Different signalling in infarcted and non‐infarcted areas of rat failing hearts: A role of necroptosis and inflammation

Abstract Necroptosis has been recognized in heart failure (HF). In this study, we investigated detailed necroptotic signalling in infarcted and non‐infarcted areas separately and its mechanistic link with main features of HF. Post‐infarction HF in rats was induced by left coronary occlusion (60 minutes) followed by 42‐day reperfusion. Heart function was assessed echocardiographically. Molecular signalling and proposed mechanisms (oxidative stress, collagen deposition and inflammation) were investigated in whole hearts and in subcellular fractions when appropriate. In post‐infarction failing hearts, TNF and pSer229‐RIP3 levels were comparably increased in both infarcted and non‐infarcted areas. Its cytotoxic downstream molecule p‐MLKL, indicating necroptosis execution, was detected in infarcted area. In non‐infarcted area, despite increased pSer229‐RIP3, p‐MLKL was present in neither whole cells nor the cell membrane known to be associated with necroptosis execution. Likewise, increased membrane lipoperoxidation and NOX2 levels unlikely promoted pro‐necroptotic environment in non‐infarcted area. Collagen deposition and the inflammatory csp‐1‐IL‐1β axis were active in both areas of failing hearts, while being more pronounced in infarcted tissue. Although apoptotic proteins were differently expressed in infarcted and non‐infarcted tissue, apoptosis was found to play an insignificant role. p‐MLKL‐driven necroptosis and inflammation while inflammation only (without necroptotic cell death) seem to underlie fibrotic healing and progressive injury in infarcted and non‐infarcted areas of failing hearts, respectively. Upregulation of pSer229‐RIP3 in both HF areas suggests that this kinase, associated with both necroptosis and inflammation, is likely to play a dual role in HF progression.


| INTRODUC TI ON
Because adult cardiomyocytes are terminally differentiated and have a very low rate of cell cycle re-entry and proliferation, 1 the heart possesses a very limited capacity to regenerate. This feature is important in particular for conditions characterized by a loss of the functional myocardium such as myocardial infarction (MI) which can progress to heart failure (HF). Although excessive research has been undertaken to investigate certain cell death modalities in such damaged heart, it is still unclear which of them underlie its phenotypes.
Recent studies investigating newly recognized cell death modalities have suggested the possible involvement of necroptosis in the pathogenesis of post-MI HF. [2][3][4][5] The presence and interaction of the main necroptotic proteins have been detected in advanced human HF, 5 while both genetic knock-out and pharmacological inhibition of necroptotic signalling has been shown to alleviate deleterious cardiac phenotypes in HF, including contractile dysfunction, remodelling and inflammation. 2,3,6 Signalling of necroptosis, a form of regulated cell death resembling morphological features of necrosis 7-10 proceeds due to a formation of the necrosome complex containing phosphorylated protein kinases RIP1 and RIP3 to further phosphorylate MLKL, a terminal pro-necroptotic protein. 11 As a result, such phosphorylated MLKL molecules at Thr357 and/or Ser358 oligomerize and translocate into the plasma membrane causing the execution of the necrotic-like cell death by inducing alterations in ion homoeostasis and plasma membrane disruption. [12][13][14][15] Recently, it has been suggested that RIP3 and MLKL activation can also be associated with inflammation via the activation of NLRP3 inflammasome and IL-1β independently on the necroptotic cell loss. [16][17][18] In view of the fact that post-MI HF is positive for main necroptotic proteins, 2,5,6 we tested a hypothesis that necroptosis is responsible for cell loss of the infarcted myocardium. Likewise, we hypothesized that necroptosis in the infarcted area, due to disruption of the plasma membrane leading to the release of intracellular content and pro-inflammatory mediators, causes diffuse pro-inflammatory and/or pro-necroptotic injury of the surrounding non-infarcted tissue. These cellular events can further promote fibrotic healing of the injured zone, as well as interstitial fibrosis of the non-infarcted tissue with subsequent changes in left ventricle geometry, and cardiac remodeling. In addition, a disproportionate accumulation of the released inflammatory mediators can contribute to the progressive impairment of both cardiac contraction and relaxation and thereby underlie other phenotypes of HF. Prague, Czech Republic) housed in a room with constant temperature of 22°C, and 12h:12h light/dark cycle were fed with a standard pellet diet and tap water ad libitum. After incubation period, rats were randomly assigned into two groups ( Figure 1): sham-operated animals (Sham, n = 9) and group subjected to MI with subsequent development of HF (n = 10). 19 In anesthetized open-chest animals (sodium pentobarbital, 60 mg/kg i.p.), MI was induced by ligation of the left coronary artery 1-2 mm distal to the left atrial appendage for 1 hour. After this period, ligation was released.

| Animal model and study design
Sham-operated rats underwent chest surgery without occlusion.
After chest closure, all spontaneously breathing animals recovering from anaesthesia were housed in separate cages with given analgesia (ibuprofen, 20 mg/day p.o.) for another 3 days. Mortality of rats with MI was 43%. Echocardiography was performed 3 days before and 42 days after the surgical procedure by using GE Vingmed System Five (GE Vingmed Ultrasound) and FPA 10 MHz probe (GE Vingmed Ultrasound). Animals were anesthetized with 2% isoflurane (Forane; Abbott Laboratories) mixed with room air.
Left ventricular (LV) systolic (LVDs) and diastolic diameters (LVDd) were directly measured, from which fractional shortening (FS) was derived according to the formula FS (%) = [100 × (LVDd-LVDs)/ (LVDd)]. At the end of experiment the animals were sacrificed by pentobarbital overdose; blood samples were taken from the right ventricular cavity and hearts rapidly excised and washed in icecold PBS. Whole free LV wall of sham-operated animals was harvested while in the HF group, the LV was dissected into infarcted area (HFi) and remaining non-infarcted tissue (HFni) ( Figure 2) and stored at −80°C till further molecular and cellular analyses ( Figure 1).

| Western blotting
Left ventricular tissue samples were processed for immunoblotting analysis by SDS-PAGE and Western Blotting as described previously. 5 Post-electrophoresis, proteins were transferred onto PVDF membranes (Immobilon-P, Merck Millipore) and incubated with primary Immunoblotting analysis was performed separately for comparison of HFni versus HFi and Sham versus HFni, and expression of the particular protein was adjusted to its 100% reference expression in HFni.

| Phos-tag™ SDS-PAGE
Phos-tag™ acrylamide (Wako Pure Chemical Industries, Ltd.) was used for detection of mobility shift of phosphorylated proteins by SDS-PAGE. 22,23 Briefly, Phos-tag™ acrylamide (60 μmol/L) and ZnCl 2 (120 μmol/L) were added to a standard gel solution, which was polymerized by addition of ammonium persulfate and TEMED. To neutralize the effect of divalent metal chelators in the lysates used for Phos-tag™ SDS-PAGE, these were supplemented with ZnCl 2 at a concentration equal to concentrations of previously added chelating agents. All subsequent procedures including electrophoresis, transfer and immunodetection were carried out by following standard protocol for Western blotting as described above.

| Subcellular fractionation
All procedures were performed on ice with pre-chilled instruments, reagents and centrifuge (4°C) following a protocol described in detail in Szobi et al. 24 Briefly, after homogenization of tissues in imidazole buffer (pH = 7.6) containing 600 mmol/L sucrose, the suspension was centrifuged at 800 g for 20 minutes. The supernatant was centrifuged at 800 g for 20 minutes first, then at 10 000 g for an additional 10 minutes. Finally, it was enriched with CaCl 2 to achieve a concentration of ~10 mmol/L and centrifuged at 20 000 g for 20 minutes. The remaining pellet referring to the membrane fraction was re-suspended in buffer containing detergents and EDTA/EGTA and incubated for 60 minutes while the supernatant representing the cytoplasmic fraction was further processed with detergents according to the above described Western blotting protocol.

| Measurement of thiobarbituric acid reactive substances
For the measurement of lipid peroxidation in total lysates and membrane subcellular fraction, thiobarbituric acid reactive substances

| Statistical analysis
The results are expressed as means ± standard error of means (SEM). Two-tailed unpaired Student's t test with or without Welch's correction was used for evaluation of differences between two groups while one-way ANOVA analysis with Tukey's post hoc tests was used for evaluation of group differences in variables with normal distribution between three groups. In case of non-normally distributed data, Mann-Whitney U test was used instead. All analyses were performed with GraphPad Prism 7.00 for Windows (GraphPad Software). Differences between groups were considered significant when P < 0.05. HFni-non-infarcted tissue; HFi-infarcted tissue. Data are presented as mean ± SEM, n = 7-9 per group. *P < 0.05 vs Sham, # P < 0.05 vs HFni content and fibrotic healing supported echocardiographic findings on cardiac remodelling due to ongoing fibrosis ( Figure 3D-E).

| Analysis of necroptotic signalling
In the infarcted zone, RIP1 levels were not detected, and no significant difference in the levels of total RIP3 kinase between infarcted and non-infarcted areas was found ( Figure 4B-C). Notably, the levels of pSer229-RIP3, a direct upstream of MLKL activation, were unchanged in the infarcted zone when compared to the non-infarcted area, however being upregulated when compared to sham group ( Figure 4D,I). Expression of the terminal necroptotic pore-forming protein MLKL was equal in both infarcted and non-infarcted areas ( Figure 4E). Of note, a phosphorylated form of MLKL, which is almost exclusively associated with MLKL oligomers formation in the cell membrane during necroptosis 10,25 was found to be heavily present in the infarcted zone ( Figure 4F).
In the non-infarcted zone, both RIP1 and RIP3 kinase expression was found to be unchanged compared with sham-operated group ( Figure 4G-H). However, in this particular area of the failing hearts,

| Apoptotic markers expression
Because there is a caspase-8-dependent interlink between necroptotic and apoptotic signalling, 26 we also analysed some well-established markers of apoptotic cell death. In the infarcted zone, levels of active caspase-8 were reduced despite its increased zymogene expression, suggesting a reduction in the caspase-dependent signalling ( Figure 7A). However, downstream caspase-7 and caspase-3 expression levels were increased relative to the non-infarcted area ( Figure 7B,C). Contrary to these changes, apoptotic PARP1 cleavage was the lowest in the non-infarcted zone compared with other groups, and the Bcl-2/Bax ratio showed a massive drop due to both a reduction in Bcl-2 expression and an upregulation of Bax ( Figure 7D,E).
The non-infarcted zone was characterized by a different profile of these apoptotic markers compared with the infarcted one.
Expression of almost all analysed pro-apoptotic markers (active form of caspase-8, executioner caspases −3 and −7 as well as specific apoptotic p25 fragment of PARP1) was reduced in the non-infarcted zone compared with the sham-operated group ( Figure 7F-I). Being significantly decreased, the Bcl-2/Bax ratio was not in line with these findings ( Figure 7J). Altogether, these data do not support the view that apoptotic cell death plays a significant role in post-MI failing hearts.

| Inflammatory response
Association between necroptosis and inflammation in the failing hearts was investigated by analysing expression of TNF, which has been proved to participate in both these processes, 27

| Oxidative stress
Because pro-necroptotic RIP3 activation has been shown to be linked with increased ROS production, 3 we also assessed oxidative stress associated with membrane lipid peroxidation and levels of NOX2, one of main pro-oxidant membrane-bound enzymes in the heart 30 known to influence cardiac remodelling and chronic HF progression. 31 In the membrane fraction of non-infarcted tissue, where necroptosis is being executed in, 13,14 both these markers were higher than in the sham group ( Figure 9A,B).

| D ISCUSS I ON
In this study, the canonical necroptotic signalling axis has been investigated for the first time in the infarcted and non-infarcted area of rat hearts separately, and thereby, a novel mechanistic insight into its potential involvement in the pathogenesis of HF of ischaemic aetiology has been delineated. We have found that p-MLKL, which has been shown to be related to necroptosis execution, 10 On the other hand, pSer229-RIP3 serving as an upstream protein of MLKL was upregulated in both areas of failing hearts. Likewise, TNF and csp-1-IL-1β axis, a pro-inflammatory downstream pathway of RIP3, 29,32 were also found to be activated in both areas. It seems that in the infarcted area the activated RIP3 proceeds to MLKL signalling to further terminate in pro-necroptotic events and inflammation while in the non-infarcted area this activation of RIP3 can solely mediate pro-inflammatory phenotype without necroptosis execution.
Excess cardiomyocyte damage occurring as a consequence of MI contributes to HF progression. 33 In these cardiac pathologies, many conventional and less known cell death modalities have been identified, 33 including necroptosis. 34   In the present study, we have evaluated this approach by separately analysing these two particular LV areas of rat hearts 42 days post-MI.
In the heavily scarred infarcted area, we were unable to assess RIP1 expression, which was, however, similarly expressed in non-infarcted tissue of failing hearts when compared to sham group. Because the infarcted tissue contained high levels of collagen, a decreased cellularity resulting in reduced RIP1 content could be an explanation for this difficulty. However, as other proteins of interest were easily detected in this area, we believe that specific RIP1-associated cellular modifications could be rather a reason for such an absence of its signal.
In support, we used various antibodies recognizing C-or N-terminal of RIP1 (CST #3493; BD 610459), which were unable to detect any single signal for this protein, thereby arguing for most likely RIP1 degradation in the scared tissue. This finding is contrast with our previous study employing human post-ischaemic failing hearts, in which, however, individual analysis of the ischaemic/infarcted versus nonischaemic/non-infarcted area was not performed and highly fibrotic tissue was excluded for any molecular investigation. 5 On the other hand, these data by themselves do not rule out active necroptosis in it because necroptosis can also occur independently of RIP1 in some F I G U R E 8 Analysis of proinflammatory signalling in left ventricle lysates. A-F, quantification and G, representative immunoblots and total protein staining of procsp1, csp1, proIL-1β, IL-1β and TNF in sham-operated group (Sham), non-infarcted (HFni) and infarcted (HFi) tissue of failing hearts. Data are presented as mean ± SEM; n = 6-9 per group; *P < 0.05 vs Sham; # P < 0.05 vs HFni It is very likely that which scenario occurs strongly depends on cellular conditions and pathological settings. 16,17,49  When discussing other programmed cell death modalities with respect to the pathogenesis of failing hearts, apoptosis is of prime importance as it has been considered for decades to be one of the major cell death modes contributing to myocyte death. 58 In our study, however, apoptosis signalling does not seem to be significantly involved in tissue damage either in infarcted or in non-infarcted LV parts of failing hearts. These findings are in accordance with our results from a study employing end-stage human failing hearts, and also with another work using a similar model of rat HF. 2,5 In contrast, some papers studying RIP3-mediated necroptosis have reported increased apoptosis in HF. 3,59 This controversy only strengthens the fact that elevation of apoptosis as a major contributor to myocyte loss in HF progression is still a rather unresolved question requiring future clarification. To help clarify this issue, improvement of methods used to assess apoptosis (protein expression versus DNA laddering assessment and TUNEL assay) as well as considering different apoptosis rate between various heartresiding cell types besides cardiomyocytes could provide some useful approaches for dealing with these discrepancies. 58,[60][61][62] Interestingly, the apparent increase in executioner csp-3 and csp-7 expression in infarcted tissue did not result in increased PARP1 cleavage. Although we did not seek a detailed explanation for these findings, the previously described non-apoptotic roles of caspases could provide a feasible link for future investigation. 63,64 Irrespective of this, it is likely that observations in apoptotic protein expression shown here do not suggest an exclusive role of apoptosis in the pathomechanisms of heart failure, at least at this stage.

| CON CLUS IONS
Taken together, although TNF and csp-1-IL-1β-associated inflammation has been present in both infarcted and non-infarcted LV area of failing hearts, these particular myocardial parts are characterized by a different profile of cell death proteins. The canonical necroptotic signalling involving pSer229-RIP3 and p-MLKL was found in the infarcted area only. On the other hand, the non-infarcted tissue has been found to exhibit pSer229-RIP3 activation with possible resultant pro-inflammatory rather than pro-necroptotic signalling.
Thus, it seems that this kinase might propagate different cellular responses, such as concurrent necroptosis and inflammation vs inflammation without necroptosis execution, in these particular parts of post-MI failing hearts, demonstrating a novel insight on a proposed dual role of pSer229-RIP3 in HF progression ( Figure 10).
Accordingly, these results have also highlighted a potential for selective pharmacological targeting of RIP3 to affect the two important processes contributing to progressive nature of HF.

| Study limitations
Although we have provided novel findings about necroptosis signalling in the infarcted and non-infarcted area of failing hearts and proposed dual role of pSer229-RIP3 kinase, our study has some limitations. First, the study has mainly a descriptive character and pharmacological or genetic modulations targeting certain proteins in necroptosis and/or inflammation signalling would be desirable to confirm our original findings. Next, it would also be interesting to perform subcellular fractionation in the infarcted tissue which was, F I G U R E 1 0 Proposed view of RIP3-associated necroptosis and inflammation in infarcted and non-infarcted area of failing heart however, not subjected to this procedure because of technical issues (a small mass of scarred tissue not allowing to separate subcellular compartments). Lastly, additional biochemical or molecular experiments could contribute to the clarification of the proposed RIP3-NLRP3-mediated inflammation.

CO N FLI C T O F I NTE R E S T
The authors confirm that there are no conflicts of interest.

DATA AVA I L A B I L I T Y
The data that support the findings of this study are available from the corresponding author upon reasonable request.