CLPP loss confers sex-specific protection from renal ischemia reperfusion injury

Abstract

Renal ischemia-reperfusion injury (IRI) is initiated by a burst of mitochondrial reactive oxygen species generated when ischemia-accumulated succinate is rapidly oxidised upon reperfusion, driving reverse electron transport and triggering lipid peroxidation-dependent tubular cell death. Here, we identify the mitochondrial matrix protease CLPP as an unexpected regulator of renal redox resilience. CLPP deficiency reduced succinate-driven mitochondrial H₂O₂ production and protected kidneys from functional decline, tubular necrosis, mitochondrial ultrastructural damage, and lipid peroxidation following IRI. Protection was most pronounced in females, in which CLPP loss established a constitutive antioxidant and detoxification programme prior to injury, characterised by increased glutathione and γ-glutamylcysteine abundance, enhanced NADPH-generating pentose phosphate pathway activity, sustained sulphur amino acid metabolism, and elevated capacity for lipid aldehyde detoxification. During reperfusion, these adaptations preserved thiol-redox buffering, limited 4-hydroxynonenal accumulation, and reduced oxidative modification of cysteine residues on proteins involved in complement, coagulation, fibrinolysis, and inflammatory signalling, thereby restricting ferroptotic lipid damage and downstream thrombo-inflammatory amplification. Although reduced reverse electron transport-derived ROS contributed to this phenotype, it was insufficient to account for the sex bias. While CLPP deficiency amplified an endogenous female antioxidant programme, male kidneys, lacking equivalent glutathione-centred preconditioning, showed only marginal benefit. These findings establish mitochondrial proteostasis as a determinant of sex-dependent redox adaptation and identify CLPP as a candidate target for preconditioning against predictable ischaemic insults, including kidney transplantation and cardiac surgery.

Graphical abstract

1. Introduction

Ischemia-reperfusion injury (IRI) is fundamentally a disease of disrupted redox homeostasis, and the kidney, owing to its high mitochondrial content, dependence on oxidative phosphorylation for tubular transport, and limited glycolytic reserve, is among the most vulnerable organs to this insult [1]. Renal IRI contributes substantially to acute kidney injury arising during major surgery, haemodynamic instability, and sepsis, and is an unavoidable component of organ preservation and transplantation, where the severity of warm and cold ischemia directly influences delayed graft function, acute rejection, and long-term allograft survival [[1], [2], [3]]. Severe or recurrent injury can promote progression from acute to chronic kidney disease. Despite its broad clinical relevance, therapeutic options remain limited, reflecting the complexity of a rapidly evolving and redox-driven injury cascade [4].

A central event in IRI is the abrupt shift from reductive stress during ischemia to an oxidative burst upon reperfusion [5]. During ischemia, impaired oxidative phosphorylation causes ATP depletion, ionic imbalance, and extensive metabolic remodelling [6]. Reducing equivalents accumulate as electron transport becomes constrained, while TCA cycle and purine nucleotide metabolism favour succinate accumulation [6]. Upon reperfusion, rapid succinate reoxidation by complex II generates a highly reduced ubiquinone pool and a large proton-motive force, conditions that favour reverse electron transport (RET) through complex I and drive a burst of superoxide and hydrogen peroxide production at the mitochondrial matrix [6]. Together with calcium overload and rapid pH normalisation, this early oxidant surge promotes mitochondrial permeability transition, loss of membrane potential, and structural damage [6,7]. Critically, however, biological outcome is determined not only by the magnitude of ROS production, but by the capacity of the tissue to buffer, metabolise, and repair the resulting oxidative damage [5].

Renal tubular cells are particularly vulnerable to membrane lipid oxidation. Iron-dependent propagation of phospholipid peroxidation can overwhelm glutathione- and GPX4-dependent hydroperoxide detoxification, initiating ferroptosis, an oxidative, iron-dependent form of regulated cell death that has emerged as a dominant mode of acute tubular injury [8]. The susceptibility of tubular cells to ferroptosis is governed by the availability of glutathione and its biosynthetic precursors, particularly cysteine and γ-glutamylcysteine, and by NADPH regeneration through the pentose phosphate pathway, which sustains reductive capacity for antioxidant recycling [8]. Lipid peroxidation also generates reactive electrophilic aldehydes, including 4-hydroxynonenal (4-HNE), that covalently modify proteins, propagate damage beyond the initial oxidant event, and contribute to microvascular and metabolic dysfunction [8]. The consequences extend beyond tubular injury: oxidised lipids and damaged proteins act as danger signals that promote complement activation, coagulation, leukocyte recruitment, and microvascular dysfunction. Renal IRI should therefore be understood as a self-amplifying redox cascade in which mitochondrial oxidant production, antioxidant buffering, lipid peroxide detoxification, and inflammatory amplification are tightly and reciprocally interconnected.

Given the central role of RET-driven ROS production in IRI, therapeutic strategies have focused on modulating the mitochondrial redox environment. Mitochondria-targeted antioxidants such as MitoQ, SS-31, and SkQ1 have shown promise in preclinical models and early clinical studies [9,10]. Inhibition of succinate dehydrogenase using malonate derivatives has been explored to limit succinate accumulation and blunt the reperfusion ROS burst [9]. Non-pharmacological approaches including hypoxic preconditioning, caloric restriction, and therapeutic hypothermia have demonstrated protective effects through overlapping mechanisms [10,11]. However, challenges related to specificity, toxicity, timing, and delivery continue to hinder clinical translation, and strategies that engage endogenous antioxidant programmes before injury occurs remain underexplored.

The maintenance of mitochondrial proteostasis has emerged as a critical and complementary determinant of mitochondrial resilience under conditions of metabolic stress. Mitochondrial proteostasis encompasses the coordinated network of chaperones, proteases, and stress-responsive transcriptional programmes that ensure the folding, assembly, and timely removal of proteins within the mitochondrial compartment. The ATP-dependent protease CLPXP, composed of the proteolytic subunit CLPP and the chaperone CLPX, regulates a defined set of mitochondrial proteins rather than functioning as a general system for bulk degradation of damaged proteins [12]. One of its established functions is the selective turnover of subunits within the NADH-oxidizing N-module of respiratory complex I [13]. This salvage pathway enables replacement of N-module components while preserving the remainder of the complex [13]. Loss of CLPP disrupts this process, resulting in N-module accumulation and tissue-dependent remodelling of respiratory-chain organization. Because complex I is a major source of superoxide during RET, CLPP-dependent regulation of its composition may influence mitochondrial ROS production during reperfusion.

Here, we investigated whether CLPP loss alters susceptibility to renal IRI by modifying complex I-dependent ROS production. We examined the effects of CLPP deficiency on respiratory-chain organization, succinate-driven mitochondrial H₂O₂ generation, renal injury, and the molecular response to reperfusion. Our findings identify CLPP as a regulator of renal responses to ischemic injury and reveal an unexpected link between mitochondrial respiratory-chain remodelling and redox homeostasis.

2. Results

2.1. CLPP loss causes mild perturbation of mitochondrial function in the kidney

To evaluate the consequences of CLPP deficiency in the kidney prior to IRI, we first characterized baseline molecular changes in CLPP-deficient kidneys. Consistent with findings in other CLPP-deficient tissues, analysis of whole-kidney protein extracts demonstrated significant accumulation of established CLPP targets in both female and male mice (Fig. 1A and S1A) [12,14,15]. Although this phenotype was evident in both sexes, the accumulation of several CLPP substrates was modestly more pronounced in CLPP-deficient females (Fig. 1A and S1A).

Fig. 1.

CLPP deficiency alters renal mitochondrial homeostasis and stress responses

A, Steady state levels of different proteins known to be CLPP substrates in male and female WT and CLPP KO. HSC70 is used as loading control (n = 3).; B,C Steady state levels of OXPHOS complexes in the kidney mitochondria of male and female WT and CLPP KO, analysed by solubilization with 1% dodecylmaltoside (DDM) separated using blue native–PAGE followed by Western blot for CI (NDUFV2 (B) and NDUFA9 (C)), CV (ATP5A (C)), CIII (UQCRC1 (C)), CIV (COX4I1 (C)) and CII (SDHA (B & C).; D, Steady state levels of OXPHOS supercomplexes in the kidney mitochondria of male and female WT and CLPP KO, analysed by solubilization with 4% digitonin separated using blue native–PAGE followed by Western blot for CI (NDUFA9), CIII (UQCRC1), CIV (COX4I1), CV (ATP5A), and CII (SDHA) (n = 3).; E, Steady state levels of different proteins associated with the ISR. HSC70 is used as loading control (n = 3).; F, OCRs of isolated mitochondria from WT and CLPP KO mice of kidney and heart stimulated with succinate and rotenone (n = 6, 3 males and 3 females) and H₂O₂ production of isolated mitochondria from kidney and heart of WT and CLPP KO fed with succinate and rotenone to induce RET. RET calculated as difference from succinate and rotenone. Data are presented as mean ± s.d. P values are indicated above each condition compared to different groups *p < 0.05, determined using paired student's t-test.; G, Volcano plot of kidney whole tissue proteome analysis (red = p-value < 0.05, Log2FC > 0.6; blue = p-value < 0.05, Log2FC < −0.6) (n = 4).

In heart and brown adipose tissue (BAT), CLPP-deficiency causes mild-to-moderate OXPHOS dysfunction, while much milder effect was detected in liver and white adipocyte tissue (WAT) of CLPP KO animals [12,16,17]. To evaluate the extent of respiratory chain dysfunction in CLPP-deficient kidneys, we analysed the levels and composition of individual OXPHOS complexes by blue native PAGE (BN-PAGE), which resolves assembly intermediates, fully assembled complexes, and supercomplexes [18]. While BN-PAGE analysis demonstrated accumulation of the CI N-module (Fig. 1B), consistent with activation of the CI salvage pathway [13], the overall abundance of fully assembled CI and steady-state levels of complexes II-V (CII-CV) remained unchanged (Fig. 1B–C). We did observe accumulation of F1 subassemblies of CV together with increased levels of free SDHA from CII in CLPP-deficient kidney mitochondria (Fig. 1C), consistent with results in other tissues [12,16,17]. Analysis of mitochondrial supercomplex organisation revealed only a mild reduction in CI- and CIV-containing supercomplexes in CLPP KO animals (Fig. 1D). CI and CIV in-gel activity assays corroborated these findings (Fig. S1B), confirming that supercomplex integrity was only subtly compromised.

Consistent with these structural observations, CLPP deficiency did not significantly reduce oxygen consumption rates (OCRs) in the kidney (Fig. S1C), indicating that mitochondrial respiratory function is largely preserved in renal tissue. Increased mitochondrial protein acetylation, which commonly accompanies more severe OXPHOS dysfunction through an elevated NADH/NAD⁺ ratio and reduced SIRT3-dependent deacetylation, was not detected in CLPP-deficient kidneys from either sex (Fig. S1D), further supporting the relatively mild nature of the respiratory defect. In agreement with mild OXPHOS impairment, CLPP-deficient kidneys displayed mild activation of the mitochondrial integrated stress response (ISRmt), evidenced by increased ALDH18A1, MTHFD2, and SHMT2, proteins involved in proline-to-glutamine conversion and one-carbon metabolism (Fig. 1E and S1E) [15,19]. As observed for CLPP substrate accumulation, some aspects of this response were modestly more pronounced in females, with higher levels of selected ISRmt-responsive proteins and CV F1 subassemblies (Fig. 1C and E and S1E). Thus, while the characteristic mitochondrial proteostasis and OXPHOS-associated phenotype of CLPP deficiency was present in both sexes, its molecular manifestations were somewhat more pronounced in females.

We previously showed that N-module accumulation in CLPP-deficient cells limits oxidative stress and reduces hydrogen peroxide (H₂O₂) production under both basal and stress conditions [13]. To determine whether CLPP deficiency similarly suppresses mitochondrial ROS production in vivo, we measured ROS generation under forward and reverse electron transport (RET) conditions using Amplex Red in an OROBOROS Oxygraph-2k system, enabling simultaneous assessment of ROS production and OCR (Fig. 1F and S1F). CLPP deficiency reduced succinate-driven H₂O₂ production in both heart and kidney (Fig. 1F and S1F). RET-associated H₂O₂ production, calculated as the difference between succinate-stimulated H₂O₂ generation before and after rotenone addition, was likewise reduced in CLPP KO animals of both sexes (Fig. 1F and S1F). Together, these findings demonstrate that CLPP loss is associated with reduced RET-dependent mitochondrial ROS production in both heart and kidney, supporting a link between CLPP-dependent respiratory-chain remodelling and mitochondrial ROS generation.

To characterise the effect of CLPP loss on the renal proteome, we performed label-free quantitative proteomics on whole-kidney lysates from CLPP KO female mice. CLPP deficiency induced selective baseline proteomic remodelling predominantly affecting pathways linked to mitochondrial proteostasis, stress adaptation, and metabolic regulation (Fig. 1G and S2A-S2B, Supplemental Table 1). Proteins associated with mitochondrial translation and respiratory chain maintenance, including established CLPP-regulated factors (e.g., GFM1, TRAP1, OAT, and MTHFD2) together with multiple mitoribosomal components, were consistently increased, indicating altered mitochondrial protein quality control and adaptive remodelling of the mitochondrial proteome [12]. In parallel, antioxidant and detoxification pathways, particularly glutathione-dependent stress responses involving GSTA3 and GSTA4 [20], were enhanced. Pathway enrichment analysis (GO-BP - Gene Ontology-Biological Processes; KEGG - Kyoto Encyclopedia of Genes and Genomes) [21] further demonstrated enrichment of mitochondrial and cytosolic translation, iron-sulphur cluster assembly, xenobiotic detoxification, amino acid metabolism, and cofactor biosynthesis among upregulated proteins (Fig. S2A–S2B, Supplemental Table 1). Conversely, proteins and pathways associated with tubular transport, complement and coagulation cascades, wound healing, peroxisomal function, and lipid and branched-chain amino acid metabolism were reduced, indicating coordinated suppression of inflammatory, tissue-remodelling, and lipid-handling programmes (Fig. S2A–S2B, Supplemental Table 1). Together, these findings indicate that CLPP deficiency establishes a constitutive mitochondrial stress-adaptive state in the kidney characterised by broad metabolic and immune remodelling.

2.2. CLPP deficiency protects against renal IRI in a sex-dependent manner

Given that CLPP deficiency reduces mitochondrial ROS production under both basal and RET conditions, we hypothesized that CLPP may represent a therapeutic target in pathophysiological settings driven by oxidative stress, such as renal IRI. To test this, CLPP-deficient and control mice were subjected to unilateral renal IRI following contralateral nephrectomy, with 40 min of ischemia and 24 h of reperfusion (Fig. 2A).

Fig. 2.

CLPP deficiency protects from renal ischemia–reperfusion injury in a sex-dependent manner

A, Scheme of ischemia reperfusion injury.; B, Blood urea nitrogen (BUN), creatinine and cystatin c values determined of serum of female and male WT and KO mice that underwent IR and controls (Sham). 24 h after IRI. (n = 6-22).; C, Representative PAS-stainings of female WT and KO kidney IR-treated and control animals. (Scale bar: 250 μm).; D, Evaluation of PAS stainings by trained nephropathologist using the composite damage score and individual plotting of the necrosis score. 24 h after IRI. (n = 6-22).; E, Steady state levels of proteins related to necrosis in WT and CLPP KO before and after 24h of IR, and its quantification. HSC70 is used as loading control. (n = 3).; F, CI (NDUFV2) in kidney mitochondria of WT and CLPP KO before and after 30 min of IR, analysed by solubilization with 1% dodecylmaltoside (DDM) separated using blue native–PAGE followed by Western blot for CI (NDUFV2) and CII (SDHA).; G, Representative transmission electron microscopy images of female WT and KO kidney before IR and after 30 min and 24 h IR. (Scale bar: 1 μM). Data are presented as mean ± s.d. P values are indicated above each condition compared to different treatment groups *p < 0.05, **p < 0,01, ***p < 0.001, ****p < 0.0001, were determined by one-way ANOVA with Turkey's multiple comparisons test.

At 24 h post-IRI, plasma levels of blood urea nitrogen (BUN), creatinine, and cystatin C were markedly elevated compared with sham-operated controls, confirming substantial impairment of renal function (Fig. 2B). Cystatin C was included as an additional marker of glomerular filtration because it is less influenced by non-renal factors, potentially modulated by CLPP-deficiency, such as muscle mass or protein intake [22].

Consistent with our hypothesis, female CLPP KO mice exhibited significantly attenuated increases in BUN, creatinine, and cystatin C compared with wild-type controls following IRI (Fig. 2B). In contrast, male CLPP KO mice showed only very mild, if any improvement across these parameters, indicating a sex-dependent effect of CLPP deficiency on susceptibility to renal injury (Fig. 2B). Histological assessment of periodic acid–Schiff (PAS)-stained kidney sections supported these findings (Fig. 2C). A composite tubular injury score [23], further supported these findings showing reduced injury in female but not male CLPP KO mice (Fig. 2D). Given these functional and histological differences, we next examined whether altered cell-death pathways contributed to the protective phenotype.

TUNEL staining showed a trend toward reduced cell death in both female and male CLPP KO kidneys, although this did not reach statistical significance in either group and did not segregate with the functional or histological outcomes (Fig. S3A–S3C). Consistent with this, we found no evidence for differential activation of apoptosis following IRI. Total caspase-3 and caspase-9 levels were comparable between female WT and CLPP KO mice, cleaved caspase-3 was not detected, and BID, BCL2, and BCL-XL levels remained unchanged (Fig. S3D).

We next examined components of the necroptotic pathway. IRI increased RIPK1 and RIPK3 abundance in both WT and CLPP KO kidneys, with a more pronounced increase in CLPP-deficient animals (Fig. 2E). However, because RIPK1 and RIPK3 participate in several signalling pathways, their increased abundance alone does not establish enhanced necroptotic execution. Notably, phosphorylation of MLKL, a downstream marker of necroptotic execution, was reduced in female CLPP KO kidneys both at baseline and following IRI compared with WT controls (Fig. 2E). Thus, despite increased RIPK1 and RIPK3 abundance, CLPP-deficient kidneys showed reduced downstream MLKL activation, arguing against enhanced necroptosis as the cause of tubular injury and consistent with the reduced histological injury observed in female CLPP KO mice.

Given that ROS bursts during IRI can damage OXPHOS complexes and trigger the degradation and replacement of damaged respiratory-chain components [24], we next investigated whether the protective effect of CLPP deficiency might involve altered OXPHOS dynamics. BN-PAGE analysis at 24 h post-injury did not reveal additional differences between control and CLPP-deficient mitochondria (Fig. S4A), although this time point may not capture early reperfusion-associated changes. Notably, the same analysis in male kidneys at 24 h post-IRI likewise revealed no substantial genotype-dependent differences in OXPHOS complex or supercomplex organization, and CI and CIV in-gel activities were also comparable between male control and CLPP-deficient mitochondria (Fig. S4B–4C). Thus, at this later time point, IRI did not induce a discernible genotype-specific OXPHOS response in either sex. Because the baseline mitochondrial phenotype associated with CLPP loss was somewhat more pronounced in females, and the protective effect of CLPP deficiency following IRI was also strongest in females, we focused the analysis of early reperfusion-associated OXPHOS changes on female kidneys at 30 min post-reperfusion.

BN-PAGE followed by immunoblotting of CI subunits previously identified as CLPP targets [13] demonstrated persistent N-module accumulation, comparable to baseline conditions (Fig. 2F and S4D). Likewise, supercomplex abundance, elevated CV F1 subassemblies, and increased free SDHA levels observed in CLPP-deficient mitochondria at baseline remained unchanged after IRI (Fig. S4D–S4E). In-gel activity assays for CI and CIV showed reduced activity in CLPP-deficient mitochondria consistent with lower expression levels (Fig. S4F). However, these deficits were not further exacerbated by IRI (Fig. S4F). Together, these findings indicate that mitochondrial OXPHOS composition and supercomplex organization are largely preserved during IRI, and that the pre-existing mitochondrial phenotype in CLPP-deficient kidneys is not further remodelled by reperfusion stress.

To directly assess mitochondrial ultrastructure, transmission electron microscopy (TEM) was performed at 30 min and 24 h post-reperfusion (Fig. 2G). In sham-operated animals, mitochondria from both WT and KO kidneys appeared elongated with densely packed, well-organized cristae, indicative of intact mitochondrial architecture. This morphology was maintained at 30 min post-reperfusion in both genotypes, suggesting that early reperfusion does not immediately disrupt mitochondrial structure. In contrast, pronounced structural abnormalities were evident at 24 h post-reperfusion in WT kidneys, where mitochondria appeared swollen and rounded with disorganized or disrupted cristae, consistent with advanced mitochondrial damage (Fig. 2G). Strikingly, mitochondria from CLPP KO kidneys largely retained their elongated morphology and cristae organization at this time point, closely resembling sham controls.

Collectively, these results demonstrate that CLPP deficiency confers protection against renal IRI in a sex-dependent manner, associated with preserved mitochondrial architecture and reduced functional decline despite only modest baseline alterations in mitochondrial function.

2.3. CLPP deficiency suppresses inflammatory and complement-coagulation amplification in female kidneys following IRI

To define how CLPP deficiency in female mice reshapes the renal response to IRI, we performed label-free proteomic analysis at early (30 min) and late (24 h) reperfusion time points (Fig. 3A). Although almost 200 proteins were significantly altered between CLPP-deficient and wild-type kidneys at 30 min post IRI, most changes reflected baseline adaptations associated with CLPP loss [[12], [13], [14],17], with only a small subset uniquely regulated by IRI (Fig. S5A–S5B, Supplemental Table 1). Consistently, sham versus IRI comparisons in female mice of both genotypes showed only modest proteomic shifts at this early stage (Supplemental Table 1), indicating that the acute injury response remained limited on the level of proteome. By contrast, the 24 h reperfusion revealed a pronounced genotype-dependent response. Kidneys from female WT mice exhibited extensive proteomic remodelling following IRI (504 significantly altered proteins), whereas CLPP-deficient kidneys displayed a substantially attenuated response (314 proteins), despite approximately half of the changes already being present at baseline (Fig. 3B–S5C-S5E).

Fig. 3.

Female CLPP-deficient kidneys exhibit an attenuated inflammatory proteomic response to ischemia–reperfusion injury

A, Scheme of individual time points used during this study.; B, Volcano plot of proteomics changes between KO and WT after 24 h IR, as identified on proteomics analysis (red = p-value < 0.05, Log2FC > 0.6; blue = p-value < 0.05, Log2FC < −0.6).; C-E, G, Heatmap of Z-score normalised proteins of Kidney damage factor proteins and inflammatory-stress responses (C), Complement factors (D), TCA Cycle proteins (E) and Serin and 1C Metabolism (G) identified in the kidney of WT and KO before and after 24 h IR.; F, Dot-plot of GO Biological Process Enrichment analysis of proteins, significantly changed in kidney whole tissue proteome analysis at 30 min vs 24h IR; number of upregulated proteins indicated in brackets; dot size represents gene ratio; color gradient represents -log₁₀ FDR.; H, Representative IHC staining of female WT and KO kidney IR-treated (24 h) and control animals for CD45 and F4/80. (Scale bar: 1 mm).

Wild type kidneys of female mice followed a canonical trajectory of IRI progression. Whereas early changes were modest, a fully developed injury phenotype emerged by 24 h post IR. Tubular injury markers including HAVCR1/KIM1, LCN2, and CALB1 were strongly induced, consistent with overt tissue damage (Fig. 3C). In parallel, inflammatory and neutrophil-associated proteins (e.g., MPO and S100A9) were markedly increased, reflecting inflammatory recruitment and amplification. This response was accompanied by induction of oxidative stress markers, together with strong activation of complement and coagulation pathways and extracellular matrix remodelling (Fig. 3C-D and S5F). At the level of metabolism, renal IRI induced partial suppression of fatty acid oxidation, whereas TCA cycle alterations were comparatively very mild (Fig. 3E and S5G). Collectively, WT kidneys transitioned from early metabolic perturbation to a coordinated program of inflammatory activation, thrombo-inflammatory signalling, and metabolic suppression by 24 h (Fig. 3F).

CLPP-deficient kidneys of female mice diverged from this trajectory at both baseline and during injury, as reflected by distinct GO-BP signatures following early and late reperfusion (Fig. 3F). Under sham conditions, KO kidneys already displayed substantial alterations in mitochondrial proteostasis and metabolic organisation, consistent with a constitutively pre-stressed mitochondrial state (Fig. 1G). Following IRI, these kidneys retained the ability to initiate a core injury response, as evidenced by induction of canonical damage markers including HAVCR1/KIM1 and HMOX1 (Fig. 3C). However, downstream propagation of the response was markedly altered. Direct comparison of CLPP KO and WT kidneys after IRI revealed consistent attenuation of pro-inflammatory and injury-amplifying mediators, together with diminished neutrophil activation markers (Fig. 3C). In parallel, factors associated with epithelial identity and cytoprotective signalling, including HNF4A and the vitamin D receptor [25], were relatively preserved or increased in KO kidneys, suggesting a more controlled response to injury (Fig. 3C). In contrast to findings in female WT kidneys, key components of mitochondrial oxidative metabolism, including representative fatty acid oxidation enzymes and TCA cycle proteins, were downregulated, indicating suppression of oxidative metabolism, whereas stress-adaptive pathways including serine and one-carbon metabolism remained elevated, consistent with sustained metabolic adaptation rather than recovery (Fig. 3E–G and S5G).

A particularly striking feature of the CLPP-deficient response in female mice was suppression of complement and coagulation/fibrinolysis signalling that was clearly observed already at the baseline (Fig. 3D and S5F). Components of these pathways, that was strongly induced in WT kidneys after IRI, were consistently reduced in CLPP deficient female animals (Fig. 3D and S5F). Given the established role of complement and coagulation cascades in amplifying IRI through inflammatory and microvascular injury [26,27], their attenuation provides a mechanistic basis for the reduced injury progression observed in CLPP-deficient kidneys. This effect was accompanied by lower levels of DAMP-associated proteins, further supporting reduced activation of damage-amplifying feedback loops (Fig. 3C). Consistent with this, immunohistochemical analysis demonstrated markedly reduced accumulation of CD45⁺ leukocytes in KO kidneys compared with WT after IRI, whereas macrophage numbers remained largely unchanged, suggesting selective suppression of early inflammatory cell recruitment (Fig. 3H).

Taken together, these findings demonstrate that CLPP deficiency reshapes the renal response to IRI in female mice, such that pre-existing adaptations constrain downstream inflammatory amplification without abolishing injury initiation, identifying CLPP as a key modulator of IRI progression.

2.4. Loss of CLPP activates transcriptional programme that preconditions against renal IRI

To define the transcriptional programmes underlying protection of CLPP-deficient kidneys from IRI, we performed RNA-seq analysis of WT and CLPP-deficient kidneys from female mice 24 h after injury. Differential expression analysis identified approximately 1100 significantly altered transcripts between two genotypes upon renal IRI, majority of which were upregulated in kidneys of CLPP-deficient female mice relative to WT controls (Fig. 4A, Supplemental Table 2). Functional GO-BP enrichment analysis, revealed broad immunometabolic remodelling in response to ischemic stress (Fig. S6A, Supplemental Table 2). A closer inspection of transcripts downregulated upon IRI in CLPP KO vs. WT mice, revealed coordinated suppression of inflammatory amplification pathways, including neutrophil activation, complement-coagulation signalling, and fibrotic tissue remodelling (Fig. 4B–Supplemental Table 2), in agreement with previously observed proteomics changes (Fig. 3C-D, Supplemental Table 1).

Fig. 4.

Baseline NRF2 activation preconditions CLPP-deficient kidneys against ischemia–reperfusion injury

A, Volcano plot of genes changed in between KO and WT after 24 h IR, as identified on RNA-seq (yellow = p-value < 0.05, Log2FC > 0.6; blue = p-value < 0.05, Log2FC < −0.6).; B-D, Heatmap of Z-score normalised genes of Inflammation/Kidney damage (B), Innate Immunity (C) and Lipid metabolic process (D) identified in the kidney of WT and KO before and after IR.; E, GSEA enrichment plots of the GO-BP term “Detoxification” of KO vs WT in Sham (black) and after 24 h IR (red).; F, Dot-plot of GO Biological Process Enrichment analysis of genes significantly changed in kidney RNA-seq analysis of KO vs WT; number of upregulated proteins indicated in brackets; dot size represents gene ratio; color gradient represents -log₁₀ FDR.; G, Enrichment analysis of Transcription factors determined from the significantly upregulated genes, determined using ENRICHR.; H, GSEA enrichment plots of NRF2 targets of KO vs WT in Sham (black) and after 24 h IR (red).; I, Heatmap of Z-score normalised genes of ATF4 targets identified in the kidney of WT and KO before and after IR.; J, Steady state level of proteins of NRF2 pathway in male and female WT and KO animals and its quantification. Data are presented as mean ± s.d. P values are indicated above each condition compared to different treatment groups *p < 0.05, **p < 0,01, ***p < 0.001, ****p < 0.0001, were determined by one-way ANOVA with Turkey's multiple comparisons test.

The most prominent signature among transcripts increased in CLPP-deficient kidneys after IRI was enrichment of innate immunity, in particular interferon-stimulated genes (ISGs) (Fig. 4C). Although CLPP deficiency has previously been linked to cGAS–STING-mediated interferon signalling through mtDNA release [28], ISG activation in CLPP KO kidneys became apparent following IRI. In wild-type female kidneys, ISG transcripts decreased markedly after injury, whereas CLPP-deficient kidneys retained or further elevated ISG expression relative to baseline (Fig. 4C). Beyond the ISG response, transcripts increased in CLPP-deficient kidneys were enriched for lipid metabolism and adipokine-related programmes, together with epithelial transport and differentiation markers (Fig. 4D and S6B). Collectively, these findings indicate that CLPP deficiency does not globally suppress renal stress responses, but instead rewires the post-ischemic transcriptional programme.

Although ischemia-reperfusion triggered extensive transcriptional remodelling in both genotypes, several observations suggested that these post-ischemic changes were unlikely to represent the primary protective mechanism in CLPP-deficient kidneys. Most notably, many stress-responsive pathways, particularly antioxidant and detoxification programmes, were equally or more strongly induced in wild-type kidneys despite substantially worse functional and structural outcomes (Fig. 4E and S6C-S6D). This pattern indicates that acute transcriptional induction primarily reflects injury burden and compensatory stress signalling rather than effective protection. Instead, the critical distinction between genotypes appeared to lie in their baseline transcriptional organisation before injury, suggesting that the protective phenotype of CLPP-deficient kidneys in female mice is driven by preconditioning rather than enhanced acute stress responses (Fig. 4E and S6C).

To define this preconditioned state, we next compared baseline kidneys between genotypes. Baseline transcriptomic analysis identified more than 800 significantly altered transcripts in CLPP-deficient female kidneys, revealing extensive remodelling prior to injury (Fig. S6E). Enriched GO-BP categories included xenobiotic metabolism, glutathione detoxification, arachidonic acid and lipid metabolism, transmembrane transport, and hypoxia-response pathways (Fig. 4F). Transcription factor enrichment analysis identified NRF2/NFE2L2 as the dominant baseline regulator (Fig. 4G-H). In contrast, regulators enriched after IRI were predominantly associated with lipid metabolism and immune signalling, including PPARG and STAT1 (Fig. S6F). NRF2/NFE2L2-driven transcriptional remodelling was not dominant in CLPP-deficient heart or brown adipose tissue [15,19] (Fig. S6G), indicating that this adaptive response is not a universal consequence of CLPP loss but rather a kidney-specific transcriptional signature.

Although ATF4 target genes showed modest baseline enrichment, several lines of evidence argued against classical mitoISR signalling as the dominant protective mechanism: hallmark features of robust mitoISR activation, including increased EIF2α phosphorylation, were absent, and only a subset of serine, proline, and one-carbon metabolism proteins were selectively upregulated (Fig. 4I–S6H). Instead, the dominant transcriptional signature was NRF2-driven, encompassing antioxidant and detoxification genes including Slc7a11, Hmox1, Gsta1, Gsta4, Mgst2 and Nqo1 (Fig. 4E and H, and S6C).

To investigate the upstream basis of this response, we examined major regulatory components of the NRF2 pathway. Steady-state NRF2 and KEAP1 protein levels were unchanged between genotypes and sexes, whereas GSTA1 was preferentially increased in CLPP-deficient females and NQO1 was increased in both sexes. In contrast, other NRF2-associated antioxidant proteins examined, including GCLC and FSP1, were unchanged (Fig. 4J and S7A). Total GSK3β abundance and basal GSK3β phosphorylation were also unchanged, while IRI increased GSK3β phosphorylation similarly in WT and CLPP-deficient female kidneys (Fig. S7A–S7B). Global protein ubiquitination was not increased by CLPP deficiency, with P62/SQSTM1 showing a modest reduction (Fig. S7C–S7D). Inverse redox-shift analysis [13] did not reveal a robust difference in KEAP1 oxidation between genotypes, owing to substantial variability across samples (Fig. S7E).

Together, these findings support selective enhancement of NRF2-associated antioxidant responses in CLPP-deficient kidneys, with the strongest induction of several protective components in females, but do not identify a single dominant upstream mechanism responsible for this response. Rather, the data indicate that CLPP deficiency establishes a constitutive antioxidant and detoxification programme in female kidneys prior to injury, associated with an NRF2-centred transcriptional signature and with reduced inflammatory amplification following reperfusion.

2.5. Sex-biased antioxidant metabolism defines the protective response to CLPP deficiency in renal IRI

Given the pronounced female-biased protection observed in CLPP-deficient kidneys, we next investigated whether mitochondrial proteostasis interacts with endogenous sex-specific stress programmes to shape the renal response to ischemic injury. Previous studies have shown that wild type female mice are also more resistant to renal IRI than male mice, in part through enhanced resistance to oxidative stress and ferroptosis [27,[29], [30], [31]]. Consistent with this, the protective effect of CLPP deficiency was markedly stronger in females, prompting us to dissect sex-specific proteomic, metabolomic, and redox adaptations under basal and post-ischemic conditions. To determine whether this sex bias could simply reflect altered systemic sex hormone levels, we measured circulating estradiol in females and testosterone in males. Neither differed significantly between WT and CLPP-deficient mice (Fig. S8A), arguing against altered circulating sex hormone levels as a simple explanation for the female-biased renal phenotype.

Under baseline conditions, WT kidneys displayed pronounced sexual dimorphism, with 243 proteins differing between males and females (Fig. 5A and S8B, Supplemental Table 3). Female WT kidneys exhibited a constitutive antioxidant and detoxification programme centred on a lipid peroxidation defence axis (Fig. 5A). This programme closely aligns with established mechanisms of renal IRI, in which oxidative stress, 4-hydroxynonenal (4-HNE) accumulation, glutathione-dependent detoxification, and ferroptotic lipid damage are major determinants of tubular injury [27,29,31].

Fig. 5.

CLPP deficiency reinforces female antioxidant metabolism to protect against renal ischemia–reperfusion injury

A, Volcano plot of proteins changed between WT females and WT males, as identified on proteomics (red = p-value < 0.05, Log2FC > 0.6; blue = p-value < 0.05, Log2FC < −0.6).; B, Volcano plot of proteins changed between KO females and KO males, as identified on proteomics (red = p-value < 0.05, Log2FC > 0.6; blue = p-value < 0.05, Log2FC < −0.6).; C, Heatmap of Z-score normalised genes, of up and downregulated proteins that are only significant in KO females.; D, Pie charts representing the significantly up- (red) and downregulated (blue) proteins, in the indicated conditions after IR.; E, Heatmap of Z-score normalised genes of upregulated proteins that are only significant in KO females after IR.; F, Correlation matrix of kidney damage markers, comparing KO Females vs other groups.; G/H, Heatmap of Z-score normalised metabolites of up (G) and downregulated (H) metabolites, that are significant in KO females vs WT females after IR.; I, Heatmap of Z-scores normalised proteins of downregulated proteins that are only significant changed in KO Females after IR. J, Overlapped GSEA Enrichment Plots of the GO-BP Glutathione Metabolic process, comparing KO/WT Sham (black) vs IR (red); K, Immunohistochemistry staining of 4-Hydroxynonenal, (Scale bar: 250 μM).; L, Heatmap of Z-score of significantly changed protein oxidation residues normalised to total protein changes as detected on cysteine centred redox proteomics. Sorted by significantly altered expression of cysteine residues as detected in WT males after IR.

CLPP deficiency markedly reduced overall sexual dimorphism, with only 99 proteins differing between sexes in CLPP KO kidneys compared to 243 in WT, indicating that CLPP loss partially overrides the physiological sex-specific renal proteome (Fig. 5B and S8B, Supplemental Table 3). This convergence was not symmetric. Many antioxidant pathways constitutively enriched in WT female kidneys overlapped with the NRF2-centred programme induced by CLPP deficiency, suggesting that CLPP loss amplifies endogenous female antioxidant metabolism rather than establishing an entirely novel stress state. While male CLPP KO kidneys exhibited changes consistent with reduced metabolic resilience and impaired baseline tubular integrity, including altered FABP1, ACOX1, TMEM260, FAM114A1, CAPN5, and BLTP2 expression, female CLPP KO kidneys displayed a distinct signature of coordinated suppression of mitochondrial and peroxisomal lipid catabolism together with selective induction of detoxification, one-carbon metabolism, and metabolite clearance pathways (Fig. 5C and S8C). Collectively, these baseline adaptations indicate that CLPP-deficient female kidneys are constitutively primed against IRI through a female-specific adaptive programme that channels mitochondrial stress toward cytoprotection rather than injury amplification.

To determine whether this sex-biased response was a general consequence of CLPP deficiency, we additionally analysed baseline heart proteomes from male and female WT and CLPP-KO mice (Fig. S8D–S8E, Supplemental Table 4). Sex-dependent differences on the level of proteome were also evident in heart, but their functional organisation differed substantially from kidney (Supplemental Table 4). WT hearts showed pronounced sex differences in complement- and immune-related proteins, whereas the more modest differences between female and male CLPP-KO hearts involved mitochondrial translation, aerobic respiration, muscle-associated and selected metabolic processes (Supplemental Table 4). Notably, the antioxidant and detoxification programmes that distinguished CLPP-deficient female kidneys were not enriched in heart (Supplemental Table 4), and the marked increase in GSTA1 observed in female CLPP-deficient kidneys (Fig. 4J) was not reproduced in the other CLPP-deficient tissues examined (Fig. S8F). Together, these findings indicate that although sex influences the proteomic response to CLPP loss in multiple tissues, the coordinated antioxidant/detoxification programme observed in female kidneys is tissue-specific.

Following IRI, CLPP-deficient female kidneys exhibited less than half the number of IRI-induced proteomic changes compared to all other groups (187 vs. 429-442 proteins), reflecting not a weaker response but a fundamentally redirected one favouring tissue preservation over inflammatory amplification (Fig. 5D–Supplemental Table 3). This redirected response sustained the cytoprotective programme established at baseline, while male KO kidneys largely failed to mount an equivalent response (Fig. 5E-F and S9A-S9B). The overall proteomic differences between CLPP-deficient females and males after IRI recapitulated those between WT females and males, indicating that CLPP loss amplifies pre-existing female cytoprotective adaptations rather than conferring an independent male-protective programme (Fig. 5E and S9A-S9C, Supplemental Table 3). Nevertheless, CLPP deficiency narrowed the sex gap in IRI susceptibility, such that male KO kidneys outperformed WT females, suggesting that even partial engagement of this adaptive programme confers meaningful protection (Fig. 5E and S9C).

Metabolomic profiling revealed a striking female-specific protective phenotype in CLPP KO kidneys, characterised by constitutively enhanced antioxidant buffering capacity that persisted and intensified following IRI (Fig. 5G–I, S9D-S9E, Supplemental Table 5). At baseline, KO females exhibited marked elevation of glutathione, γ-glutamylcysteine, NADPH, and pentose phosphate pathway intermediates, indicating sustained glutathione biosynthesis and reducing power for redox homeostasis (Fig. 5G–Supplemental Table 5). Following IRI, oxidised glutathione remained strongly elevated in CLPP KO females, consistent with active glutathione recycling rather than antioxidant depletion under stress (Fig. 5H and S9D-S9E). Metabolites linked to sulphur amino acid and one-carbon metabolism were similarly maintained, supporting continued cysteine supply and metabolic flexibility during injury (Fig. 5H and S9D-S9E). Consistent with this, transcriptional analysis demonstrated selective baseline upregulation of glutathione metabolism genes in female CLPP-deficient kidneys, whereas renal IRI broadly induced these pathways across both genotypes (Fig. 5J).

By contrast, metabolites reflecting mitochondrial amino acid remodelling, including homocitrulline, homoarginine, and 2-hydroxyglutarate, were elevated in both sexes and often more strongly in CLPP-deficient males, yet did not confer comparable protection (Fig. 5H). This dissociation suggests that the protective phenotype is not a generic consequence of CLPP-deficiency-induced mitochondrial stress, but specifically depends on female-biased antioxidant metabolism. Furthermore, oxidative damage markers including allantoin, N-acetylhistidine, creatinine, and methionine sulfoxide were markedly blunted in CLPP-deficient females and not males, whereas metabolites associated with nucleotide salvage and membrane repair were selectively increased, suggesting enhanced recovery capacity (Fig. 5H and I).

Functional and redox analyses supported this model. Consistent with reduced susceptibility to lipid peroxidation, ACSL4 levels were lower in CLPP-deficient female kidneys, whereas FSP1 and GCLC were unchanged (Figure S9F, Figure S7A). Histological assessment of 4-HNE accumulation demonstrated markedly reduced lipid peroxidation in female CLPP KO kidneys after IRI, whereas male kidneys of both genotypes displayed extensive 4-HNE positivity (Fig. 5K). Cysteine-centred redox proteomics further revealed that, although basal oxidation profiles were similar across groups, female CLPP KO kidneys exhibited markedly reduced oxidation of proteins involved in complement activation, fibrinolysis, coagulation, and inflammatory signalling after IRI (Fig. 5L–Supplemental Table 6). Instead, oxidised proteins in female CLPP-deficient kidneys were preferentially associated with cell-matrix adhesion and structural remodelling, suggesting a shift from destructive thrombo-inflammatory signalling toward tissue stabilization and repair.

Taken together, these findings support a model in which CLPP deficiency establishes a kidney-specific preconditioning state whose protective output is strongly shaped by sex. In females, CLPP loss reinforces endogenous antioxidant metabolism, expanding glutathione biosynthesis and sulphur amino acid support before injury occurs. When IRI triggers oxidative stress, this pre-existing metabolic reserve prevents severe antioxidant depletion, limits lipid and protein oxidation, and suppresses complement-coagulation-driven inflammatory amplification. In males, by contrast, CLPP deficiency predominantly exposes stress-induced mitochondrial vulnerability without establishing a comparable antioxidant reserve. Thus, mitochondrial proteostasis, biological sex, and antioxidant metabolism converge to determine renal resilience to ischemic injury.

3. Discussion

This study identifies CLPP, a mitochondrial matrix serine protease, as a sex-specific modulator of renal resilience to IRI. Our findings demonstrate that CLPP deficiency establishes a constitutive preconditioned state in the kidney that protects specifically female mice from the oxidative, inflammatory, and metabolic consequences of IRI. Mechanistically, this state is characterised by NRF2-centred antioxidant remodelling, suppression of lipid peroxidation and downstream inflammatory amplification, and preservation of proximal tubular metabolic integrity. These findings expand the known roles of CLPP beyond mitochondrial protein quality control and metabolic regulation to encompass a broader immunometabolic programme leading to preconditioning with direct consequences for organ injury susceptibility [[12], [13], [14], [15], [16], [17]].

In the kidney, CLPP deficiency reproduced core mitochondrial phenotypes previously described in heart and brown adipose tissue, including accumulation of canonical CLPP substrates, altered CI assembly, and mild activation of the mitochondrial integrated stress response [[12], [13], [14],16]. Importantly, however, respiratory function was only modestly impaired in renal tissue relative to heart, a finding that likely reflects the kidney's remarkable cellular heterogeneity, as it harbours one of the most diverse arrays of specialized cell types in the body, each adapted to distinct oxygen gradients and osmotic environments along the nephron. This metabolic zonation, spanning the oxygen-rich cortex to the hypoxic and hyperosmotic medulla, where anaerobic glycolysis rather than oxidative phosphorylation serves as the principal energy source, is associated with markedly divergent bioenergetic strategies across nephron segments and likely contributes to the kidney's comparatively lower reliance on mitochondrial respiration under physiological conditions [32].

One of the most important mechanistic findings of this study is that CLPP-deficient mitochondria produce significantly less H₂O₂ under succinate-driven RET conditions in both heart and kidney. This observation is particularly relevant to IRI, where ischaemic succinate accumulation followed by rapid oxidation during reperfusion generates a burst of RET-derived ROS at complex I, representing a major upstream driver of oxidative injury [6]. The reduction in RET-driven ROS production in CLPP-deficient mitochondria is likely attributable to altered complex I assembly state associated with N-module accumulation [13]. Functionally, this phenotype resembles previously described CI manipulations that suppress RET-dependent ROS generation while preserving forward electron transport [33]. Since RET-derived ROS is a major trigger of lipid peroxidation and ferroptotic tubular injury during reperfusion [6], these findings provide a direct mechanistic link between CLPP deficiency and the protected renal phenotype observed after IRI. In this respect, the CLPP KO phenotype resembles established preconditioning paradigms, in which chronic subthreshold stress induces durable antioxidant and metabolic adaptations that subsequently confer tolerance to acute injury [27,31,34,35]. Unlike pharmacological antioxidants [[36], [37], [38]], which must be administered acutely, CLPP deficiency establishes this protection constitutively through chronic mitochondrial remodelling and stress adaptation.

Our study suggests that NRF2/NFE2L2 acts as a dominant regulator of the preconditioned state in CLPP-deficient female kidneys. Comparative analysis across tissues demonstrated that CLPP deficiency engages NRF2-centred antioxidant programmes prominently in kidney and heart, whereas brown adipose tissue instead showed activation of immune and antigen-presentation pathways without substantial NRF2 target induction [14,15,19]. This tissue specificity has important implications for therapeutic development, particularly given ongoing efforts to pharmacologically target CLPP in oncology. The temporal organisation of NRF2 activation proved mechanistically critical.

In wild type kidneys, antioxidant genes were induced predominantly after IRI, whereas CLPP-deficient kidneys already displayed increased expression of several NRF2-associated targets at baseline. This pattern is consistent with a preconditioned antioxidant state established before reperfusion injury, in which prior low-level stress generates durable cytoprotective adaptations [27]. Although ATF4-related pathways showed partial engagement, classical ISR signalling did not emerge as the dominant protective mechanism [39]. Rather, the transcriptional and metabolic data point to an NRF2-associated antioxidant programme as a major component of this preconditioned state. Our additional analyses did not allow us to distinguish between the two major canonical routes of NRF2 activation, involving KEAP1-dependent redox regulation or GSK3β-dependent control of NRF2 stability. This may reflect the subtle, chronic nature of the basal response and the cellular heterogeneity of the kidney, where regulatory changes confined to specific cell populations may be difficult to resolve in whole-tissue analyses. Together, these findings support a model in which CLPP deficiency establishes a selective antioxidant preconditioning state that may contribute to protection from lipid peroxidation and ferroptotic injury during reperfusion.

The most striking and biologically important finding of this study is the pronounced female specificity of the protective phenotype, while male CLPP KO mice showed only marginal, if any, benefits across functional, histological, and molecular readouts. Wild-type female kidneys already display a constitutive antioxidant and detoxification programme that is absent in WT males, centred on a lipid peroxidation defence axis. This programme aligns well with recently described mechanisms by which oestradiol signalling inhibits ferroptosis and protects against AKI, establishing a mechanistic continuum between sex hormone signalling, NRF2 activation, and ferroptosis resistance in the kidney [30,40]. The NRF2 programme induced by CLPP deficiency overlaps extensively with this endogenous female antioxidant programme, suggesting that CLPP loss does not establish an entirely novel cytoprotective state, but rather amplifies and extends a pre-existing female-specific protective network. Consistent with this interpretation, CLPP deficiency markedly reduced overall sexual dimorphism at the proteomic level, indicating that CLPP loss partially overrides the physiological sex-specific renal proteome.

In several aspects of the IRI response, CLPP KO males adopted an intermediate phenotype positioned between WT males and WT females, and in some datasets even exceeded the protective state observed in WT females. At the same time, male CLPP deficiency clearly failed to establish the coordinated cytoprotective remodelling observed in females. While CLPP KO females exhibited a markedly attenuated proteomic response to IRI, male KO kidneys showed a substantially stronger injury response, indicating that mitochondrial stress alone is insufficient to redirect the canonical injury programme. Several mitochondrial stress-associated metabolites (e.g. homocitrulline, homoarginine, and 2-hydroxyglutarate) were elevated in both sexes, in some cases more strongly in males, yet failed to confer meaningful protection outside the female context. Together, these findings suggest that the protective effects of CLPP deficiency depend not simply on mitochondrial stress signalling itself, but on its integration into a pre-existing female-specific antioxidant and ferroptosis-resistant state.

Converging evidence from our metabolomic, proteomic, and histological analyses identifies ferroptosis as the principal injury pathway attenuated in CLPP KO female kidneys. Ferroptosis is recognised as a dominant form of tubular cell death in renal IRI, mechanistically linking mitochondrial ROS production, lipid peroxidation, inflammatory amplification, and subsequent fibrotic remodelling [41,42]. The protective programme induced by CLPP deficiency interrupts this cascade at multiple levels simultaneously: sustained glutathione and NADPH availability support continued GPX4-dependent detoxification of phospholipid hydroperoxides, elevated GSTA4 and AKR1C13 enhance clearance of reactive lipid aldehydes, and reduced accumulation of 4-HNE after IRI provides direct in vivo evidence of attenuated lipid peroxidation. Importantly, protection from ferroptosis in renal IRI appears to depend less on baseline antioxidant abundance than on the capacity to maintain glutathione recycling and thiol buffering during reperfusion [42]. In this context, the markedly enhanced glutathione, pentose phosphate pathway, and sulphur amino acid metabolism observed in CLPP KO female kidneys likely establish a metabolic configuration that sustains antioxidant flux under oxidative stress rather than simply increasing resting antioxidant pools. The redox proteomics data further suggest that this improved thiol-buffering environment limits oxidative activation of downstream inflammatory cascades. Proteins involved in complement activation, coagulation, and fibrinolysis showed markedly reduced cysteine oxidation in CLPP KO female kidneys after IRI, consistent with lower ROS burden and reduced propagation of redox-sensitive inflammatory amplification pathways [[41], [42], [43]].

Beyond attenuating the acute destructive phase of IRI, CLPP KO females also mounted a distinct post-injury recovery programme absent in all other groups. Particularly notable was the selective induction of the vitamin D axis and HMGCS2-dependent ketogenesis after reperfusion, both of which have recently emerged as active reno-protective pathways in AKI [25,44]. Together with preservation of proximal tubular identity markers, these findings suggest that CLPP deficiency in females not only limits ferroptotic injury during reperfusion, but also preserves the metabolic and transcriptional integrity required for efficient tubular recovery.

Finally, the present findings position CLPP as a potentially relevant therapeutic target in renal IRI and related disorders characterised by pathological mitochondrial ROS generation. While pharmacological activation of CLPP has primarily been explored in oncology as a strategy to induce catastrophic mitochondrial proteotoxic stress in tumour cells [45], our data suggest that inhibition of CLPP activity may conversely establish a protective preconditioned state in the kidney by attenuating RET-driven ROS production and enhancing endogenous antioxidant defences. Recent findings linking elevated CLPP expression to kidney injury severity in diabetic kidney disease further support the idea that CLPP-dependent mitochondrial proteostasis contributes broadly to renal vulnerability under metabolic stress conditions [46]. Although CLPP loss-of-function in humans causes Perrault syndrome and CLPP-deficient mice recapitulate key features of the human disorder, whether the sex-dependent renal protection identified here is conserved in humans will require further investigation.

The pronounced sex dependence of the protective phenotype represents both a mechanistic insight and an important translational consideration. Our findings suggest that the beneficial effects of CLPP inhibition depend on integration into a pre-existing NRF2- and ferroptosis-resistant transcriptional environment, which is substantially more developed in female kidneys. Whether pharmacological NRF2 activation or ferroptosis-targeted therapies could extend this protective window to males remains an important question for future studies.

Collectively, our data support a model in which chronic mitochondrial proteostatic stress induced by CLPP deficiency establishes a constitutive NRF2-centred preconditioned state that selectively synergises with endogenous female antioxidant programmes to suppress ferroptotic injury during renal reperfusion. By linking mitochondrial proteostasis, RET-derived ROS production, ferroptosis resistance, and sex-dependent renal resilience, this work identifies the CLPP-RET-NRF2 axis as a potentially targetable pathway for limiting acute kidney injury and improving recovery after ischaemic stress.

4. Material and methods

4.1. Animal handling and experiments

All mice in this study were bred and raised in the CECAD in vivo research facility. Animals were housed in groups in ventilated cages at 22-24 °C in 12h light/dark cycle with ad libitum access to drinking water and normal chow diet. Mice were killed by cervical dislocation at the age of 12-17 weeks.

The mice within this study were derived from a conditional Clpp gene targeting, prior described in Ref. [12]. Clpp deficient (−/−) mice were obtained from intercrosses of Clpp±mice on a C56BL/6 TacN background. Mice were genotyped using the following primers (5' - 3′): (Clpp fw - GTGGATGATG GTCAGTAGAATCC; Clpp rev - CCCAGACATGATTCCTAGCAC; Clpp ko - TGTGCA TTCTTACCATAGTCTGC). Our study examined male and female animals, and sex-dimorphic effects are reported. Experiments were approved and conducted in compliance with the guidelines and approval by the local governmental authorities (Landesamt für Natur, Umwelt und Verbraucherschutz Nordrhein-Westfalen; LANUV).

4.2. Ischemia-reperfusion injury (IRI)

Twelve-to-fifteen weeks old WT or CLPP KO mice were used for the renal IRI surgery. Warm renal IRI model was performed on temperature-controlled heating pad (Harvard apparatus, Holliston, MA) to keep the body temperature at 32 °C. Anaesthesia by ketamine (100 mg/kg) and xylazine (10 mg/kg) was applied via intraperitoneal injection. Afterwards, the right kidney was removed and the left renal artery was clamped for 40 min. The left kidney was harvested either 30 min or 24 h post-ischemia.

4.3. RNA isolation and sequencing

Total RNA was isolated from kidney tissue using TRIzol reagent according to the manufacturer's instructions. Approximately 50 μg tissue was homogenized in 1 mL TRIzol with ceramic beads using a FastPrep homogenizer. RNA pellets were resuspended in RNase/DNase-free H₂O and further purified using the RNA Clean & Concentrator Kit, including DNase I digestion. RNA sequencing was performed in-house using 1 μg total RNA for library preparation with the NEBNext RNA Prep Kit, followed by paired-end 2 × 50 bp sequencing on an Illumina NovaSeq 6000, as previously described [15]. Gene Ontology analyses were performed using DAVID v2024q4 by over-representation analysis and Gene Set Enrichment Analysis [47,48].

4.4. Protein isolation and Western blot analyses

Kidney tissue (∼50 mg) was homogenized in ice-cold organ lysis buffer using ceramic beads and a FastPrep homogenizer. Lysates were cleared by centrifugation at 20,000×g for 45 min at 4 °C, and supernatants were collected [15]. Protein lysates (30–50 μg) were mixed with Laemmli buffer, separated on 8–15% polyacrylamide gels, and transferred to nitrocellulose membranes. Transfer efficiency was assessed by Ponceau S staining. Membranes were blocked in milk-, BSA-, or fish gelatin-based blocking buffer, incubated with primary antibodies overnight at 4 °C, followed by HRP-conjugated secondary antibodies for 1h. Signals were detected using ECL western blotting substrate.

Antibodies used in western blot and BN-PAGE analyses

Name	Supplier	Cat Nr.
ACSL4	Santa Cruz	Sc271800
ALDH18A1	Sigma	HPA01260
ATP5A1	Abcam	ab14748
BID	Santa Cruz	Sc373939
BCL2	Santa Cruz	Sc492
BCL-XL	Cell Signaling	2764
CASPASE 3 (cleaved)	Cell Signaling	9661, 9662
CASPASE 9	Cell Signaling	9502
CLPP	Sigma	WH0008192M1
CLPX	Sigma	HPA040262
MT-CO1	Molecular Probes	459600
COX4I1	Molecular Probes	A21348
ERAL1	Proteintech	29465-1-AP
EFG1	Abcam	ab173529
EIF2a	Abcam	ab26197
p-EIF2a	Abcam	ab32157
FSP1	Proteintech	20886-1-AP
GCLC	Proteintech	12601-1-PBS
GPX4	Abcam	ab125066
GSK3B	Cell signaling	9832
Phospho-GSK3B	Cell signaling	9323
GSTA1	Proteintech	14475-1-AP
HSC70	Santa Cruz	sc-7298
KEAP1	Abcam	Ab119403
Lysine-Ac	Abcam	AB21623
MLKL-phospho	Abcam	Ab196436
MRPS35	Proteintech	16457-1-AP
MTHFD2	Abcam	ab56772
NDUFA9	Molecular Probes	459100
NDUFS2	Abcam	ab96160
NDUFV2	Proteintech	15301-1-AP
NRF2	Abcam	ab31184
NQO1	Proteintech	67240-1-Ig
OAT	Proteintech	ab137679
RIPK1	Cell signaling	4926
RIPK1-phospho	Cell signaling	31122
RIPK3	Cell signaling	10188
RPS6	Cell signaling	2217
pRP-S6	Cell signaling	2211
SDHA	Molecular Probes	459200
SHMT2	Sigma Aldrich	AV46129
SIRT3	Abcam	Ab86671
SQSTM1	Abcam	Ab91526
TRAP1	BD Bioscience	612344
Ubiquitin	Cell Signaling	3936
UQCRC1	Molecular Probes	459140
HRP linked anti-mouse IgG	Sigma	A 4416
HRP linked anti-rabbit IgG	Sigma	A 6154

4.5. Isolation of mitochondria from kidney

Freshly dissected tissues were placed in mitochondrial isolation buffer (100 mM Sucrose, 50 mM KCl, 1 mM EDTA, 20 mM TES, 0.2% fatty acid-free BSA, pH7.2) and homogenized using a Potter S homogenizer at 1000 rpm for 15 strokes. Homogenates were centrifuged at 8,500g for 5 min at 4 °C, and the pellet was resuspended and cleared by centrifugation at 800g for 5 min. The mitochondria-containing supernatant was collected, and mitochondria were pelleted at 8,500g for 5 min at 4 °C and resuspended in mitochondrial isolation buffer without BSA.

4.6. Blue native polyacrylamide electrophoresis (BN-PAGE)

BN-PAGE was performed using the NativePAGE Novex Bis-Tris Gel System according to the manufacturer's instructions. Mitochondria (10–30 μg) were solubilised with either 4% digitonin to preserve supercomplexes or 1% n-dodecyl-β-D-maltoside (DDM) to resolve individual complexes, and separated on 4–16% native acrylamide gels. Proteins were transferred to PVDF membranes and immunodecorated with the indicated antibodies.

For in-gel activity assays, BN-PAGE gels were briefly washed and incubated either in complex I activity buffer for 1h at 37 °C or in complex IV activity buffer overnight. Gels were destained for 2h at room temperature in 50% methanol/10% acetic acid, rehydrated in ddH₂O, and documented using a V800 transparency scanner.

4.7. Inverse redox shift assay

The inverse redox-shift assay was adapted from a previously published protocol [13]. Approximately 5 mg of tissue was used per sample. For steady-state samples, tissue was incubated in ice-cold PBS containing 20 mM N-ethylmaleimide (NEM) for 30 min on ice to block reduced thiols. Samples were then homogenized with ceramic beads using a FastPrep homogenizer. Lysates were cleared by centrifugation at 20,000 × g for 45 min at 4 °C, and proteins in the supernatant were precipitated overnight in ice-cold 8% trichloroacetic acid (TCA). Protein pellets were collected by centrifugation at 20,000 × g for 15 min at 4 °C, washed with ice-cold 5% TCA, and resuspended by sonication in 1× Laemmli buffer containing 10 mM tris(2-carboxyethyl)phosphine (TCEP). Samples were incubated for 15 min at 45 °C with vigorous agitation and subsequently cooled to room temperature. Minimum-shift samples were modified with 15 mM NEM, whereas maximum-shift and steady-state samples were modified with 15 mM mm(PEG)24. Samples were incubated for 1 h at room temperature in the dark and subsequently analysed by SDS–PAGE and immunoblotting.

4.8. H₂O₂ measurement using Amplex Red via Oxygraph

Oxygen consumption with parallel fluorescent detection measurement of H₂O₂ production was measured at 37 °C, using 50 or 100 μg of heart and kidney mitochondria diluted in 2.1 ml of respiratory buffer (120 mM Sucrose, 50 mM KCl, 20 mM Tris-HCl, 4 mM KH₂PO₄, 2 mM MgCl₂, 1 mM EGTA, 0.25 mg/ml BSA; pH7.1) in an Oxygraph-2k (OROBOROS INSTRUMENTS, Innsbruck, Austria). To study RET-initiated ROS production SUIT-26 protocol from OROBOROS was followed. After calibration with H₂O₂, mitochondria Amplex Red, SOD (used at 5U/ml, Sigma Aldrich) and horse radish peroxidase (final concentration 1U/ml, Sigma Aldrich) were added to the cuvette. Maximal RET was induced by succinate addition and quantified as the ratio between the signal after rotenone addition and the signal obtained with succinate alone.

4.9. Periodic acid-Schiff (PAS) staining

Kidneys were fixed in 4% formaldehyde, paraffin-embedded, and cut into 2 μm sections. Periodic acid–Schiff staining was performed using standard protocols. Acute tubular injury was assessed by an experienced nephropathologist using a scoring system based on the percentage of tubules showing vacuolization, epithelial flattening, brush border loss, nuclear loss, and necrosis [23].

4.10. Transmission electron microscopy (TEM) image acquisition

Freshly isolated mouse kidney pieces (∼1 mm³) were fixed in 2% glutaraldehyde/2% formaldehyde in 0.1 M cacodylate buffer, pH 7.3, for 48 h at 4 °C in the dark. Samples were processed by the in-hose Imaging Facility, post-fixed in 2% osmium tetroxide, dehydrated through an ethanol series, transferred to propylene oxide, and embedded in epoxy resin. Ultrathin sections (70 nm) were cut using an EM-UC6 ultramicrotome, contrasted with uranyl acetate and lead citrate, and imaged at room temperature on a JEM 2100 Plus transmission electron microscope equipped with a OneView 4K camera using DigitalMicrograph software at 80 kV.

4.11. Immunohistochemistry

After deparaffinization, heat-induced antigen retrieval was performed in citrate buffer (pH 6.0). Endogenous peroxidase activity and nonspecific binding were blocked using BLOXALL and animal-free blocker, respectively. Sections were incubated overnight with primary antibodies against CD45, F4/80, or 4-HNE, followed by ImmPRESS HRP Goat Anti-Rat IgG Polymer and ImmPACT DAB detection. Samples were counterstained with hematoxylin, mounted in HIGHDEF IHC mount, and scanned at 20× magnification using an SCN4000 slide scanner. Histological sections were visualized with NDP Viewer.

Antibodies used in Immunohistochemistry staining

Antibody	Supplier	Cat Nr.
CD45	eBioscience	14-0451-082
F4/80	BioRad	MCA497R
4-HNE	Abcam	Ab46545

4.12. TUNEL staining

TUNEL staining was performed using the DeadEnd Fluorometric TUNEL System according to the manufacturer's instructions. Images were acquired on a Zeiss AxioObserver microscope equipped with an AxioCam MR 702 camera, using a Plan-Apochromat 20×/0.8 objective and ZEN 2.6 software. TUNEL-positive cells were quantified using Fiji/ImageJ.

4.13. Whole proteome analyses

For label-free proteomics, kidney tissue (30–50 mg) was homogenized in 8 M urea buffer containing protease inhibitors using ceramic beads and a FastPrep homogenizer. Lysates were cleared by centrifugation, and protein concentration was determined by Bradford assay.

For redox proteomics, kidney tissue (30–50 mg) was homogenized in 8 M urea buffer containing protease inhibitors and 100 mM NEM. Lysates were cleared by centrifugation, incubated for 30 min at 50 °C, and proteins were precipitated with 20% TCA. Pellets were washed with 10% TCA and ice-cold acetone to remove excess NEM, resuspended in 8 M urea containing 5.5 mM TCEP, and incubated for 30 min at 50 °C in the dark to reduce disulfide bonds. Protein concentration was then determined by Bradford assay.

For both workflows, 50 μg protein per sample was processed for in-solution digestion. Proteins were alkylated with chloroacetamide, digested with Lys-C, diluted to 2 M urea with 50 mM TEAB, and digested overnight with trypsin. Digestion was stopped with 1% formic acid, and peptides were loaded onto SDB-RPS StageTips before submission to the in-house Proteomics Facility for analysis as previously described [13].

4.14. RNA sequencing

1 μg of purified RNA validated on a tape station (Agilent) was submitted to Cologne Centre for Genomics for further processing. Libraries were prepared using the Stranded mRNA Preparation Kit (Illumina) starting with 500 ng total RNA. After poly-A selection (poly-T oligo-attached magnetic beads), mRNA was purified and fragmented using divalent cations under elevated temperature. The RNA fragments underwent reverse transcription using random primers. This was followed by second-strand cDNA synthesis with DNA Polymerase I and RNase H. After end repair and A-tailing, indexing adapters were ligated. Purified and amplified products (12 PCR cycles) were used to create the final cDNA libraries. After library validation and quantification (Agilent Tape Station), equimolar amounts of the library were pooled. The pools were quantified using the Peqlab KAPA Library Quantification Kit and the Applied Biosystems 7900HT Sequence Detection System. Sequencing was performed on an Illumina NovaSeq 6000 sequencing instrument with a PE100 protocol. Data were analysed using R packages DESeq2 for analysis, dplyr for data wrangling and ggplot2 for visualization.

4.15. Metabolite extraction and LC-MS analysis

Frozen kidney tissue (10–30 mg) was transferred to pre-cooled homogenization tubes containing ceramic beads. Metabolites were extracted using methanol:acetonitrile (50:30:20, LC-MS grade) supplemented with 0.5 μM valine-d8 as internal standard, using 25 μL extraction solvent per mg tissue. Samples were homogenized using a Precellys 24 tissue homogenizer, shaken for 15 min at 4 °C, and centrifuged at 13,000 rpm for 30–40 min at 4 °C. The upper 80% of the supernatant was transferred to fresh tubes for LC-MS analysis.

Metabolites were separated on a Vanquish LC system using a SeQuant ZIC-pHILIC column and a gradient of 20 mM ammonium carbonate with 0.05% ammonium hydroxide and InfinityLab Deactivator additive as mobile phase A, and 95% acetonitrile/5% water as mobile phase B. Chromatography was performed at 0.20 mL/min with the column maintained at 40 °C, autosampler at 4 °C, and an injection volume of 5 μL. The LC system was coupled to an Orbitrap Exploris 240 mass spectrometer operated in polarity-switching MS1 mode with H-ESI ionization, a scan range of 60–900 Da, and 120,000 resolution.

Pooled quality-control samples were prepared by combining 5 μL from each extract and were analysed repeatedly throughout the run together with blanks. Samples were injected in randomized order, with QC pools and blanks included every five injections. Instrument stability was monitored by seven consecutive QC injections at the beginning of the run, requiring a coefficient of variation below 10% for the valine-d8 internal standard. Pooled samples were additionally analysed by MS2 in positive and negative ionization modes for metabolite annotation based on fragmentation patterns and comparison with an in-house library.

4.16. Hormone detection using mass-spec

Serum samples were analysed using an adapted version of a validated UHPLC-HRMS/MS method [49]. Briefly, for testosterone, 100 μL of serum was diluted with 200 μL of acetic acid (1%). Subsequently, 5 μL of the internal standard solution (20 ng/mL in MeOH) was added. After vortexing, the samples were equilibrated for 5 min at room temperature, and 400 μL of acetonitrile was added to the sample and vortexed before the sample was shaken well in a ThermoMixer (Eppendorf AG, Hamburg, Germany) at room temperature at 600 rpm for 10 min. Finally, the sample was centrifuged at 17,000 × g for 7 min, and 150 μL of the supernatant was transferred into a glass injection vial. A volume of 10 μL was injected into the LC-HRMS/MS system for analysis.

For 17β-estradiol (E2) quantification, an adapted version of the protocol described by Keski-Rahkonen (2015) was applied [50]. Briefly, 50 μL of the internal standard solution (0.2 ng/mL in MeOH) was added to 200 μL of serum. The sample was equilibrated for 15 min at 4 °C. For phase separation, 1 mL of TBME was added to the sample and centrifuged. Subsequently, the upper phase was transferred into a fresh 1.5 mL Eppendorf tube and evaporated to dryness using a vacuum centrifuge at 45 °C. Afterwards, derivatization was performed by reconstituting the dry residue with 20 μL of DMIS (1 mg/mL in acetone) and 30 μL of NaHCO3 (50 mM in water) in a ThermoMixer (Eppendorf AG, Hamburg, Germany) at 65 ◦C and 600 rpm for 15 min. Finally, 30 μL was injected into the LC-HRMS/MS system. External serum reference calibrators and quality controls (Chromsystems, Munich, Germany) were used for quantification. The calibration curves for testosterone and E2 ranged from 0,055 to 11,6 ng/mL and from 0,025 to 5,24 ng/mL, respectively.

4.17. Statistical analyses

Statistical significance of Tables with two independent groups was analysed for using two-tailed unpaired Student's t-test. Tables with more than two groups were analysed using one-way analysis of variance (ANOVA) followed by Tukey's post hoc test. All p-values below 0.05 were considered statistically significant; *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. All data are presented as mean ± SD.

5. Software

Statistical analysis and data visualization were performed using Microsoft Excel, Graphpad Prism 6 (Graphpad Software), Enrichr-Webtool, DAVID Bioinformatics Resources [47] and GSEA_4.4.0 [48]. Adobe Illustrator 2025 and Adobe Photoshop 2025 were utilized for figure making. Text editing as well as literature research was supplemented using ChatGPT and Perplexity AI. Endnote20 was used as reference editor software.

CRediT authorship contribution statement

Lea Isermann: Conceptualization, Investigation, Supervision, Validation, Visualization, Writing – original draft, Writing – review & editing. Pedro Rojas-Morales: Conceptualization, Investigation, Validation, Formal analysis. Katharina Senft: Formal analysis, Investigation. Nidhi Nair: Formal analysis, Investigation, Validation. Linda Baumann: Investigation, Formal analysis. Marijana Croon: Formal analysis, Conceptualization, Investigation. Viktoria Katharina Lauterbach: Formal analysis, Investigation. Heike Göbel: Formal analysis, Investigation. Francesco Landini: Formal analysis, Investigation. Chiara Tuma: Investigation, Formal analysis. Ignacio Sottorff-Neculhueque: Investigation, Methodology, Formal analysis. Mario Thevis: Supervision, Methodology, Conceptualization. Michael Ignarski: Formal analysis, Investigation, Supervision. Sebastian Brähler: Conceptualization, Methodology, Resources. Jan Riemer: Supervision, Methodology. Roman-Ulrich Müller: Conceptualization, Supervision, Methodology. Aleksandra Trifunovic: Supervision, Formal analysis, Funding acquisition, Conceptualization, Validation, Visualization, Writing – original draft, Writing – review & editing.

Declaration of generative AI and AI-assisted technologies in the manuscript preparation process

During the preparation of this work the author(s) used Claude in order to generate the dot-plots in the figures. Additionally, Perplexity AI was used to support literature search and ChatGPT for text editing. After using these tools/services, the authors reviewed and edited the content as needed and take full responsibility for the content of the published article.

Funding support

Research reported in this publication was supported by funding to A. T. from the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation) SFB1218 (Project No. 269925409) and TR1018/12-1. L.I was funded through a DFG-funded CCRC Graduate Program (GRK 2407).

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A. Supplementary data

The following are the Supplementary data to this article:

Table 1

Proteomics analysis of CLPP KO and WT females at baseline and, 30 min and 24 h post IRI. Includes also GOBP and KEGG analyses,

Table 2

RNAseq analysis of CLPP KO and WT females, baseline and 24h post IRI. Includes GOBP analyses,

Table 3

Proteomics analysis of heart of CLPP KO and WT female and males at baseline Includes GOBP analyses.

Table 4

Proteomics analysis of CLPP KO and WT female and males at baseline and 24h post IRI. Includes GOBP analyses,

Table 5

Metabolomics analysis of CLPP KO and WT female and males at baseline and 24h post IRI.

Table 6

Redox-proteome analysis of CLPP KO and WT female and males at baseline and 24h post IRI.