Gender-Specific Gene Regulation of Ferroptosis in Non-Utilized Liver Donors

TL;DR

This study explores how sex influences gene regulation of ferroptosis in non-used donor livers, finding no major sex-based differences in most genes, except for a potential trend in one gene in female livers.

Contribution

The study is the first to systematically investigate sex-based differences in ferroptosis gene regulation in non-utilized human donor livers using ex vivo normothermic machine perfusion.

Findings

Normothermic machine perfusion of non-utilized donor livers was feasible.

No significant sex-based differences were found in MDA levels or most ferroptosis-related gene expression after 6 hours.

ACSF2 expression was higher in female livers, but not significant after multiple testing correction.

Abstract

Background/Objectives: Females are generally more resistant to ischemia-related ferroptosis than males, due to differences in iron metabolism, antioxidant pathways, and sex hormone-mediated regulation of ferroptosis suppressors. This has not been systematically studied in a human donor liver model. To investigate the effect of sex on ferroptosis and oxidative stress pathways in non-utilized donor livers (NDLs), we assessed patterns of gene expression in NDLs under ex vivo normothermic machine perfusion (NMP). Methods: We utilized the PROTECT dual-circuit ex vivo NMP system to assess three male and two female NDLs undergoing 6 h NMP. Perfusate and tissue samples were collected at baseline and 6 h of NMP. Malondialdehyde (MDA) levels were quantified as biochemical markers of iron overload and lipid peroxidation, respectively. Ferroptosis-related gene expression was assessed using molecular assays. Comparisons between male and female NDLs were used to determine the influence of sex on ferroptosis and oxidative injury during NMP. Results: NMP was successfully performed on NDLs (n = 5) from three male (56.3 ± 5.7 years) and two female donors (46.5 ± 0.7 years, p = 0.15). The fold-change in the oxidative stress marker MDA was comparable between female (1.2 ± 0.6) and male (1.0 ± 0.4) NDLs after 6 h NMP (p = 0.76). All livers showed upregulation of ferroptosis-related genes (Hypoxia-inducible factor 1 alpha, Iron Responsive Binding Elements 2, Ribosomal Protein L8, Ferritin Heavy Chain 1, Acyl-CoA synthetase family member 2, ATP synthase membrane subunit c locus 3, Heme-oxygenase 1, NAD(P)H Quinone Dehydrogenase 1, Tetratricopeptide Repeat Domain 35, Nuclear Factor Erythroid 2 Related Factor 2). ACSF2 expression was significantly higher in female NDLs compared with males undergoing 6 h NMP (3.6 ± 3.0 vs. 1.0 ± 0.7-fold change, p = 0.04). There were no sex-based significant differences observed in the expression of other ferroptosis-related genes (HIF-1α, IREB2, RPL8, FTH-1, ATP5G3, HO-1, NQO1, TTC35, and NRF2) between male and female NDLs. No gene reached statistical significance after false-discovery-rate (FDR) correction. Conclusions: Normothermic machine perfusion of NDLs was feasible, and no sex-related differences were observed in MDA levels or most ferroptosis-related gene expression after 6 h. Although ACSF2 showed higher expression in female livers, this was not significant after multiple testing correction, highlighting the need for larger studies to explore sex-dependent ferroptosis signaling during liver preservation.

Figures

Full text

1. Introduction

1.1. Physiological Interplay Between Sex, Ferroptosis, and Hepatic Ischemia–Reperfusion Injury (IRI)

Sex significantly influences susceptibility to ferroptosis and ischemia–reperfusion injury (IRI) in transplant organs, with male organs generally more vulnerable than females, most notably in hepatic IRI [1]. This sex-based dimorphism is driven by higher male expression of the denticleless E3 ubiquitin protein ligase (DTL), which regulates the degradation of Prospero homeobox 1 (PROX1), a transcription factor involved in organ development [1]. This degradation leads to increased polyunsaturated fatty acid (PUFA) levels and enhanced ferroptosis in hepatocytes, resulting in more severe liver injury [1]. Evidence for sex-specific effects is strongest in the liver, where the DTL-PROX1 axis appears central. In contrast, the current kidney and lung transplantation literature does not clearly delineate sex-based differences in ferroptosis or IRI, although ferroptosis broadly contributes to graft injury in these organs [2,3].

1.2. Sex-Based Differences in Vulnerability to Ferroptosis in Hepatic IRI

In hepatic IRI, males demonstrate higher expression of DTL, resulting in enhanced degradation of PROX1, increased polyunsaturated fatty acid levels, and an exacerbation of ferroptosis. In contrast, female livers show lower DTL expression, conferring relative resistance to ferroptosis and IRI [1]. Additionally, male hepatocytes have higher mitochondrial iron and reactive oxygen species (ROS) accumulation, with increased expression of iron importers (TfR1, Mfrn1) and lower ferritin (FTH1), further predisposing them to ferroptosis. Female hepatocytes exhibit enhanced iron storage and reduced iron import, contributing to ferroptosis resistance [3]. In the kidney, female sex confers protection against ferroptosis in proximal tubular cells following ischemic injury, mediated in part by upregulation of the NRF2 antioxidant pathway. Pharmacological activation of NRF2 in males can recapitulate the female-resistant phenotype [4]. Furthermore, estradiol and its metabolites directly inhibit ferroptosis through both non-genomic (radical trapping antioxidant activity) and genomic mechanisms (ESR1-mediated suppression of pro-ferroptotic pathways), explaining the reduced sensitivity of female kidneys to ischemic injury [5]. Sex hormones regulate ferroptosis surveillance mechanisms: estrogen receptor activation upregulates anti-ferroptotic enzymes (e.g., MBOAT1), while androgen receptor signaling modulates distinct ferroptosis suppressors (e.g., MBOAT2), further contributing to sex differences in ferroptosis susceptibility [6].

1.3. Ferroptosis, IRI, and Implications in Liver Transplantation

Marginal donor livers (MDLs), including those with steatosis or procured after circulatory death, are increasingly utilized to address the organ shortage crisis. However, MDLs are highly vulnerable to IRI, a key driver of early allograft dysfunction. Ferroptosis, an iron-dependent regulated form of cell death, has emerged as a central mechanism underlying IRI. Biological sex significantly influences ischemia-related ferroptosis, with males generally exhibiting greater susceptibility to ferroptotic cell death in ischemic injury compared to females. This sexual dimorphism has been demonstrated in multiple organ systems, including the liver and kidney [1,2,3,4,5,6].

In summary, females are generally more resistant to ischemia-related ferroptosis than males, due to differences in iron metabolism, antioxidant pathways, and sex hormone-mediated regulation of ferroptosis suppressors [1,2,3,4,5,6]. This has not been systematically studied in human donor livers. Considering this, we investigated the effects of sex on ferroptosis and oxidative stress pathways by leveraging ex vivo normothermic machine (NMP) perfusion in non-utilized donor livers (NDLs). This preliminary study seeks to further the understanding of sex-specific influences on liver allograft outcomes and enhance transplantation strategies.

2. Materials and Methods

2.1. Donors and Livers

Following Institutional Review Board and Institutional Biosafety Committee approval (No. 2018-00040), human NDLs were procured with informed donor consent for research purposes. All livers were deemed unsuitable for transplantation. The study complied with the Declarations of Helsinki and Istanbul. Anonymized donor information was provided by the Mid-America Transplant Center.

2.2. Normothermic Machine Perfusion of Donor Livers

Each liver was surgically split into right and left lobes and cannulated according to Couinaud and Bismuth classifications, as previously described [7]. The portal vein and hepatic artery of each lobe were cannulated and perfused using two independent circuits within the PROTECT system [7] (Figure 1). After successful cannulation, perfusion parameters were stabilized. Mean arterial pressure (MAP) was maintained between 65 and 80 mmHg, and portal vein pressure remained steady between 10 and 15 mmHg. Whole blood flow was adjusted to 0.8–1.2 L/min, with an arterial-to-portal flow ratio of 1:4. Baseline and endpoint (6 h) samples were collected from each lobe.

2.3. Hematoxylin and Eosin (H&E) Staining

Liver biopsies were fixed in 10% buffered formalin, embedded in paraffin, and stained using the Scytec H&E kit (NC0510871, Fisher, Waltham, MA, USA). A board-certified pathologist, blinded to group assignment, assessed the samples for steatosis, inflammation, and injury.

2.4. RNA Extraction and Real-Time PCR Analysis

RNA extraction was performed using TRIzol™ (Invitrogen, Carlsbad, CA, USA) at Saint Louis University. cDNA was synthesized using the Verso cDNA Synthesis Kit (Thermo Fisher, Vilnius, Lithuania). Primers for ferroptosis-related genes (HO-1, HIF-1α, RPL8, IERB2, ACSF2, ATP5G3, FTH1, NQO1, TTC35, NRF2) were designed via Integrated DNA Technologies. Real-time PCR was performed using the CFX Connect Real-Time Detection System (Bio-Rad, Hercules, CA, USA) and iTaq™ Universal SYBR^®^ Green Supermix (Bio-Rad, Hercules, CA, USA). Each assay was performed in triplicate. Gene expression analysis was compared at baseline and after 6 h of NMP to evaluate the effect of sex.

2.5. Assessment of Lipid Peroxidation (MDA Assay)

Lipid peroxidation was evaluated using the Thiobarbituric Acid Reactive Substances (TBARS) assay kit (MAK085, Sigma-Aldrich, St. Louis, MO, USA). Malondialdehyde (MDA) is a stable end product of lipid peroxidation that serves as one of the most widely used biomarkers for oxidative stress in these tissue samples. Tissues were homogenized in lysis buffer with butylated hydroxytoluene (BHT), centrifuged, and the supernatant was subsequently incubated with TBA at 95 °C for 60 min. Following centrifugation, absorbance at 532 nm was measured using a Synergy 2 Microplate Reader (Biotek Instruments, Winooski, VT, USA).

2.6. Statistical Analysis

All statistical analyses were performed using R (version 4.4.3, R Foundation for Statistical Computing, Vienna, Austria). Continuous variables (age, body mass index (BMI), cold ischemia time, and fold changes in gene expression and MDA level) were summarized as mean, standard deviation, median, and interquartile range. Because of the small sample size (5 livers: 2 female and 3 male), non-parametric methods were used for group comparisons. For each gene (HO-1, HIF-1α, RPL8, IREB2, ACSF2, ATP5G3, FTH1, NQO1, TTC35, NRF2) and for MDA fold change, we compared fold-change values between male and female livers using the Wilcoxon rank-sum test. All analyses were performed at the biological level, with each donor liver representing one independent observation. Comparisons between female and male livers used the Wilcoxon rank-sum test applied to per-donor median values, yielding an effective sample size of N = 2 female and N = 3 male livers. All tests were two-sided. To account for multiple testing across markers, p-values were adjusted using the false discovery rate (FDR) method. A p-value < 0.05 was considered statistically significant in the primary (unadjusted) analysis, and FDR-adjusted p-values were reported to aid interpretation in this exploratory study. Descriptive plots (boxplots with overlaid data points) were generated to visualize the distribution of fold-change values by sex for each marker.

3. Results

3.1. Donor and Liver Characteristics

All five livers (n = 5) were procured from donation after circulatory death (DCD), a group associated with increased susceptibility to IRI. The mean donor age for 3 male and 2 female donors is 56.3 ± 5.7 years and 46.5 ± 0.7 years, respectively (p = 0.15). Mean donor BMI is 31.6 kg/m^2^ (range: 23.8–36.2), reflecting a cohort with mild to moderate obesity. BMI and cold ischemia time were comparable between sex groups (p = 0.77 and p = 0.14, respectively). Computed tomography (CT) prior to procurement identified macrovesicular steatosis in liver A, while no other significant abnormalities were detected. The average cold ischemic time (CIT), defined as the interval between cross-clamping and the initiation of perfusion, was 5.8 h (range: 3–9), within acceptable ranges for ex vivo perfusion protocols. Characteristics of the donor population and non-utilized donor liver grafts are shown in Table 1.

3.2. Histological Assessment via H&E Staining

To establish baseline histological features and evaluate early injury, Hematoxylin and Eosin (H&E) staining was performed on pre- and post-perfusion liver samples (Figure 2). Liver A exhibited severe (>30%) macro vesicular steatosis, a known risk factor for perfusion-related injury and impaired recovery. In contrast, livers B and D showed only mild to moderate steatosis. Inflammation was typically mild and localized to zone 3 (centrilobular region) across all samples, consistent with early post-reperfusion changes. No significant hepatocellular necrosis or architectural disruption was observed, likely due to the relatively short duration of normothermic perfusion. These findings provided a foundational context for subsequent analyses of iron metabolism and ferroptosis signaling.

3.3. Ferroptosis-Related Gene Expression Patterns

Transcriptional responses to iron chelation were quantified via mRNA expression patterns of key ferroptosis-associated genes, including HO-1, HIF-1α, NQO1, FTH1, RPL8, IREB2, ATP5G3, ACSF2, TTC35, and NRF2 (Table 2). At baseline, there were no significant sex-based differences in the expression of ferroptosis-related genes, including HO-1, HIF-1α, RPL8, IREB2, ACSF2, ATP5G3, FTH1, NQO1, TTC35, and NRF2. After 6 h of NMP, ferroptosis-related genes (HIF-1α, IREB2, RPL8, FTH1, ACSF2, ATP5G3, HO-1, NQO1, TTC35, and NRF2) were upregulated across all livers (Figure 3). ACSF2 expression was significantly higher in female NDLs compared with males undergoing 6 h of NMP (3.6 ± 3.0 vs. 1.0 ± 0.7 fold-change, p = 0.04). Although not statistically significant, ATP5G3 (female: 1.8 ± 1.1 vs. male: 1.0 ± 0.7) and FTH-1 (female: 1.3 ± 0.7 vs. male: 2.3 ± 0.6) exhibited sex-related differences, with both comparisons trending toward significance (p = 0.09 for each). There were no sex-based significant differences observed in the expression of other ferroptosis-related genes (HIF-1α, IREB2, RPL8, HO-1, NQO1, TTC35, and NRF2) between male and female NDLs after 6 h of NMP. After correction for multiple testing using the FDR method, none of the observed differences remained statistically significant.

3.4. Lipid Peroxidation (MDA Assay)

Prior to perfusion, analysis of MDA levels at baseline showed no significant differences between male and female donor livers. Following 6 h of NMP, the fold-change in the oxidative stress marker MDA was comparable between female (1.2 ± 0.6) and male (1.0 ± 0.4) NDLs (p = 0.77) (Figure 4).

4. Discussion

4.1. Sex-Specific Differences in Liver Ferroptosis and IRI

In this exploratory cohort, we did not observe statistically robust sex-related differences in MDA levels or ferroptosis-related gene expression fold changes after 6 h of NMP; however, ACSF2 and, to a lesser extent, ATP5G3 and FTH1, emerged as potential candidates for further investigation in larger studies.

Sex differences in susceptibility to ferroptosis and IRI have been most clearly demonstrated in hepatic models, where males exhibit greater vulnerability than females. This sexual dimorphism has been linked to differential expression of the denticleless E3 ubiquitin protein ligase (DTL), which promotes PROX1 degradation, increases PUFA levels, and enhances ferroptosis in hepatocytes, resulting in more severe liver injury in males [1].

Although no unified pathway directly links acyl-CoA synthetase family member 2 (ACSF2), DTL, and PROX1, ACSF2 plays a direct role in PUFA metabolism by enhancing PUFA uptake and activation, suggesting a potential intersection with ferroptosis-related lipid homeostasis. In contrast, while DTL regulates protein degradation via ubiquitination, there is currently no evidence that it directly modulates ACSF2 or PUFA levels. Evidence for sex-specific ferroptosis mechanisms in kidney and lung transplantation remains limited, despite established roles for ferroptosis in graft injury across organs [3]. An overview of prior studies examining sex differences in hepatic IRI is provided in Table 3.

4.2. Understanding Interplay Between Sex-Specific Genetic Factors and Hormonal Signaling

Sex is a significant biological determinant of susceptibility to hepatic IRI and ferroptosis, arising from interactions among hormonal signaling, genetic sex determinants, iron metabolism, and cell-death pathways relevant to NDL. Prior studies show that males generally experience more severe hepatic IRI than females, though this advantage is age dependent and may be attenuated by hormonal fluctuations, reduced estrogen receptor abundance, and hepatic macrosteatosis [18,19]. Evidence suggests that estrogen-independent mechanisms also contribute to early female resilience following IRI [14].

In this study of non-utilized human livers undergoing 6 h of NMP, we did not observe statistically robust sex-related differences in biochemical or transcriptional markers of IRI. Lipid peroxidation, assessed by serum MDA fold change, showed no detectable sex-related differences, consistent with emerging evidence that sex effects on IRI are context-dependent and may be attenuated in ex vivo perfusion settings. Limited sample size likely reduced power to detect subtle biological differences.

Expression changes in ferroptosis and stress-response genes showed substantial overlap between sexes after correction for multiple testing. Although no gene reached statistical significance, ACSF2 demonstrated the strongest sex-associated trend, with higher induction in female livers, while ATP5G3 and FTH1 showed intermediate differences. These genes are involved in lipid metabolism, mitochondrial function, and iron handling, processes central to ferroptotic susceptibility, and warrant validation in larger cohorts. The remaining genes, including HO-1, HIF-1α, IREB2, NQO1, TTC35, and NRF2, showed no evidence of differential regulation between sexes under the conditions studied.

Male hepatocytes are generally more vulnerable to ferroptosis due to higher mitochondrial iron and ROS and preferential upregulation of iron importers, whereas females exhibit higher baseline FTH1 expression, conferring greater resistance. Given its role in iron sequestration and oxidative stress buffering, FTH1 remains a relevant candidate for further study in marginal livers undergoing NMP.

Females also demonstrate higher hepatic expression of NRF2-regulated antioxidant genes, and NRF2 activation is a recognized protective mechanism against ferroptosis [22,23]. During NMP, differential reliance on FTH1 induction in males and NRF2-driven antioxidant pathways in females may reflect these baseline differences [24,25,26].

Genetic sex determinants further modulate hepatic resilience. The sex-determining Region Y (SRY) gene on the Y chromosome enhances inflammation and oxidative stress by suppressing β-catenin/FOXO signaling and activating NF-κB and TLR4 pathways, directly increasing male vulnerability to IRI [17]. KCNE4 deletion reveals additional sex-dependent signaling differences, as in males, it is known to suppress protective RISK/SAFE pathways, while in females, it enhances protective GSK-3β inhibitory phosphorylation [16]. Similarly, estrogen sulfotransferase (EST), induced by oxidative stress via Nrf2, decreases active estrogen signaling and disproportionately increases injury in females, while EST deletion worsens injury in males [15]. Together, these findings highlight that sex differences in hepatic resilience are not explained by a single hormone or receptor but arise from multi-layered interactions between hormonal pathways, genetic sex factors, and cellular stress responses.

4.3. Assessing Influence of Iron Metabolism and Liver Injury

Sex differences in hepatic IRI are closely linked to ferroptosis and iron metabolism. Multi-omics studies identify the DTL-PROX1 axis as a key mediator of sex-dimorphic ferroptosis, with higher DTL expression in males promoting PROX1 degradation, increasing PUFA availability, and heightening ferroptotic susceptibility [1]. Sex-specific differences in iron handling further contribute to this dimorphism: male hepatocytes exhibit greater mitochondrial iron import, higher TfR1 expression, and increased ROS generation, whereas females show higher ferritin expression and stronger antioxidant defenses, conferring relative resistance to ferroptosis [3]. Aging and metabolic status may modify these effects, as studies suggest sex-dependent alteration in hepatic iron storage and lipid peroxidation [20,21]. Together, these differences in iron homeostasis help explain sex-based variability in ferroptotic injury and are relevant to assessing donor liver vulnerability.

4.4. Sex Differences in Liver Graft Injury and Relevance to Transplant Outcomes

Taken together, the literature demonstrates that sex differences in hepatic IRI and ferroptotic injury arise from an intricate interplay of hormonal status, sex chromosome–linked regulators, iron metabolism, and cell-death signaling pathways, including MAPK, β-catenin/FOXO, RISK/SAFE, and DTL-PROX1-mediated ferroptosis. These mechanisms have direct relevance for NDLs, as donor sex, age, metabolic status, iron load, and receptor expression may influence ferroptotic sensitivity and post-reperfusion viability. Incorporating sex-specific ferroptosis pathways into donor assessment and preservation strategies may improve risk stratification and support targeted interventions to safely expand the use of marginal livers.

4.5. Study Limitations and Future Directions

Limitations of this study include its preliminary, pilot design based on findings reported from five NDLs. Cohort size in this study was limited by scarce availability of non-utilized human donor livers as well as the resource-intensive and costly nature of NMP-focused experiments. Due to small sample size, generalizability and statistical power is reduced, and further studies should be performed to investigate causation. Future studies can build upon the observational nature of this preliminary study by expanding the sample size to enhance the representativeness of a study population and, in turn, improve generalizability to clinical settings. Secondly, it is important to acknowledge variability in donor characteristics, including cause of death, race, and factors such as the hepatitis C virus, hypertension, and diabetes mellitus status. With only five donors, the ability to assess these factors as potential confounders is limited. Third, our study utilizes MDA concentrations as a lipid peroxidation marker to investigate sex-based differences. Due to limited tissue availability, additional assays, such as GPX4, ACSL4, and iron-handling proteins, including transferrin receptor (TfR) and ferroportin, could not be performed. Future studies incorporating these measures, as well as correlating ferroptosis-related gene expression data with protein-level analysis, will be essential to validate and extend these preliminary findings.

Our findings provide preliminary analyses on the influences of sex in ferroptosis-related gene expression patterns under NMP conditions. Results of this observational study provide a hypothesis-generating guide for future research, and future studies will be important to validate these initial trends.

5. Conclusions

In this small exploratory cohort, we did not observe statistically robust sex-related differences in MDA levels or ferroptosis-related gene expression changes after 6 h of NMP. Although ACSF2 and, to a lesser extent, ATP5G3 and FTH1, showed nominal sex-associated expression differences, these findings did not remain significant after correction for multiple testing and should be interpreted cautiously. Overall, this study is hypothesis-generating and underscores the need for larger, adequately powered investigations to determine whether sex-dependent ferroptosis signaling contributes meaningfully to IRI during liver preservation. Future studies integrating interactomic and network-based approaches may provide critical causal insight into the sex-specific regulation of ferroptosis during IRI, enabling identification of key regulatory nodes and therapeutic targets that are not apparent from single-gene analyses.