Gelatin, dextran, bromoacetic acid, silver nitrate, anhydrous acetonitrile, dichloromethane, N-hydroxysuccinimide (NHS), 1-ethyl-3-(3-dimethylaminopropyl) carbodiimide (EDC), 2-(N-morpholino) ethanesulfonic acid (MES), sodium periodate, Pluronic F127 (PF127), 4-carboxyphenylboronic acid, 4-(dimethylamino)pyridine (DMAP), and tetrahydrofuran were purchased from Adamas (Shanghai, China). ISO-1 was obtained from MedChemExpress (Shanghai, China). Cell Counting Kit-8(CCK-8 assay kit), ROS assay kit, and NO assay kit were purchased from Beyotime Biotech Inc. (Shanghai, China). Dulbecco's modified Eagle's medium (DMEM) media, fetal bovine serum (FBS), and penicillin-streptomycin double antibiotics were purchased from Gibco (New York, NY, USA).

Synthesis of Gel–NO: Bromoacetic acid (2.00 g, 14.3 mmol) and silver nitrate (3.70 g, 21.8 mmol) were stirred in anhydrous acetonitrile (80 mL) at 70 ℃ for 18 h. After filtration, the product was concentrated under vacuum, and dichloromethane (100 mL) was added. The mixture was incubated at room temperature for 2 h, filtered again, and the dichloromethane was removed to obtain the yellow oily nitroxy acetic acid. 2-(Nitroxy) acetic acid (3.5 mL), NHS (6.3 g), and EDC (10.5 g) were dissolved in 100 mL of MES buffer solution (pH 6.0) to activate the carboxyl group. After 2 h, gelatin (5.5 g) was added and stirred for 48 h, and then dialyzed with deionized water (DW) and lyophilized. Synthesis of ODex: 10.0 g of dextran and 8.0 g of sodium periodate were dissolved in DW and stirred at 37 ℃ for 4 h. Then, 1.2 mL of ethylene glycol was added to the above solution and stirred for 2 h to mix evenly. Finally, it was dialyzed with DW for 48 h and lyophilized to obtain ODex. Proton nuclear magnetic resonance (¹H NMR) confirmed the synthesis of these substances.

20 g of PF127 was dissolved in dichloromethane, followed by the addition of 1.5 g of 4-carboxyphenylboronic acid, 1.92 g of EDC, and 611 mg of DMAP, and reacted under nitrogen protection to form PF127B. Synthesis of PF127B nanoparticles: 100 mg of PF127B was dissolved in 5 mL of tetrahydrofuran and processed with a rotary evaporator at 40 ℃ for 15 min and dried in an oven at 40 ℃ for 20 min to obtain PF127BA nanoparticles. PF127B@ISO-1 nanoparticles: 100 mg of PF127B and 10 mg of ISO-1 were each dissolved in 5 mL of tetrahydrofuran, with subsequent steps being the same as the synthesis of PF127B nanoparticles. The particle size of PF127BA and PF127B@ISO-1 was measured by dynamic light scattering (DLS, Malvern Instruments Ltd., Zetasizer, Malvern, UK). Particles were observed using a transmission electron microscope (TEM, Hitachi, Ltd., H600, Tokyo, Japan).

The lyophilized polymer was dissolved in phosphate buffered saline (PBS), and the Gel–NO and ODex solutions were mixed in equal proportions to form a hydrogel rapidly. 1 mL of hydrogel containing PF127BA (0.2 mg) or PF127B@ISO-1 (0.2 mg) was abbreviated as GNO–OD HG@ NPs or G_NO–OD HG@ISO-1 NPs. The three groups of hydrogels were named as follows: G_NO–OD HG, Hydrogel grafted with NO; G_NO–OD HG@ NPs, Hydrogel with empty nanoparticles; G_NO–OD HG@ISO-1 NPs, hydrogel encapsulating ISO-1 loaded nanoparticles.

The rheological behavior was measured using a rheometer (Anton Paar GmbH, MCR302, Graz, Austria). Frequency sweeps were performed with oscillation frequencies ranging from 0.1 to 10 rad/s and a fixed strain of 1%. Strain sweeps ranged from 0.01% to 1000% with a fixed oscillation frequency of 1 rad/s. The fracture structure of the lyophilized hydrogel was observed using scanning electron microscopy (SEM, Hitachi, Ltd., S4800, Tokyo, Japan).

In vitro drug release was detected using UV–Vis and the Griess assay kit (NO assay kit, Beyotime Biotech Inc., China). G_NO–OD HG@ISO-1 NPs were placed in dialysis bags containing 10 mL of PBS (pH = 7.4) and 10 mL of PBS (pH = 6.5) + 1 mmol/L H₂O₂ solution (n = 3). On Days 0, 0.25, 0.5, 1, 2, 3, 4, 7, and 14, 2 mL of dialysate was taken and the release of ISO-1 was measured by UV spectrophotometry, and the release of NO was measured using the Griess assay kit.

HT22 and BV2 were purchased from Procell Life Science & Technology Co., Ltd. (Wuhan, China). Cells were routinely cultured with complete medium (DMEM high glucose, 10% FBS, 1% penicillin/streptomycin). The complete medium was replaced with D-glucose-free DMEM and placed in a container with 37 ℃, 5% CO₂, 94% N₂, and 1% O₂ for 2 h, then replaced with complete medium and placed in a regular environment to establish the OGD model. Control group cells were cultured routinely.

CCK-8 and Live/Dead staining evaluated cell compatibility and metabolic activity. First, 1 mL of hydrogel was immersed in DMEM complete medium (10 mL) for 48 h to obtain the dialysis solution; next, HT22 and BV2 cells were seeded in a 96-well plate (3 × 10³ cells/well) and subjected to OGD, then replaced with hydrogel dialysis solution and treated for 48 h. Cell viability was measured using the CCK-8 assay kit. FDA/PI staining kit (BestBio, Nanjing, China) was used, and cell survival was observed under a fluorescence microscope (Leica, DMI 4000, Wiesbaden, Germany).

BV2 cells were seeded in a 6-well plate. After 48 h of OGD and hydrogel treatment, the cell supernatant was collected, and the protein concentrations of IL-6, IL-10, and MIF were quantitatively determined using an ELISA assay kit (Boster, Wuhan, China) according to the instructions.

The photothrombotic cerebral ischemia model was used to induce focal cortical ischemia. All experimental procedures were executed according to the protocols approved by the Animal Ethics Committee of Sichuan University (No.20240226245). Male Sprague Dawley (SD) mice were used for the animal experiments (Ensiweier Experimental Animals Co., Ltd., Beijing, China). First, mice were anesthetized with isoflurane, and 1% Bengal Rose (10 μL/g body weight, Aladdin, Shanghai, China) was injected intraperitoneally; 15 min later, the mice were fixed in a stereotactic apparatus (RWD Life Science, Shenzhen, China), and a 3 mm bone window was made 2 mm lateral to the anterior fontanelle with a trephine. Then, the motor cortex area was irradiated for 8 min through the bone window using a 100-mW cold light illuminator, and the incision was sutured after irradiation. During the recovery period, the health status of the mice was monitored daily.

After 3 days of completing the cerebral ischemia model, the mice were anesthetized with isoflurane and fixed on a stereotactic apparatus. Hydrogel was injected into the central area of the infarction zone 2 mm below the endocranium using a micro syringe, with 2 μL injected into the infarction area of each mouse. After the injection, the needle was slowly withdrawn after staying for 5 min, and the incision was disinfected with iodine and sutured. Experimental animals were randomly divided into the following 5 groups: (a) Sham operation group (sham), (b) Photothrombotic ischemic stroke model (PT), (c) Ischemic stroke model injected with Gel–NO hydrogel group (G_NO–OD HG), (d) Ischemic stroke model injected with Gel–NO+empty PF127BA nanoparticle hydrogel group (G_NO–OD HG@ NPs), (e) Ischemic stroke model injected with Gel–NO+PF127B@MIF inhibitor (ISO-1) nanoparticle hydrogel group (G_NO–OD HG@ISO-1 NPs).

Neurological function was measured using the modified neurological severity score (mNSS), rotarod test, and open field test after stroke in mice. Mice were trained for 3 consecutive days before stroke modeling for mNSS, rotarod test, and open field test. Neurobehavioral tests were performed on Days 0, 1, 7, 14, and 21 after modeling. The mNSS ranges from 0 to 18 points, reflecting changes in motor, sensory, balance, and reflex functions. Higher scores indicate more severe neurological deficits, while lower scores indicate neurological function recovery. The rotarod test was performed on a fixed-speed rotarod device to assess the mouse's motor coordination and balance ability; the stay time was recorded by the sensor. The longer time mice stay, the better the coordination and balance ability they have. Open field test: The experimental animals were placed in a 25 cm × 25 cm × 25 cm empty box, with a single mouse placed in the center of the open field and allowed to move freely for 5 min. The activity distance and time in the central and peripheral areas were analyzed to evaluate the spontaneous activity of the experimental animals.

On Days 1 and 21 after mouse modeling, the mice were continuously anesthetized with an animal isoflurane gas anesthesia device. The anesthetized mice were placed on the imaging stage of the laser speckle blood flow instrument (RWD Life Science, RFLSI ZW, Shenzhen, China), the entire brain of the mouse was exposed, the position was adjusted, and the automatic focusing mirror was used to make the image clear and complete. The parameters of the laser speckle blood flow instrument were adjusted, the laser viewing light source was 780 nm; the irradiation area was through a charge-coupled device (CCD), the exposure time was 5 ms, the frame frequency was 30 fps, the imaging speed was 200 frames per second, and a 1280 × 960-pixel image was obtained. The region of interest (ROI) was outlined to observe vascular changes and read blood perfusion values.

Brain tissue was taken for RNA-seq testing 21 days after hydrogel injection. Total RNA was isolated using the RNeasy mini kit (Qiagen, Dusseldorf, Germany). The cBot and the library generated the gene cluster and then sequenced it on the Illumina NovaSeq 6000 (Santiago, CA, USA). Library construction and sequencing were completed by Metware Biotechnology Co., Ltd. (Wuhan, China).

Values are expressed as mean ± standard deviation (SD). ANOVA statistical methods were used to compare differences between groups. Data analysis was performed using GraphPad Prism software (Santiago, CA, USA). ∗P < 0.05, ∗∗P < 0.01, ∗∗∗P < 0.001, and ∗∗∗∗P < 0.0001.

ODex and NO grafted gelatin (Gel–NO) were crosslinked through the Schiff base reaction to form a hydrogel (G_NO–OD HG). To ascertain the optimal ratio for the synthesis of hydrogels, we took both injectability and drug release into account, ultimately selecting a ratio of oxidized dextran to gelatin of 1:1 for our experiment (Supporting Information Fig. S1). In simple terms, dextran was oxidized with sodium periodate to form ODex. Gel–NO was synthesized by the reaction of 2–(nitroxy) acetic acid with gelatin under a catalytic system of EDC/NHS and nitrogen protection (Supporting Information Fig. S2A and B). The successful synthesis of ODex and Gel–NO were confirmed by ¹H NMR hydrogen spectroscopy. The characteristic proton signal at 5.00 ppm is the evidence for the synthesis of Gel–NO (Fig. 1A). The successful synthesis of ODex was evidenced by a distinctive peak at 8.31 ppm (Fig. 1B). Concurrently, PF127 was reacted with carboxylic acidphenylboronic acid under a catalytic system of EDC/DMAP and under nitrogen protection to form PF127BA (Supporting Information Fig. S2C). The characteristic peak at 7.63 ppm indicated the successful synthesis of PF127BA (Fig. 1C). Due to the hydrophobicity of ISO-1, ISO-1 was encapsulated within PF127BA nanoparticles using the thin-film hydration method³⁴. The size and shape of the nanoparticles were measured using DLS and TEM. The size of PF127BA and PF127B@ISO-1 nanoparticles was shown in Fig. 1D, with average diameters of 101.33 nm and 79.29 nm, respectively. The particle size of drug-loaded nanoparticles is slightly smaller than that of blank nanoparticles, which may be attributed to the increased density of the nanoparticles upon drug loading. This enhancement in density could facilitate the aggregation of nanoparticles during the fabrication process, resulting in reduced particle size³⁵. The diminished particle diameter potentially augments their diffusivity and penetrating capability within biological systems. TEM images showed that both PF127BA and PF127B@ISO-1 nanoparticles are uniformly sized spherical particles (Fig. 1E). The sol-gel transition was demonstrated using the inverted vial method, and the injectability of the hydrogel was observed by injecting it with a syringe, indicating that the synthesized hydrogel has good gelation and injectability (Fig. 1F). The internal fracture structure of the three freeze-dried hydrogels was captured using SEM, showing similar structural information with honeycomb-like pores of similar sizes (Fig. 1G). Frequency sweep results showed that within the frequency range of 0.1–100 rad/s, the storage modulus (G′) is always greater than the loss modulus (G′′), indicating the gel state under test conditions (Fig. 1H). The modulus of the hydrogel at 1–2 kPa is comparable to that of the mouse cerebral cortex³⁶. To determine the ISO-1 release efficiency under physiological and pathological conditions, tests were conducted in PBS pH = 7.4 and PBS pH = 6.5 with 1 mmol/L H₂O₂, respectively. The drug loading of ISO-1 NPs was 43.87 ± 4.04%. The degree substitution of NO donor was determined to be 29.89%. The release curve shows that the final release rate of ISO-1 in G_NO–OD HG@ISO-1 NPs in PBS pH = 7.4 after 14 days is 54.67%, while in PBS pH = 6.5 with 1 mmol/L H₂O₂, the final release rate is 70.69% (Fig. 1I). It can be observed that the release of ISO-1 conforms more closely to the first-order rate model under both conditions (Supporting Information Table S1). Thus, it can be seen that the release efficiency of ISO-1 in an inflammatory microenvironment is higher than in a physiological environment. Similarly, the release of NO was also monitored, with Fig. 1J showing the release of NO from the G_NO–OD HG@ISO-1 NPs hydrogel. Similar to the release of ISO-1, the cumulative concentration of NO released under the condition of PBS pH = 6.5 with 1 mmol/L H₂O₂ is higher than in PBS pH = 7.4. This is attributed to the rapid degradation of the hydrogel under the inflammatory microenvironment. Under the condition of PBS pH = 6.5 with 1 mmol/L H₂O₂, the cumulative concentration of NO released on the 14th day is 42.54 μmol/L, while in PBS pH = 7.4, the cumulative concentration is 33.65 μmol/L (Fig. 1J). The condition of PBS pH = 6.5 with 1 mmol/L H₂O₂ is similar to the pathological microenvironment after cerebral infarction, indicating that injecting the hydrogel designed in this study into the infarction area has good drug release efficiency and a modulus that matches the rat cerebral cortex, making it feasible for in situ injection in the stroke cavity³⁷.

To investigate the potential influence of the hydrogel on ischemic injury, we established an OGD model using HT22 and BV2 cell lines to simulate ischemic damage. The three types of hydrogels were co-cultured with HT22-/-BV2 cells for 48 h, then the cell compatibility and metabolic activity in HT22 and BV2 cells were assessed using CCK-8 and live/dead staining and utilizing the 2′,7′–dichlorodihydrofluorescein diacetate (DCFH-DA) probe method to quantify the ROS content within cells (Fig. 2A–E). The live/dead cell staining results showed that after 48 h of co-culture with G_NO–OD HG, G_NO–OD HG@ NPs, and G_NO–OD HG@ISO-1 NPs, the group co-cultured with G_NO–OD HG@ISO-1 NPs had a higher number of live cells and fewer dead cells for both HT22 and BV2 cells, indicating that the hydrogels have good cell compatibility and proliferative capacity (Fig. 2A and B). In terms of cell viability, the HT22 cell viability in the OGD group decreased to 47.66%, and the BV2 cell viability decreased to 59.79%. After 48 h of co-culture with G_NO–OD HG, G_NO–OD HG@ NPs, and G_NO–OD HG@ISO-1 NPs, the HT22 cell viability showed a significant improvement compared to the OGD group, with 52.74%, 60.96%, and 88.64%, respectively. The BV2 cell viability also showed a significant improvement compared to the OGD group, with 65.58%, 71.90%, and 79.04%, respectively (Fig. 2C and D). This suggested that these hydrogels had good cell compatibility and could improve the activity of HT22 and BV2 affected by OGD.

ROS are primarily produced due to the ischemic and glucose-deprived environment that simulates ischemic stroke, leading to energy metabolism imbalance, and ROS are also one of the mechanisms that further affect the progression of ischemic stroke³⁸. DCFH-DA is a probe that can detect ROS, exhibiting green fluorescence. DCFH-DA staining results showed significant green fluorescence in the OGD group, indicating a large amount of ROS production. However, in the co-culture with HT22 and BV2 cells and G_NO–OD HG, G_NO–OD HG@ NPs, and G_NO–OD HG@ISO-1 NPs, there was a decrease in both the number of ROS and fluorescence intensity, indicating that the hydrogel had some scavenging effect on ROS induced by OGD (Fig. 2D and E).

To explore the hydrogel's effect on controlling inflammation, we co-cultured G_NO–OD HG, G_NO–OD HG@ NPs, and G_NO–OD HG@ISO-1 NPs with BV2 cells for 48 h. We examined the expression of MIF, which is specifically inhibited by ISO-1, and found that G_NO–OD HG and G_NO–OD HG@ NPs showed a decrease compared to the OGD group, while the G_NO–OD HG@ISO-1 NPs group showed a significant decrease, with MIF levels approaching those of the normal control group (Fig. 2F). Additionally, the expression levels of IL-6 were measured and found to have decreased in all three hydrogel treatment groups compared to the OGD group. We also examined the level of IL-10 and found that the expression level of IL-10 was increased, but not statistically significant (Fig. 2G and H). This indicates that the hydrogel has a good anti-inflammatory effect, and G_NO–OD HG@ISO-1 NPs can specifically inhibit MIF, while the decrease in MIF in other hydrogels may be due to their universal anti-inflammatory effect^39-41.

We established a PT stroke model in mice to explore the in vivo therapeutic effects of the hydrogel on ischemic injury. 3 days after the model establishment, different hydrogels or saline were injected via a stereotactic apparatus into the brain. The specific experimental route of animal studies in this research is shown in Fig. 3A. Mice were randomly divided into 5 groups: sham operation group (sham), saline-injected post-stroke group (PT group), G_NO–OD HG-injected post-stroke group (G_NO–OD HG), G_NO–OD HG@ NPs-injected post-stroke group (G_NO–OD HG@ NPs), and G_NO–OD HG@ISO-1 NPs-injected post-stroke group (G_NO–OD HG@ISO-1 NPs). To verify the success of the model, we observed the PT and sham groups of mice through hematoxylin and eosin (H&E) staining. We found that under H&E staining, the PT group exhibited numerous cellular vacuoles and shrunken nuclei, indicating the presence of infarcts (Supporting Information Fig. S3).

To further explore the underlying mechanisms, we focused on anti-inflammatory, angiogenic, and neuroprotective pathways, detecting relevant proteins at corresponding time points through immunofluorescence (IF). During the first week, we assessed apoptosis in brain tissue cells using TUNEL expression (Fig. 3B). Compared to the sham group, the number of TUNEL-positive cells significantly increased in the PT group (0.12 ± 0.21 vs. 7.50 ± 1.29, P < 0.001), and after treatment with G_NO–OD HG, G_NO–OD HG@ NPs, and G_NO–OD HG@ISO-1 NPs, the number of TUNEL-positive cells decreased compared to the PT group (6.07 ± 0.94, 5.12 ± 1.15, 3.93 ± 0.36), shown in Fig. 3C. Additionally, we examined 4-hydroxynonenal (4-HNE), a highly toxic lipid peroxidation end product associated with ferroptosis³⁷. The results showed a significant increase in 4-HNE levels in the PT group compared to the sham group (9.17 ± 1.25 vs. 1.91 ± 0.21, P < 0.001). After treatment with G_NO–OD HG (5.24 ± 0.82), G_NO–OD HG@ NPs (4.52 ± 0.82), and G_NO–OD HG@ISO-1 NPs (3.57 ± 0.62), the number of 4-HNE-positive cells significantly decreased, especially in the G_NO–OD HG@ISO-1 NPs group (Fig. 3D and E). This suggested that NO and ISO-1 exert neuroprotective and antioxidant effects, inhibiting neuronal apoptosis and oxidation to a certain extent^42,43.

Consequently, we performed H&E staining and Nissl body staining on mouse brain sections at 7 and 21 days post-modeling. According to our experimental results, in H&E staining, it can be observed that on Day 7, during the early phase of cerebral infarction, compared to the PT group, the brain tissue exhibited cortical neuronal cell shrinkage, visible necrotic neurons, a reduction in the neuronal count, and vacuole formation between tissues (Fig. 3F). In contrast, the treatment groups showed milder neuronal damage. By Day 21, the ischemic injury in the PT group's brain tissue had alleviated compared to Day 7. On Day 21, ischemic injuries in all treatment groups had also significantly improved, especially in the G_NO–OD HG@ISO-1 NPs group.

Regarding neuronal restoration, Nissl body staining revealed that on Day 7, compared to the sham group, the PT group exhibited a reduction in the number of Nissl bodies in the cerebral cortex, with deeply stained and condensed nuclei and disordered arrangement. In the treatment groups, however, the number of Nissl bodies increased (Fig. 3G). By Day 21, there was a slight recovery in the morphology and quantity of Nissl bodies in the PT group, but there was still a significant difference compared to the sham group. The treatment groups showed a marked increase in the number of Nissl bodies compared to Day 7 and were greater than those in the PT group.

In the second week of treatment, we assessed the inflammatory level by observing proteins such as CD86, CD206, MIF, MMP-9, IL-6 and IL-10 (Fig. 4 and Supporting Information Fig. S4). CD86 and CD206 are surface-specific markers of the activated microglial M1 pro-inflammatory phenotype and M2 anti-inflammatory phenotype, respectively^44,45. IF results showed that the number of CD86 and CD206 positive cells significantly increased in the PT group compared to the sham group, and after treatment with G_NO–OD HG, G_NO–OD HG@ NPs, and G_NO–OD HG@ISO-1 NPs, both CD86 and CD206 significantly decreased compared to the PT group, especially in the G_NO–OD HG@ISO-1 NPs treatment group (Fig. 4A, B, G and J). In the PT, G_NO–OD HG, G_NO–OD HG@NPs, and G_NO–OD HG@ISO-1 NPs groups, the numbers of CD86 protein-positive cells were 6.19 ± 0.74, 5.48 ± 0.90, 2.26 ± 0.74, and 1.67 ± 0.55, respectively. And, the numbers of CD206 protein-positive cells were 3.33 ± 0.21, 1.79 ± 0.36, 1.19 ± 0.41, and 0.71 ± 0.36, respectively. These results indicated a significant modulation of the immune response as evidenced by the expression of the CD86 and CD206 co-stimulatory molecules. This indicated that NO and ISO-1 exhibited excellent anti-inflammatory effects. Similarly, the expression of MIF and MMP-9 also decreased in the treatment groups with G_NO–OD HG, G_NO–OD HG@ NPs, and G_NO–OD HG@ISO-1 NPs, with the most significant decrease in the G_NO–OD HG@ISO-1 NPs group. In the PT, G_NO–OD HG, G_NO–OD HG@NPs, and G_NO–OD HG@ISO-1 NPs groups, the numbers of MIF protein-positive cells were 8.69 ± 1.49, 6.91 ± 1.49, 6.19 ± 1.61, and 4.64 ± 0.94, respectively. And, the numbers of MMP-9 protein-positive cells were 12.02 ± 1.44, 10.83 ± 2.38, 9.88 ± 0.90, and 7.86 ± 1.99, respectively (Fig. 4C, D, H and K). This suggested that G_NO–OD HG@ISO-1 NPs had a more pronounced specific inhibitory effect on MIF and MMP-9. To corroborate the results of the protein content of IL-6 and IL-10 detected in cell experiments, we observed the expression of these two proteins in mouse brain sections through IF staining. The results were consistent with the cell experiments, showing that the protein expression of IL-6 (Fig. 4E and L) in the brain sections of mice with cerebral infarction was higher compared to the sham group, while the expression of IL-6 in the three treated groups decreased. Conversely, compared to the sham group, the expression of IL-10 (Fig. 4F and I) was lower in the PT group, and the expression of IL-10 in the three treated groups was higher than in the PT group, especially in the G_NO–OD HG@ISO-1 NPs group.

In the third week of treatment, we further assessed the proteins related to angiogenesis, CD31, α-SMA, eNOS, PDGFR, and VEGF (Fig. 5 and Supporting Information Fig. S5). In the PT, G_NO–OD HG, G_NO–OD HG@NPs, and G_NO–OD HG@ISO-1 NPs groups, the numbers of CD31 protein-positive cells were 11.90 ± 1.25, 13.33 ± 1.35, 14.17 ± 3.08, and 14.64 ± 2.50, respectively. In the PT, G_NO–OD HG, G_NO–OD HG@NPs, and G_NO–OD HG@ISO-1 NPs groups, the numbers of α-SMA protein-positive cells were 3.45 ± 0.90, 5.60 ± 0.21, 6.43 ± 1.29, and 7.02 ± 1.49, respectively. It can be observed that the levels of these two proteins increased in the treatment groups with G_NO–OD HG, G_NO–OD HG@ NPs, and G_NO–OD HG@ISO-1 NPs (Fig. 5A and B), indicating that the designed hydrogels can promote angiogenesis during the third week. We also detected the expression of endothelial nitric oxide synthase (eNOS, Fig. 5C), platelet-derived growth factor receptor (PDGFR, Fig. 5D), and vascular endothelial growth factor (VEGF, Fig. 5E) and found that all three proteins were highly expressed in the G_NO–OD HG@ISO-1 NPs group. eNOS is an enzyme expressed in vascular endothelial cells that catalyzes the conversion of L-arginine to NO. The NO produced by eNOS can increase vascular relaxation, thereby improving cerebral blood flow, which has a positive effect on the recovery of blood flow in the infarcted area. During the repair process after cerebral infarction, eNOS may be involved in promoting new blood vessel formation, helping to restore blood supply to damaged brain tissue. In cerebral infarction, eNOS may also interact with inflammatory cytokines such as IL-6, affecting inflammatory responses and cellular damage. PDGFR mainly plays a role in the proliferation and migration of vascular smooth muscle cells and pericytes, contributing to vascular reconstruction and repair after cerebral infarction. VEGF is a group of signaling proteins with angiogenic effects, playing a key role in neovascularization, vascular repair, and vascular maintenance. The increase in these three proteins in the G_NO–OD HG@ISO-1 NPs group indicates that hydrogel treatment promotes the process of angiogenesis.

Therefore, in vivo, we used a laser speckle blood flow imager to observe cerebral ischemia and angiogenesis, on the first day after modeling, compared to the sham group, other groups showed significant cerebral ischemic areas with obvious vascular damage. On the Day 21 after modeling, compared to Day 1, the cerebral ischemic areas decreased in all groups, while the PT group showed less reduction, with a large part of the ischemic area still present, while the treatment groups G_NO–OD HG, G_NO–OD HG@ NPs, and G_NO–OD HG@ISO-1 NPs showed a significant improvement in cerebral ischemia and re-established a rich blood supply (Fig. 5F). These results indicate that treatment of cerebral infarction via the designed hydrogel can promote angiogenesis and reestablish blood perfusion. Such effects are also conducive to the recovery of neurological function in the later stages of cerebral infarction^46-48.

In terms of neuroprotection, we examined the expression of doublecortin (DCX, Fig. 6A) and Nestin (Fig. 6B) and found that both were highly expressed in the G_NO–OD HG@ISO-1 NPs group. DCX is a protein associated with microtubules and plays an important role in neuronal migration and plasticity. DCX helps in the reorganization and functional recovery of damaged neural networks, protecting neurons from cell death caused by ischemic injury. Nestin is an intermediate filament protein expressed in neural stem cells and progenitor cells. The upregulation of Nestin is usually associated with the activation and proliferation of neural precursor cells, which can differentiate into neurons and glial cells, participating in the process of neural regeneration and repair after cerebral infarction. Nestin is also expressed in vascular endothelial cells and during angiogenesis, especially in the formation of new blood vessels. Nestin-positive cells may be involved in the vascular reconstruction of the infarcted area, helping to improve blood flow in damaged brain tissue. This indicates that after hydrogel treatment, the related processes of neural regeneration and repair in mice were activated.

Neurological functional recovery analyses included mNSS, rotarod test, and open field test. Neurological functions were dynamically monitored from before surgery to 3 weeks after hydrogel injection. The mNSS score is a comprehensive assessment that includes motor, sensory, reflex, and balance functions⁴⁹. The lower the score is, the better the neurological function of mice. After the model establishment, there was a significant neurological deficit in mice, which began to recover after treatment with G_NO–ODHG, G_NO–OD HG@ NPs, and G_NO–OD HG@ISO-1 NPs (Fig. 6C). On Day 21, the G_NO–OD HG@ISO-1 NPs treatment group showed the best recovery of neurological function, with a significant difference from the PT group. The rotarod test measures the time mice can stay on a rotating rod, with a longer duration indicating better neurological function⁵⁰. On Day 21, the G_NO–OD HG@ISO-1 NPs treatment group had the longest duration on the rotarod, indicating the best recovery of neurological function (Fig. 6D). Finally, the open field test was conducted to assess the spontaneous behavioral and exploratory abilities of mice with ischemic stroke⁵¹. Fig. 6E recorded the movement trajectories of different groups in the field. Compared to the PT group, the G_NO–OD HG@ISO-1 NPs group showed a significant increase in movement distance (Fig. 6F) and movement speed (Fig. 6G), which means the mice's spontaneous movement and exploratory abilities were substantially improved. The other two groups (G_NO–OD HG and G_NO–OD HG@ NPs) also showed improved spontaneous movement abilities compared to the PT group. In summary, the neurological deficits in mice were partially recovered after treatment with G_NO–OD HG, G_NO–OD HG@ NPs, and G_NO–OD HG@ISO-1 NPs, especially in the G_NO–OD HG@ISO-1 NPs group, indicating that NO and ISO-1 significantly contributed to the recovery of neurological function.

On Day 21 post-treatment, brain tissue from the cerebral infarction area was harvested for mRNA sequencing. The gene expression profiles of the PT group and the composite hydrogel group (G_NO–OD HG@ISO-1 NPs) were analyzed to assess the effects of NO and ISO-1 at the genetic level. As shown in the figure, there were 375 differentially expressed genes between these two groups, with 241 upregulated and 134 downregulated (Supporting Information Fig. S6). The Kyoto Encyclopedia of Genes and Genomes (KEGG) database was used for enrichment pathway analysis. The inflammation and immune-related pathways included Th1 and Th2 cell differentiation, viral protein interaction with cytokine and cytokine receptor, Th17 cell differentiation, TNF signaling pathway, antigen processing and presentation, cytokine-cytokine receptor interaction, chemokine signaling pathway, and cell adhesion molecules. Neurological function-related pathways included the dopaminergic synapse and neuroactive ligand-receptor interaction (Supporting Information Fig. S7A). The KEGG classification showed that the differentially expressed genes were enriched in pathways related to cellular processes, environmental information processes, human diseases, metabolism, and organ systems (Fig. S7B). Some regulated genes were involved in various GO terms, such as chemokine activity, response to interferon-gamma and interferon-beta, Major histocompatibility complex (MHC) and MHC class II protein complex binding, and neurotransmitter transport (Fig. S7C). The differentially expressed genes were enriched in many GO terms, categorized as biological process (BP), cellular component (CC), and molecular function (MF) (Fig. S7D). These results indicated G_NO–OD HG@ISO-1 NPs exerted multiple functions to improve ischemic injury. We conducted an analysis of specific genes related to mechanisms such as angiogenesis, anti-inflammation, and neuroprotection using transcriptomic data. We found that some genes were highly expressed in the treatment group (Fig. 7). However, due to the limitation of sample size, the results obtained still require further validation.