Liproxstatin-1 protects the mouse myocardium against ischemia/ reperfusion injury by decreasing VDAC1 levels and restoring GPX4 levels

a b s t r a c t
Ferroptosis is a distinct iron-dependent mechanism of regulated cell death recognized in cancer and ischemia/reperfusion (I/R) injury of different organs. It has been reported that molecules such as liproxstatin-1 (Lip-1) inhibit ferroptosis and promote cell survival however, the mechanisms underlying this action are not clearly understood. We investigated the role and mechanism of Lip-1 in reducing cell death in the ischemic myocardium. Using an I/R model of isolated perfused mice hearts in which Lip-1 was given at the onset of reperfusion, we found that Lip-1 protects the heart by reducing myocardial infarct sizes and maintaining mitochondrial structural integrity and function. Further investigation revealed that Lip-1-induced cardioprotection is mediated by a reduction of VDAC1 levels and oligo- merization, but not VDAC2/3. Lip-1 treatment also decreased mitochondrial reactive oxygen species production and rescued the reduction of the antioxidant GPX4 caused by I/R stress. Meanwhile, mito- chondrial Ca2+ retention capacity needed to induce mitochondrial permeability transition pore opening did not change with Lip-1 treatment. Thus, we report that Lip-1 induces cardioprotective effects against I/R injury by reducing VDAC1 levels and restoring GPX4 levels.

1.Introduction
Ferroptosis is a uniquely regulated mechanism for cell death that results from iron-dependent lipid peroxidation [1]. Ferroptosis particularly differs from other forms of regulated cell death in that it is caspase-independent and has neither nuclear morphology al- terations nor aberrant mitochondrial morphology [1,2]. Com- pounds that selectively induce ferroptosis include erastin, buthioninesulfoxomine and other Ras-selective lethal small mole- cules, while inhibitors include ferrostatins and liproxstatin-1,which reduce accumulation of ROS from lipid peroxidation (reviewed in Ref. [3]). In particular, liproxstatin-1 (Lip-1) is a spi- roquinoxalinamine derivative, which potently and specifically in- hibits ferroptosis [4]. However, although models of ferrostain-1 action have been proposed [5], the mechanism of Lip1 protection, is yet to be elucidated [6]. Nevertheless, ferroptosis is commonly observed in cancer cells and has also been shown in brain, kidney and liver cells, especially after ischemia/reperfusion (I/R) injury, hence ferroptosis is a recognized significant cause of cell death [7]. Devolution of mitochondrial structure is a key ferroptosis characteristic. Although ferroptosis involves production of reactive oxygen species (ROS) from extra-mitochondrial iron-mediated lipid peroxidation, it results in shrinkage of mitochondria, reduction in cristae and rupture of the outer mitochondrial membrane (OMM) [4,8].

Additionally, erastin-initiated ferroptosis is also mediated by a direct interaction between erastin and the OMM protein, voltage- dependent anion channel (VDAC) isoforms 2 and 3 [9]. And up- stream of mitochondria, glutathione peroxidase 4 (GPX4), is a vital antioxidant enzyme that regulates ferroptosis by sensing andtransducing oxidative stress [10]. GPX4 operates by catalyzing the conversion of reduced glutathione (GSH) to oxidized glutathione (GSSG) in a reaction that also reduces phospholipid hydroperoxides to less harmful alcohols [3,11]. Recently, knockdown of GPX4 in a glutathione-independent manner was shown to lead to mito- chondrial morphology destruction and an increase in mitochon- drial ROS production [12].Further, mitochondria are a critical target of cardiac I/R damage, particularly through VDAC1 regulation, and opening of the mito- chondrial permeability transition pore (mPTP), an event known to cause cell death [13e15]. And although ferroptosis is a recognized contributor to cell death during cardiac I/R injury (reviewed in Ref. [7]), little is known about the protective mechanisms utilized by ferroptosis inhibitors, especially Lip-1.In this study, the objective was to investigate whether Lip-1, a potent inhibitor of ferroptosis, can protect the heart against I/R injury. Following this, we sought to determine the mechanism by which such cardioprotection may be elicited. We found that Lip-1 indeed mitigates cardiac damage caused by I/R insult via the reduction of VDAC1 protein levels, without affecting VDAC 2/3. Further, we show an increase in GPX4 protein levels accompanied by a reduction of mitochondrial ROS production by the NADH- ubiquinone oxidoreductase (ETC complex I). All these Lip-1 ef-fects are elicited without impacting the amount of Ca2+ required totrigger opening of the mPTP.

2.Materials and methods
Male adult mice (C57BL/6J, Jackson Labs) 9e12 weeks old were used. Protocols were approved by the UT Health Science Center at San Antonio Institutional Animal Care and Use Committee and conformed to the Guide for the Care and Use of Laboratory Animals: Eighth Edition (2011) from the National Research Council.All reagents were purchased from Sigma Aldrich, unless other- wise stated. Mice were anesthetized with ketamine (80 mg/kg i.p.) and xylazine (8 mg/kg i.p.) as previously described [15]. Hearts were excised and arrested in cold (4C) Krebs Henseleit (KH) buffer (in mM): glucose 11, NaCl 118, KCl 4.7, MgSO4 1.2, KH2PO4 1.2,NaHCO3 25 and CaCl2 3, pH 7.4. Hearts were retrograde perfused (3 ml/min) through the aorta with KH buffer bubbled with 95% O2/5% CO2 at 37 ◦C on the Langendorff apparatus. Following 30 minequilibration, normothermic ischemia was induced by stopping buffer follow for 35 min. Reperfusion followed with KH buffer + Lip-1 (200 nM) or vehicle (DMSO), for 30 min (to isolatemitochondria) or 2 h (for infarct size measurement). This protocoltypically results in ~50% infarct size. Sham hearts were not sub- jected to I/R.Following 2 h reperfusion, hearts were sliced into 4 transverse sections parallel to the atrio-ventricular groove as described in Ref. [16]. Samples were then incubated for 10 min with 2% triphe-nyltetrazolium chloride at 37 ◦C to identify viable (red) frominfarcted (white) heart tissue.

Adobe Photoshop CS6 planimetry was used to quantify necrotic area by a blinded investigator.TEM was used for ultrastructural analysis of tissue samples aspreviously described [14]. Samples were fixed in 4% formaldehyde with 1% glutaraldehyde overnight at 4 ◦C before being washed andpost-fixed for 2 h at room temperature in 2% osmium tetroxide. Sections were dehydrated in a graded alcohol series thenembedded in Eponate 12 medium and cured for 48 h at 60 ◦C.Sections were then sliced, mounted and stained with uranyl acetate and lead citrate and then viewed on a JEOL 1230 electron microscope.As recently described in Ref. [17], we loaded equal concentra- tions of lysed tissue in 4e20% Tris-glycine gels (Bio-Rad). We car- ried out electrophoresis for 120 min at 90 V, then transfer onto nitrocellulose membranes for 80 min at 90 V. Membranes were blocked using 5% blocker solution (Bio-Rad), then probed overnightat 4 ◦C using primary antibodies for the following targets: VDAC1(Santa Cruz Biotechnology, sc390996), VDAC2 (Cell Signaling Technology, 9412s), VDAC3 (LifeSpan BioSciences, LS-C80113), cyclophilin D (Thermo Fisher, 455900), GPX4 (Abcam, ab125066) and GAPDH (Cell Signaling Technology, 97166S). Ponceau staining was also used to determine total protein quantities. Visualization was done on an Odyssey CLx system using IRDye secondary anti- bodies (LI-COR, 926e32211 and 926e68070).

Crude mitochondria were isolated from ischemic (30 min reperfusion) and sham hearts as described in Ref. [14]. Tissue was minced and homogenized in isolation buffer A (in mM): sucrose 70, mannitol 210, EDTA 1 and Tris-HCl 50, pH 7.4, at 0.1 g of tissue/ml of buffer. The homogenate was centrifuged at 3000 rpm for 3 min, then the supernatant for 10 min at 13,000 rpm in a Galaxy 20R centrifuge (VWR). The resultant mitochondrial pellet was resus- pended in isolation Buffer B (in mM): sucrose 150, KCl 50, KH2PO4 2, succinic acid 5 and Tris/HCl 20, pH 7.4. Concentration was deter- mined using the DC assay kit (Bio-Rad).Mitochondria ROS production was measured spectrofluoro- metrically at 560/590 nm (excitation/emission) in 100 mg of mito- chondrial protein in a buffer: (in mM): 20 Tris, 250 sucrose, 1 EGTA,1 EDTA, and 0.15% bovine serum albumin adjusted to pH 7.4 at 30 ◦C with continuous stirring. Amplex red dye (1 mM) (Thermofisher)and horseradish peroxidase (0.345 U/mL) were used to monitor H2O2 production, an analog for ROS generation [15]. H2O2 levels were calculated using a standard curve of H2O2 concentration and fluorescence intensity and a sodium salt of glutamate/malate (3 mM) was used to activate ETC complex I.Mitochondrial resistance to Ca2+ overload-induced formation of the mitochondrial permeability transition pore (mPTP) was measured as described in Refs. [18,19].

Isolated mitochondria (500 mg) were suspended in isolation Buffer B with the dye, calcium green-5N (0.1 mM) (ThermoFisher). Samples were placed in afluorescence spectrophotometer (Hitachi F2710) at 500/530 nm (excitation/emission) at 30 ◦C and incubated for 90 s followed by injections of CaCl2 (10 mmoles) pulses at 60 s intervals. Pulses induce a fluorescence peak of extra-mitochondrial Ca2+ + dye, which returns to baseline as mitochondria absorb Ca2+. Calcium uptake reduces as more Ca2+ pulses are added, and eventually, alarge release of mitochondrial Ca2+ occurs, signaling opening of themPTP. Calcium retention capacity (CRC) was defined as the amount of Ca2+ required to initiate mPTP opening and was expressed in nmol of CaCl2 per mg of mitochondrial protein.Data is shown in bar graphs expressed as means with error bars that are the standard errors of the mean (±SEM). Student’s t-test was used for comparisons using Prism 6 (Graphpad Software). A difference of p < 0.05 was considered to be statistically significant.

3.Results
Inhibiting ferroptosis using Lip-1 reduces I/R damage in the liver [4]. We found that administering Lip-1 at the start of reperfusion induced cardiprotective effects against I/R injury in mice (Fig. 1). Myocardial infarct size measured at the end of reperfusion was significantly reduced in the I/R + Lip-1 group compared to the I/Rcontrol group (53% in I/R versus 30% in I/R + Lip-1) (Fig. 1B).Further, analysis of tissue structure using electron microscopy revealed that Lip-1 protective effects directly result in protection of mitochondrial structural integrity, as well as preservation of cardiac contraction machinery (Fig. 1C).Erastin, a potent ferroptosis instigator, has been reported to actdirectly via binding to VDAC2/3 in BJeLR cells [9]. To assess the protective Lip-1 effects on mitochondria, we assessed expression levels of all the VDAC isoforms 1, 2 and 3. We found that although VDAC 2 (Fig. 2EeF) and VDAC 3 (Fig. 2GeH) levels remained the same after 2 h reperfusion, levels of VDAC1 were significantlyreduced in the I/R + Lip-1 group (Fig. 2AeB). Further, we alsofound that VDAC1 oligomerization, an occurrence tied to increasing cell death [20], was actually reduced with Lip-1 treat- ment (Fig. 2D).

Taken together, these results suggest that Lip-1 treatment leads to the increased degradation of VDAC1 and its oligomers.Lip-1 is known to inhibit ferroptosis by preventing lipid ROS build-up and cell death in inducible Gpx4—/— mice [4]. As shown in Fig. 3A, we found that I/R decreases levels of the cytosolic antiox- idant, GPX4, responsible for removing lipid peroxides, but that Lip- 1 treatment restores GPX4 levels. Furthermore, in order to assess ifLip-1 had any effects on mitochondrial ROS production, we stim- ulated complex I (using glutamate/malate) in isolated mitochon- dria. Here, we found that post-ischemic Lip-1 treatment results in significantly reduced generation of ROS from this complex (253 pmol/min/mg of mito protein for I/R versus 153 pmol/min/mgof mito protein for I/R + Lip-1) (Fig. 3B). From our previous studies,we have shown that reduction of ROS from complex I following I/R is cardio-beneficial [15], hence these results are in line with the notion of Lip-1 actions resulting in cardioprotection.Opening of the mPTP results in mitochondrial depolarization and subsequently cell death by apoptosis and necrosis. Previously, VDAC was postulated to be a component of the pore itself, but it has since been shown to be dispensable for mPTP formation [21]. In light of the protection of mitochondrial cristae and reduction of VDAC1 levels we observed, we next sought to assess mPTP openingin Lip-1 treated hearts. As shown in Fig. 3C, we did not observe any significant changes in the Ca2+ required to trigger mPTP opening in isolated mitochondria (145 nmol/mg of mito protein for I/R versus 160 nmol/mg of mito protein for I/R + Lip-1). From this, we concluded that Lip-1 protection of mitochondria does not involvedelaying mPTP opening.

4.Discussion
In this study, we have shown that post-ischemic treatment of isolated mice hearts with the ferroptosis inhibitor, liproxstatin-1, results in cardioprotective effects by reducing myocardial infarct size and protection of mitochondrial structural integrity and func- tion. On examining the mechanisms involved in Lip1-induced cardioprotection after I/R, we showed that protection of mito- chondria is associated with a significant decrease in the pro-cell death VDAC1, but does not involve the ferroptotic targets VDAC 2/3. Further, we showed that Lip-1 treatment leads to an increase in levels of the antioxidant GPX4 and a reduction of mitochondrial ROS production by complex I, without delaying Ca2+-induced mPTPopening.Induction of cell death is carried out by a variety of regulated pathways such as caspase-mediated (apoptosis), ROS-induced(oxeiptosis), inflammasome-activated (pyroptosis), and iron/lipid dependent (ferroptosis) pathways (reviewed in Ref. [22]). Despite ferroptosis being a recognized contributor to cell death in cardiac I/ R injury [23], understanding the mechanisms of anti-ferroptotic compounds has largely gone understudied, with a few studies focusing on ferrostatin-1 [22]. Here, we show that post-ischemic application of Lip-1 to mice myocardium results in decreased I/R injury as shown by decreased infarct sizes (Fig. 1). Given that ferrostatin-1 and Lip-1 both reduce lipid peroxidation, this result is in line with Lip-1 reducing ferroptosis following I/R injury in other organs such as the liver and kidneys [4,24].

Furthermore, our studies remarkably found that protection of mitochondrial struc- ture is an important outcome for Lip-1 cardioprotective effects (Fig. 1), contrary to the often-implied suggestions that mitochon- drial damage is a secondary component of ferroptosis [3,9]. This observation led us to conclude that Lip-1 cardioprotection against ferroptosis is facilitated via direct mitochondrial effects.The potent ferroptosis-inducer, erastin, both interferes with thecystine/glutathione antiporter system (system X—) and binds directly with the mitochondrial outer membrane (OMM) channel, VDAC2/3, leading to OMM rupture [9]. In evaluating the impact ofLip-1 protection of mitochondria, we measured protein levels of the VDAC isoforms 1, 2 and 3, but notably observed decreases only in VDAC1 content and VDAC 1 oligomerization (Fig. 2). The role of open/closed VDAC in cardiac I/R has long been debated, but generally VDAC1 is thought to promote injury, while VDAC2 may beprotective [25]. Increasing intracellular Ca2+ enhances VDAC1oligomerization and subsequently apoptosis [20], and our results show that Lip-1 reduces VDAC1 oligomerization. Knockdown ofVDAC1, a highly Ca2+-permeable channel, decreases cell death by reducing low-amplitude apoptotic Ca2+ signals being conveyed from the endoplasmic reticulum to mitochondria via the VDAC1/GRP75/IP3R1 complex [26,27]. Hence, it is plausible that Lip-1 cardioprotection is mediated by the decline of VDAC1 expression, without affecting VDAC 2/3.Lip-1 actions are presumed to be downstream of the cytosolic GPX4 antioxidant system that also converts reduced GSH to oxidized GSSH [3]. Adding to the complexity of ferroptosis path- ways, mitochondria have recently been suggested to be important for cysteine-deprivation-induced ferroptotsis, but dispensable for GPX4 inhibition-induced ferroptosis [12,28].

Here it is worth noting that cysteine is the precursor to GSH, the substrate of GPX4; andthat Lip-1 was discovered via its inhibition of ferroptosis in a Gpx4—/— mouse model [4]. In our studies however, Lip-1 treatment resulted in increased GPX4 levels in the 2 h reperfusion window(Fig. 3), suggesting that Lip-1 actions may not be restricted to downstream lipid peroxidation outcomes. Further, as shown in Fig. 3B, we also found that Lip-1 treatment reduced mitochondrial ROS production by the NADH-ubiquinone oxidoreductase (complex I). This is in contrast to studies that have not observed mitochon- drial complex I ROS changes in erastin-induced or ferrostatin- inhibited ferroptosis [1,5]. While it has also been shown that erastin-insitgated ferroptosis can also increase intra-mitochondrial ROS in the form of lipid peroxides [29]. Overall, it appears that the impact of GPX4 and ferroptosis on mitochondrial ROS may be dependent on the particular trigger or inhibitor of ferroptosis being studied.Lastly, opening of the mPTP in cardiac I/R injury because of Ca2+overload and/or increased ROS production causes cell death [17]. Our results show that decreased mitochondrial ROS and VDAC1 levels do not coincide with effects on Ca2+ needed to induce mPTP opening (Fig. 3C). In fact, Lip-1 treatment showed no impact on mitochondrial calcium retention capacity, which along withreduction of VDAC1, is in line with reports discounting VDAC as being part of the mPTP [21]. Ferroptosis inhibition by ferrostatin-1 has not been linked with delaying mPTP opening except with the addition of mPTP inhibitors, similar to the no-effect on Ca2+- induced mPT pore opening we observed with Lip-1 (Fig. 3C).

In conclusion, our study is the first to demonstrate the post- ischemic cardioprotective effects of Lip-1 in the myocardium against I/R injury via its reduction of infarct size and preservation of mitochondrial structure and function. The mechanism of this pro- tection involves reducing VDAC1 levels and oligomerization, without affecting VDAC2/3, and reduction of mitochondrial ROS generation. These Liproxstatin-1 effects are complemented by the increase in
GPX4 levels and no changes to Ca2+ levels required to induce mPTP opening.