2.1 Animals

All animal experiments were carried out in accordance with the National Institutes of Health (NIH) Guide for the Care and Use of Laboratory Animals and were approved by the Institutional Animal Care and Use Committee at Chongqing Medical University. The study used adult male Sprague Dawley (SD) rats weighing 280–300 g. The rats were housed in standard laboratory conditions with a temperature maintained at 22°C and relative humidity of 40%–60%. A 12-h light/dark cycle was provided. The animals were kept in ventilated cages with bedding material regularly replaced and with free access to standard rodent chow and water.

2.2 SAH Model and SAH Grading

After isoflurane anesthesia, the rats were intubated and placed on mechanical ventilation. A 4-0 monofilament punctured the bifurcation where the anterior and middle cerebral arteries meet by inserting it through the left external carotid artery into the internal carotid artery. The same surgical procedure was performed in the sham group, but the arterial puncture was omitted. Vital signs such as respiration, heart rate, skin color, and pedal reflex were monitored every 5 min to ensure appropriate anesthesia depth and to minimize distress. All procedures were performed according to previously established methods [24].

The severity of SAH was evaluated in a blinded assessment by one independent investigator using a scoring system based on previous studies [25]. In summary, the basal cistern was divided into six regions, and each was assigned a score from 0 to 3 depending on the extent of blood clot formation in the subarachnoid space. The maximum score was 18, with 0 indicating no blood. Rats with scores below 9 were excluded to ensure consistency in SAH severity across all subjects included in further analyses.

In this study, 424 rats were used, of which 322 rats underwent SAH induction. The overall mortality rate for SAH rats was 16.8% (54 out of 322). None of the rats died in the sham or naïve groups. Additionally, 14 SAH rats were excluded from the analysis due to low SAH grading scores. SAH grades were not significantly disparate among each group of SAH, as shown in Table 1.

2.3 Cell Culture

Primary neuron culture and hemoglobin (Hb) stimulation were used as an in vitro model. Primary neurons were extracted from P1 SD rat pups. Briefly, whole brain tissues were removed aseptically from the pups' heads, and cortex tissues were separated from the brains. After removing the meninges and blood vessels, cortex tissues were lysed with 0.25% trypsin for 10 min at 37°C. Then, the lysis was terminated, and the tissues were dispersed by a Pasteur pipette. Neurons were seeded on poly-l-lysine-coated plates with a neurobasal medium (Gibco, MD, USA), which contained 2% B₂₇ (Gibco) and 0.5 mM glutamine, and were cultured in an incubator at 37°C, with 5% CO₂. The medium was half changed every 3 days. The cells were then subjected to the in vitro experiments after 7 days of culture.

For in vitro model, rat primary neurons were seeded onto six-well plates. Once the cultures reached maturity, the medium was fully replaced with a fresh medium containing 25 μM Hb (9008-02-0, Sigma, USA) and incubated for another 24 h.

2.4 Drug Administration

2.5 Experiment Design

The study was organized into six distinct sub-experiments, with subjects randomly and blindly assigned to different groups (Figure 1). If any death or exclusion occurred, an additional animal was allocated as a replacement to ensure that each group maintained consistent numbers and the sample size was adequate for statistical analysis.

2.6 Neurobehavioral Test

Short-term neurological function was evaluated using two established tests: the modified Garcia test, scoring from 0 to 18, and the beam balance test, scoring from 0 to 4. The Garcia test measures response, alertness, coordination, and motor abilities, whereas the beam balance test focuses on complex movements and coordination [28]. Higher scores indicate superior performance. All the tests were performed by one independent investigator who was blinded to the group assignments and interventions.

2.7 Immunofluorescence Staining

Following established protocols, double immunofluorescence staining was performed [29]. Brain tissues were initially fixed in formalin for 24 h and then dehydrated through a sucrose gradient. Samples were embedded in OCT and sectioned into 10-μm-thick coronal slices. After rinsing in 0.01 M PBS, sections were blocked with 5% donkey serum at room temperature for 2 h before an overnight incubation at 4°C with primary antibodies: anti-Metrnl (1:50, orb185592, Biorbyt, MA, UK), anti-p-C-KIT (1:50, BS-12916R, Thermo Fisher Scientific, Carlsbad, USA), anti-4HNE (1:50, MA5-27570, Thermo Fisher Scientific, Carlsbad, USA), anti-GPX4 (1:100, ab125066, Abcam, MA, USA), and anti-NeuN (1:100, ab177487, Abcam, MA, USA). The next day, slices were incubated with secondary fluorophore-conjugated antibodies (1:500, Jackson ImmunoResearch) for 1 h. After incubation, a fluorescence microscope was utilized to visualize and photograph the slices. 4HNE- or GPX4-positive cells were identified and quantified at 200× magnification. For each group containing four rats, the average cell numbers were calculated from four slices per rat.

2.8 FJC Staining

We used FJC staining with a ready-to-dilute kit (TR-100-FJ, Biosensis, USA) to identify degenerating neurons in the basal temporal cortex, adhering to the prior protocol [30]. FJC-positive cells in the ipsilateral temporal cortex were identified and quantified using a fluorescence microscope at 200× magnification. For each group containing four rats, the average cell numbers were calculated from four slices per rat.

2.9 Mitochondrial Membrane Potential Analysis

The mitochondrial membrane potential of primary neurons was measured using the TMRE kit (T669, Thermo Fisher Scientific, Carlsbad, USA) according to the manufacturer's instructions. Briefly, primary neurons were incubated with the TMRE working solution for 30 min at 37°C. After incubation, the cells were washed three times with serum-free DMEM to remove excess dye. The cells were observed and photographed using a fluorescence microscope at 400× magnification. The mean fluorescence intensity was analyzed by ImageJ software. The average intensity was calculated from six slides per group.

2.10 Nissl Staining

Nissl staining was conducted to examine neuronal damage in the ipsilateral hippocampus on the 28th day after SAH. A 0.5% crystal violet solution was used for staining, following established protocols [31]. The slices were observed and photographed at 200× magnification using a light microscope. Surviving neurons were characterized by pale nuclei, large cell bodies, and abundant cytoplasm. In contrast, neurons with shrunken cytoplasm and condensed staining were considered injured [31]. In the long-term study, for each group containing six rats, the average surviving neurons in hippocampal areas, including cornu ammonis (CA)1, CA3, and the dentate gyrus (DG), were quantified from four slices per rat.

2.11 ELISA Assay

Lipid peroxidation was assessed using the MDA Assay Kit (MBS162002, MyBioSource, CA, USA), and GSH levels were determined using the GSH Assay Kit (Beyotime, Shanghai, China), both in accordance with the manufacturer's protocols.

2.12 Western Blot Analysis

Western blot was conducted as in the previous studies [25, 27]. Brain tissues or primary neurons underwent lysis in RIPA buffer to isolate proteins. In all, 10% of SDS-PAGE gels were used for electrophoresis and proteins were then transferred to PVDF membranes (Millipore, USA). The membranes were treated with 5% skimmed milk at room temperature for 60 min and then incubated with primary antibodies at 4°C overnight. The primary antibodies used are listed as follows: anti-Metrnl (1:500, ab235775, Abcam, MA, USA), anti-p-C-KIT (1:500, GTX25616, GeneTex, CA, USA), anti-C-KIT (1:500, ab256345, Abcam, MA, USA), anti-p-AMPK (1:500, 2535, Cell Signaling Technology, MA, USA), anti-AMPK (1:500, 2532, Cell Signaling Technology, MA, USA), anti-Nrf2 (1:1000, ab137550, Abcam, MA, USA), anti-SLC7A11 (1:1000, ab175186, Abcam, MA, USA), anti-GPX4 (1:1000, ab125066, Abcam, MA, USA), anti-FSP1 (1:1000, sc-377,120, Santa Cruz Biotechnology, TX, USA), anti-FTH (1:1000, sc-376,594, Santa Cruz Biotechnology, TX, USA), anti-ACSL4 (1:1000, MA5-42523, Thermo Fisher Scientific, CA, USA), and anti-β-actin (1:3000, Santa Cruz Biotechnology, TX, USA). The following day, the membrane was treated with horseradish peroxidase (HRP)–linked secondary antibodies for 60 min at room temperature. The detection of immunoreactive signals was achieved using enhanced chemiluminescence reagents (Millipore, USA). Subsequently, the signals were quantified with Image J software, with β-actin used as a reference for normalization.

2.13 Transmission Electron Microscopy

Ipsilateral basal temporal cortex tissue was fixed for 2 h in an electron microscopy fixative solution containing 2.5% glutaraldehyde. Following dehydration, embedding, and sectioning at 100 nm, the samples were stained with 4% uranyl acetate for 20 min and 0.5% lead citrate for 5 min. The sections were then examined using a TEM (Hitachi H 7500, Japan). At a magnification of 3000×, the assessment of damaged neuronal mitochondria involved observing shrunken structures, vanished cristae, and denser membranes. For each group containing four rats, the average percentage of impaired neuronal mitochondria was counted in four sections per rat.

2.14 Statistical Analysis

Statistical assessments were performed using GraphPad Prism 8 software (San Diego, CA, USA). Protein levels, immunostaining, ELISA, Nissl staining, TEM, and neurological scores underwent the Shapiro–Wilk test for normal distribution and Levene's test for variance consistency. Subsequent statistical processing was conducted using one-way ANOVA and Tukey's post hoc test. For data that did not conform to a normal distribution or exhibited nonhomogeneous variance, the Mann–Whitney U test or Kruskal–Wallis test was applied. The results are presented as mean ± SD, with a p-value < 0.05 indicating statistical significance.

3.1 Time Course of Endogenous Protein Levels and Cellular Localization of Metrnl and p-C-KIT in Ipsilateral Hemisphere After SAH

The levels of endogenous Metrnl and p-C-KIT in the left cerebral hemisphere were measured using Western blot. Compared to the sham group, Metrnl levels significantly increased 6 h and peaked at 24 h after SAH (p < 0.05, Figure 2a,b). Similarly, p-C-KIT levels showed significant upregulation, which peaked at 24 h after SAH (p < 0.05, Figure 2a,c). Furthermore, immunofluorescence staining demonstrated that Metrnl and p-C-KIT co-localized with neurons in the cortex of the left hemisphere (Figure 2d).

3.2 Endogenous Brain Metrnl Knockdown Exacerbated Neurological Deficits and Neuronal Degeneration After SAH

To investigate the role of endogenous Metrnl in ferroptosis following SAH, Metrnl shRNA was administered 2 weeks before SAH induction. Western blot analysis confirmed a significant decrease in Metrnl protein levels after Metrnl knockdown (p < 0.05, Figure 3a,b). Neurobehavioral assessments, including modified Garcia and beam balance tests, revealed significantly lower scores in SAH rats compared to the sham group (p < 0.05, Figure 3c,d). Moreover, Metrnl knockdown exacerbated neurobehavioral deficits relative to the SAH group or the SAH + scr shRNA group (p < 0.05, Figure 3c,d).

FJC staining revealed that SAH led to neuronal degeneration. Metrnl knockdown further worsened neuronal degeneration compared to the SAH group or the SAH + scr shRNA group (p < 0.05, Figure 3e,f).

3.3 Endogenous Brain Metrnl Knockdown Exacerbated Neuronal Ferroptosis After SAH

Metrnl protein levels at 24 h after SAH were markedly lower in the SAH + Metrnl shRNA group compared to the SAH group and the SAH + scr shRNA group (p < 0.05, Figure 4a,b). Compared to the sham group, SAH rats showed significantly higher expressions of ACSL4 and MDA but lower levels of SLC7A11, GPX4, FSP1, FTH, and GSH at 24 h after SAH (p < 0.05, Figure 4c–g). When comparing the SAH group to the SAH + scr shRNA group, there was no significant difference observed in the protein levels of Metrnl and ferroptosis indicators (p > 0.05, Figure 4a–i), including SLC7A11, GPX4, FSP1, FTH, ACSL4, MDA, and GSH. Metrnl knockdown significantly exacerbated the SAH-induced increase in ACSL4 and MDA, and further decreased the expression of SLC7A11, GPX4, FSP1, FTH, and GSH compared to the SAH group or the SAH + scr shRNA group (p < 0.05, Figure 4a–i).

To explore the specific location where ferroptosis happened, we performed immunofluorescence experiments to co-stain neurons with GPX4 and 4HNE. Our findings indicated that the knockdown of Metrnl decreased GPX4-positive neurons in the ipsilateral cortex at 24 h after SAH, compared to the SAH + scr shRNA group (p < 0.05, Figure 4j,k). In addition, the level of 4HNE was increased in the ipsilateral cortex after SAH. Consistently, Metrnl knockdown led to an increase in the number of 4HNE-positive neurons, compared to both the SAH group and the SAH + scr shRNA group (p < 0.05, Figure 4l,m).

Mitochondria in ferroptosis can often exhibit morphological changes such as condensation and shrinkage, a reduction or disappearance of cristae, and outer membrane damage. In our study, all these characteristics could be observed in the SAH group under TEM, which were further intensified by Metrnl knockdown (p < 0.05, Figure 4n,o).

3.4 Exogenous r-Metrnl Treatment Attenuated Neurological Deficit and Neuronal Degeneration After SAH

After SAH induction, marked neurological impairments were displayed compared to the Sham group. Intranasal administration of r-Metrnl (4 μg/kg and 12 μg/kg) significantly improved neurological function 24 h after SAH compared to the SAH + PBS group (p < 0.05, Figure 5a,b). Based on the neurobehavioral assessments, the medium dosage of r-Metrnl (4 μg/kg) was selected for further experiments. FJC staining revealed that SAH-induced neuronal damage was significantly mitigated by r-Metrnl treatment (p < 0.05, Figure 5c,d).

3.5 Exogenous r-Metrnl Treatment Attenuated Neuronal Ferroptosis After SAH

Western blot analysis revealed that compared to the SAH + PBS group, r-Metrnl treatment significantly decreased the level of the pro-ferroptosis protein ACSL4 and increased the expression of SLC7A11, GPX4, FSP1, and FTH at 24 h after SAH (p < 0.05, Figure 6a–g). Moreover, ELISA results demonstrated that r-Metrnl treatment significantly improved GSH levels and reduced MDA expression compared to the SAH + PBS group (p < 0.05, Figure 6h,i).

Immunofluorescence analysis of GPX4 in the ipsilateral cortex indicated that SAH induction significantly reduced GPX4-positive neurons in the SAH group compared to the Sham group, and r-Metrnl restored the expression of GPX4 (p < 0.05, Figure 6j,k). Furthermore, r-Metrnl treatment significantly decreased the number of 4HNE-positive neurons 24 h after SAH compared to the untreated SAH group (p < 0.05, Figure 6l,m). Intranasal r-Metrnl treatment alleviated the morphological changes of SAH-induced mitochondrial atrophy, including reduced mitochondrial membrane density and cristae loss, as observed under TEM (p < 0.05, Figure 6n,o).

3.6 Intranasal r-Metrnl Decreased Neuronal Degeneration in the Hippocampus Region on 28 Days After SAH

Nissl staining performed 28 days after SAH revealed increased neuronal loss and morphological shrinkage within the CA1 and CA3 regions of the ipsilateral hippocampus in the SAH + PBS group compared to the sham group (p < 0.05, Figure 7a–c). In contrast, the SAH + r-Metrnl group exhibited reduced neuronal degeneration, underscoring the neuroprotective impact of r-Metrnl treatment (p < 0.05, Figure 7a–c).

3.7 R-Metrnl Attenuated Neuronal Ferroptosis via C-KIT/AMPK/Nrf2 Signaling Pathway After SAH

To investigate the neuroprotective mechanism of r-Metrnl, we examined the involvement of the C-KIT receptor tyrosine kinase and AMPK pathways. The C-KIT inhibitor ISCK03 and the AMPK inhibitor dorsomorphin were used. As shown in Figure 8, p-C-KIT, p-AMPK, Nrf2, ACSL4, and MDA levels were significantly increased, and the expression of SLC7A11, GPX4, FSP1, FTH, and GSH was markedly reduced in the SAH group compared to the sham group (p < 0.05, Figure 8a–k). Treatment with r-Metrnl significantly upregulated p-C-KIT, p-AMPK, Nrf2, SLC7A11, GPX4, FSP1, FTH, and GSH, while decreasing ACSL4 and MDA levels compared to the SAH + PBS group at 24 h after SAH (p < 0.05, Figure 8a–k). However, ISCK03 administration prior to SAH induction caused a significant reduction in p-C-KIT level compared to the SAH + r-Metrnl + DMSO (i.c.v.) group (p < 0.05, Figure 8a,b). Furthermore, co-administration of ISCK03 or dorsomorphin led to elevated ACSL4 and MDA levels and reduced p-AMPK, Nrf2, SLC7A11, GPX4, FSP1, FTH, and GSH compared to the SAH + r-Metrnl + DMSO (i.c.v.) group (p < 0.05, Figure 8a–k). Immunofluorescence analysis of 4HNE showed that the level of lipid peroxidation in the SAH + r-Metrnl + ISCK03 group and the SAH + r-Metrnl + dorsomorphin group was significantly higher than that in the SAH + r-Metrnl + DMSO (i.c.v.) group (p < 0.05, Figure 8l,m).

To further explore downstream molecules in this pathway, we used the selective Nrf2 inhibitor ML385 that was administered intraperitoneally 1 h before SAH induction. ML385 significantly reduced the expression of Nrf2, SLC7A11, GPX4, FSP1, FTH, and GSH, and increased ACSL4 and MDA levels 24 h after SAH (p < 0.05, Figure 9a–i). In addition, the level of lipid peroxidation of the SAH + r-Metrnl + ML385 group was significantly higher than that of the SAH + r-Metrnl + DMSO (i.p.) group (p < 0.05, Figure 9j,k).

3.8 R-Metrnl Inhibits Ferroptosis via C-KIT/AMPK/Nrf2 Signaling Pathway in Hb-Stimulated Primary Neurons

Primary neuron culture experiment was performed to verify whether the effects of Metrnl were mediated by attenuating neuronal ferroptosis through the C-KIT/AMPK/Nrf2 signaling pathway. Consistent with the results in vivo, Hb stimulation significantly increased p-C-KIT, p-AMPK, Nrf2, ACSL4, and MDA levels, and the expression of SLC7A11, GPX4, FSP1, FTH, and GSH was markedly reduced (p < 0.05, Figure 10a–k). Compared to the Hb + PBS group, the expression of p-C-KIT, p-AMPK, Nrf2, SLC7A11, GPX4, FSP1, FTH, and GSH increased in the Hb + r-Metrnl treated group (p < 0.05, Figure 10a–k). Meanwhile, the expression of ACSL4 and MDA decreased (p < 0.05, Figure 10a–k). ISCK03 treatment significantly decreased p-C-KIT expression compared to the Hb + r-Metrnl + DMSO group. Administration of ISCK03 or dorsomorphin reversed the effects of r-Metrnl on p-AMPK, Nrf2, SLC7A11, GPX4, FSP1, FTH, ACSL4, MDA, and GSH expression (p < 0.05, Figure 10a–k).

Meanwhile, the TMRE probe test for mitochondrial membrane potential consistently showed that decreased mitochondrial membrane potential after Hb stimulation was restored by r-Metrnl administration. These effects were reversed by the inhibition of C-KIT or AMPK (p < 0.05, Figure 10l,m).