Inflammation Impairs Poststroke Recovery by Disrupting Iron Homeostasis in Brain

Nestin-Cre-mediated Fpn1 knockout impairs sensorimotor functional recovery after cerebral ischemia

To evaluate functional recovery after stroke, we conducted modified neurological severity score (mNSS), rotarod tests, adhesive removal tests, and gait analysis tests in mice with or without conditional Fpn1 knockout before distal middle cerebral artery occlusion (dMCAO) and on days 1, 3, 7, and 14 after the operation (Supplementary Fig. S1C). At day 1 after dMCAO, mNSS scores did not differ significantly between groups. However, during the recovery phase, Fpn1^Nestin-CKO dMCAO mice exhibited significantly higher mNSS scores than Fpn1^flox/flox dMCAO group on days 3 and 7. On day 14, mNSS scores in both groups approached sham levels (Fig. 1A). The rotarod test showed the residence time of Fpn1^flox/flox dMCAO mice was significantly shorter than that of the Fpn1^Nestin-CKO dMCAO group on day 1. However, on day 14, the residence time of Fpn1^Nestin-CKO dMCAO group was significantly shorter compared with the Fpn1^flox/flox dMCAO group (p < 0.05, Fig. 1B). The adhesive removal test showed that on day 14, both the contact time and removal time for the left and right limbs in the Fpn1^flox/flox dMCAO group returned to a control level, while these the parameters of left forelimb (LF)/right forelimb (RF)-contact time and LF-removal time in the Fpn1^Nestin-CKO dMCAO group remained significantly higher than the control level (p < 0.05, Fig. 1C–F). The results of gait analysis experiments reveal that on day 14, the step amplitude of LF, RF, left hindlimb and right hindlimb, and the standing width of forelimb standing and hind limb standing in the Fpn1^Nestin-CKO dMCAO group were all different from those of the Fpn1^flox/flox dMCAO group, with shorter stride, increased forelimb and hindlimb standing width (p < 0.05, Fig. 1G–L). In general, these results demonstrate that Nestin-Cre-mediated Fpn1 knockout delays sensorimotor functional recovery following cerebral ischemia.

OPEN IN VIEWER

FIG. 1.

Nestin-Cre-mediated Fpn1 knockout impairs sensorimotor functional recovery after cerebral ischemia. (A) Quantification of mNSS in each group before dMCAO and on days 1, 3, 7, and 14 after the operation. (B) Quantification of Rotarod test in each group before dMCAO and on day 1, 3, 7, and 14 after the operation. †: Fpn1^flox/flox dMCAO vs. Fpn1^flox/flox Sham, ^†p < 0.05, ^†††p < 0.001. ‡: Fpn1^Nestin-CKO dMCAO vs. Fpn1^Nestin-CKO Sham, ^‡p < 0.05, ^‡‡p < 0.01, ^‡‡‡p < 0.001. *: Fpn1^Nestin-CKO dMCAO vs. Fpn1^flox/flox dMCAO, *p < 0.05. (C–L) Quantification of adhesive removal test (C–F), gait analysis (G–L) in each group before dMCAO and on days 1, 3, 7, and 14 after the operation. *p < 0.05, **p < 0.01, ***p < 0.001. LMN and two-way repeated measures ANOVA test were used. Data are expressed as the mean ± SD. Their statistics were shown in Supplementary Table S4. dMCAO, distal middle cerebral artery occlusion; mNSS, modified neurological severity score; SD, standard deviation.

Fpn1 knockout aggravates neuronal injury, BBB disruption, and enhances apoptosis- and necroptosis-associated signaling after cerebral ischemia

Neuronal injury after ischemic stroke involves multiple forms of cell death-related signaling. In the present study, we performed hematoxylin and eosin (H&E) and Nissl staining in the peri-infarct cortex to further validate tissue injury histologically. H&E staining revealed more severe histopathological damage in the peri-infarct cortex of Fpn1^Nestin-CKO dMCAO mice compared with Fpn1^flox/flox dMCAO mice. Quantitative analysis further confirmed a significant increase in tissue injury in Fpn1^Nestin-CKO dMCAO mice (Fig. 2A–B). Nissl staining showed a marked reduction in surviving neurons in the peri-infarct cortex, which was significantly more pronounced in Fpn1^Nestin-CKO dMCAO mice than in Fpn1^flox/flox dMCAO mice (Fig. 2C–D). In addition, BBB integrity was further evaluated by IgG staining. Increased IgG leakage was observed in the injured cortex of Fpn1^Nestin-CKO dMCAO mice, indicating exacerbated BBB disruption and secondary neurovascular injury following cerebral ischemia (Fig. 2E–F). To investigate the underlying mechanisms, we mainly evaluated markers related to apoptosis- and necrosis-associated signaling. The expression of receptor-interacting protein kinases RIP1 and RIP3 was significantly increased in the ipsilateral cortex of Fpn1^Nestin-CKO dMCAO group compared with the Fpn1^flox/flox dMCAO mice. Both RIP1 and RIP3 content on the infarct side in each group were significantly higher than that on the contralateral side (CS) (Fig. 2G–H). Notably, phosphorylation of mixed lineage kinase domain-like protein, a key executioner of necroptosis, was increased in Fpn1^Nestin-CKO dMCAO mice, further supporting enhanced necroptosis-associated signaling after cerebral ischemia (Fig. 2G–H).

OPEN IN VIEWER

FIG. 2.

Fpn1 deficiency aggravates neuronal injury, blood–brain barrier disruption, and apoptosis- and necroptosis-related signaling after cerebral ischemia. (A) Representative H&E staining images showing histopathological alterations in peri-infarct cortex (scale bar = 100 μm). (B) Quantification of the percentage of damaged neurons in the peri-infarct cortex based on H&E staining. (C) Representative Nissl staining images showing neuronal loss and structural changes in peri-infarct cortex (scale bar = 100 μm). (D) Quantification of the neuronal density of 5 adjacent histological sections of 4 μm (20 μm total distance) stained with Nissl in the peri-infarct cortex. (E) Representative images showing the extravasation of plasma IgG (green) into the brain parenchyma on day 14 after dMCAO (scale bar = 1 mm). (F) Volume of leakage of endogenous IgG. (G) Representative Western blot images showing RIP1, RIP3, Bcl2, BAX, p-MLKL, and MLKL content in the ipsilateral (IS) and contralateral (CS) sides of the cerebral cortex on day 14 after dMCAO. (H) Quantification of RIP1 and RIP3 content, Bcl2/BAX, and p-MLKL/MLKL content ratio (n = 6; β-actin was used as an internal control). T-tests or two-way repeated measures ANOVA test was used. The data are expressed as the mean ± SD, *p < 0.05, **p < 0.01, ***p < 0.001. Their statistics were shown in Supplementary Table S5. H&E, hematoxylin and eosin; p-MLKL, phosphorylation of mixed lineage kinase domain-like protein; RIP1/3, receptor-interacting protein 1/3.

Apoptosis-related signaling was also assessed. The Bcl2/BAX ratio on the infarct side in the Fpn1^Nestin-CKO dMCAO group was significantly more reduced than that in the Fpn1^flox/flox dMCAO group (Fig. 2G–H). However, cleaved caspase3 levels did not differ significantly between groups (Supplementary Fig. S2A–S2B), suggesting that classical caspase-dependent apoptosis may not be prominently altered.

To further explore whether ferroptosis-related pathways were involved, we examined several ferroptosis-related proteins, including acyl-CoA synthetase long-chain family member 4 (ACSL4), glutathione peroxidase 4 (GPX4), and nuclear receptor co-activator 4 (NCOA4) (Doll et al., 2019). ACSL4 content on the infarct side was significantly increased compared with the CS in both groups, whereas no significant change was observed between groups. The other protein content of GPX4 and NCOA4 on the infarct side did not differ significantly from that on the CS in either group, and no significant difference was observed between the infarct sides of Fpn1^Nestin-CKO dMCAO and Fpn1^flox/flox dMCAO mice (Supplementary Fig. S2C-S2F). Taken together, these results indicate that the present data do not provide evidence for enhanced ferroptosis in Nestin-Cre-mediated Fpn1 knockout mice after cerebral ischemia.

Fpn1 knockout impairs myelin-associated repair, synaptic remodeling, and endogenous repair-related responses after cerebral ischemia

To evaluate myelin-related repair histologically, we performed immunofluorescence staining for Olig2 and myelin basic protein (MBP) in the peri-infarct cortex. Olig2-positive area increased on the injured side after dMCAO in both groups, but the increase was less pronounced in Fpn1^Nestin-CKO dMCAO than in Fpn1^flox/flox mice (Fig. 3A,C). In contrast, MBP-positive area was reduced after dMCAO, and this reduction was more pronounced in Fpn1^Nestin-CKO dMCAO (Fig. 3B,D).

OPEN IN VIEWER

FIG. 3.

Fpn1 deficiency impairs myelination, synaptic plasticity, and neurotrophic signaling in the cortex after cerebral ischemia. (A–B) Immunofluorescence staining of Olig2 (A) and MBP (B) in the cortex of dMCAO mice. DAPI was used for nuclear staining (scale bar = 50 or 20 μm). (C–D) Quantification of the Olig2 (C) and MBP (D) covered area percent in cortex was presented on the bar charts (n = 6). (E) Representative Western blot images showing Galc, MBP, PSD95, SYN, pro-BDNF, and m-BDNF content in the IS and CS of the cerebral cortex on day 14 after dMCAO. (F–K) Quantification of Galc (F), MBP (G), PSD95 (H), SYN (I), pro-BDNF (J), and m-BDNF (K) content (n = 6; β-actin or GAPDH was used as an internal control). (L) Relative mRNA abundance of BDNF in the cerebral cortex on day 14 after dMCAO (n = 6; β-actin was used as an internal control). Two-way repeated measures ANOVA test was used. The data were expressed as the mean ± SD, *p < 0.05, **p < 0.01, ***p < 0.001. Their statistics were shown in Supplementary Table S6. Galc, galactosyl ceramidase; m-BDNF, mature BDNF; MBP, myelin basic protein; mRNA, messenger ribonucleic acid; pro-BDNF, pro-brain-derived neurotrophic factor; PSD95, postsynaptic density protein-95; SYN, synaptophysin.

To further assess myelin and synaptic remodeling, we examined the expression of myelin sheath- and synapse-related proteins in the cerebral cortex on day 14 after cerebral ischemia by Western blot analysis (Fig. 3E). Galactosyl ceramidase (Galc) is an enzyme involved in the metabolism of myelin lipids, whereas MBP is a major structural component of the myelin sheath. Galc content was significantly reduced in the injured side of Fpn1^Nestin-CKO dMCAO group compared with the Fpn1^flox/flox dMCAO group (Fig. 3F), whereas MBP content did not differ significantly between groups (Fig. 3G). Synaptic integrity was evaluated by postsynaptic density protein-95 (PSD95) and synaptophysin (SYN). PSD95 content was significantly decreased in the injured cortex of Fpn1^Nestin-CKO dMCAO group compared with both the CS and the Fpn1^flox/flox dMCAO group (Fig. 3H). Similarly, the SYN content was reduced after dMCAO in both groups, and significantly lower in Fpn1^Nestin-CKO dMCAO group than in Fpn1^flox/flox dMCAO group (Fig. 3I). Although pro-brain-derived neurotrophic factor content did not differ significantly between groups (Fig. 3J), mature BDNF content was significantly lower in the injured cortex of Fpn1^Nestin-CKO dMCAO mice than in the Fpn1^flox/flox dMCAO group (Fig. 3K). Consistently, BDNF messenger ribonucleic acid (mRNA) abundance was reduced after dMCAO and showed a greater decrease in the Fpn1^Nestin-CKO dMCAO group (Fig. 3L).

To investigate endogenous repair-related cellular responses, we examined cell proliferation marker protein Ki-67 (Ki67) and 5-bromodeoxyuridine (BrdU) and neurogenesis-associated marker doublecortin (DCX) level by immunofluorescence and Western blot. Immunofluorescence staining showed reduced Ki67- and DCX-positive areas in the Fpn1^Nestin-CKO dMCAO group compared with Fpn1^flox/flox dMCAO group (Fig. 4A–B, 4D–E). BrdU incorporation, reflecting DNA synthesis in proliferating cells, was also decreased in the Fpn1^Nestin-CKO dMCAO (Fig. 4C,F). Western blot analysis further showed that Ki67 and DCX content were significantly lower in the Fpn1^Nestin-CKO dMCAO group than in the Fpn1^flox/flox dMCAO group (Fig. 4G–I). Finally, Golgi staining showed that a significant reduction in dendritic spine density on the injured side after dMCAO, with a further decrease observed in Fpn1^Nestin-CKO dMCAO group compared with Fpn1^flox/flox dMCAO group (Fig. 4J–K). Taken together, these findings indicate that Nestin-Cre-mediated Fpn1 knockout is associated with impaired cellular proliferation, reduced neurogenesis-related responses, and attenuated dendritic spine remodeling after cerebral ischemia. These alterations may contribute to compromised structural plasticity and delayed functional recovery following stroke.

OPEN IN VIEWER

FIG. 4.

Fpn1 deficiency impairs neurogenesis and dendritic spine remodeling after cerebral ischemia. (A–C) Immunofluorescence staining of Ki67 (A), DCX (B), and BrdU (C) in the cortex of dMCAO mice. DAPI was used for nuclear staining (scale bar = 50 or 20 μm). (D–F) Quantification of the Ki67 (D), DCX (E), and BrdU (F) covered area percent in cortex (n = 6). (G) Representative Western blot images showing Ki67 and DCX content in the IS and CS of the cerebral cortex on day 14 after dMCAO. (H–I) Quantification of Ki67 (H) and DCX (I) content (n = 6; β-actin was used as an internal control, relative to Fpn1^flox/flox CS). (J) Images of stained cortical dendritic spines on day 14 after dMCAO. (K) Quantification of dendritic spine density (n = 9). T-tests or two-way repeated measures ANOVA test was used. The data were expressed as the mean ± SD, *p < 0.05, **p < 0.01, ***p < 0.001. Their statistics were shown in Supplementary Table S7. BrdU, 5-bromodeoxyuridine; DCX, doublecortin; Ki67, nuclear proliferation-associated antigen.

Fpn1 knockout disrupts brain iron homeostasis and aggravates iron-related changes in the ischemic cortex

To explore whether Fpn1 deficiency alters iron homeostasis after stroke, we measured the total iron content in the cerebral cortex by inductively coupled plasma mass spectrometry (ICP-MS). Under baseline conditions, Fpn1 knockout significantly decreased the total cortical iron content (Supplementary Fig. S3A). However, after dMCAO, the total iron content on the injured side increased significantly in both Fpn1^flox/flox and Fpn1^Nestin-CKO groups, indicating disturbed local iron homeostasis in the ischemic cortex (Fig. 5A).

OPEN IN VIEWER

FIG. 5.

Fpn1 deficiency disrupts iron homeostasis in the cortex after cerebral ischemia. (A) Total iron content in the cortex on day 14 after dMCAO was detected by ICP-MS (n = 6; relative to Fpn1^flox/flox CS). (B) Relative mRNA abundance of Hepcidin in the cerebral cortex on day 14 after dMCAO (n = 9; β-actin was used as an internal control, relative to Fpn1^flox/flox CS). (C) Representative Western blot images showing FtL, FtH, FPN1, TfR1, and pro-hepcidin content in the IS and CS of the cerebral cortex on day 14 after dMCAO. (D) Quantification of FtL, FtH, FPN1, TfR1, and pro-hepcidin content (n = 6; β-actin or GAPDH was used as an internal control, relative to Fpn1^flox/flox CS). (E) Relative mRNA abundance of FtL, FtH, and Fpn1 in the cerebral cortex on day 14 after dMCAO (n = 6; β-actin was used as internal control, relative to Fpn1^flox/flox CS). Two-way repeated measures ANOVA test was used. The data were expressed as the mean ± SD, *p < 0.05, **p < 0.01, ***p < 0.001. Their statistics were shown in Supplementary Table S8. FPN1, ferroportin1; FtL/H, ferritin light chain/heavy chain; ICP-MS, inductively coupled plasma mass spectrometry; TfR1, transferrin receptor 1.

Consistent with the increased cortical iron content, ferritin light chain (FtL) and ferritin heavy chain (FtH) content were elevated in the injured side compared with the CSs in both groups. Although absolute FtL and FtH levels in Fpn1^Nestin-CKO dMCAO group were lower than those in Fpn1^flox/flox dMCAO group, the relative increase between the injured and CS in the Fpn1^Nestin-CKO dMCAO group was more pronounced in the Fpn1^flox/flox dMCAO group (Fig. 5C–E). At the transcriptional level, the Fpn1 mRNA abundance on the injured side was upregulated, with FPN1 content on the injured side in Fpn1^Nestin-CKO dMCAO mice showing a lower level than that in the Fpn1^flox/flox dMCAO group (Fig. 5C–E). Transferrin receptor (TfR1) content, a key mediator of cellular iron uptake, on the injured side of Fpn1^Nestin-CKO dMCAO group was lower than that on the opposite side, and there was an increase compared with the Fpn1^flox/flox dMCAO group (Fig. 5C–D). Hepcidin, a central regulator of iron metabolism produced by the liver and glial cells, controls iron efflux by promoting FPN1 internalization and degradation (Gao et al., 2019). The relative mRNA abundance of hepcidin on the infarcted side was higher than that on the CS and was further elevated in the Fpn1^Nestin-CKO dMCAO group compared with the Fpn1^flox/flox dMCAO group (Fig. 5B). Nestin-Cre-mediated Fpn1 knockout led to an increase in pro-hepcidin content on the injured side after cerebral ischemia (Fig. 5C–D). Finally, we examined divalent metal transporter 1 (DMT1), including both +IRE (iron-responsive element) and −IRE isoforms. No significant differences were observed in DMT1 content between groups (Supplementary Fig. S3B–S3D). Overall, these findings indicate that Nestin-Cre-mediated Fpn1 knockout is associated with dysregulated iron homeostasis during the postischemic phase, characterized by enhanced iron accumulation and altered iron-handling responses in the ischemic cortex.

Increased endothelial iron content is associated with aggravated BBB disruption after cerebral ischemia

Disruption of iron homeostasis has been implicated in the regulation of BBB integrity (You et al., 2022). To determine whether altered iron levels contribute to BBB dysfunction after stroke, we first assessed the content of tight-junction proteins by Western blot analysis (Fig. 6A). Compared with the Fpn1^flox/flox dMCAO group, the zonula occludens-1 (ZO-1), occludin, and claudin-5 content were all diminished on the injured side of Fpn1^Nestin-CKO dMCAO group (Fig. 6B–D). Immunofluorescence analysis further demonstrated that the ZO-1 positive cell numbers were a marked reduction in the injury side of Fpn1^Nestin-CKO dMCAO mice (Fig. 6E–F). Ultrastructural examination by transmission electron microscopy (TEM) revealed damaged cerebral microvascular endothelial cells after dMCAO, with more severe endothelial damage observed in Fpn1^Nestin-CKO dMCAO, including increased intercellular gaps between endothelial cells and surrounding structures (Fig. 6G).

OPEN IN VIEWER

FIG. 6.

Fpn1 deficiency impairs blood–brain barrier integrity and alters endothelial iron metabolism after cerebral ischemia. (A) Representative Western blot images showing ZO-1, occludin, and claudin-5 content in the IS and CS of the cerebral cortex on day 14 after dMCAO. (B–D) Quantification of ZO-1 (B), occludin (C), and claudin-5 (D) content (n = 6; β-actin or GAPDH was used as an internal control, relative to Fpn1^flox/flox CS). (E) Double immunofluorescence staining of CD31 (red), ZO-1 (green), and DAPI (blue) in the cortex of dMCAO mice (scale bar = 20 μm). (F) The ratio of CD31⁺ZO1⁺ cell to CD31⁺ cells was presented on the graph (n = 6). (G) TEM showing cerebral microvessels in the cortex of dMCAO mice (scale bar = 2 μm). (H) Double immunofluorescence staining of CD31 (red), FPN1 (green), and DAPI (blue) in the cortex of dMCAO mice (scale bar = 20 μm). (I) The ratio of CD31⁺FPN1⁺ cells to CD31⁺ was presented on the graph (n = 6). (J) Double immunofluorescence staining of CD31 (red), FtL (green), and DAPI (blue) in the cortex of dMCAO mice (scale bar = 20 μm). (K) The ratio of CD31⁺FtL⁺ cells to CD31⁺ cells was presented on the graph (n = 6). T-tests or two-way repeated measures ANOVA test was used. The data were expressed as the mean ± SD, *p < 0.05, p**< 0.01, ***p < 0.001. Their statistics were shown in Supplementary Table S9. TEM, transmission electron microscopy; ZO-1, zonula occludens-1.

In addition, double immunofluorescence staining revealed that reduced CD31/FPN1 co-localization and increased CD31/FtL co-localization in the injury side of Fpn1^Nestin-CKO dMCAO mice (Fig. 6H–K), suggesting altered endothelial iron handling following cerebral ischemia. Thus, these findings indicate that Fpn1 deficiency is associated with increased endothelial iron content and exacerbated BBB disruption following cerebral ischemia.

BBB disruption is associated with aggravated gliosis and inflammatory responses after cerebral ischemia

Astrocytes and microglia play central roles in coordinating neuroinflammatory responses after brain injury. To assess glial activation, we examined the expression of glial fibrillary acidic protein (GFAP) and ionized calcium-binding adaptor molecule 1 (Iba1) in the cortex after dMCAO by Western blot analysis (Fig. 7A). Compared with the CS, GFAP content was significantly increased in the injured side in both groups, and was further elevated in the Fpn1^Nestin-CKO dMCAO group relative to the Fpn1^flox/flox dMCAO group (Fig. 7B). Similarly, Iba1 content was significantly higher in the Fpn1^Nestin-CKO dMCAO group than in the Fpn1^flox/flox dMCAO group (Fig. 7C). Immunofluorescence staining further demonstrated an increased GFAP-positive areas in the peri-infarct region of Fpn1^Nestin-CKO dMCAO group compared with the Fpn1^flox/flox dMCAO group (Fig. 7D–E). The results of immunohistochemical staining showed that, compared with the Fpn1^flox/flox dMCAO group, an increased number of Iba1-positive cells in the peri-infarct region of the Fpn1^Nestin-CKO dMCAO mice. (Fig. 7F–G). These results suggest that Nestin-Cre-mediated Fpn1 knock out aggravated the gliosis associated with cerebral ischemia recovery.

OPEN IN VIEWER

FIG. 7.

Fpn1 deficiency enhances gliosis and inflammatory responses in the cortex after cerebral ischemia. (A) Representative Western blot images showing GFAP and Iba1 content in the IS and CS of the cerebral cortex on day 14 after dMCAO. (B–C) Quantification of GFAP (B) and Iba1 (C) content (n = 6, GAPDH was used as an internal control, relative to Fpn1^flox/flox CS). (D) Immunofluorescence staining showing GFAP (green) and DAPI (blue) in the cortex of dMCAO mice (scale bar = 50 or 20 μm). (E) Quantification of the GFAP covered area percent in cortex (n = 6). (F) Immunohistochemical staining showing Iba1 in the cortex of dMCAO mice (scale bar = 50 or 20 μm). (G) Quantification of the Iba1 covered area percent in cortex (n = 6). (H) Relative mRNA abundance of IL-6, IL-1β, TNF-α, and iNOS mRNA in the cerebral cortex on day 14 after dMCAO (n = 6; β-actin was used as an internal control, relative to Fpn1^flox/flox CS). (I) Relative mRNA abundance of TGF-β, Arg1, IL-33, and IL-10 in the cerebral cortex on day 14 after dMCAO (n = 6; β-actin was used as an internal control, relative to Fpn1^flox/flox CS). T-tests or two-way repeated measures ANOVA test was used. The data were expressed as the mean ± SD, *p < 0.05, **p < 0.01, ***p < 0.001. Their statistics were shown in Supplementary Table S10. GFAP, glial fibrillary acidic protein; IL-10, interleukin-10; IL-1β, interleukin-1β; IL-33, interleukin-33; IL-6, interleukin-6; iNOS, inducible nitric oxide synthase; TNF-α, tumor necrosis factor-α.

To further characterize the inflammatory milieu, we quantified the mRNA abundance of cytokines and immune-related markers in the cortex at day 14 post-dMCAO using real-time quantitative polymerase chain reaction (PCR). Pro-inflammatory mediators, including interleukin-6 (IL-6), interleukin-1β (IL-1β), tumor necrosis factor-α (TNF-α), and inducible nitric oxide synthase, were significantly upregulated in the injured side compared with the CS in both groups, with a further increase observed in Fpn1^Nestin-CKO dMCAO mice compared with Fpn1^flox/flox dMCAO group (Fig. 7H). We also assessed markers associated with anti-inflammatory and reparative responses. The mRNA levels of transforming growth factor-β, arginase 1 (Arg1), interleukin-33, and interleukin-10 were elevated in the injured side after dMCAO and were further increased in Fpn1^Nestin-CKO dMCAO mice (Fig. 7I). Arg1, a key enzyme in arginine metabolism, is commonly associated with alternatively activated macrophages/microglia and is regulated, at least in part, by the JAK-STAT3 signaling pathway. Together, these findings indicate that Nestin-Cre-mediated Fpn1 knockout is associated with enhanced reactive gliosis and an overall amplification of inflammatory responses in the ischemic cortex, characterized by concurrent upregulation of both pro-inflammatory and reparative mediators, suggesting a heightened but potentially dysregulated immune response following cerebral ischemia.

AG490 exerts effects on JAK/STAT3 signaling and promotes repair-related processes after cerebral ischemia

To examine the role of JAK/STAT signaling in the poststroke pathology, we first assessed pathway activation by examining phosphorylation of STAT3 and JAK in the cortex after dMCAO. Consistent with activation of inflammatory signaling after ischemia, the p-STAT3 and p-JAK content were significantly increased in the infarcted side compared with the CS in both groups. Moreover, both p-STAT3 and p-JAK content were further elevated in Fpn1^Nestin-CKO mice relative to Fpn1^flox/flox mice, indicating enhanced activation of the JAK/STAT3 signaling pathway in the context of Fpn1 deficiency (Fig. 8A–C).

OPEN IN VIEWER

FIG. 8.

AG490 suppresses JAK-STAT3 activation and reduces hepcidin content after cerebral ischemia in Fpn1-deficient mice. (A) Representative Western blot images showing p-Stat3, Stat3, p-JAK1, and JAK1 content in the IS and CS of the cerebral cortex on day 14 after dMCAO. (B–C) Quantification of p-Stat3/Stat3 (B) and p-JAK1/JAK1 (C) content (n = 6, β-actin or GAPDH was used as an internal control, relative to Fpn1^flox/flox CS). (D) Representative Western blot images showing p-Stat3, Stat3, and pro-hepcidin content in the IS and CS of the cerebral cortex on day 14 after dMCAO with AG490 intervention. (E–F) Quantification of p-Stat3/Stat3 (E) and pro-hepcidin (F) content (n = 4, β-actin or GAPDH was used as an internal control, relative to Fpn1^flox/flox dMCAO + DMSO). Two-way repeated measures ANOVA test was used. The data were expressed as the mean ± SD, *p < 0.05, **p < 0.01, ***p < 0.001, ns: p > 0.05, ^▲: the ratio of protein levels in Fpn1^flox/flox dMCAO + DMSO to Fpn1^flox/flox dMCAO + AG490 vs. the ratio of protein levels in Fpn1^Nestin-CKO dMCAO + DMSO to Fpn1^Nestin-CKO dMCAO + AG490, ^▲▲▲p < 0.001. Their statistics were shown in Supplementary Table S11.

To further investigate the functional significance of JAK/STAT3 signaling, we administered AG490, a well-characterized tyrosine kinase inhibitor that suppresses JAK/STAT signaling by inhibiting STAT3 phosphorylation. Behavioral assessments were performed before dMCAO and on days 1, 3, 7, and 14 after surgery to evaluate the effects of AG490 treatment on poststroke functional recovery. AG490 treatment was associated with improved neurological performance, including reduced mNSS scores, increased rotarod performance, and improved sensorimotor function in the adhesive removal test. No significant differences were observed in gait parameters (Supplementary Fig. S4A-S4L). At the molecular level, AG490 effectively reduced p-STAT3 levels after dMCAO (Supplementary Fig. S5A-S5B). This was accompanied by a decrease in pro-hepcidin, with a more pronounced reduction observed in Fpn1^Nestin-CKO mice (Fig. 8D–F).

Under pathological conditions, pro-inflammatory cytokines such as IL-6 can activate the JAK-STAT3 pathway, thereby amplifying inflammatory signaling cascades (Johnson et al., 2018). AG490, a tyrosine kinase inhibitor, suppresses JAK-STAT3 signaling by inhibiting STAT3 phosphorylation and has been shown to attenuate inflammatory responses (Chen et al., 2017). In Fpn1^flox/flox mice, AG490 treatment significantly increased Arg1, FTL, and FTH content. However, in Fpn1^Nestin-CKO mice, AG490 treatment significantly decreased Arg1 content, whereas FTL and FTH did not change significantly. No significant changes in IL-1β, TNF-α, and FPN1 content were observed in either genotype. Notably, when comparing AG490 treated-to-untreated ratios, the reductions in p-STAT3 and pro-hepcidin content were more pronounced in Fpn1^Nestin-CKO mice than in Fpn1^flox/flox mice, whereas the increases in TNF-α, Arg1, FTL, and FTH content were greater in Fpn1^flox/flox mice (Supplementary Fig. S6A-S6H).

AG490 treatment exerted differential effects on myelin- and synapse-related proteins between genotypes. In Fpn1^flox/flox mice, AG490 treatment reduced Galc and MBP content. However, in Fpn1^Nestin-CKO mice, Galc content was decreased, whereas MBP content remained unchanged. Regarding synaptic markers, PSD95 content in Fpn1^Nestin-CKO mice did not differ from that in control mice following AG490 treatment. However, SYN content increased in Fpn1^Nestin-CKO mice but decreased in Fpn1^flox/flox mice. Further comparison between genotypes showed that the reduction in Galc and MBP content was more pronounced in Fpn1^flox/flox mice than in Fpn1^Nestin-CKO mice, whereas the increase in SYN content was greater in Fpn1^Nestin-CKO mice (Supplementary Fig. S7A-S7F).

AG490 also induced changes in proliferation- and gliosis-related markers. In Fpn1^flox/flox mice, AG490 treatment was associated with decreased Ki67 and NeuN, accompanied by increased GFAP and Iba1 content. In contrast, in Fpn1^Nestin-CKO mice, AG490 treatment was associated with increased Ki67 and NeuN content and decreased Iba1 content, with no significant change in GFAP content. Consistently, genotype comparison showed more pronounced increases in Ki67 and NeuN content and greater reductions in GFAP and Iba1 content in Fpn1^Nestin-CKO mice than those in Fpn1^flox/flox mice (Supplementary Fig. S7G-S7L). Taken together, inhibition of JAK-STAT signaling with AG490 reduced p-STAT3 and hepcidin levels and was associated with changes in iron-related and repair-associated responses. These effects were more evident in Fpn1-deficient mice than in control mice, indicating that the impact of JAK-STAT signaling may depend on the underlying state of iron homeostasis.

Animals

Mice were kept under 12-h light and 12-h dark cycles at 22 ± 1°C and 45%−55% relative humidity and were provided feed and drink ad libitum. All procedures were carried out in accordance with the Guidelines for the Care and Use of Laboratory Animals issued by the National Institutes of Health and approved by the Animal Care and Use Committee of Hebei Science and Technology Bureau, China. All experiments in the study were approved by the Animal Care and Use Committee of Hebei Normal University (authorization number: 2023LLSC027). We generated Nestin-Cre recombinase mice on a C57BL/6 genetic background and Fpn1^flox/flox mice on a 129/SvEvTac genetic background. Progeny of Fpn1^flox/W with Nestin-Cre recombinase were obtained and mated to produce the next generation, from which genotypes were identified by PCR amplification. Fpn1^flox/flox mice showed only a band around 500 bp bands (Supplementary Fig. S1A), while only a band around 100 bp bands was found in Nestin-Cre homozygous mice (Supplementary Fig. S1B), and the primer sequences used were shown in Supplementary Table S1. Three-month-old male Fpn1^flox/flox mice and Fpn1^flox/flox mice with Nestin-Cre recombinase (Fpn1^Nestin-CKO mice) were selected for experiments. A total of 190 male mice were used in this study. Mice were allocated to the Sham, dMCAO, and AG490 intervention groups as described below: Fpn1^flox/flox Sham (n = 15): Fpn1^flox/flox mice subjected to sham surgery, Fpn1^Nestin-CKO Sham (n = 15): conditional Fpn1 knockout mice subjected to sham surgery, Fpn1^flox/flox dMCAO (n = 30): Fpn1^flox/flox mice subjected to dMCAO surgery, Fpn1^Nestin-CKO dMCAO (n = 30): conditional Fpn1 knockout mice subjected to dMCAO surgery. Mice were allocated to the Sham, dMCAO, and AG490 intervention groups as follows: Sham (n = 9), dMCAO (n = 14), Sham + AG490 (n = 7), dMCAO + AG490 (n = 10). For AG490 intervention, dMCAO groups were further subdivided into: Fpn1^flox/flox dMCAO + DMSO (n = 15): Vehicle control (intraperitoneal injection of DMSO), Fpn1^Nestin-CKO dMCAO + DMSO (n = 15): Vehicle control, Fpn1^flox/flox dMCAO + AG490 (n = 15): AG490 treatment (3 mg/kg, i.p.), Fpn1^Nestin-CKO dMCAO + AG490 (n = 15): AG490 treatment. Sample sizes were determined based on power analysis (α = 0.05, power = 0.8) using historical data from pilot studies.

All animal experiments were conducted using randomized and blinded methods.

dMCAO model

The dMCAO model was generated by electrocoagulation under microscopy (Guo et al., 2022), where the right focal cortex was infarcted. The specific steps were as follows: (i) The mice were fasted for 18 h before surgery and anesthetized with pentobarbital sodium (20 mg/kg, intraperitoneally). To minimize pain and distress, animals received peri- and postoperative analgesia (buprenorphine, 0.05–0.1 mg/kg, subcutaneously). Before surgery, the surgical area was prepared and disinfected with povidone-iodine. (ii) After conventional skin preparation and disinfection, the mice were fixed on the operating table in a supine position. A ∼1 cm incision was cut at the center of the neck, then the neck subcutaneous muscle and fascia were separated to fully expose the neck triangle on the right side, the right common carotid artery (CCA) was carefully separated and permanently ligated, and the neck skin incision was then closed. (iii) The mice were immobilized on their left side, a 1 cm skin incision was made between the external canthus and the auditory meatus of the right eye to expose the temporalis muscle, which was carefully separated to expose skull, then the right middle cerebral artery cortical branches were identified, a miniature cranial drill was used to create a ∼1.5 mm diameter hole, and the cortical branch of the right middle cerebral artery was exposed. (iv) The cortical branch of the right middle cerebral artery was cauterized with a coagulation pen, minimizing damage to the surrounding brain tissue as much as possible. (v) After confirming under the microscope that no reperfusion occurred in the arterial branch, the skin was sutured. (vi) The mice were maintained in a clean, sterile cage at 37°C with controlled humidity until they recovered from anesthesia, after which they were provided with soft food and allowed free access to food and water. Animals were monitored daily for general condition, mobility, and signs of pain or distress. All surgical procedures were performed under aseptic conditions. No prophylactic antibiotics were routinely administered in accordance with institutional guidelines. dMCAO group: the dMCAO mouse model was established by the above methods. Sham group: the model establishment procedures were the same as the dMCAO group, but the right CCA was separated without ligation, the skull was exposed and drilled, but the cortical branches of the right middle cerebral artery were not cauterized. Exclusion criteria: (i) Intraoperative subarachnoid hemorrhage. (ii) Blood flows after two rounds of electrocoagulation. For experiments, the mice in each group were graded and euthanized, and any tissue or organs were collected at the corresponding time points.

Neuroethological tests

The following four behavioral tests were performed on mice 1 day before dMCAO and 1, 3, 7, and 14 days after dMCAO, by researchers who were unaware of the experimental groups.

Staining of brain tissue sections

Frozen section: After thoroughly perfusing the brain with saline, the tissue was fixed in 4% paraformaldehyde, dehydrated in 30% sucrose, and sectioned coronally with an ultrathin frozen slicer to a thickness of 15 μm.

Paraffin sections: After euthanizing the mice, brain tissue was extracted and placed in a 4% (w/v) paraformaldehyde solution for fixation at 4°C for 24 h. Following fixation, the tissue underwent routine ethanol gradient dehydration, followed by clearing, paraffin embedding, and sectioning. Continuous coronal sections with a thickness of 4 μm were prepared.

Golgi staining

Golgi staining was performed using GolgiStain Kit (Shanghai Jiemei Gene, GMS80020.2) according to the manufacturer’s instructions. Mice were anesthetized with pentobarbital sodium, and the brains were removed and immersed in 4% paraformaldehyde at 4°C for 24 h. In the dark, brain tissues were immersed in a mixture of 7 mL GENMED fixative solution A and 7 mL GENMED fixative solution B and incubated at room temperature for 14 days. The tissue was then transferred to 30% sucrose solution C and kept at 4°C in the dark for 48 h. The brains were sectioned at 50 μm using a microtome (VT1000S, Leica, Wetzlar, Germany). Sections were imaged using a bright-field microscope (Axio Imager, Zeiss). Dendritic spine density was quantified using ImageJ software by counting the number of spines per unit length of dendrite.

Transmission electron microscopy

Brain tissue samples from the peri-infarct cortex were rapidly dissected and cut into small blocks (∼1 mm³), then fixed in 2.5% glutaraldehyde prepared in 0.1M phosphate buffer (PB, pH 7.4) at 4°C overnight. After primary fixation, the samples were rinsed three times with 0.1M PB for 15 min each. The tissues were then postfixed in 1% osmium tetroxide for 4 h in the dark under a fume hood, followed by rinsing twice with 0.1M PB for 20 min each. Samples were dehydrated through a graded acetone series, including 50%, 70%, 90%, and 95% acetone for 20 min each at 4°C, followed by 100% acetone for 20 min twice at room temperature. For embedding, samples were sequentially infiltrated with acetone and embedding medium mixtures at room temperature as follows: acetone: embedding medium = 3:1 for 3–4 h, 1:1 for 5–6 h, and 1:3 overnight. The samples were then transferred into pure embedding medium and incubated at 37°C overnight, followed by embedding in flat molds and polymerization at 65°C for 48 h. After polymerization, the samples were sectioned at 100 nm, lifted onto 3-mm copper grids, and stained for 30 min in 1.5% aqueous uranyl acetate and for 10 min in lead citrate. The sections were finally dried and viewed on a Hitachi H7000 TEM (Bruker, Germany).

RNA isolation and real-time quantitative PCR

Total RNA was extracted from brain tissues using TRIzol-based method (Thermo Fisher, 15596018) following a standard multistep protocol. Briefly, tissue samples were homogenized in TRIzol reagent, followed by phase separation with chloroform. After centrifugation, the aqueous phase containing RNA was transferred to a new tube, and RNA was precipitated using isopropanol. The RNA pellet was washed with 75% ethanol, air-dried, and resuspended in RNase-free water. RNA concentration and purity were assessed using NanoDrop. RNA purity was evaluated by measuring the A260/A280 ratio, and only samples meeting the quality criteria were used for reverse transcription before downstream analysis. Using a Prime Script RT Kit (Takara, rr047a), total RNA (1 μg) was reverse transcribed into cDNA. Using a Bio Rad CFX Connect Real-Time system, SYBR Green PCR Master Mix (CWBIO, CW0957) was used for PCR amplification. The cycle parameters were as follows: 95°C for 10 min, then 40 cycles of 95°C for 15 s and 60°C for 1 min. Relative mRNA abundance was calculated using the 2^−ΔΔCT method, with β-actin as the internal reference gene and the Fpn1^flox/flox CS group as the calibrator, and primers used were shown in Supplementary Table S2.

Determination of iron content in tissues by ICP-MS

After perfusion with saline, the mouse brain tissue was extracted, weighed, and transferred into 1.5 mL metal-free digestion tubes (NEST, 615601). No predrying or ashing step was performed, and wet tissue weight was used directly. Tissue samples were digested in 500 μL of 65% nitric acid (Thermo Fisher, HPLC grade, 010984.AK) overnight at room temperature. On the following day, the tube caps were opened in a fume hood, and the remaining acid was evaporated in a 90°C metal bath for 20 min. Subsequently, 500 μL of H₂O₂ (Sigma, #323381) was added, and the samples were heated in a metal bath at 70°C for 15 min and then at 100°C until the remaining acid had evaporated. In the blank control, only nitric acid was added to the tube. Finally, the digested residue was reconstituted to a final volume of 1 mL with deionized water (resistivity ≥18.2 MΩ·cm).

Iron concentrations were measured by ICP-MS using an iCAP RQplus system (Thermo Fisher Scientific). An internal standard (Indium, SCP Science, 140-051-491) was added to all standards and samples to correct for instrumental drift and matrix effects. Calibration standards were prepared by serial dilution of a certified iron standard solution (SCP Science, 700-101-121) in the same acid matrix as the samples to generate a calibration curve covering (concentration range, 0, 5, 10, 50, 100, 200 μg/L). Calibration blanks and standards were measured together with the samples under the same conditions. Iron concentrations were calculated from the calibration curve and normalized to [wet tissue weight/dry tissue weight].

Western blot analysis

Protein lysis buffer was prepared by adding 10 μL phenylmethanesulfonylfluoride to 1 mL radioimmunoprecipitation assay buffer (50.0 mM Tris–HCl, pH 7.4, 150.0 mM NaCl, 1% NP40, 0.1% SDS) at 4°C, which was then immediately added to mouse brain tissue at a ratio of 1 mg tissue to 15 μL buffer. The cerebral cortex tissue was cut into pieces on ice, and ultrasonicated (power: 560 W, 10 s of sonication and 10 s of rest, repeated six times) until the liquid became clear without any lumps. The mixture was centrifuged at 12,000 × g at 4°C for 20 min, and the supernatant was collected. The protein concentration in the supernatant was determined using a BCA protein concentration determination kit (Yeasen, 20201ES76). Protein samples of 25 µg were resolved by sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS–PAGE) and then transferred to nitrocellulose membranes. After blocking with 5% nonfat milk in TBST for 90 min at room temperature, membranes were incubated overnight at 4°C with primary antibodies diluted in TBST. Next, the membranes were washed four times for 15 min with TBST (pH 7.6, 137.0 mM NaCl, 0.05% Tween-20), after which HRP-conjugated goat antimouse IgG or HRP-conjugated goat antirabbit IgG diluted in TBST was added for 90 min at room temperature. The membranes were again washed four times for 15 min with TBST. The protein bands were visualized by chemiluminescence imaging, and the relative content was normalized to β-actin or glyceraldehyde-3-phosphate dehydrogenase (GAPDH) content. The Antibodies information and working dilutions were presented in Supplementary Table S3.

Drug injection

On days 1–5 after the operation, 20 mg/mL BrdU solution (Sigma, B9285) was injected intraperitoneally at a dose of 50 mg/kg, and the control group was injected with the same dose of normal saline. 20 mM AG490 (MCE, HY-12000) or dimethyl sulfoxide (DMSO) was injected intraperitoneally at 3 mg/kg 1 h before the operation and on days 1, 3, 5, 7, 9, 11, and 13 after the operation. The Fpn1^flox/flox mice and Fpn1^Nestin-CKO mice were divided into four groups: Fpn1^flox/flox dMCAO + DMSO group, Fpn1^flox/flox dMCAO + AG490 group, Fpn1^Nestin-CKO dMCAO + DMSO group, and Fpn1^Nestin-CKO dMCAO + AG490 group (Supplementary Fig. S1C–S1D).

Statistical analysis

The densities of Western blot image bands were quantified using Image J software, and the protein content was normalized to their respective values of β-actin or GAPDH bands. The positive signals of the target proteins in the images of the cortex obtained through immunofluorescence/immunohistochemical staining were quantified using Image J software based on the coverage area and quantity of the same molecule in the same area size of the region of interest in different groups. The proportions obtained were then analyzed. The Graph Pad Prism 8.0 was used to generate graphs.

The experimental data were processed and analyzed using SPSS 26.0 software. Normality and homogeneity of variance were assessed using the Shapiro-Wilk and Levene’s tests, respectively. Depending on data distribution and variance equality, appropriate parametric or nonparametric tests were applied, with Welch’s correction used when variances were unequal. For behavioral experiments with repeated measurements over time, we first applied linear mixed-effects models to evaluate overall group × time interactions. Subsequently, two-way analysis of variance (ANOVA) was performed at each individual time point to further examine group differences. Homogeneity of variance was assessed using Levene’s test at each time point, followed by appropriate post hoc comparisons: Tukey’s test was used when variances were equal, whereas Games–Howell correction was applied when variances were unequal. The data of >2 variables were analyzed using two-way ANOVA, with time as a within-subject factor and treatment/genotype as between-subject factors. Student’s t-test was used to analyze the difference between the two groups. The quantitative data are presented as the mean ± standard deviation. The difference of p < 0.05 was considered to be statistically significant. Electronic laboratory notebook was not used.