compared to controls (caspase-8 shRNA: 10 ± 2 vs. control shRNA: 1 ± 0, p = 0.04, n = 3/group; Figure 1C). Inhibition of caspase-8 in the donor renal allograft appears, therefore, to increase tissue necrosis and subsequent release of HMGB1. As increased mononuclear infiltrates in caspase-8 shRNA-treated kidneys were observed, it is possible that reduced survival was related to augmented rejection due to increased necrosis and HMGB1.
RIPK3 is regulated by pro-inflammatory cytokines in renal TEC
TNF-α is expressed by infiltrating cells as well as kidney parenchymal cells during acute kidney injury [49, 50]. Soluble TNF-α engagement with surface TNFR1 can therefore induce caspase-8-mediated apoptosis or necroptosis via RIPK1/3 if caspase-8 is inhibited [23, 24, 45, 47, 51-53]. Expression of RIPK3 was confirmed in untreated primary culture TEC (Figure 2A) and following exposure to TNF-α and interferon gamma (IFN-γ), which up-regulated RIPK3 mRNA (1 ± 0 vs. TNF-α: 1.5 ± 0.2, p = 0.007, n = 3) and was maximal upon combined application of both. Similarly, RIPK3 protein was constitutively expressed in resting TEC but only modestly increased in cytokine exposed cells (Figure 2B). Therefore, RIPK3 is basally expressed in renal TEC as in most cells, and expression is required for necroptosis [29, 32].
RIPK1/3-mediated necroptosis regulates HMGB1 release in TEC
RIPK3−/− and wild-type TEC express both TNFR1 and TNFR2 surface receptors (not shown). To bias toward TNFR1-related death rather than enhanced survival via TNFR2 , we exposed TEC to hTNF-α, which has a greater affinity for TNFR1 than TNFR2 in murine cells , as well as CHX to enhance CICD [34, 55]. We also tested the RIPK1 inhibitor Nec-1 as well as RIPK3−/− TEC for survival and release of HMGB1 following TNF-α with caspase inhibition [56-58]. Wild-type TEC increased PI positivity (8.3 ± 0.7% vs. 25.4 ± 3.2%, p = 0.01, n = 3) with CHX and hTNF-α at 24 h. z-VAD-fmk (zVAD) increased the number of PI positive cells (25.4 ± 3.2% vs. 40.2 ± 1.6%, p = 0.01, n = 3; Figure 3A) consistent with caspase independent RN. Nec-1, which blocks necroptosis via RIPK1 [34, 59], modestly reduced PI positivity (40.2 ± 1.6% vs. 32.9 ± 2.2%, p = 0.03, n = 3). Addition of Nec-1 alone did not have an effect on TEC apoptosis (not shown), as shown in previous studies . In contrast, RIPK3−/− TEC were completely resistant to CHX and hTNF-α-induced necrosis compared to wild-type TEC (10.8 ± 2.3% vs. 40.2 ± 1.6%, p = 0.0005, n = 3) and did not change with zVAD. Interestingly, death induced without caspase inhibition, using only CHX and TNF-α treatment, was also abolished in RIPK3−/− TEC. This latter finding is in line with in vivo reports of observed protective effects with Nec-1 without caspase inhibition, as reviewed by Linkermann et al . Consistent with PI results and a necroptosis mechanism, HMGB1 release into supernatant was greater with zVAD-treated TEC compared to hTNF-α/CHX-treated TEC and could not be detected in the supernatant from RIPK3−/− TEC. Importantly, HMGB1 release from necrotic cells  was nearly completely absent in Nec-1-treated TEC (Figure 3B) and intracellular HMGB1 from cell lysates remained unchanged in all treatment groups. Exposure of RIPK3−/− tubular cells to extremely high concentrations of hTNF-α (300 ng/mL) can induce apoptosis as detected by Annexin-V (Figure 3C), yet no detectable PI positivity was induced nor was HMGB1 was found in the supernatant (Figure 3D). This suggests that in the absence of necroptosis as a death pathway, apoptosis can occur in RIPK3−/− TEC but without the release of HMGB1 or PI positivity. These data demonstrate that caspase inhibition in TEC results in necroptosis, which could account for shRNA transplant results.
RIPK3−/− mice are resistant to kidney injury after renal IRI
The RIPK1 inhibitor Nec-1 can ameliorate kidney injury in a mouse model . We therefore extended those studies to include the participation of RIPK3 in kidney injury. RIPK3 expression was detected at low levels in naïve murine kidney sections but was expressed at higher levels at 4 h and persisted as long as 48 h post-IRI (Figure 4A). RIPK3 expression was ubiquitously expressed in both proximal and distal tubules (indicated by arrows), which is consistent with our results using primary culture TEC. Kidney function post-IRI was assessed by serum creatinine measurements at 24, 48 and 72 h in both wild-type (B6) and RIPK3 mice and compared to naïve mice. Interestingly, both were equivalently elevated at 24 h post-IRI. However, a clear benefit of RIPK3 absence in renal IRI was observed at 48 h (61 ± 24 vs. 137 ± 26 µmol/L, p = 0.03, n = 7/group; Figure 4B). In our model, 48 h post-IRI consistently represents a maximum injury time point, as mice have recovered sufficiently postprocedure to exclude hydration as a variable. Wild-type mice had increased acute tubular necrosis and a greater injury score as compared to RIPK3−/− after 48 h of renal IRI (2.5 ± 0 vs. 1.5 ± 0.2, p = 0.02, n = 4–7/group; Figure 4C). Our data demonstrate that inhibition of RIPK3 can ameliorate acute kidney injury similar to that observed with RIPK1 inhibition .
Using EHD, tissue necrosis was easily detected in wild-type kidneys 48 h after renal IRI but was nearly undetectable in RIPK3−/− kidneys (Figure 5A). Kidney sections also showed more areas of tubular injury and necrosis (arrows) in wild type as compared to RIPK3−/−. EHD was quantitated in Figure 5B and confirmed markedly decreased levels of necrosis in RIPK3−/− kidneys 48 h after renal IRI (1 ± 0 vs. 18 ± 2, p < 0.0001, n = 4/group). HMGB1 increased in wild-type kidneys at 48 h post-IRI (0.1 ± 0.01 vs. 0.3 ± 0.04, p = 0.02, n = 3; Figure 5C) but did not increase in RIPK3−/− kidneys.
RIPK3−/− donor kidneys are resistant to allograft dysfunction and inflammation
Chronic allograft injury is a major complication associated with kidney transplantation , and previous studies have suggested that fibrosis with persistent inflammation is important [2, 62, 63]. We therefore tested whether the absence of RIPK3 and its resultant effect on necroptosis in a donor kidney could improve function or long-term survival following allotransplantation. Serum creatinine levels of RIPK3−/− grafts were lower than wild type at study end (31 ± 0.6 vs. 86 ± 24 µmol/L, p = 0.03, n = 8–9/group; Figure 6A) and had reduced inflammation and histological injury (21.4 ± 1.2 vs. 11.2 ± 0.2, p = 0.006, n = 4–5/group; Figure 6B and C). Wild-type grafts also had greater neutrophil infiltration, fibrosis, tubilitis and vascular injury (Figure 6D). Sections from long-term (>100 days) grafts were found to have equivalent expression of CD3 positive infiltrates in both wild-type and RIPK3−/− kidneys (Figure 6C). Kidney HMGB1 was less in RIPK3−/− kidneys at both days 8 and 100 posttransplant (p = 0.04, n = 3/group; Figure 7A).
Inhibition of RIPK3 in kidney allografts prolongs long-term survival
RIPK3−/− kidney graft recipients achieved greater rejection-free survival to day 100 (90 vs. 23%, p = 0.002, n = 10–17/group; Figure 7B) and survived longer than those that received wild-type (B6) kidneys (56.7 ± 9.4 vs. 94.1 ± 2.1 days, p = 0.0006, n = 10–17/group). Body weight loss was maintained in both groups of mice at day 100. Wild-type mice terminated prior to day 100, however, had weight loss from baseline that was greater than RIPK3−/− recipients (7.8 ± 4.5% vs. 1.2 ± 1.2%, n = 4–5/group), which was consistent with rejection, as terminated wild-type kidney recipients had histological rejection and higher serum creatinine levels than RIPK3−/− kidney recipients (148 ± 51 vs. 29 ± 6 µmol/L, p = 0.03, n = 5–7/group). Mice that survived to day 100 had similar kidney function in each group (wild type: 53.2 ± 17.7 vs. RIPK3−/−: 27.2 ± 6.6 µmol/L, n = 5/group, p = ns). As CD3+ mononuclear cell infiltration and tubulitis were observed in RIPK3−/− kidneys, rejection was clearly not prevented. However, there was a marked absence of fibrosis and reduction in vascular injury in RIPK3−/− kidneys (Figure 6C and D).
- Top of page
- Materials and Methods
Although acute rejection rates and 1-year graft survivals have improved in kidney transplantation, long-term survival has not substantially changed or is improving very slowly [64-66]. Although reasons are most certainly multifactorial and biologically complex, there has been little attention directed to donor organ factors that might contribute to long-term graft loss. Kidney transplantation is invariably associated with organ damage, which includes IRI. IRI triggers a cascade of linked innate and adaptive immune responses that propagate injury, kill parenchymal cells and promote antibody- and cell-mediated rejection [9, 63, 67, 68]. The form of cell death may be an early variable that directs the outcome of immune responses. The current understanding of cell death mechanisms has greatly expanded beyond apoptosis to regulated necrosis, a broad category that includes necroptosis, pyroptosis and others. As a result, various assays have been used to differentiate between cell death subroutines including cell viability, morphology, quality of DNA fragmentation, loss of membrane integrity assays, TUNEL positivity, PARP1 cleavage, caspase activation, labeling with Annexin or PI [57, 69] and many others . A powerful approach in labeling the type of cell death utilizes pharmacologic inhibitors that target specific pathways, RNA knockdown strategies and animal models with genetic deletions .
Apoptosis is a classic form of PCD [12, 13]. In previous studies, we have shown that caspase inhibition using shRNA to silence caspase-8 or transgenic overexpression of the endogenous caspase-8 inhibitor c-FLIP can protect renal TEC against TNF-α-induced apoptosis in vitro or ischemic kidney injury in vivo [4, 36]. Pan-caspase inhibition of both initiator and effector caspases can reduce cold preservation injury due to apoptosis in liver endothelial cells and transplanted islet cells [71, 72]. However, reports in other models and, in particular, renal IRI has not confirmed a benefit using caspase inhibition  perhaps as variability exists in the specificity of caspase inhibitors . Long-term studies to evaluate the effect of donor organ apoptosis inhibition in kidney transplants have been hampered as the embryonic lethality of caspase-8 deficiency precludes the use of animal models  and the duration of gene silencing with small interfering RNA is limited. Results in the present study in which donor caspase-8 RNA silencing worsened kidney transplants are consistent with recent insights into new forms of regulated cell death in which caspase-8 silencing triggers pro-inflammatory necroptosis [47, 56, 75]. These results support that short-term outcomes in an IRI model can differ from long-term outcomes in a transplant model, depending on targets of cell death.
While apoptosis generates membrane-bound apoptotic bodies that sequester cellular contents [73, 76], necrosis results in loss of membrane integrity and the release of HMGB1 and other CDAMP that promote inflammatory responses [56, 69] through interaction with TLRs, as well as other innate receptors that are ubiquitously expressed within the kidney [77-79] and on a wide variety of immune cells including dendritic cells. Maximal protection in IRI has been observed when HMGB1 was not able to engage kidney TLR4 [80, 81]. Thus, HMGB1 release may contribute to the propagation of kidney injury.
Ligand engagement of death receptor (DR) family members (CD95/Fas, TNFR1 and TRAIL) normally results in apoptosis and inhibition of caspase-8 might be expected to be of benefit in IRI in which acute loss of cells reduces organ function . However, recent studies using TNF-α indicate that a primary function of caspase-8 may be also to block necroptosis [22, 23, 30]. Genetic deletion of caspase-8 is embryonically lethal due to unrestricted necroptosis in the developing yolk sack . Similarly, caspase-8 inhibition could be detrimental in adult tissue if DR-mediated apoptosis was “replaced” by necroptosis [32, 82]. Necroptosis may be particularly relevant to kidney injury in that TEC produce TNF-α as well as respond to TNF-α .
In this article, we found that pan-caspase inhibition resulted in necrosis of cytokine exposed TEC in vitro. As the presence of PI positive labeling by flow cytometry alone may not exclusively define necrotic cells as opposed to cells undergoing late apoptosis, we also measured HMGB1 release to confirm necrosis in TEC and grafts [57, 58]. Loss of plasma membrane integrity during necrosis results in nuclear HMGB1 moving to the extracellular space. As HMGB1 is not actively synthesized during this process, the resulting HMGB1 found in supernatant is presumed to be due to release of HMGB1 from nuclei. In the present article, we confirm that TEC undergo necroptosis, which can be abrogated by Nec-1. We also show that TEC express abundant levels of RIPK3 protein, which increased within hours of ischemic injury. RIPK3−/− TEC exposed to extremely high concentrations of hTNF-α to TEC can undergo Annexin-V positive apoptosis without PI positive necrosis or the release of HMGB1. Our data suggest that in the absence of the necroptosis pathway, apoptosis can still occur in TEC.
Consistent with our in vitro results, the worsening of graft survival following caspase inhibition likely represents enhanced alloimmune responses following necroptosis and the release of CDAMP from kidney cells [17, 84]. In testing kidney samples, protein was isolated using a lysis buffer that excludes nuclear protein including HMGB1. Therefore, the HMGB1 we detected in whole kidney lysates was released from either kidney parenchymal cells or infiltrating cells. We utilized EHD to measure tissue necrosis, which is reproducible and quantifiable but is limited by a lack of specific cellular detail. Increased HMGB1 and EHD positivity in shRNA caspase-8 silenced kidneys was consistent with necroptosis and observed worsened kidney function. While not tested here, the release of CDAMP such as HMGB1 likely promoted alloimmune responses that reduced survival. In marked contrast, HMGB1 was lower, EHD was absent and kidney function was better using RIPK3−/− mice in our IRI model. Unlike RIPK1, RIPK3 cannot participate in NF-kappaB signaling, and kidney function and development are normal in RIPK3−/− mice. Protection in RIPK3−/− mice was very similar to IRI studies in which both kidney and infiltrating cells were exposed to Nec-1 , suggesting that the beneficial effect of RIPK1 and RIPK3 inhibition was due to elimination of kidney necroptosis.
These data provide the first unequivocal demonstration of RIPK3-mediated necroptosis in both renal IRI and transplantation. While both apoptosis and necrosis appear to contribute to kidney dysfunction in short-term IRI models, our results suggest that there may be a transitional phase following IRI in which alloimmune activation can be promoted by necroptosis as well as loss of anti-inflammatory responses generated by apoptosis [12, 13, 76, 85, 86]. HMGB1 and TLR4 expression peaks with 5–10 days in renal IRI [80, 81]. In blocking caspase-8 and augmenting necroptosis, prolonged HMGB1 and CDAMP release may have increased the participation of IRI relevant infiltrating T cells , NK cells  and other effectors. Additional studies will be required to delineate the immune effector cells involved. Furthermore, the relative impact of various forms of cell death may vary in different solid organs, generating organ-specific apoptosis–necroptosis “equilibriums” and both may need to be targeted for maximal protection .
This article provides the first report of the benefit of kidney RIPK3 deletion and necroptosis elimination on long-term allograft survival. Despite preserved function reflected by serum creatinine and prolonged survival of RIPK3−/− kidney recipients, tissue injury with HMGB1 release and CD3+ cellular infiltration were observed in the grafts of long-term survivors. Clearly the elimination of necroptosis and reduced CDAMP release early posttransplant did not result in tolerance per se. It, therefore, may be speculated that the benefit of RIPK3 deletion is related to an attenuation but not elimination of alloimmune responses. While HMGB1 levels were consistently less than wild-type controls at every time point, significant levels of HMGB1 were detected in late day 100 RIPK3−/− allografts, which may have been derived from parenchymal cells undergoing nonnecroptosis death or secretion from viable infiltrating cells [89, 90]. Notably however, the near complete absence of fibrosis in RIPK3−/− kidneys suggests that necroptosis may play a critical role in long-term allograft injury that results in scarring.
In summary, we show for the first time that RIPK3 regulates necroptosis in the kidney and that this has a major impact on renal IRI and kidney transplant survival. As well we have demonstrated that inhibition of caspase-8 within TEC eliminates a key regulatory role for this enzyme in controlling RIPK1/3-mediated necroptosis, and therapeutic strategies may require control of multiple pathways. We suggest that reduction of necroptosis in donor organs will have a profound benefit in graft function and survival. More efficacious forms of Nec-1  and the possibility of targeting RIPK3  will greatly advance such strategies and may represent a paradigm shift in modifying organ injury to dampen alloimmune responses.
- Top of page
- Materials and Methods
We thank Genentech and Dr. Francis K. M. Chan for providing breeder pairs of RIP3−/− mice to establish a colony and Ms. Pamela Gardner for administrative support. This work was supported by grants from the Canadian Institutes of Health Research (XTW-90932, MOP-111180, MOP-115048), Kidney Foundation of Canada (AMJ, Z-XZ) and the Program of Experimental Medicine (POEM) at LHSC.
- Top of page
- Materials and Methods
The authors of this manuscript have no conflicts of interest to disclose as described by the American Journal of Transplantation.
- Top of page
- Materials and Methods
- 76Epidermal growth factor and perlecan fragments produced by apoptotic endothelial cells co-ordinately activate ERK1/2-dependent antiapoptotic pathways in mesenchymal stem cells. Stem Cells 2010; 28: 810–820., , , et al.