Icariin-loaded liposomal nanoparticles alleviate streptozotocin-induced diabetic hepatopathy associated with modulation of ferroptosis-related and inflammatory signaling Authors: Badriah A. Hifni, Shaza A. Alyamani, Nasser S. Alqahtani, Ayat B. Al-Ghafari, Huda A. Al Doghaither, Ekramy M. Elmorsy, Manal S. Fawzy, Gehad E. Elshopakey Published: 2026 ABSTRACT BackgroundDiabetic hepatopathy is a major complication of type 2 diabetes mellitus characterized by metabolic dysregulation, oxidative stress, inflammation, and emerging evidence of ferroptosis-mediated liver injury. Icariin (IC), a bioactive flavonoid, exhibits hepatoprotective potential; however, its clinical applicability is limited by poor bioavailability. Liposomal nanoformulation represents a potential approach to modify IC delivery and enhance its pharmacological efficacy.MethodsNinety rats were randomly assigned to six groups (n = 15/group): Control, IC, IC-LNPs, NA–STZ, NA–STZ + IC, and NA–STZ + IC-LNPs. Type 2 diabetes was induced by nicotinamide (120 mg/kg, i.p.) followed 15 min later by streptozotocin (60 mg/kg, i.p.). Five days after diabetes confirmation, diabetic treatment groups received IC or IC-LNPs at an IC-equivalent dose of 80 mg/kg/day, i.p., for eight consecutive weeks. Hepatic injury was assessed through biochemical parameters, oxidative stress markers, inflammatory cytokines, ferroptosis-related markers, and histopathological and ultrastructural analyses.ResultsNA–STZ administration induced significant metabolic dysfunction, including hyperglycemia, insulin depletion, dyslipidemia, weight loss, and elevated hepatic enzyme levels. IC treatment significantly improved these parameters, while IC-LNPs produced a more pronounced protective effect. Diabetic induction was associated with reduced hepatic NRF2 and HO-1 levels, accompanied by reduced levels of GSH, SOD, CAT, and GPX, and increased oxidative stress markers (MDA and H₂O₂). IC-LNPs more effectively restored antioxidant defenses and redox homeostasis than free IC. Moreover, NA–STZ increased NF-κB and COX-2 expression and the inflammatory mediators TNF-α, IL-6, and IL-1β, which were markedly reduced by IC-LNP treatment. Importantly, diabetic hepatopathy was associated with alterations in ferroptosis-related markers, evidenced by iron accumulation (Fe²⁺), upregulation of ACSL4 and TFR1, and downregulation of GPX4 and SLC7A11. IC-LNPs significantly attenuated these alterations, consistent with modulation of ferroptosis-related processes. Histopathological and ultrastructural analyses confirmed severe hepatic damage in diabetic rats, whereas IC-LNPs markedly preserved hepatic architecture.ConclusionIC-LNPs provided greater protection than free IC against NA–STZ-induced diabetic hepatopathy, associated with attenuation of oxidative stress and inflammation and modulation of NRF2/HO-1, NF-κB/COX-2, and ferroptosis-related markers. These findings support the potential of liposomal IC as a preclinical approach for diabetic liver injury; however, pharmacokinetic, biodistribution, safety, and further mechanistic studies are required before its therapeutic or translational potential can be established.Supplementary InformationThe online version contains supplementary material available at https://doi.org/10.1186/s40360-026-01213-0. Introduction Diabetes mellitus is a chronic metabolic disorder characterized by persistent hyperglycemia resulting from insulin resistance and/or impaired insulin secretion [1]. Among its different forms, type 2 diabetes mellitus (T2DM) represents the predominant subtype worldwide and constitutes a major global public health challenge [2]. The global burden of diabetes continues to rise steadily, largely due to the development of progressive complications affecting multiple organs and physiological systems [3]. Among these complications, hepatic dysfunction has attracted increasing attention because the liver plays a central role in maintaining glucose and lipid homeostasis [4]. In T2DM, insulin resistance in peripheral tissues together with impaired hepatic glucose metabolism contributes to persistent hyperglycemia [2]. This metabolic disturbance is associated with profound alterations in lipid, carbohydrate, and protein metabolism [1, 5]. Consequently, excessive accumulation of lipids and glucose metabolites within hepatocytes promotes oxidative stress, mitochondrial dysfunction, and inflammatory responses, ultimately leading to hepatic injury [6, 7]. Ferroptosis has recently emerged as a critical mechanism underlying diabetes-associated liver injury [8]. It is an iron-dependent regulated form of cell death characterized by excessive lipid peroxidation and disruption of cellular antioxidant defenses [9]. This process is primarily mediated by glutathione depletion and inactivation of glutathione peroxidase 4 (GPX4) [10, 11]. Ferroptosis is closely linked to excessive reactive oxygen species (ROS) generation and intracellular iron accumulation [12]. Moreover, ferroptotic signaling can activate pro-inflammatory pathways, particularly nuclear factor-kappa B (NF-κB), thereby aggravating hepatic inflammation and cellular damage in diabetes [13]. In contrast, the nuclear factor erythroid 2–related factor 2 (NRF2) pathway represents a major endogenous cytoprotective mechanism that preserves redox homeostasis through the regulation of antioxidant and detoxifying enzymes [14]. Activation of NRF2 signaling enhances the expression of several downstream antioxidant mediators, including heme oxygenase-1 (HO-1), which plays a pivotal role in limiting oxidative stress–mediated cellular injury [15]. Therefore, targeting oxidative stress and ferroptosis-related pathways may represent a promising therapeutic strategy for mitigating diabetes-induced hepatic injury. There has been significant fascination with utilizing natural products and their bioactive components as origins of pharmacologically active substances that may provide health benefits. The diverse biological functions of these substances, such as antioxidant, anti-inflammatory, metabolic, and cytoprotective properties, could facilitate the concurrent regulation of interrelated pathogenic mechanisms associated with chronic metabolic diseases [16, 17]. Plant-based polyphenols and flavonoids have demonstrated potential in mitigating oxidative and inflammatory damage thru the regulation of cellular redox balance, inflammatory agents, and metabolic signaling pathways [18–20]. Icariin (IC), a prenylated flavonoid glycoside isolated mainly from Epimedium brevicornu, has attracted considerable attention because of its diverse pharmacological activities [21]. Traditionally, IC has been widely used in Chinese medicine for strengthening kidney function and enhancing yang activity [22]. Recent pharmacological investigations have demonstrated that IC exerts potent antioxidant, anti-inflammatory, anti-apoptotic, and metabolic regulatory effects [23, 24]. Notably, IC has shown beneficial effects in improving lipid metabolism and alleviating metabolic disorders, including diabetes and obesity [23, 25]. In addition, its phytoestrogen-like properties may contribute to the regulation of endocrine and metabolic homeostasis, further supporting its therapeutic potential against metabolic dysfunction [26]. Despite its promising pharmacological properties, the therapeutic application of IC remains limited because of its poor aqueous solubility, low oral bioavailability, and rapid metabolic elimination [27]. Collectively, these limitations may restrict its systemic exposure and therapeutic potential. Liposomal and other nanocarrier-based drug delivery systems have therefore gained substantial interest as approaches for modifying the physicochemical properties, stability, and delivery of poorly soluble bioactive compounds [28–30]. Liposomal nanoparticles (LNPs) are considered particularly promising owing to their high biocompatibility and their capacity to encapsulate both hydrophilic and lipophilic compounds [31]. Previous studies have shown that liposomal encapsulation can influence the pharmacokinetic and pharmacodynamic behavior of loaded therapeutics, although these effects depend on the physicochemical characteristics of the formulation and route of administration [32]. Accordingly, encapsulation of IC within liposomal nanoparticles may provide a formulation approach capable of enhancing its pharmacological efficacy compared with free IC. However, improvements in bioavailability, circulation time, hepatic targeting, or cellular uptake require direct pharmacokinetic and biodistribution evidence and cannot be inferred from formulation or efficacy data alone [32, 33]. Therefore, the present study was designed to evaluate the hepatoprotective effects of icariin-loaded liposomal nanoparticles (IC-LNPs) in a nicotinamide–streptozotocin (NA–STZ)-induced model of diabetic hepatopathy, with comparison to crude icariin (IC). Because both formulations were administered intraperitoneally at an equivalent IC dose, the study was designed to compare their relative hepatoprotective efficacy rather than to evaluate improvement in oral bioavailability. The study specifically aimed to determine whether liposomal nanoformulation provides greater protective efficacy than free IC in attenuating diabetes-induced hepatic injury. In addition, this work examined changes in oxidative stress, inflammatory signaling, and ferroptosis-related markers, with particular attention to NRF2/HO-1, NF-κB/COX-2, and GPX4/SLC7A11-related responses. These molecular assessments were intended to identify pathways associated with the observed hepatoprotective effects rather than to establish definitive pathway activation or causal involvement. Collectively, the study provides a preclinical comparison of free and liposomal IC in experimental diabetic hepatopathy while exploring molecular changes potentially associated with their protective effects. Materials and methods Preparation and characterizations of icariin -loaded liposomal nanoparticles (IC-LNPs) The study compounds, Icariin, cholesterol, and L-α-phosphatidylcholine, were purchased from Sigma-Aldrich (Egypt). All reagents were analytical grade and utilized without further purification. The preparation of liposomes involved blending 30 mg (0.078 mmol) of cholesterol with 70 mg (0.09 mmol) of L-α-phosphatidylcholine. Liposome formulation was prepared by combining cholesterol (30 mg; 0.078 mmol) with L-α-phosphatidylcholine (70 mg; 0.09 mmol). In preparation for thin-film hydration, L-α-phosphatidylcholine and cholesterol were dissolved in 10 mL of chloroform/methanol (2:1, v/v) using a round-bottom flask. The organic solvent mixture was removed under reduced pressure to form a thin and homogeneous lipid film. The lipid film was subsequently hydrated with 1 mL PBS (pH 7.4) containing 46.22 mg IC (approximately 68.3 µmol), corresponding to an initial IC concentration of 46.22 mg/mL. The mixture was homogenized in order to achieve complete dissolution of lipids before subsequent processing. The hydrated formulation was incubated at 37 °C for 30 min under mild shaking to facilitate hydration and incorporation of IC into the lipid vesicles. The sample was sonicated for 10 min using a SONICS Vibra-Cell system, with 10-second intervals and 30% amplitude. The experiment was carried out under ice-cold conditions to prevent heat-related destabilization. The formulation was centrifuged at 15,000 × g for 30 min to separate unencapsulated IC from the IC-LNP fraction. The resulting IC-LNP dispersion was collected for subsequent characterization. Particle size, polydispersity index, and zeta potential analysis were conducted on the obtained IC-LNP dispersion. Particle size, PDI, and zeta potential measurements were performed in triplicate (n = 3). These tests aimed to characterize IC-LNPs in terms of size distribution, stability, surface charge, and uniformity. Distilled water was used for Z-average particle size measurement. The structural features and integrity of IC-LNPs were characterized by high-resolution transmission electron microscopy (JEOL 2100, Japan). Transmission electron microscopy at 160 kV provided clear visualization of the nanoscale architecture of IC-LNPs. Encapsulation efficiency and drug loading Encapsulation efficiency of IC-LNPs was assessed based on the level of free IC detected in the supernatant. UV–Vis spectrophotometry at 260 nm was employed to determine unencapsulated drug content. The unencapsulated fraction was used to estimate the extent of IC incorporation into the lipid bilayer. Encapsulation efficiency and drug loading were derived from standard equation-based calculations. Drug loading (%) was computed as (encapsulated IC / total liposome mass) × 100, while encapsulation efficiency (%) was calculated as (encapsulated IC / total IC added) × 100. The amount of encapsulated IC was calculated by subtracting the amount of free IC detected in the supernatant from the total amount of IC initially added. The IC-equivalent concentration of the final IC-LNP formulation was determined from the encapsulated IC amount and the final formulation volume. The final IC-equivalent concentration was 40 mg/mL. This concentration provided the target dose of 80 mg/kg at an injection volume of 2 mL/kg. Fourier transform infrared spectroscopy (FTIR) analysis The chemical compatibility and possible interactions between IC and the lipid components were evaluated using Fourier Transform Infrared Spectroscopy (FTIR), where lyophilized samples of IC-LNPs, pure IC, cholesterol, and L-α-phosphatidylcholine were individually mixed with potassium bromide (KBr) and compressed into transparent pellets. The FTIR spectra were recorded under standard operating conditions over a scanning range of 4000–400 cm⁻¹. This analysis was conducted to identify characteristic functional groups of each component and to detect any potential shifts or changes in peak intensity. Such spectral variations may indicate molecular interactions between IC and the lipid matrix, as well as confirm the successful encapsulation of IC within the lipid nanoparticles. In vitro drug release study The in vitro release behavior of IC was evaluated for both the free drug (raw IC) and the IC-LNPs using a dialysis bag diffusion method. A known amount of either free IC or IC-LNPs equivalent to the same IC concentration was placed separately into dialysis membranes (MWCO 12–14 kDa). The bags were then immersed in 50 mL of phosphate-buffered saline (PBS, pH 7.4) maintained at 37 ± 0.5 °C under continuous gentle stirring. Free IC was included as a control for comparison with the release profile of IC-LNPs. At predetermined time intervals, aliquots of the release medium were withdrawn and replaced with equal volumes of fresh PBS to maintain sink conditions. The amount of released IC was quantified using UV–Vis spectrophotometry at 260 nm. The cumulative release profiles of both free IC and IC-LNPs were then constructed and compared to evaluate the effect of nanoformulation on the release pattern and sustainment of drug delivery over time. Formulation quality and injection suitability The final IC-LNP formulation was evaluated for pH, osmolarity, sterility, and endotoxin content before administration. The pH was measured using a calibrated digital pH meter. Osmolarity was determined using a calibrated osmometer. The formulation was sterilized by filtration through a sterile 0.22-µm membrane filter under aseptic conditions. Endotoxin levels were determined using a Limulus Amebocyte Lysate (LAL) assay according to the manufacturer’s instructions. Experimental animals The experimental animals consisted of ninety adult male albino rats, 8 weeks old with body weights of 220–250 g, obtained from the MERC Center, Faculty of Medicine, Mansoura University, Egypt. Male rats were specifically selected to avoid the potential influence of hormonal fluctuations associated with the estrous cycle in females, which may affect glucose metabolism, oxidative stress responses, inflammatory pathways, and susceptibility to hepatic injury, thereby reducing experimental variability. The use of male animals only is acknowledged as a limitation regarding the generalizability of the findings to both sexes. Housing conditions were strictly controlled, with temperature set at 23 ± 2 °C, humidity at 60–65%, and a 12-hour light and dark cycle. Throughout the study period, rats were given unrestricted access to a commercial pellet diet and drinking water. The study was executed in full compliance with the NIH Guide for the Care and Use of Laboratory Animals (NIH Publication No. 8023, revised 1996). The study protocol was approved by Mansoura University Animal Care and Use Committee (approval no. MU- ACUC-M.V.R.26.6.295) and was reported in accordance with the ARRIVE guidelines, where applicable. Induction of diabetes Type 2 diabetes mellitus (T2DM) was experimentally induced using the nicotinamide (NA)–streptozotocin (STZ) protocol following an overnight fast of 16 h. This experimental model was selected because it reproduces selected metabolic characteristics of T2DM, including partial pancreatic β-cell dysfunction and hyperglycemia. Nicotinamide (120 mg/kg, i.p.) dissolved in physiological saline was administered to each rat, followed 15 min later by streptozotocin (STZ, 60 mg/kg, i.p.) freshly prepared in 0.1 M sodium citrate buffer (pH 4.5). Pretreatment with nicotinamide partially protects pancreatic β-cells against STZ-induced cytotoxicity, thereby producing a moderate diabetic phenotype rather than the extensive β-cell destruction typically associated with STZ alone. To minimize STZ-induced acute hypoglycemia, rats were provided with 5% glucose solution ad libitum for 24 h, starting 6 h after STZ administration, as previously described by Demoz, Gachoki, Mungai and Negusse [34]. Fasting blood glucose (FBG) levels were measured 72 h after diabetes induction using tail vein blood samples and a glucometer. Rats with FBG levels ≥ 250 mg/dL were considered diabetic and included in the diabetic experimental groups [35]. This cutoff value was selected to ensure successful establishment of persistent hyperglycemia and experimental diabetes. Blood glucose measurements were performed using a Rightest Tango TD-4235 glucometer (Taiwan). All STZ -treated animals showed FBG levels ≥ 250 mg/dL and were enrolled in the study. Experimental design A total of six groups were randomly assigned, with 15 rats included in each group based on computer generated random numbers. Group I (Control) consisted of non-diabetic rats that received daily intraperitoneal (i.p.) injections of sterile saline. Group II (IC) included non-diabetic rats treated daily with crude icariin (IC) at a dose of 80 mg/kg, i.p. Group III (IC-LNPs) comprised non-diabetic rats administered icariin-loaded liposomal nanoparticles (IC-LNPs) daily at a dose equivalent to 80 mg/kg of free IC. The selected IC dose (80 mg/kg) was based on previous experimental studies demonstrating its significant antioxidant, anti-inflammatory, and antidiabetic activities without inducing toxic effects in rats [36]. The same IC-equivalent dose was used for the IC and IC-LNP groups to permit direct comparison of their biological effects. Group IV (NA–STZ) consisted of diabetic untreated rats and received daily saline injections following diabetes induction. Diabetic rats in Group V (NA–STZ + IC) were treated daily with crude IC (80 mg/kg, i.p.), starting five days after diabetes confirmation. Diabetic rats in Group VI (NA–STZ + IC-LNPs) received daily i.p. injections of IC-LNPs at a dose equivalent to 80 mg/kg of free IC, administered at an injection volume of 2 mL/kg, corresponding to an IC-equivalent concentration of 40 mg/mL. Treatment began five days after confirmation of diabetes. Because both IC and IC-LNPs were administered intraperitoneally, the experimental comparison was intended to evaluate the relative efficacy of the two formulations and not improvement in oral bioavailability. The five-day interval before treatment initiation was implemented to ensure stable establishment of the diabetic state prior to therapeutic intervention. The experimental treatments continued for eight consecutive weeks. The treatment duration was selected to permit evaluation of chronic diabetes-associated hepatic alterations and the effects of IC and IC-LNP treatment. At the end of the experimental period, body weights were recorded for all groups to assess diabetes-associated changes and treatment-related effects. No mortality, treatment-related adverse events, or exclusions occurred during the experimental period. All animals completed the study, and no animals or data were excluded from the analyses. Investigators responsible for histopathological scoring, immunohistochemical quantification, and other outcome assessments were blinded to group allocation. Animal handling, sample collection, and tissue processing protocols After completing the 8-week treatment period, the 15 animals in each group were allocated for endpoint assessments as follows: 10 animals per group were used for biochemical and molecular analyses, while the remaining 5 randomly selected animals per group were used for histopathological evaluation. Thus, all 90 animals included in the experimental design were accounted for in the endpoint analyses. Animals designated for biochemical and molecular analyses were fasted for 24 h before sample collection. Isoflurane was used to induce anesthesia until all reflexes were completely abolished. Euthanasia was then carried out by administering an overdose of 5% isoflurane in oxygen for 5 min. Death was confirmed by the absence of respiratory and cardiac activity. Retro-orbital venous sinus blood was collected from the 10 animals per group designated for biochemical and molecular analyses into plain tubes. Samples were allowed to clot at room temperature and then centrifuged at 4,000 rpm for 10 min to separate serum. Serum was stored at − 20 °C until biochemical analysis. Serum insulin was measured using the specified immunoassay kit (INS-5131), whereas fasting glucose (GLU-2505), AST (MET-5127), ALT (MET-5123), total cholesterol (TC; MET-5025), and triglycerides (TG; MET-5027) were determined using the corresponding commercial biochemical assay kits from Bio-Med Diagnostic (Egypt), according to the manufacturer’s instructions. Following euthanasia, the liver was carefully excised and rinsed with ice-cold 1.15% KCl solution to remove residual blood. Liver samples obtained from the 10 animals per group designated for biochemical and molecular analyses were immediately divided into appropriate portions according to the planned assays. Samples intended for biochemical analyses were stored at − 20 °C until analysis, whereas samples intended for molecular analyses were. For biochemical analyses, liver specimens were homogenized in 50 mmol/L Tris–HCl buffer (pH 7.4) under chilled conditions to obtain a uniform homogenate. The homogenates were centrifuged at 10,000 × g for 15 min at 4 °C, and the resulting clear supernatants were collected for subsequent assays and stored at − 20 °C until analysis. For histopathological evaluation, liver specimens from the remaining five randomly selected animals in each group (n = 5/group) were fixed in 10% neutral-buffered formalin and processed for routine histological examination. Histopathological scoring was therefore based on five independent biological replicates per group. Histopathological examination and scoring were performed by an investigator blinded to the experimental group allocation. The biological sample size (n) used for each endpoint is specified in the corresponding Methods subsection and figure legend, with n referring to the number of individual animals rather than technical replicates. Hepatic oxidative stress and antioxidant signaling assessment Hepatic oxidative status was assessed by measuring reduced glutathione (GSH) content and the activities of major antioxidant enzymes, including glutathione peroxidase (GPx), superoxide dismutase (SOD), and catalase (CAT). Measurements were performed using commercially available colorimetric assay kits from BioDiagnostic (Egypt): GSH (Cat. No. 2511), GPx (Cat. No. 2524), SOD (Cat. No. 2521), and CAT (Cat. No. 2517), according to the manufacturer’s instructions. GSH content was determined based on its reaction with 5,5′-dithiobis-(2-nitrobenzoic acid) (DTNB), resulting in the formation of a yellow chromogen. GPx activity was assessed by monitoring NADPH oxidation during the glutathione-dependent reduction of hydrogen peroxide, whereas SOD activity was determined based on inhibition of nitroblue tetrazolium (NBT) reduction. CAT activity was assessed by measuring the decomposition of hydrogen peroxide. Malondialdehyde (MDA), as an index of lipid peroxidation, was measured using a thiobarbituric acid-reactive substances (TBARS) colorimetric assay kit (BioDiagnostic, Egypt; Cat. No. MD 2529) according to the manufacturer’s instructions. Hepatic hydrogen peroxide (H₂O₂) levels were determined using a commercial assay kit (MyBioSource, USA; Cat. No. MBS3808898) according to the manufacturer’s protocol. Absorbance was measured at the assay-specific wavelengths of 412 nm for GSH, 340 nm for GPx, 560 nm for SOD, 510 nm for CAT, 532 nm for MDA, and 450 nm for H₂O₂. Hepatic total antioxidant capacity (TAC) was measured using a commercially available assay (MyBioSource, USA; Cat. No. MBS1600693) according to the manufacturer’s instructions. Absorbance was measured at 450 nm using a microplate reader, and concentrations were calculated from the corresponding standard curve. To assess changes associated with NRF2/HO-1 antioxidant signaling, hepatic NRF2 and HO-1 protein levels were quantified using rat-specific ELISA kits from MyBioSource (USA; NRF2, Cat. No. MBS3807961; HO-1, Cat. No. MBS2024438), according to the manufacturer’s instructions. Absorbance was measured at 450 nm, and concentrations were calculated from the respective standard curves. Because NRF2 was measured in hepatic tissue without assessment of nuclear translocation or cytosolic/nuclear fractionation, these measurements were interpreted as changes in NRF2 protein levels rather than definitive evidence of NRF2 pathway activation. For all assays, biological replicates consisted of individual liver samples from 10 animals per group (n = 10/group). Assessment of hepatic inflammatory mediators Hepatic concentrations of tumor necrosis factor-alpha (TNF-α), interleukin-1 beta (IL-1β), and interleukin-6 (IL-6) were quantified using rat-specific ELISA kits from Assay Genie (TNF-α, Cat. No. AEES00516; IL-1β, Cat. No. AEES02613; IL-6, Cat. No. RTES00015, respectively), according to the manufacturer’s instructions. Hepatic cyclooxygenase-2 (COX-2) levels were determined using a rat-specific ELISA kit from MyBioSource (USA; Cat. No. MBS266603). For NF-κB assessment, nuclear extracts were prepared from liver tissue using a commercial nuclear extraction kit (Abcam, Cambridge, UK; Cat. No. ab113474) according to the manufacturer’s instructions. NF-κB levels in the resulting nuclear extracts were subsequently quantified using a rat-specific ELISA kit (MyBioSource, USA; Cat. No. MBS453975). NF-κB measurements were normalized to the total protein content of the corresponding nuclear extracts. Absorbance was measured at 450 nm using a VersaMax microplate reader, and concentrations were calculated from the respective standard curves. The nuclear NF-κB measurements were interpreted as changes in nuclear NF-κB abundance associated with inflammatory signaling. Because phospho-p65, IκBα degradation, and NF-κB DNA-binding/transcriptional activity were not directly assessed, the findings were not considered definitive evidence of NF-κB pathway activation. All measurements were performed using liver samples from 10 individual animals per group (n = 10/group). Assessment of hepatic apoptotic markers Rat-specific ELISA kits from MyBioSource (USA) were used to quantify hepatic levels of the apoptosis-related proteins Bax (Bcl-2–associated X protein), caspase-3, and Bcl-2 (B-cell lymphoma-2), using Cat. Nos. MBS2512405, MBS729893, and MBS2515143, respectively. The assays were performed according to the manufacturer’s instructions. Protein concentrations were normalized to the total protein content of the corresponding liver samples. Absorbance was measured at 450 nm using a microplate reader, and concentrations were calculated from the corresponding standard curves. Measurements were performed using liver samples from 10 individual animals per group (n = 10/group). Assessment of hepatic ferroptosis-related markers To evaluate biochemical and molecular alterations associated with ferroptosis, hepatic levels of glutathione peroxidase 4 (GPX4), solute carrier family 7 member 11 (SLC7A11), acyl-CoA synthetase long-chain family member 4 (ACSL4), transferrin receptor 1 (TFR1), and ferrous iron (Fe²⁺) were determined. SLC7A11, ACSL4, and TFR1 were quantified using rat-specific ELISA kits from MyBioSource (USA; Cat. Nos. MBS7204860, MBS8820647, and MBS265236, respectively). GPX4 was quantified using a rat-specific ELISA kit from Biomatik (Canada; Cat. No. EKU12535). Hepatic Fe²⁺ content was determined using a colorimetric iron assay kit (Abcam, Cambridge, UK; Cat. No. ab83366). All assays were performed according to the manufacturers’ protocols, and protein-based measurements were normalized to the total protein content of the corresponding liver samples, where applicable. Measurements were performed using liver samples from 10 individual animals per group (n = 10/group). These measurements were interpreted collectively as ferroptosis-related biochemical and molecular alterations rather than definitive evidence of ferroptotic cell death, because ferroptosis-specific functional validation using lipid ROS probes or rescue experiments with ferroptosis inhibitors was not performed. ELISA Assay specifications and quality control The analytical characteristics of the commercial immunoassay kits used in this study were verified against the corresponding manufacturers’ product information and assay instructions. Available information regarding assay principle, detection range, analytical sensitivity, and intra- and inter-assay coefficients of variation is summarized in Supplementary Table S2 according to their available manufacturer sources verified data. All assays were performed according to the manufacturers’ recommended protocols, and analyte concentrations were calculated from the corresponding standard curves. For hepatic tissue measurements, concentrations were normalized to the total protein content of the respective tissue extracts, whereas nuclear Nrf2 and NF-κB levels were normalized to the total protein content of the corresponding nuclear extracts. Where specific analytical characteristics were not provided in the available manufacturer documentation, these parameters were reported as not specified rather than inferred. Molecular assessment of hepatic tissue using real-time PCR Liver tissues were homogenized using a TissueLyser II (Qiagen) in 1 mL QIAzol Lysis Reagent, followed by total RNA extraction. Chloroform was added for phase separation, and samples were centrifuged at 12,000 × g for 15 min. The RNA-containing aqueous phase was carefully collected, followed by RNA precipitation with isopropanol and centrifugation at 12,000 × g for 10 min. The resulting RNA was reconstituted in DNase/RNase-free water, and its concentration and purity were assessed spectrophotometrically. Extracted RNA was reverse-transcribed into cDNA using the iScript™ cDNA Synthesis Kit (Bio-Rad) according to the manufacturer’s instructions. Quantitative real-time PCR (RT-qPCR) was performed using iTaq Universal SYBR Green Supermix (Bio-Rad; Cat. No. 172–5121) on a Rotor-Gene Q system (Qiagen). Gene-specific primers for the oxidative stress-, inflammatory-, apoptosis-, lipid metabolism-, and ferroptosis-related genes evaluated in the study are provided in the revised Table S1. The primer table was comprehensively cross-checked to ensure that each gene reported in the Results is represented by a unique and correctly identified forward and reverse primer pair. The amplification protocol consisted of an initial denaturation at 95 °C for 3 min, followed by 40 cycles of denaturation at 95 °C for 15 s, primer-specific annealing at 58–60 °C for 30 s, and extension at 72 °C for 30 s. Melt-curve analysis was performed over 65–95 °C to assess amplification specificity. Relative gene expression was calculated using the 2^−ΔΔCt method [37], with β-actin used as the reference gene for normalization. The complete Actb forward and reverse primer sequences are provided in Table S1. RT-qPCR analyses were performed using liver samples from 10 individual animals per group (n = 10/group). Changes in NRF2-, NF-κB-, and ferroptosis-related gene expression were interpreted as transcriptional associations with these signaling processes and not, in isolation, as evidence of pathway activation or definitive ferroptotic cell death. Histological processing and semi-quantitative evaluation of liver tissue Liver specimens from five randomly selected animals per group (n = 5/group) were used for histopathological evaluation. Tissue fixation was carried out in 10% neutral-buffered formalin at a fixative-to-tissue volume ratio of approximately 20:1 for 72 h. Samples were subsequently dehydrated through graded ethanol, cleared in xylene, infiltrated with molten paraffin, and embedded in paraffin blocks. Tissue sections were prepared using a rotary microtome, deparaffinized, rehydrated through graded alcohols, and stained with hematoxylin and eosin (H&E). Histological sections were examined by light microscopy for diabetes-associated hepatic alterations, including the principal lesions observed in the experimental model. Because the NA–STZ model represents diabetic hepatopathy and was not specifically established as a nonalcoholic steatohepatitis model, histopathological findings were evaluated using a semiquantitative liver-injury scoring approach based on the observed lesions rather than interpreted diagnostically as a NASH Activity Score (NAS). The individual histopathological criteria and corresponding scores (0–3) are defined in Table S3. For each animal, three non-overlapping liver sections were examined, and four randomly selected microscopic fields per section were evaluated at ×400 magnification, resulting in 12 fields per animal. Lesion severity was graded on a 0–3 scale, and the mean of the evaluated fields was calculated to obtain a single histopathological score for each animal. Thus, the animal, rather than the individual microscopic field, represented the biological replicate (n = 5/group). Histological examination and scoring were conducted by an observer blinded to the experimental group allocation. Assessment of hepatic ultrastructure The Small liver specimens designated for ultrastructural examination were immediately fixed in 2.5% glutaraldehyde prepared in 0.1 M phosphate buffer for 24 h at 4 °C. Following primary fixation, tissues were washed and post-fixed in 1% osmium tetroxide for 2 h. Samples were dehydrated through a graded ethanol series (50–100%), treated with acetone, infiltrated with epoxy resin, and embedded. Ultrathin Sects. (60–70 nm) were prepared using an ultramicrotome, mounted on copper grids, and contrasted with uranyl acetate and lead citrate. Ultrastructural examination was performed using a JEOL 2100 transmission electron microscope (Japan) at an accelerating voltage of 160 kV. Representative digital micrographs were obtained from multiple fields to qualitatively assess hepatocellular ultrastructural alterations. TEM observations were used as supportive morphological evidence and were not treated as independent biological replicates unless quantitative ultrastructural analysis was specifically performed. Immunohistochemical assay of GPX4 and ACSL4 in liver tissue Paraffin-embedded liver sections from five animals per group (n = 5/group) were used for immunohistochemical analysis. Sections were deparaffinized in xylene (2 × 5 min) and rehydrated through graded ethanol (100–70%) followed by distilled water. Antigen retrieval was performed in 0.01 M citrate buffer (pH 6.0) using microwave heating for 10 min. Endogenous peroxidase activity was blocked with 3% hydrogen peroxide in methanol for 10 min, followed by washing with PBS. Non-specific binding was blocked with 5% bovine serum albumin for 30 min at room temperature. Sections were incubated overnight at 4 °C with primary antibodies against GPX4 (1:200; Abcam, Cat. No. ab125066) and ACSL4 (1:200; Abcam, Cat. No. ab155282). After washing with PBS, sections were incubated with a biotinylated goat anti-rabbit secondary antibody (1:500) for 30 min, followed by streptavidin–HRP for 20 min. Immunoreactivity was visualized using 3,3′-diaminobenzidine (DAB), producing brown staining in immunopositive cells. Sections were counterstained with Mayer’s hematoxylin, dehydrated, cleared in xylene, and mounted with DPX. Slides were examined under light microscopy at ×400 magnification. Five randomly selected, non-overlapping fields were analyzed per animal. Digital images were acquired under standardized imaging conditions, and the positively stained area was quantified using ImageJ software. For each marker, measurements from the five fields were averaged to generate one value per animal; therefore, n = 5/group represents five independent animals rather than individual microscopic fields. Image acquisition and ImageJ-based quantification were conducted by an investigator blinded to the experimental group allocation. Statistical analysis The required sample size was estimated using G*Power software (version 3.1.9.7) for a one-way ANOVA comparing six independent groups. Assuming a significance level (α) of 0.05, statistical power of 80%, and an expected effect size (Cohen’s f) of [insert justified effect size], the analysis indicated a minimum total sample size of [insert calculated n]. Fifteen animals were allocated to each group (total n = 90) to ensure adequate statistical power and accommodate the planned biochemical, molecular, and histopathological assessments. Initial data processing was performed using Microsoft Excel. The distribution of continuous data was assessed using the Shapiro–Wilk test, and homogeneity of variances was evaluated using Levene’s test. Differences among experimental groups were analyzed using one-way analysis of variance (ANOVA) followed by Tukey’s honestly significant difference (HSD) post hoc test for multiple pairwise comparisons when the overall ANOVA was significant. Statistical analyses were performed using SAS 2012. Data are presented consistently as mean ± standard error of the mean (SEM). The individual animal was considered the biological replicate. Unless otherwise specified, n = 10/group for biochemical and molecular analyses and n = 5/group for histopathological and immunohistochemical analyses. Technical measurements or multiple microscopic fields obtained from the same animal were not treated as independent biological replicates. Exact p values were reported where available, with p < 0.05 considered statistically significant. Where software output provided values below the reporting limit, these were reported using the corresponding threshold (e.g., p < 0.0001). GraphPad Prism 9.0 (GraphPad Software, USA) was used for graphical presentation of data. Superscript symbols used in figures were defined explicitly in each figure legend according to the corresponding comparison (e.g., ^a significantly different from the Control group; ^b significantly different from the NA–STZ group; ^c significantly different from the NA–STZ + IC group), rather than using the ambiguous expression “versus control (a), STZ (b), and crude IC (c).” Results Physicochemical characterization of IC-LNPs The IC-LNPs exhibited a predominantly spherical, well-dispersed morphology under TEM, with particle sizes ranging from 35 to 83 nm (Fig. 1A, B). DLS showed a larger mean hydrodynamic diameter of 160 nm with a PDI of 0.155 (Fig. 1C). This difference is expected because TEM measures dried particle dimensions, whereas DLS measures the hydrodynamic diameter of particles in aqueous dispersion, including the associated hydration layer. The zeta potential of IC-LNPs was − 35.65 mV (Fig. 1D), indicating a negative surface charge that may contribute to colloidal stability through electrostatic repulsion. Fig. 1(A) TEM image of IC-LNPs demonstrating nearly spherical structure; (B) particle size histogram (35–83 nm distribution); (C) volume-based particle size profile; (D) zeta potential distribution profile (A) TEM image of IC-LNPs demonstrating nearly spherical structure; (B) particle size histogram (35–83 nm distribution); (C) volume-based particle size profile; (D) zeta potential distribution profile In vitro release profile of IC and IC-LNPs The in vitro release study demonstrated a clear difference between free IC and IC-LNPs. Free IC exhibited a rapid burst release, reaching approximately 45% within the first hour and nearly complete release (about 100%) within 12 h. This indicates the absence of any sustained release mechanism and reflects the immediate diffusion of the drug into the surrounding medium. In contrast, IC-LNPs showed a markedly slower and more controlled release profile. Only a small fraction of IC was released during the initial phase, accounting for about 8% at 1 h. This was followed by a gradual and sustained release over time, reaching approximately 65% at 24 h. The release continued steadily, reaching about 92% after 72 h. This sustained pattern suggests that IC was successfully encapsulated within the lipid bilayer, which effectively delayed its diffusion into the release medium (Fig. 2A). Encapsulation efficiency and drug loading of IC-LNPs The IC-LNPs showed high encapsulation efficiency of 86.54 ± 2.52%, indicating successful entrapment of Icariin within the lipid bilayer. Drug loading was 9.63 ± 0.69%, reflecting effective incorporation of Icariin into the liposomal structure. The IC-equivalent concentration of the final IC-LNP formulation was 40 mg/mL. This concentration provided the target dose of 80 mg/kg at an injection volume of 2 mL/kg. Overall, the formulation demonstrated high encapsulation efficiency and effective drug incorporation suitable for further in vivo application. Formulation quality and injection suitability The formulation showed a pH of 7.32 and an osmolality of 297 mOsm/kg. Endotoxin content was 0.14 EU/mL. Sterility testing showed no detectable microbial growth. These findings indicated that the formulation was suitable for repeated intraperitoneal administration under the experimental conditions. FTIR analysis of IC-LNPs The FTIR spectra of free IC, cholesterol, L-α-phosphatidylcholine, and IC-LNPs were analyzed to evaluate possible chemical interactions and support the incorporation of IC into the lipid formulation. IC exhibited characteristic absorption bands at 3410–3380 cm⁻¹ corresponding to O–H stretching of phenolic groups, 2930–2850 cm⁻¹ attributed to C–H stretching, 1660–1600 cm⁻¹ assigned to C = O stretching, and 1260–1020 cm⁻¹ related to C–O–C and C–O stretching vibrations of glycosidic linkages. Cholesterol showed a broad O–H stretching band around 3420 cm⁻¹, along with C–H stretching at 2935–2865 cm⁻¹ and C–O vibrations near 1050–1000 cm⁻¹, while L-α-phosphatidylcholine displayed prominent peaks at 1740–1730 cm⁻¹ (ester C = O), 1240–1220 cm⁻¹ (P = O stretching), and 1085–1060 cm⁻¹ (P–O–C stretching). In the IC-LNPs spectrum, notable shifts and broadening were observed in the O–H stretching region (3400–3350 cm⁻¹) and carbonyl-associated bands (1735–1720 cm⁻¹), along with a reduction in the intensity of phosphate-related peaks. These spectral changes may reflect physical interactions between IC and the lipid components. The observed spectral profile was consistent with the incorporation of IC into the lipid formulation. The absence of new characteristic peaks suggested that no major new chemical groups were formed during formulation (Fig. 2B). Fig. 2(A) In vitro cumulative release profile of free Icariin (IC) and Icariin-loaded liposomal nanoparticles (IC-LNPs) was evaluated over time. The study demonstrated a rapid burst release pattern from free IC under physiological conditions. In contrast, IC-LNPs showed a sustained and controlled release behavior. The experiment was conducted in PBS (pH 7.4) at 37 ± 0.5 °C. (B) FTIR spectra of pure IC, cholesterol, L-α-phosphatidylcholine, and IC-LNPs (4000–400 cm⁻¹); IC showed characteristic bands at 3410 cm⁻¹ (O–H), 2930–2850 cm⁻¹ (C–H), 1660 cm⁻¹ (C = O), and 1260–1020 cm⁻¹ (C–O–C/C–O). Lipid components exhibited their specific functional group peaks as described. In IC-LNPs, shifts and broadening of O–H (3380 cm⁻¹) and C = O (1725 cm⁻¹) bands, along with reduced phosphate peak intensity, confirmed successful physical encapsulation and intermolecular hydrogen bonding without new covalent bond formation (A) In vitro cumulative release profile of free Icariin (IC) and Icariin-loaded liposomal nanoparticles (IC-LNPs) was evaluated over time. The study demonstrated a rapid burst release pattern from free IC under physiological conditions. In contrast, IC-LNPs showed a sustained and controlled release behavior. The experiment was conducted in PBS (pH 7.4) at 37 ± 0.5 °C. (B) FTIR spectra of pure IC, cholesterol, L-α-phosphatidylcholine, and IC-LNPs (4000–400 cm⁻¹); IC showed characteristic bands at 3410 cm⁻¹ (O–H), 2930–2850 cm⁻¹ (C–H), 1660 cm⁻¹ (C = O), and 1260–1020 cm⁻¹ (C–O–C/C–O). Lipid components exhibited their specific functional group peaks as described. In IC-LNPs, shifts and broadening of O–H (3380 cm⁻¹) and C = O (1725 cm⁻¹) bands, along with reduced phosphate peak intensity, confirmed successful physical encapsulation and intermolecular hydrogen bonding without new covalent bond formation Effect of IC and IC-LNPs on glycemic control, body weight, liver function, and lipid profile in STZ-induced diabetic rats Fasting blood glucose levels (Fig. 3A) were significantly elevated following STZ exposure compared with the normal control group (p < 0.05). Co-treatment with either IC or IC-LNPs significantly reduced blood glucose levels, with significantly lower values observed in the STZ + IC-LNPs group than in the STZ + crude IC group. In contrast, serum insulin levels (Fig. 3B) were significantly decreased following STZ induction compared with the normal control group. Administration of IC-LNPs alongside STZ significantly improved insulin levels, with significantly higher values than those observed in the STZ + crude IC group. Final body weight was significantly decreased in the STZ-treated group compared with the normal control group. Co-treatment with IC-LNPs significantly improved final body weight more effectively than crude IC. Notably, no significant difference was observed between the STZ + IC-LNPs group and the normal control group (p > 0.05; Fig. 3C). Regarding liver function, STZ exposure significantly increased serum activities of liver enzymes, including AST (Fig. 3D) and ALT (Fig. 3E), compared with the normal control group. Treatment with IC-LNPs alongside STZ significantly alleviated these elevations, restoring enzyme activities to levels that did not differ significantly from those of the normal control group (p > 0.05). Regarding the lipid profile, the STZ-treated group showed altered lipid levels compared with the normal control group (p < 0.05). Co-treatment with either crude IC or IC-LNPs significantly improved lipid parameters in STZ-induced diabetic rats. TC (Fig. 3F) and TG (Fig. 3G) levels were significantly reduced, with a greater improvement observed in the IC-LNPs-treated group compared with the crude IC group. Notably, no significant difference was observed between the STZ + IC-LNPs group and the normal control group in serum TG levels (p > 0.05). Fig. 3Effects of IC and IC-LNPs on glycemic parameters, body weight, liver enzymes, and lipid profile in NA–STZ-induced diabetic rats. (A) Fasting blood glucose (FBG); (B) serum insulin; (C) final body weight; (D) aspartate aminotransferase (AST); (E) alanine aminotransferase (ALT); (F) total cholesterol (TC); and (G) triglycerides (TG). Data are presented as mean ± SEM (n = 10 animals/group). Overall p values displayed in the panels were obtained using one-way ANOVA, followed by Tukey’s HSD test for multiple comparisons. Superscript letters indicate significant pairwise differences (p < 0.05): ^a significantly different from the Control group; ^b significantly different from the NA–STZ group; and ^c significantly different from the corresponding crude IC-treated group Effects of IC and IC-LNPs on glycemic parameters, body weight, liver enzymes, and lipid profile in NA–STZ-induced diabetic rats. (A) Fasting blood glucose (FBG); (B) serum insulin; (C) final body weight; (D) aspartate aminotransferase (AST); (E) alanine aminotransferase (ALT); (F) total cholesterol (TC); and (G) triglycerides (TG). Data are presented as mean ± SEM (n = 10 animals/group). Overall p values displayed in the panels were obtained using one-way ANOVA, followed by Tukey’s HSD test for multiple comparisons. Superscript letters indicate significant pairwise differences (p < 0.05): ^a significantly different from the Control group; ^b significantly different from the NA–STZ group; and ^c significantly different from the corresponding crude IC-treated group Effect of IC and IC-LNPs on hepatic NRF2/HO-1 and oxidative stress markers in STZ-induced diabetic rats Exposure to STZ significantly suppressed the hepatic NRF2/HO-1 signaling pathway, as demonstrated by a marked reduction in hepatic NRF2 protein levels (Fig. 4A), NRF2 mRNA expression (Fig. 4B), and HO-1 protein levels (Fig. 4C) compared with the normal control group (p < 0.05). Administration of both crude IC and IC-LNPs significantly elevated these parameters relative to the STZ group, with the nanoliposomal formulation showing greater efficacy than the crude form. Notably, no significant difference was detected between the STZ + IC-LNPs group and the normal control group regarding hepatic NRF2 protein and mRNA expression levels (p > 0.05). STZ administration also markedly impaired the hepatic antioxidant defense system, as evidenced by significant reductions in TAC (Fig. 4D), reduced GSH levels (Fig. 4E), and the activities of antioxidant enzymes, including GPX (Fig. 4F), SOD (Fig. 4G), and CAT (Fig. 4H), compared with the normal control group. Treatment with IC-LNPs significantly improved these parameters compared with crude IC administration. In particular, IC-LNPs restored hepatic GSH content and SOD activity to levels comparable to those of the normal control group, with no significant differences observed between both groups (p > 0.05). In contrast, no significant differences were detected between the diabetic group and the diabetic group treated with crude IC regarding CAT and GPX activities. Regarding oxidative stress biomarkers, STZ exposure significantly increased hepatic MDA levels (Fig. 4I) and H₂O₂ content (Fig. 4J) compared with the normal control group. Treatment with both forms of IC significantly reduced hepatic MDA and H₂O₂ levels relative to the diabetic group. However, the reductions were significantly greater in the IC-LNPs-treated group than in the crude IC-treated group. Furthermore, no significant difference was observed between the STZ + IC-LNPs group and the normal control group in hepatic H₂O₂ levels (p > 0.05). Fig. 4Effect of IC and IC-LNPs on hepatic NRF2/HO-1 signaling pathway and hepatic oxidative stress biomarkers in STZ-induced diabetic rats (A: NRF2, Nuclear factor erythroid 2–related factor 2, protein level; B: Nfe2l2 mRNA expression; C: HO-1, Heme oxygenase-1, protein level; D: TAC, Total Antioxidant Capacity; E: GSH, Reduced Glutathione; F: GPX, Glutathione Peroxidase; G: SOD, Superoxide Dismutase; H: CAT, Catalase; I: MDA, Malondialdehyde; J: H₂O₂, Hydrogen Peroxide). Data are presented as mean ± SEM (n = 10 animals/group). Overall p values displayed in the panels were obtained using one-way ANOVA, followed by Tukey’s HSD test for multiple comparisons. Superscript letters indicate significant pairwise differences (p < 0.05): ^a significantly different from the Control group; ^b significantly different from the NA–STZ group; and ^c significantly different from the corresponding crude IC-treated group Effect of IC and IC-LNPs on hepatic NRF2/HO-1 signaling pathway and hepatic oxidative stress biomarkers in STZ-induced diabetic rats (A: NRF2, Nuclear factor erythroid 2–related factor 2, protein level; B: Nfe2l2 mRNA expression; C: HO-1, Heme oxygenase-1, protein level; D: TAC, Total Antioxidant Capacity; E: GSH, Reduced Glutathione; F: GPX, Glutathione Peroxidase; G: SOD, Superoxide Dismutase; H: CAT, Catalase; I: MDA, Malondialdehyde; J: H₂O₂, Hydrogen Peroxide). Data are presented as mean ± SEM (n = 10 animals/group). Overall p values displayed in the panels were obtained using one-way ANOVA, followed by Tukey’s HSD test for multiple comparisons. Superscript letters indicate significant pairwise differences (p < 0.05): ^a significantly different from the Control group; ^b significantly different from the NA–STZ group; and ^c significantly different from the corresponding crude IC-treated group Effect of IC and IC-LNPs on hepatic NF-κB/COX-2 signaling pathway and pro-inflammatory cytokines in STZ-induced diabetic rats Induction of diabetes significantly activated the hepatic NF-κB signaling pathway, as evidenced by a significant increase in NF-κB protein (Fig. 5A) and mRNA expression (Fig. 5B), as well as COX-2 protein levels (Fig. 5C), compared with the negative control group (p < 0.05). Treatment with IC-LNPs in combination with STZ significantly reduced NF-κB and COX-2 levels, with a more pronounced effect than that observed with crude IC administration. Notably, no significant difference was observed between the STZ + IC-LNPs group and the normal control group in hepatic COX-2 levels. Regarding pro-inflammatory cytokines, STZ exposure significantly elevated hepatic TNF-α (Fig. 5D), IL-6 (Fig. 5E), and IL-1β (Fig. 5F) compared with the normal control group. Co-treatment with IC-LNPs significantly reduced the levels of these cytokines compared with the diabetic group. Notably, no significant difference was observed between the STZ group and the STZ + crude IC group in hepatic IL-6 levels. In addition, no significant difference was detected between the STZ + IC-LNPs group and the crude IC-treated group in hepatic IL-1β levels. Fig. 5Effect of IC and IC-LNPs on hepatic NF-κB/COX-2 signaling pathway, pro-inflammatory cytokines, and apoptosis markers in STZ-induced diabetic rats (A: NF-κB, Nuclear factor kappa B, protein level; B: NF-κB mRNA expression; C: COX-2, Cyclooxygenase-2 level; D: TNF-α, Tumor necrosis factor-alpha; E: IL-6, Interleukin-6; F: IL-1β, Interleukin-1 beta; G: Bax, Bcl-2–associated X protein; H: Caspase-3; I: Bcl-2, B-cell lymphoma 2). Data are presented as mean ± SEM (n = 10 animals/group). Overall p values displayed in the panels were obtained using one-way ANOVA, followed by Tukey’s HSD test for multiple comparisons. Superscript letters indicate significant pairwise differences (p < 0.05): ^a significantly different from the Control group; ^b significantly different from the NA–STZ group; and ^c significantly different from the corresponding crude IC-treated group Effect of IC and IC-LNPs on hepatic NF-κB/COX-2 signaling pathway, pro-inflammatory cytokines, and apoptosis markers in STZ-induced diabetic rats (A: NF-κB, Nuclear factor kappa B, protein level; B: NF-κB mRNA expression; C: COX-2, Cyclooxygenase-2 level; D: TNF-α, Tumor necrosis factor-alpha; E: IL-6, Interleukin-6; F: IL-1β, Interleukin-1 beta; G: Bax, Bcl-2–associated X protein; H: Caspase-3; I: Bcl-2, B-cell lymphoma 2). Data are presented as mean ± SEM (n = 10 animals/group). Overall p values displayed in the panels were obtained using one-way ANOVA, followed by Tukey’s HSD test for multiple comparisons. Superscript letters indicate significant pairwise differences (p < 0.05): ^a significantly different from the Control group; ^b significantly different from the NA–STZ group; and ^c significantly different from the corresponding crude IC-treated group Effect of IC and IC-LNPs on hepatic apoptotic markers in STZ-induced diabetic rats Induction of diabetes with STZ significantly elevated hepatic pro-apoptotic markers, including Bax (Fig. 5F) and caspase-3 (Fig. 5H), compared with the normal control group (p < 0.05). Co-treatment with either crude IC or IC-LNPs significantly reduced the levels of these markers relative to the diabetic group, with a more pronounced reduction observed in the IC-LNPs-treated group. Notably, no significant difference was detected between the STZ + IC-LNPs group and the normal control group in hepatic caspase-3 levels. In contrast, the anti-apoptotic biomarker Bcl-2 (Fig. 5I) was significantly decreased following STZ induction compared with the normal control group. Administration of IC-LNPs alongside STZ significantly improved Bcl-2 levels more effectively than crude IC treatment, with no significant difference observed between the STZ + IC-LNPs group and the normal control group. Effect of IC and IC-LNPs on hepatic iron homeostasis and ferroptosis-related markers in STZ-induced diabetic rats Hepatic Fe²⁺ levels (Fig. 6A) were significantly elevated in the NA–STZ group compared with the Control group (p < 0.05). Treatment with either form of IC significantly reduced Fe²⁺ levels, with significantly lower values in the NA–STZ + IC-LNPs group than in the NA–STZ + IC group (p < 0.05). Notably, no significant difference was observed between the NA–STZ + IC-LNPs and Control groups (p > 0.05). Similarly, hepatic ACSL4 mRNA expression (Fig. 6B) and TFR1 protein levels (Fig. 6C) and mRNA expression (Fig. 6D) were significantly increased in the NA–STZ group compared with the Control group. Both IC treatments significantly attenuated these increases, with IC-LNPs producing significantly greater effects than crude IC (p < 0.05). Regarding ferroptosis-related protective markers, NA–STZ significantly decreased hepatic GPX4 mRNA expression (Fig. 6E) and SLC7A11 protein levels (Fig. 6F) and mRNA expression (Fig. 6G) compared with the Control group (p < 0.05). IC-LNP treatment significantly increased GPX4 mRNA expression compared with the NA–STZ group, restoring it to a level that was not significantly different from the Control group (p > 0.05). Both crude IC and IC-LNPs significantly increased SLC7A11 protein and mRNA levels compared with the NA–STZ group, with significantly greater increases following IC-LNP treatment than following crude IC treatment (p < 0.05). Collectively, these findings demonstrate that IC, particularly IC-LNPs, attenuated NA–STZ-induced alterations in multiple ferroptosis-related markers. These changes support an association with modulation of ferroptosis-related processes but do not, in isolation, establish ferroptotic cell death. Fig. 6Effect of IC and IC-LNPs on hepatic iron overload and ferroptosis-related markers in STZ-induced diabetic rats (A: Fe²⁺, Ferrous iron; B: ACSL4, Acyl-CoA synthetase long-chain family member 4 mRNA expression; C: TFR1, Transferrin receptor 1 protein level; D: TFR1 mRNA expression; E: GPX4, Glutathione peroxidase 4 expression; F: SLC7A11, Solute carrier family 7 member 11 protein level; G: SLC7A11 mRNA expression). Data are presented as mean ± SEM (n = 10 animals/group). Overall p values displayed in the panels were obtained using one-way ANOVA, followed by Tukey’s HSD test for multiple comparisons. Superscript letters indicate significant pairwise differences (p < 0.05): ^a significantly different from the Control group; ^b significantly different from the NA–STZ group; and ^c significantly different from the corresponding crude IC-treated group Effect of IC and IC-LNPs on hepatic iron overload and ferroptosis-related markers in STZ-induced diabetic rats (A: Fe²⁺, Ferrous iron; B: ACSL4, Acyl-CoA synthetase long-chain family member 4 mRNA expression; C: TFR1, Transferrin receptor 1 protein level; D: TFR1 mRNA expression; E: GPX4, Glutathione peroxidase 4 expression; F: SLC7A11, Solute carrier family 7 member 11 protein level; G: SLC7A11 mRNA expression). Data are presented as mean ± SEM (n = 10 animals/group). Overall p values displayed in the panels were obtained using one-way ANOVA, followed by Tukey’s HSD test for multiple comparisons. Superscript letters indicate significant pairwise differences (p < 0.05): ^a significantly different from the Control group; ^b significantly different from the NA–STZ group; and ^c significantly different from the corresponding crude IC-treated group Effect of IC and IC-LNPs on hepatic histopathological alterations in STZ-induced diabetic rats Microscopic findings revealed intact hepatic architecture in the crude IC- and IC-LNPs-treated groups, with no detectable evidence of tissue injury or abnormalities. The normal structural arrangement of hepatic lobules was maintained throughout the examined sections. Preserved central veins and hepatocytes with characteristic morphology and eosinophilic cytoplasmic staining were clearly observed (Figs. 7A–C). Significant hepatic pathological changes were detected in STZ-induced diabetic rats, as evidenced by nuclear degeneration, hepatocellular vacuolation, and dilated congested central veins (Fig. 7D). Treatment with either crude IC or IC-LNPs substantially reversed the histopathological lesions, resulting in organized hepatocyte cords and near-normal hepatic architecture (Figs. 7E–F). Histopathological scoring revealed significantly higher liver lesion values in the diabetic model group than in the normal control group (p < 0.05). In contrast, both crude IC and IC-LNP administration significantly reduced histopathological scores (p < 0.05). This reduction indicated a stronger hepatoprotective effect in the IC-LNP-treated diabetic group (Fig. 7G). Fig. 7Representative photomicrographs of liver tissue from control and different experimental groups: (A) control group; (B) IC-treated group; (C) IC-LNPs-treated group; (D) STZ group; (E) STZ + IC group; (F) STZ + IC-LNPs group; (CV) central vein; (S) sinusoids; (H) hepatocytes; (V) vacuolated cytoplasm (red arrow); perivascular lymphocytic infiltration. All images were captured at 400× magnification, scale bar = 50 μm. (G) Semi-quantitative hepatic histopathological injury scores. Five animals per group were evaluated (n = 5/group), with three liver sections and four randomly selected microscopic fields per section examined for each animal. Field-level scores were averaged to generate a single animal-level score. Data are presented as mean ± SEM. Superscript letters indicate significant pairwise differences (p < 0.05): ^a significantly different from the Control group; ^b significantly different from the NA–STZ group; and ^c significantly different from the corresponding crude IC-treated group Representative photomicrographs of liver tissue from control and different experimental groups: (A) control group; (B) IC-treated group; (C) IC-LNPs-treated group; (D) STZ group; (E) STZ + IC group; (F) STZ + IC-LNPs group; (CV) central vein; (S) sinusoids; (H) hepatocytes; (V) vacuolated cytoplasm (red arrow); perivascular lymphocytic infiltration. All images were captured at 400× magnification, scale bar = 50 μm. (G) Semi-quantitative hepatic histopathological injury scores. Five animals per group were evaluated (n = 5/group), with three liver sections and four randomly selected microscopic fields per section examined for each animal. Field-level scores were averaged to generate a single animal-level score. Data are presented as mean ± SEM. Superscript letters indicate significant pairwise differences (p < 0.05): ^a significantly different from the Control group; ^b significantly different from the NA–STZ group; and ^c significantly different from the corresponding crude IC-treated group Effect of IC and IC-LNPs on hepatic ultrastructural alterations in STZ-induced diabetic rats Electron microscopic examination showed that hepatocytes in the control group (Fig. 8A), as well as in the crude IC-treated group (Fig. 8B), retained their typical ultrastructural organization and integrity. Similar preservation of ultrastructure was also observed in the IC-LNPs-treated group (Fig. 8C). The cells possessed centrally placed nuclei with conspicuous nucleoli, clearly outlined plasma membranes, and uniform chromatin distribution. Mitochondria appeared intact with well-preserved cristae, and the rough endoplasmic reticulum was properly arranged. Clearly defined bile canaliculi and sinusoidal spaces suggested intact liver ultrastructure and maintenance of cellular homeostasis. Diabetic hepatic tissues analyzed by TEM showed profound necrotic changes in hepatocytes, characterized by condensed chromatin within the nucleus, widespread cytoplasmic vacuolation, and prominent rough endoplasmic reticulum dilation. Mitochondria showed marked swelling and severe cristae disruption, whereas the nuclei appeared misshapen and irregular (Fig. 8D). The hepatocellular ultrastructure was substantially preserved following co-administration of crude IC or IC-LNPs, with features approaching normal organization. Hepatocytes retained normal ultrastructural features, including intact mitochondria with minimal degenerative changes, organized rough endoplasmic reticulum, and distinct nuclei and nucleoli (Figs. 8E–F). Fig. 8Liver ultrastructural representative images from control and treated groups showing: (A) control group; (B) IC-treated group; (C) IC-LNPs-treated group; (D) STZ group; (E) STZ + IC group; (F) STZ + IC-LNPs group; N: nucleus; M: mitochondria; RER: rough endoplasmic reticulum; V: cytoplasmic vacuolation; C: collagen fibers. The images illustrate representative hepatocellular ultrastructural features and were used for qualitative morphological assessment Liver ultrastructural representative images from control and treated groups showing: (A) control group; (B) IC-treated group; (C) IC-LNPs-treated group; (D) STZ group; (E) STZ + IC group; (F) STZ + IC-LNPs group; N: nucleus; M: mitochondria; RER: rough endoplasmic reticulum; V: cytoplasmic vacuolation; C: collagen fibers. The images illustrate representative hepatocellular ultrastructural features and were used for qualitative morphological assessment Effect of IC and IC-LNPs on hepatic ACSL4 immunoreactivity in STZ-induced diabetic rats ACSL4 immunostaining was not detectable in the liver sections of the control group, crude IC group, and IC-LNPs-treated group (Fig. 9A–C). In contrast, the STZ group exhibited a marked increase in ACSL4 immunoreactivity, characterized by strong brown staining predominantly localized in hepatocytes (Fig. 9D). This elevated expression indicates pronounced hepatic injury associated with STZ exposure. On the other hand, both crude IC and IC-LNPs co-administered with STZ showed a clear reduction in ACSL4 expression compared with the STZ-only group. Weak hepatocellular immunostaining was observed in the treated groups (Fig. 9E–F). The reduction in ACSL4 immunoreactivity was consistently evident across liver sections. Notably, ACSL4 expression in the treated groups appeared nearly comparable to that of the control group. A marked elevation in ACSL4 immunoscore was detected in the STZ group compared with control (p < 0.05). This increase was significantly attenuated by IC and IC–LNPs administration, particularly in the nanoliposomal group (Fig. 9G). Fig. 9Immunohistochemical detection of hepatic ACSL4 in STZ-induced diabetic rats. (A–C) Control group showing negative ACSL4 expression; (D) STZ group exhibiting strong positive staining in hepatocytes (brown, arrow), indicating hepatic injury; (E–F) treated groups demonstrating reduced ACSL4 expression with faint staining comparable to control (400× magnification). (G) Quantitative analysis of ACSL4-positive immunostaining using ImageJ. Five randomly selected, non-overlapping fields were analyzed per animal, and field-level measurements were averaged to obtain one value for each animal (n = 5 animals/group). Data are presented as mean ± SEM. Superscript letters indicate significant pairwise differences (p < 0.05): ^a significantly different from the Control group; ^b significantly different from the NA–STZ group; and ^c significantly different from the corresponding crude IC-treated group Immunohistochemical detection of hepatic ACSL4 in STZ-induced diabetic rats. (A–C) Control group showing negative ACSL4 expression; (D) STZ group exhibiting strong positive staining in hepatocytes (brown, arrow), indicating hepatic injury; (E–F) treated groups demonstrating reduced ACSL4 expression with faint staining comparable to control (400× magnification). (G) Quantitative analysis of ACSL4-positive immunostaining using ImageJ. Five randomly selected, non-overlapping fields were analyzed per animal, and field-level measurements were averaged to obtain one value for each animal (n = 5 animals/group). Data are presented as mean ± SEM. Superscript letters indicate significant pairwise differences (p < 0.05): ^a significantly different from the Control group; ^b significantly different from the NA–STZ group; and ^c significantly different from the corresponding crude IC-treated group Effect of IC and IC-LNPs on Hepatic ACSL4 immunolocalization in STZ-induced diabetic rats GPX4 immunostaining was strongly positive in the liver sections of the control, crude IC, and IC-LNPs-treated groups (Fig. 10A–C), indicating abundant hepatocellular expression. In contrast, the STZ group exhibited a marked reduction in GPX4 immunoreactivity, characterized by weak (mild) cytoplasmic staining in hepatocytes (Fig. 10D). Co-treatment with either crude IC or IC-LNPs resulted in a clear restoration of GPX4 expression compared with the STZ group. The treated groups showed markedly increased GPX4 immunostaining within hepatocytes (Fig. 10E–F), indicating strong immunoreactivity. Overall, IC and IC-LNPs effectively reversed STZ-induced downregulation of hepatic GPX4 expression. STZ exposure resulted in a significant decrease in GPX4 immunoscore versus control (p < 0.05). IC and IC–LNPs significantly improved this parameter, with IC–LNPs reaching levels not significantly different from control (Fig. 10G). Fig. 10R Immunohistochemical detection of hepatic GPX4 in STZ-induced diabetic rats. (A–C) Liver sections from the control group exhibited strong positive GPX4 immunoreactivity within hepatocytes. (D) The STZ group showed reduced (mild) GPX4 immunostaining in hepatocytes. (E–F) In the treated groups, GPX4 immunostaining was markedly increased, indicating strong immunoreactivity (×400 magnification). (G) Quantitative analysis of GPX4-positive immunostaining using ImageJ. Five randomly selected, non-overlapping fields were analyzed per animal, and field-level measurements were averaged to obtain one value for each animal (n = 5 animals/group). Data are presented as mean ± SEM. Superscript letters indicate significant pairwise differences (p < 0.05): ^a significantly different from the Control group; ^b significantly different from the NA–STZ group; and ^c significantly different from the corresponding crude IC-treated group R Immunohistochemical detection of hepatic GPX4 in STZ-induced diabetic rats. (A–C) Liver sections from the control group exhibited strong positive GPX4 immunoreactivity within hepatocytes. (D) The STZ group showed reduced (mild) GPX4 immunostaining in hepatocytes. (E–F) In the treated groups, GPX4 immunostaining was markedly increased, indicating strong immunoreactivity (×400 magnification). (G) Quantitative analysis of GPX4-positive immunostaining using ImageJ. Five randomly selected, non-overlapping fields were analyzed per animal, and field-level measurements were averaged to obtain one value for each animal (n = 5 animals/group). Data are presented as mean ± SEM. Superscript letters indicate significant pairwise differences (p < 0.05): ^a significantly different from the Control group; ^b significantly different from the NA–STZ group; and ^c significantly different from the corresponding crude IC-treated group Discussion Diabetic rats exhibited marked hyperglycemia accompanied by reduced insulin levels, confirming successful establishment of the diabetic state. Persistent hyperglycemia is a major driving force in diabetic hepatopathy because it promotes excessive ROS generation through several interconnected mechanisms, including glucose autoxidation, mitochondrial electron transport chain dysfunction, polyol pathway activation, and advanced glycation end-product (AGE) formation. These processes collectively impair insulin signaling, promote hepatic lipid accumulation, and induce hepatocellular dysfunction through oxidative and inflammatory stress pathways [38, 39]. IC-LNPs produced greater glycemic improvement than crude IC, suggesting that liposomal formulation enhanced the pharmacological efficacy of IC under the present experimental conditions. However, because pharmacokinetic and cellular uptake studies were not performed, this greater efficacy cannot be directly attributed to enhanced bioavailability or cellular uptake. Improved insulin sensitivity may additionally suppress hepatic gluconeogenesis and lipogenesis, thereby reducing glucotoxicity- and lipotoxicity-mediated oxidative injury [40, 41]. The marked elevation of ALT and AST levels in diabetic rats further reflects hepatocellular membrane disruption and mitochondrial injury induced by chronic hyperglycemia and lipid peroxidation. Oxidative stress–mediated mitochondrial dysfunction is known to impair ATP production and destabilize hepatocyte membrane integrity, thereby facilitating leakage of hepatic enzymes into circulation [42, 43]. The reduction in liver enzyme activities following IC-LNP treatment was consistent with attenuation of hepatocellular injury and was supported by the histopathological and ultrastructural findings. Similarly, the improvement in serum lipid profile following IC-LNP administration may indicate improved hepatic lipid homeostasis associated with better glycemic control and attenuation of oxidative stress [44]. Since hepatic lipid accumulation itself amplifies ROS generation and inflammatory signaling, correction of dyslipidemia may have contributed indirectly to the observed hepatoprotective effects. Oxidative stress appears to represent a central upstream event linking hyperglycemia to inflammation, apoptosis, and ferroptosis-related alterations in diabetic liver injury. Diabetic rats demonstrated marked elevations in H₂O₂ and MDA levels together with depletion of endogenous antioxidant systems, including GSH, SOD, CAT, GPx, and TAC. Persistent ROS accumulation can oxidize membrane phospholipids, proteins, and mitochondrial DNA, thereby amplifying mitochondrial dysfunction and triggering inflammatory and cell death pathways [6, 45, 46]. Importantly, IC-LNPs markedly restored antioxidant defenses while reducing oxidative stress markers, consistent with improved hepatic redox homeostasis. These changes were accompanied by increased hepatic NRF2 and HO-1 levels, suggesting modulation of NRF2/HO-1-associated antioxidant signaling. NRF2 is a master transcriptional regulator of antioxidant defense genes, and its activation promotes transcription of cytoprotective enzymes involved in ROS detoxification, glutathione synthesis, iron homeostasis, and maintenance of mitochondrial function [14, 47]. However, because nuclear NRF2 translocation or other direct measures of NRF2 activation were not assessed, the present findings demonstrate an association with NRF2/HO-1 signaling rather than definitive pathway activation. This response may contribute to the observed attenuation of oxidative, inflammatory, apoptotic, and ferroptosis-related alterations [12, 48]. The inflammatory response observed in diabetic liver tissue was evidenced by significant elevation of TNF-α, IL-1β, IL-6, NF-κB, and COX-2. Chronic oxidative stress can activate NF-κB signaling through ROS-dependent phosphorylation pathways, resulting in transcription of pro-inflammatory cytokines and amplification of hepatic inflammation [49]. Importantly, inflammatory cytokines themselves further stimulate ROS generation and mitochondrial dysfunction, thereby establishing a self-perpetuating oxidative stress–inflammation cycle that exacerbates hepatic injury [50]. The marked reduction in inflammatory mediators by IC-LNPs is consistent with attenuation of this inflammatory response. This effect was accompanied by increased NRF2/HO-1 levels and reduced nuclear NF-κB levels, suggesting coordinated modulation of redox-sensitive antioxidant and inflammatory signaling [51]. However, because phospho-p65, IκBα degradation, and NF-κB DNA-binding/transcriptional activity were not assessed, these findings should be interpreted as changes associated with NF-κB signaling rather than definitive evidence of pathway activation or inhibition. Thus, IC-LNPs may exert hepatoprotective effects through concurrent modulation of oxidative and inflammatory processes. Apoptosis-related alterations were markedly activated in diabetic liver tissue, as demonstrated by increased Bax and caspase-3 expression together with reduced Bcl-2 levels. Oxidative stress–induced mitochondrial dysfunction may contribute to this process. Excess ROS can induce mitochondrial membrane depolarization, opening of mitochondrial permeability transition pores, and release of cytochrome c into the cytosol, ultimately activating the intrinsic apoptotic pathway [7, 52]. In addition, chronic hyperglycemia and inflammatory cytokines further sensitize hepatocytes to apoptosis through disruption of mitochondrial bioenergetics and activation of stress-related signaling pathways [53, 54]. The restoration of Bcl-2 and reduction of Bax and caspase-3 following IC-LNP treatment were consistent with attenuation of apoptosis-related signaling. Together with the ultrastructural findings, these changes suggest that reduced oxidative stress and mitochondrial injury may contribute to the observed effects of IC-LNPs [55, 56]. An important finding of the present study was the modulation of multiple ferroptosis-related markers in diabetic liver tissue. Ferroptosis is strongly linked to oxidative stress and iron dysregulation and is characterized by iron-dependent phospholipid peroxidation. In diabetic rats, reduced GPX4 and SLC7A11 expression together with increased ACSL4, TFR1, and Fe²⁺ levels were consistent with alterations associated with ferroptosis. However, these markers alone do not definitively establish ferroptotic cell death. Mechanistically, impaired SLC7A11 expression can reduce cystine uptake and glutathione synthesis, resulting in diminished GPX4-dependent detoxification of lipid hydroperoxides. Concurrently, increased ACSL4 promotes incorporation of polyunsaturated fatty acids into membrane phospholipids, increasing susceptibility to lipid peroxidation, whereas elevated TFR1 can enhance intracellular iron availability and promote Fenton reaction–mediated ROS generation [9, 57]. Collectively, these processes are recognized contributors to cellular susceptibility to ferroptosis [58, 59]. Importantly, IC-LNPs attenuated these ferroptosis-related alterations, increasing GPX4 and SLC7A11 while reducing ACSL4, TFR1, and Fe²⁺ levels. These findings are consistent with improved glutathione-dependent antioxidant defense and iron homeostasis and reduced susceptibility to lipid peroxidation. The relationship between NRF2 signaling and ferroptosis is particularly relevant because NRF2 regulates several genes involved in glutathione metabolism, iron homeostasis, and antioxidant defense [60]. The concurrent increase in NRF2/HO-1 and GPX4/SLC7A11-related responses therefore suggests a potential association between NRF2-related antioxidant signaling and attenuation of ferroptosis-associated alterations. Nevertheless, because nuclear NRF2 translocation and functional ferroptosis validation using lipid ROS assays or ferroptosis inhibitors were not performed, a causal NRF2-mediated anti-ferroptotic mechanism cannot be established from the present findings [61]. The histopathological and ultrastructural findings were consistent with the biochemical and molecular observations. Severe hepatocellular degeneration, inflammatory infiltration, sinusoidal congestion, and lipid vacuolization in diabetic rats indicate extensive metabolic and oxidative hepatic injury. At the ultrastructural level, mitochondrial shrinkage, disrupted cristae, and membrane alterations are morphological changes that may occur in association with ferroptotic injury, although they are not specific to ferroptosis. The marked preservation of hepatic architecture and mitochondrial ultrastructure following IC-LNP treatment was consistent with attenuation of hepatocellular and mitochondrial injury. However, these morphological findings alone cannot establish ferroptotic cell death or its prevention. Study Limitations Despite the promising findings, several limitations should be acknowledged. First, the study was limited to an experimental animal model, and therefore extrapolation of the findings to human diabetic hepatopathy requires further translational investigation. Second, although the study demonstrated significant changes in NRF2/HO-1, NF-κB, and ferroptosis-related markers, pathway activation and causal involvement were not directly established. Nuclear NRF2 translocation, phospho-p65/IκBα assessment, pathway inhibitors, and gene-silencing approaches were not performed. Third, ferroptosis was inferred from changes in Fe²⁺, ACSL4, TFR1, GPX4, and SLC7A11 rather than functionally validated using lipid ROS assays or ferroptosis-specific rescue experiments with inhibitors such as ferrostatin-1 or liproxstatin-1. Fourth, pharmacokinetic, biodistribution, hepatic accumulation, and cellular uptake studies were not performed; therefore, the greater efficacy of IC-LNPs cannot be attributed directly to enhanced bioavailability, hepatic delivery, or cellular uptake. Moreover, because IC and IC-LNPs were administered intraperitoneally, the present study does not provide evidence of improved oral bioavailability. Fifth, the absence of a blank-liposome control prevents complete exclusion of potential contributions of the liposomal carrier itself to the observed effects. Sixth, additional mitochondrial functional measurements, including mitochondrial membrane potential and ATP production, would provide deeper insight into mitochondrial involvement. Finally, only male rats were included, limiting assessment of potential sex-related differences in treatment responses. Conclusion In conclusion, icariin exerted significant hepatoprotective effects against NA–STZ-induced diabetic hepatopathy, with IC-LNPs generally producing greater protective effects than crude IC under the present experimental conditions. IC-LNPs more effectively ameliorated metabolic disturbances, including hyperglycemia, insulin depletion, dyslipidemia, and elevated hepatic enzyme activities. These effects were accompanied by improved antioxidant status and changes in NRF2/HO-1-associated signaling, together with reductions in inflammatory mediators and nuclear NF-κB levels. IC-LNPs also attenuated multiple ferroptosis-related alterations, including increased Fe²⁺, ACSL4, and TFR1 and decreased GPX4 and SLC7A11. Histopathological and ultrastructural findings further demonstrated substantial preservation of hepatic architecture following IC-LNP treatment. Collectively, these findings indicate that the hepatoprotective effects of IC-LNPs are associated with coordinated modulation of oxidative stress, inflammatory responses, and ferroptosis-related processes. Further pharmacokinetic, mechanistic, safety, and translational studies are required before IC-LNPs can be considered a therapeutic strategy for diabetic liver injury. Supplementary Information Below is the link to the electronic supplementary material. Supplementary Material 1 Supplementary Material 1