In situ implantable DNA hydrogel for diagnosis and therapy of postoperative rehemorrhage following intracerebral hemorrhage surgery

Abstract

Postoperative rehemorrhage following intracerebral hemorrhage surgery is intricately associated with a high mortality rate, yet there is now no effective clinical treatment. In this study, we developed a hemoglobin (Hb)–responsive in situ implantable DNA hydrogel comprising Hb aptamers cross-linked with two complementary chains and encapsulating deferoxamine mesylate (DFO). Functionally, the hydrogel generates signals upon postoperative rehemorrhage by capturing Hb, demonstrating a distinctive “self-diagnosis” capability. In addition, the ongoing capture of Hb mediates the gradual disintegration of the hydrogel, enabling the on-demand release of DFO without compromising physiological iron-dependent functions. This process achieves self-treatment by inhibiting the ferroptosis of neurocytes. In a collagenase and autologous blood injection model–induced mimic postoperative rehemorrhage model, the hydrogel exhibited a 5.58-fold increase in iron absorption efficiency, reducing hematoma size significantly (from 8.674 to 4.768 cubic millimeters). This innovative Hb-responsive DNA hydrogel not only offers a therapeutic intervention for postoperative rehemorrhage but also provides self-diagnosis feedback, holding notable promise for enhancing clinical outcomes.

In situ implantable DNA hydrogel enables self-diagnosis and self-therapy for postoperative rehemorrhage.

INTRODUCTION

Intracerebral hemorrhage (ICH) represents a critical medical condition, characterized by the sudden and spontaneous leakage of blood into the brain parenchyma (1). It stands as the second most common subtype of stroke, drawing significant attention due to its exceptionally high mortality and morbidity rates (2, 3). Surgical interventions, including hematoma evacuation (4), are commonly used in the clinical management of ICH, holding the potential to prevent herniation and alleviate the pathophysiological impact of hematoma on surrounding tissues, thereby promptly saving lives (5–7). However, the risk of postoperative rehemorrhage, stemming from surgical trauma, site infections, or primary vascular risk factors, remains prevalent (~21.5%) (8). In addition, despite the benefits of reducing mass effect, improving intracranial pressure, and enhancing brain perfusion, the infiltrative nature of surgery often leaves residual products, affecting patient prognosis and potentially leading to rebleeding in severe cases (9, 10).

The occurrence of rehemorrhage may not be easily detected, particularly when the patient’s consciousness has not fully recovered after the initial surgery, resulting in missed opportunities for salvage and consequential fatal outcomes and a mortality rate as high as 42.6 to 83.3% (11). Moreover, the opportunity for a second craniotomy after rebleeding is extremely slim, and it can inflict greater suffering upon patients. Therefore, prompt treatment for postoperative rehemorrhage is imperative. Recent studies have shown that ferroptosis, an iron-dependent cell death pathway (12, 13), emerges as the primary factor causing nerve injury after hemorrhage (14). Iron chelators, such as deferoxamine mesylate (DFO) and minocycline (15–17), exhibit therapeutic advantages in treating ICH by sequestering excess iron, thereby preventing oxidative stress and mitigating cellular damage (18–20). This targeted intervention holds promise for restoring cellular homeostasis and improving outcomes in the context of brain hemorrhage therapy. For example, Jiayu Xie’s study showed that administering DFO significantly reduced the damage induced by ferroptosis after ICH by regulating iron metabolism, amino acid metabolism, and lipid peroxidation processes (21). However, despite significant progress in rodent studies, the clinical application of iron chelators remains limited (22), primarily due to challenges in controlling dosage and timing of administration, especially in complex cases of rehemorrhage.

Herein, we engineered a hemoglobin (Hb)–responsive in situ implantable DNA hydrogel precisely tailored to the specific pathological nuances of ICH (Fig. 1). Through the incorporation of Hb aptamers cross-linked with two complementary chains and encapsulating DFO, the hydrogel exhibited optimal injectability and biocompatibility, allowing it to fill irregular void spaces through localized injection after hematoma aspiration in clinic. Functionally, the hydrogel generates signals after responding Hb upon postoperative rehemorrhage through fluorescence resonance energy transfer, showcasing a unique “self-diagnosis” capability. In addition, the ongoing capture of Hb mediates the gradual disintegration of the hydrogel, enabling the on-demand release of DFO without compromising physiological iron-dependent functions. This process achieves self-treatment by inhibiting the ferroptosis of neurocytes. Its ability to perceive and respond to pathological cues ensures precise drug release in rebleeding microenvironments, aligning seamlessly with the pathological context for improved postoperative recovery. This innovative Hb-responsive hydrogel not only provides an effective therapeutic intervention for postoperative rehemorrhage but also delivers invaluable self-diagnosis feedback, promising enhanced clinical outcomes, serving as a beacon of hope for optimizing postoperative outcomes in the realm of ICH.

Fig. 1. In situ implantable DNA hydrogel for diagnosis and therapy of postoperative rehemorrhage following ICH surgery.

(A) Schematic diagram to show the construction and properties of hemoglobin (Hb)–responsive DNA hydrogel. Apt-linker, aptamer linker. (B) Schematic of the hematoma aspiration in clinic and implantation of the DNA hydrogel. (C) Self-diagnosis and self-treatment of the DNA hydrogel at rehemorrhage.

RESULTS

Synthesis and characterization of the DNA hydrogel

In this study, we initiated the preparation by dispersing anchor DNA strands in an acrylamide solution, followed by a polymerization reaction to create an acrylamide polymer with multiple DNA strands. Using the principle of Watson-Crick base pairing, DNA cross-linkers capable of complementary interaction with the anchor DNA were used to connect the acrylamide polymers, resulting in the synthesis of the DNA hydrogel. To identify specific toxic proteins of ICH lesion site, we used Hb aptamer as the cross-linker in the DNA hydrogel, connecting two other DNA chains. In the presence of Hb, the aptamer competes with the targets and prefers to form target-aptamer complexes instead of aptamer-DNA hybridization. This process results in the disassembly of the cross-linked aptamer-A-B duplex, leading to the release of drugs from the hydrogel (Fig. 2A). The number of hybrid nucleotides between strand B and the Hb aptamer linker influences the aptamer-Hb interactions, subsequently affecting the dissociation of strand B and the sensitivity of the DNA hydrogel. To address this, we designed 11 complementary oligonucleotides covering different regions of the aptamer (Fig. 2B). As illustrated in Fig. 2C, we used streptavidin-modified magnetic beads to investigate the sensitivity of various combinations to Hb response. On the basis of the results from polyacrylamide gel electrophoresis (PAGE), the design of B-7 exhibited the most significant response to Hb. Consequently, we selected the B-7 for subsequent experiments. To further demonstrate the responsiveness of the DNA strands combinations to Hb, strand B and the aptamer linker were labeled with a fluorophore (FAM) and a quencher (BHQ1), respectively. Upon the formation of the DNA complex, the fluorescence of FAM was quenched because of its proximity to the quencher group. Treatment with Hb restored the fluorescence of FAM, indicating that the de-hybridization of the double-strand DNA was triggered by the competitive binding of Hb to its specific aptamer (Fig. 2D). As depicted in Fig. 2E, the fluorescence recovery of FAM increased rapidly after different concentrations of Hb treatment, suggesting the dissociation of the DNA complex.

Fig. 2. Synthesis of the DNA hydrogel.

(A) Schematic diagram of Hb-responsive sol. (B) Design schemes for different B-chains. (C) Magnetic bead experiment to investigate the responsiveness of different B-chains by nondenaturing polyacrylamide gel electrophoresis (PAGE). (D) The principle of fluorescence assay for detecting Hb-responsive sol efficacy. The fluorescence intensity is the evaluation criteria of sol efficacy. The higher the fluorescence intensity, the higher the sol efficiency. (E) Fluorescence intensity spectra of S-B under different Hb concentrations. (F) Agarose gel to verify the cross-linking between S-A, S-B, and acrylamide chain. (G) Scheme of the DNA hydrogel formation. Schematic diagram of (C) was created with BioRender.com. bp, base pairs; a.u., arbitrary units.

In the current approach, two acrydite-modified oligonucleotides, namely, S-A and S-B, are individually copolymerized with acrylamide (4%, w/v) and incorporated into the polyacrylamide chains (namely, PS-A and PS-B, respectively). The Fourier transform infrared analysis showed that the characteristic absorption peaks of PS-A and PS-B are 3407, 2938, 1667, and 1452 cm⁻¹, due to the existence of N─H bond, ─CH2, C═O bond, and C─N bond bending in polyacrylamide, and the peak of the C═C in acrylamide vanished at 1620 to 1680 cm⁻¹ (fig. S1A). In ¹H–nuclear magnetic resonance assay, the peaks at 1.65 and 2.24 parts per million were assigned to the stretching vibrations of C─H and ─CH2 in polyacrylamide, respectively (fig. S1B). Moreover, as depicted in Fig. 2F, we used agarose electrophoresis to confirm the ligation between polyacrylamide and oligonucleotides. The results revealed distinct hysteresis in the DNA bands, indicating the successful connection of S-A and S-B to polyacrylamide chains, thereby impeding the electrophoretic migration of oligonucleotides.

Subsequently, we introduced the Hb aptamer to bind with PS-A and PS-B incorporated into the polyacrylamide mixture, resulting in the formation of the hydrogel (Fig. 2G, sequences in table S1). As illustrated in Fig. 3A, following the addition of the linker chain, the DNA hydrogel (named Gel) was successfully synthesized, demonstrating certain injectability and plasticity (notably, the formation of the word “ZZU” by extruding the hydrogel from a syringe). However, for hydrogels intended for use in deep brain ICH injuries, stringent property requirements must be met, including moderate toughness, low swelling levels, shape adaptability, and injectability, to enable filling of variously shaped pathological cavities and achieve minimally invasive administration. These properties are influenced by the concentration of the aptamer linker.

Fig. 3. Characterization of the DNA hydrogel.

(A) The images of the gelation progress and injectability of the hydrogel. (B) Time sweeps showing the changes of storage moduli (G′) and loss moduli (G″) of the hydrogel and brain against the gelling time at 1% dynamic strain and 1-Hz frequency. (C) G′ and G″ of the DNA hydrogel under strain amplitude sweep (strain range from 0.1 to 100%). (D) G′ and G″ of the DNA hydrogel under angular frequency sweep (frequency range from 0.1 to 10 rad/s). (E) The swelling performance of different aptamer-linker concentrations hydrogel immersed in phosphate-buffered saline (PBS) buffer. (F) Hydrogels formed by different concentrations of aptamer linker showing different performances upon the same concentration of Hb. Images of decomposed DNA hydrogel and percentage of sol after addition of Hb when different concentrations of aptamer linker were used as linker. (G) The encapsulation efficiency of DFO in hydrogel at different aptamer-linker concentrations. (H) The viscosity changes of the DNA hydrogel (“0.7” group) at the shear rate of 1 1/s to 100 1/s. (I) G′ and G″ of the DNA hydrogel (0.7 group) after being extruded from a syringe. (J) Confocal laser scanning microscopy (CLSM) image revealing the co-localization between drug (rhodamine B) and hydrogel (CY5). Data are means ± SD. n = 3. ns, no significant difference; ***P < 0.001 and ****P < 0.0001. Statistical analysis was performed using one-way analysis of variance (ANOVA) followed by Tukey’s post hoc test (K and L).

To obtain the optimal Gel, we conducted optimization experiments with three well-formed gels prepared from different aptamer concentrations (0.3, 0.7, and 1 mM). First, the 0.3 and 0.7 mM concentrations exhibited excellent injectability compared to 1 mM (Fig. 3A). As reported, soft matrix materials with 100- to 500-Pa modulus were suitable for neuron cell growth (23). The modulus properties were further characterized using rheology. As shown in Fig. 3B, the 0.7 mM concentration demonstrated a storage modulus more suitable for the brain environment (206 ± 5 Pa), akin to brain tissue (200 ± 14 Pa). Moreover, the hydrogel’s storage modulus consistently exceeded its loss modulus in different oscillation conditions, indicating that the DNA hydrogel exhibits relatively stable structure (Fig. 3, C and D, and fig. S2). In addition to comparable modulus properties, the swelling performance of the hydrogel is crucial when it serves as a brain implant. A hydrogel with a nearly ≈100% swelling ratio can mitigate mass effects induced by hematoma, edema, or implants, preventing brain midline shift and secondary damage. As depicted in Fig. 3E, the swelling ratios of hydrogels with 0.3, 0.7, and 1 mM aptamer linkers were 103.4 ± 5.1%, 128.1 ± 6.9%, and 149.0 ± 7.2%, respectively. In addition, we introduced 2 mM Hb to observe the dissociation rate of the Gel, revealing that the hydrogel with a 0.7 mM linker exhibited moderate sensitivity compared to other groups (Fig. 3F).

For the degradation assay in vitro, 0.5 ml of DNA hydrogel was added to 0.5 ml of phosphate-buffered saline (PBS) solution for 7 days at 37°C. The remaining mass of the DNA hydrogel was recorded. As illustrated in fig. S3, after 7 days, the DNA hydrogel can sustain about 88.30% (1 mM), 81.62% (0.7 mM), and 68.71% (0.3 mM) content, respectively, which exhibited significantly extended degradation under normal physiological conditions, making it suitable for a variety of biological processes and particularly as a carrier material for postoperative implantation requiring prolonged durations. In addition, to simulate responsive degradation under conditions of ICH, the DNA hydrogel was added to the PBS solution containing 2 mM Hb for 7 days at 37°C. As shown in fig. S4, all hydrogels degraded gradually over time in Hb solution. At 7 days, the degradation rate of 0.7 mM DNA hydrogel attained 13.09%, which indicated that DNA hydrogel can respond to ICH microenvironment to achieve controlled drug release.

Next, we assessed the potential of the three concentrations of the gel as carriers for drug delivery. As shown in Fig. 3G, the 0.7 and 1 mM concentrations demonstrated similar drug encapsulation efficiency, while the 0.3 mM concentration seemed unable to tightly enclose drugs due to its relatively large aperture, which was also confirmed by images of scanning electron microscopy (SEM; fig. S5). On the basis of this analysis, the 0.7 mM hydrogel was selected as the optimized hydrogel for subsequent experiments. Therefore, we further investigated the viscosity of DNA hydrogel with 0.7 mM linker strand, spanning a range from 1 to 100 1/s. This result unveils a pronounced shear-thinning behavior, marked by a reduction in viscosity under shear stress application (Fig. 3H). Such behavior signifies the adaptive rheological properties of the hydrogel, showcasing exceptional injectability. Moreover, to explore whether mechanical properties are influenced during injection, we investigate the changes of rheological property of DNA hydrogel after injection. As illustrated in Fig. 3I, there is no discernible alteration in the storage modulus and loss modulus of the DNA hydrogel compared with that before injection (Fig. 3B), suggesting that the injection procedure did not exert a notable impact on its mechanical properties. Last, confocal laser scanning microscopy (CLSM) investigations confirmed the encapsulation of the drug in Cyanine5 (Cy5)–labeled hydrogel, as indicated by uniform red fluorescence emission from Rhodamine B (Fig. 3J).

Hb-responsive drug on-demand release from the DNA hydrogel in vitro

Following this, we investigated the stimuli responsiveness of the DNA hydrogel. Initially, SEM images confirmed the presence of microporous channels and an enlarged pore size after Hb treatment (Fig. 4A). To gain physical insights into the DNA hydrogel in its solution state, rheological studies were conducted. The DNA hydrogel treated with Hb showed a low storage modulus of 16 Pa, indicating the gel-to-sol transition of the DNA hydrogel (Fig. 4B). In addition, CLSM results demonstrated a significant decrease in drug concentration after treatment with Hb (Fig. 4C).

Fig. 4. Hb-responsive drug release from the DNA hydrogel in vitro.

(A) Corresponding photos and SEM images of the hydrogel with or without Hb treatment. (B) Representative rheological measurements with or without Hb treatment. (C) CLSM image of the hydrogel that was treated with Hb. Green, model drug, rhodamine B; red, hydrogel labeled with Cy5. (D) Schematic diagram of percentage of sol detection. (E) The percentage of sol detection of the hydrogel under different aptamer linker. (F) The percentage of sol detection of the hydrogel with Hb or bovine serum albumin (BSA) treatment. (G) Images of the hydrogel under different Hb concentrations and the change of percentage of sol. (H) The percentage of sol detection of the hydrogel at different times. (I) The drug release kinetics curves of DFO from hydrogel with 2 mM Hb treatment at different times. (J) The drug release kinetics curves of DFO from hydrogel after incubation with different Hb concentration. (K) Schematic of drug release kinetics curves of DFO on-demand in vitro. (L) The drug release kinetics curves of DFO with or without Hb. Data are means ± SD. n = 3 independent samples. *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001. Statistical analysis was performed using one-way ANOVA followed by the unpaired two-tailed Student’s t-test (E, F) and Tukey’s post hoc test (G).

To mimic the ICH environment, the hydrogel was co-incubated with Hb solution, and the Hb content in the supernatant at different times was detected to obtain the sol ratio by ultraviolet/visible (UV-vis) spectroscopy (Fig. 4D). A mutated linker was designed to form the no-responsive hydrogel as a control group (no-res. gel). As shown in Fig. 4E, no gel-to-sol transition was observed for the hydrogel linked by the mutated linker after adding 2 mM Hb solution compared with the responsive gel group (res. gel). This suggests that the observed gel-to-sol transition described above was caused by Hb/aptamer interactions.

Furthermore, to demonstrate the selectivity of the smart DNA hydrogel system, bovine serum albumin (BSA) was applied as a negative control. Equal concentrations of Hb solution and BSA solution were separately added to the smart hydrogel system. The results indicated that the smart hydrogel retained aptamer selectivity with a highly specific response to the target Hb (Fig. 4F).

Next, we investigated the sensitivity of this DNA hydrogel to Hb. As shown in Fig. 4G, the hydrogel color and percentage of sol gradually increased as the Hb concentration increased. It is reasonable to expect that more aptamer linkers would form Hb-Apt-linker complexes with increasing Hb concentrations, resulting in an increase in the dissociation of the hydrogel. We also monitored the disintegration of the hydrogel over time after adding 2 mM Hb. Figure 4H suggested that the hydrogel was almost entirely degraded within 120 min.

In addition, we examined the drug release behavior triggered by Hb. As illustrated in Fig. 4 (I and J), the release profiles for DFO corresponded to changes in pathological cue concentration, with more drug release observed at higher Hb concentrations, reaching saturation within 120 min. To simulate the postoperative rehemorrhage environment, the Hb solution was intermittently introduced to the hydrogel and removed after 20 min to investigate the controllability of hydrogel drug release (Fig. 4K). As depicted in Fig. 4L, the drug gradually released from the hydrogel after the addition of Hb. However, the rate of drug release rapidly decreased after the removal of Hb, suggesting that drugs can be actively controlled and released from the DNA hydrogel based on the content of Hb. This provides a foundation for the controlled and targeted treatment of rehemorrhage in vivo.

Protective effect of the DNA hydrogel via confrontation ferroptosis and Hb toxicity in vitro

The potential therapeutic application of the DNA hydrogel in ICH therapy was initially assessed in vitro at the cellular level. First, we evaluated the potential toxicity of the hydrogel. As depicted in Fig. 5A and fig. S6, the hydrogel exhibited no obvious cytotoxicity to various brain cells at different concentrations (1 to 6%), suggesting the excellent biocompatibility of the hydrogel. Hence, for subsequent biological assays, 6% concentration of the DNA hydrogel was used, with a DFO concentration of 80 μM, calculated on the basis of encapsulation efficiency (Fig. 3G). In addition, the long-term cytotoxicity of the DNA hydrogel and the biocompatibility of degradation products were also detected. As depicted in fig. S7, the hydrogel demonstrated no significant cytotoxicity within 72 hours, affirming the long-term biocompatibility. Moreover, the results illustrated in fig. S8 indicated that the degradation products did not exert a significant negative effect on cell viability, laying a solid foundation for in vivo applications.

Fig. 5. Protective effect of the DNA hydrogel via confrontation ferroptosis and Hb toxicity in vitro.

(A) Cell viability assays of HT-22 cells after incubation with different concentrations of the DNA hydrogel for 24 hours. (B) Schematic of the transwell for coculture of hydrogel and HT-22 cells to investigate the ferroptosis. (C) Representative CLSM images showing intracellular ferrous iron in the above-treated cells, as reflected by FerroOrange probes. (D) Determination of the intracellular ferrous iron concentration with the Ca-AM probes in the HT-22 cells through flow cytometry. (E) Representative Western blotting bands showing the level of GPX-4 in HT-22 Cells. (F) Representative CLSM images of reactive oxygen species (ROS) staining of HT-22 with different treatments. (G) Malondialdehyde (MDA) content in the HT-22 with different treatments (n = 3). (H) Ferrous iron content in the HT-22 with different treatments. (I) Schematic of Hb capture by hydrogel to prevent the HT-22 cells death. (J) Representative fluorescent images of Calcein-AM (Ca-AM) and propidium iodide (PI)–costained HT-22 cells after different treatments. (K) Schematic of protective effect of hydrogel via confrontation ferroptosis and Hb toxicity. (L) The viability of HT-22 cells was measured by a cell counting kit–8 (CCK-8) assay after treatment with ferric ammonium citrate (FAC; 150 μM) and Hb (100 μM) (n = 3). (M) Lactate dehydrogenase (LDH) activity detection of HT-22 cells via LDH assay (n = 3). The results are shown as the means ± SD. *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001. Statistical analysis was performed using one-way ANOVA followed by Tukey’s post hoc test (G, H, L, and M). HT-22, mouse hippocampal neurons. Schematic diagrams of (B), (I), and (K) were created with BioRender.com.

Ferroptosis, a form of programmed cell death, is characterized by mitochondrial cristae and outer membrane rupture with condensed mitochondrial membrane densities, the accumulation of lipid peroxidation products, and lethal reactive oxygen species (ROS) originating from iron metabolism (24). Recent studies have verified that ferroptosis is intricately involved in ICH (25, 26). Therefore, the next evaluation focused on the therapeutic effect of the hydrogel in counteracting ferroptosis (Fig. 5B). FerroOrange, an Fe(II)-selective fluorescent probe, was used to detect intracellular Fe(II). As illustrated in Fig. 5C and fig. S9, the fluorescence intensity of FerroOrange in the PBS group increased significantly compared with that in the control group, demonstrating the successful construction of the ferric ammonium citrate (FAC)–induced iron overload model. However, the significantly decreased fluorescence intensity of FerroOrange in the gel treatment group indicated that the gel could effectively alleviate iron overload. Similarly, intracellular ferrous iron was also determined by flow cytometry analysis using Calcein–Acetoxymethyl ester (Ca-AM) (Fig. 5D).

Glutathione peroxidase 4 (GPX4) plays a crucial role in protecting cells against oxidative stress and lipid peroxidation damage (27, 28). The detection of GPX4 activity could provide valuable insights into the mechanisms of cell defense against ferroptosis. FAC-induced ferroptosis resulted in the down-regulation of GPX4 protein. However, the gel reversed the GPX4 expression (Fig. 5E and fig. S10). Elevated levels of ROS in cellular environments are also closely associated with ferroptosis. We also used FAC to stimulate HT-22 cells and detected intracellular ROS levels using the 2′,7′-dichlorodihydrofluorescein diacetate (DCFH-DA) assay. As expected, a faint signal was observed in the gel group compared with that in the PBS group, indicating that the hydrogel could suppress intracellular oxidative stress (Fig. 5F and fig. S11). Malondialdehyde (MDA) is a by-product of lipid peroxidation, and its accumulation indicates increased oxidative damage and lipid peroxidation, signifying ferroptosis. As shown in Fig. 5G, the gel group significantly reduced the accumulation of lipid peroxidation products. 4-Hydroxynonenal (4-HNE) is a well-known by-product of lipid peroxidation and an important marker of ferroptosis. Therefore, we measured the content of 4-HNE by the commercial kits. As expected, saline group leads to a significant increase in the 4-HNE level compared with the sham group, whereas gel significantly protected against the 4-HNE elevation elicited (fig. S12). Last, the protective effect of gel was detected via a cell counting kit–8 (CCK-8) assay. As depicted in fig. S13, the cell viability of the gel group significantly improved, indicating that the gel could effectively reduce cell injury caused by FAC through the controlled release of DFO.

Iron is essential for several physiological processes, including mitochondrial function, cell signaling, cell division, and myelination (29, 30). A challenge in the therapeutic use of iron chelators is to reduce pathological iron accumulation without disrupting physiological iron-dependent functions (31). Nonselective metal chelators may prevent injury, but they may also have dose-limiting toxicities through nonspecific effects on iron-dependent processes in the cell. Therefore, controlling the on-demand release of DFO is a major challenge. In this study, sols are stimulated by toxic proteins Hb; therefore, the release of DFO will be controlled on the basis of the amount of Hb. As shown in Fig. 5H, the release of DFO is positively correlated with the concentration of Hb. Notably, compared with the pure DFO group, the concentration of iron ions will be controlled near normal levels, rather than excessive chelation of intracellular iron affecting normal cell activity.

Hb is a toxic protein after ICH and also a trigger for the sol. Therefore, we assume that the DNA hydrogel could specifically capture the Hb to restrain subsequent toxicity (Fig. 5I). Hb and different gel were incubated with HT-22 cells, and the cell viability of HT-22 cells was measured after 24 hours. The percentage of propidium iodide (PI)–positive (dead) cells was significantly reduced in the gel-treated group, suggesting that nerve cells were effectively rescued by capturing Hb (Fig. 5J and fig. S14). Above all, the hydrogel could capture the toxic protein Hb to prevent nerve death, and, then, the DFO was released from the hydrogel to mitigate ferroptosis (Fig. 5K). Thus, to investigate the neuroprotection of the hydrogel, it was measured with the CCK-8 assay and lactate dehydrogenase (LDH) assay. The results showed that the gel group could reduce nerve cell death and inhibit ferroptosis, suggesting the excellent neuroprotection of the gel (Fig. 5, L and M).

The stability, specificity, and biosafety of the DNA hydrogel in vivo

Having confirmed the precise stimuli responsiveness of the DNA hydrogel and controlled release of DFO in vitro, the stability and specificity of the DNA hydrogel were next evaluated in vivo. Cy5-labeled hydrogel was implanted into healthy mice, with a no-res. gel as the control group to investigate the stability and specificity of the gel (Fig. 6, A and B). Figure 6C and fig. S15 demonstrated that these implants could remain in mice for at least 6 days under normal physiological conditions. In addition, to validate the stimuli responsiveness in vivo, gel and no-res. gel were injected into mice brain in a hemorrhagic pathological environment. The results showed that the gel disintegrated gradually on the second day and almost completely degraded on the sixth day. However, the fluorescence intensity was still high for the no-res. gel, indicating that the gel responded to the hemorrhagic environment (Fig. 6D and fig. S16).

Fig. 6. The stability, specificity, and biosafety of hydrogel in vivo.

(A) Schematic of the experimental procedure for evaluating the stability and specificity of the hydrogel. (B) Schematic of the time schedule in vivo. (C) Representative near-infrared images of Cy5-labeled hydrogel in healthy mice at different times (n = 5). (D) Representative near-infrared images of Cy5-labeled different hydrogel in ICH mice at different times (n = 5). (E) Interleukin-6 (IL-6) and (F) tumor necrosis factor–α (TNF-α) concentrations in the peripheral blood of mice subjected to PBS and hydrogel treatments, as measured by the enzyme-linked immunosorbent assay (ELISA; n = 3). (G) Change in body weight of mice between PBS and hydrogel groups (n = 5). (H) Hematoxylin and eosin (H&E) staining and (I) terminal deoxynucleotidyl transferase–mediated deoxyuridine triphosphate nick end labeling staining of brain slices from mice subjected to PBS and hydrogel injection. Data are means ± SD. Statistical analysis was performed using the unpaired two-tailed Student’s t test (E to G).

For the successful delivery of the hydrogel to the injured brain, implantation of a chemically biocompatible hydrogel with nontoxic degradation products reduces the infiltration of activated macrophages and microglia into the site of injury (32–34). Hence, we observed the biosafety via the injection of hydrogel in healthy mice. Interleukin-6 (IL-6) and tumor necrosis factor–α (TNF-α) concentrations in the peripheral blood of mice were measured by enzyme-linked immunosorbent assay (ELISA). The results indicated no obvious difference between the gel and PBS groups (Fig. 6, E and F). Mice both exhibited a slight change in body weight without a significant difference (Fig. 6G). During the experimental period, we monitored the behavior of mice and found that the long-term implantation of DNA hydrogel had no obvious adverse effects on mice (fig. S17). In addition, compared with the PBS groups, brain tissue from mice treated with the gel showed no damage [as demonstrated by hematoxylin and eosin (H&E) staining] (Fig. 6H). Moreover, terminal deoxynucleotidyl transferase–mediated deoxyuridine triphosphate nick end labeling staining confirmed the excellent biosafety of the gel (Fig. 6I). The above results manifested optimal biocompatibility of the DNA hydrogel in vivo.

Hydrogel treatment efficacy in vivo

Encouraged by the prominent performance in vitro, the therapeutic potential of the gel in vivo was investigated. Figure 7 (A and B) demonstrates the flowchart of the entire animal study and photo images of the surgical process. The stereotactic injection location was 0.5 mm anterior, 2.85 mm lateral to bregma, and 3.2 mm ventral to the dura mater. The concentration of hydrogel used is 500 μl/kg (with a dose of DFO at 435 μg/kg, calculated on the basis of the dose required for DFO and the brain’s capacity to hold the hydrogel).

Fig. 7. Hydrogel treatment efficacy in vivo.

(A) Schematic of the design of the animal experiment. (B) Schematic diagram of constructing a cerebral hemorrhage model. (C) Prussian blue staining and GPX4 of the pathological examination in brain with different treatments. (D) Semiquantitative analysis of GPX4 staining using ImageJ software (n = 5). MFI, mean fluorescence intensity. (E) MDA content in the brains of mice subjected to different treatments. (F) Change in body weight of mice in the different groups during the whole animal experiment (n = 5). (G) Representative T2*-weighted images of saline, DFO, no-res. gel, blank gel, or gel at different times. (H) Quantification of hematoma volumes on T2*-weighted images at 3 days after ICH. (I) Representative images of brain slices from mice with ICH treated with saline, DFO, no-res. gel, blank gel, or gel. (J) Hb concentration in the brains of mice subjected to different treatments measured using Drabkin’s reagent (n = 3). (K) Neuronal nuclear antigen (NeuN) staining of the pathological examination in brain with different treatments. (L) H&E staining of the pathological examination in brain with different treatments. Data are means ± SD. *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001. Statistical analysis was performed using one-way ANOVA followed by Tukey’s post hoc test (D to G, I, and J)..

Ferroptosis, a newly discovered process of cell death with a vital role in ICH, involves iron ions and the expression of GPX4, serving as important indicators (35). Therefore, Prussian blue’s staining and immunofluorescent staining of GPX4 in the brains of different groups were performed. As shown in Fig. 7C, the fewer blue spots in the gel group compared to those in the DFO group initially indicate enhanced iron removal efficiency (fig. S18). Meanwhile, the intensity of GPX4 was markedly diminished in the saline group compared to that in the sham group, whereas the gel robustly suppressed the reduction of GPX4 intensity (Fig. 7D). Moreover, we examined the production of MDA and ROS in the brains of mice with ICH. The results showed that the MDA and ROS levels in mice treated with the gel were much lower than those in control mice (Fig. 7E and fig. S19). Subsequently, mice treated with the gel exhibited a slight change in body weight, but mice treated with saline demonstrated a rapid weight loss (Fig. 7F).

Hematoma size is an important determinant and biomarker to predict the prognosis of patients with ICH in the clinic (36). Resistance to ferroptosis reportedly facilitates hematoma absorption in mice with ICH (37). Therefore, we assessed the efficiency of hematoma clearance via Gel treatments. As expected, the amount of iron deposition strongly correlated with the brain lesion size on T2*-weighted magnetic resonance imaging (MRI) scan, and the mean hematoma volume was about the same in each group at day 1. However, at day 3, the hematoma volumes of the hydrogel group were significantly smaller than that of the saline group (4.768 ± 0.362 mm³ versus 8.674 ± 0.716 mm³) (Fig. 7, G and H). Meanwhile, the coronal brain slices showed a similar tendency (Fig. 7I). In addition, the Hb content in the brain was also measured using Drabkin’s reagent, and the tendency of Hb content was consistent with that of hematoma size (Fig. 7J and fig. S20). We next examined brain tissue histology. Compared with the control groups, brain tissue from mice treated with the gel not only showed minimal damage [neuronal nuclear antigen (NeuN) staining] (Fig. 7K) but also exhibited the greatest number of neurons (as demonstrated by H&E staining and Nissl’s staining) (Fig. 7L and fig. S21).

Postoperative rehabilitation in ICH mouse model in vivo

Following cerebral hemorrhage, behavioral assessments in mouse models are essential for evaluating the effectiveness of therapeutic interventions. These assessments objectively measure cognitive and motor functions, providing valuable insights into the impact of treatments. Continual monitoring of behavioral changes in mice not only tracks rehabilitation progress but also contributes crucial data for scientific analysis, optimizing experimental outcomes.

A detailed procedure for postoperative rehabilitation training was designed (Fig. 8A). First, all animals underwent preoperative exercise for at least 3 days before the experiment. Then, the ICH model was constructed by injecting collagenase into the brain. Last, postoperative rehabilitation tests were performed to assess the therapeutic effects of the gel. In brief, functional rehabilitation was assessed in mice with ICH according to modified neurological severity scores, including scores for neurological deficit tests, the beam balance test, the corner test, and the adhesive removal test (Fig. 8, B to K). Mice subjected to gel treatment could walk in a straight line and showed a lower neurological severity score than mice in the other groups (a higher score indicates a more severe injury). The beam balance test, the corner test, the elevated body swing test, and the adhesive removal test also reflected the favorable postoperative rehabilitation in mice treated with the gel.

Fig. 8. Postoperative rehabilitation in ICH mouse model in vivo.

(A) Schematic of time schedule for postoperative rehabilitation training. (B) Schematic of behavioral testing including beam balance test, elevated body swing test, corner test, and adhesive removal test. (C) Representative beam balance images of sham and ICH mice treated with saline, DFO, no-res. gel, blank gel, or gel. Postoperative rehabilitation tests were conducted to validate the treatment effect of the gel, including the beam balance test (D), elevated body swing test (E), Longa test (F), corner test (G), and sticky tape test (H to K) (n = 15). Data are means ± SD. n = 15 mice per group. *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001. Statistical analysis was performed using one-way ANOVA followed by Tukey’s post hoc test (D to K).

Self-diagnosis and self-treatment of DNA hydrogel for postoperative rehemorrhage

Postoperative rehemorrhage commonly occurs in patients who undergo surgery within 4 hours (~40%), and a relationship between rebleeding and mortality in a surgery group has been shown (38, 39). Therefore, it is of great urgency to perceive postoperative rehemorrhage and remedy it without delay.

Timely and accessible neuroimaging is a critical step in the diagnostic workup of patients presenting with suspected acute brain injury such as ICH. In this study, we reedited the DNA sequences using Cy3 to label Hb aptamer and BHQ2 to label strand B to achieve self-diagnosis. Upon the formation of the three-strand complex, the fluorescence of Cy3 was quenched because it was close to the quencher group. The fluorescence of Cy3 was restored, suggesting that postoperative rehemorrhage occurred (Fig. 9A). First, the collagenase ICH model was used to detect the self-diagnosis ability of the DNA hydrogel. As shown in fig. S22, compared with the healthy mice group, the obvious fluorescence was observed in the ICH mice groups. To further investigate the detection sensitivity of the Gel, we also established an autologous blood injection ICH model to validate the change of fluorescence signal under different rebleeding volumes. The fluorescence recovery of Cy3 increased rapidly with different rebleeding volumes, suggesting that the dissociation of the three-strand complex was dependent on the rebleeding volume (Fig. 9, B and C). We also used T2-weighted MRI scan; the hematoma volume lesion showed a hyperintense area in the right brain and an increase in signal intensity with the volume gain of autologous blood injection, which was consistent with fluorescence imaging results (Fig. 9, D and E).

Fig. 9. Self-diagnosis and self-treatment of hydrogel for postoperative rehemorrhage.

(A) Schematic of gel detection strategy for sensing rebleeding in vivo. (B) Representative near-infrared images to validate the signal intensity of different rebleeding volume. (C) Quantitative analysis of fluorescence intensity (n = 5). (D) Bleeding volume was evaluated by magnetic resonance T2-weighted imaging. (E) Quantification analysis of bleeding volume (n = 5). (F) Schematic of the design of the rebleeding experiment. (G) MDA content in the brains of mice with autologous blood injection treated with saline, DFO, no-res. gel, blank gel, or gel (n = 3). (H) Hb concentration in the brains of mice with autologous blood injection treated with saline, DFO, no-res. gel, blank gel, or gel (n = 3). (I) Representative images of brain slices from mice with autologous blood injection treated with saline, DFO, no-res. gel, blank gel, or gel. (J) Semiquantitative analysis of hematoma size using ImageJ software (n = 5). (K) Representative images of H&E staining of the pathological examination in brain from mice with autologous blood injection treated with saline, DFO, no-res. gel, blank gel, or gel. Data are means ± SD. *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001. Statistical analysis was performed using the unpaired two-tailed Student’s t test (C) and one-way ANOVA followed by Tukey’s post hoc test (G, H, and J).

To investigate the treatment efficacy of the hydrogel after rebleeding, a series of animal experiments were involved (Fig. 9F). In brief, we first injected 5 μl of autologous blood to establish the ICH model and implanted the hydrogel. Fifteen microliters of autologous blood was reinjected to imitate postoperative rehemorrhage, and tissue analyses were performed to assess the therapeutic effects three days later. As shown in Fig. 9G, the MDA concentration in mice treated with the gel was much lower than that in control mice. The Hb content in the brain was also measured using Drabkin’s reagent. As expected, the tendency of Hb content was consistent with that of hematoma size (Fig. 9H). In addition, as shown in Fig. 9 (I and J), the gel induced more significant hematoma elimination than blank or no-res. gel. In addition, compared with the control groups, brain tissue from mice treated with the gel showed minimal damage (as demonstrated by H&E staining) (Fig. 9K).

A comprehensive assessment of Gel biosafety was performed, including a tissue section analysis of major organs (including the heart, liver, spleen, lungs, and kidneys) by H&E staining (fig. S23) and hematology analysis (fig. S24). The results clearly demonstrated that the gel had good biocompatibility and had a negligible influence on mice. Collectively, these results highlighted that the gel was a safe and efficacious therapy in mice with ICH.

DISCUSSION

Postoperative rehemorrhage after ICH surgery is a severe complication, posing a significant threat to patients with elevated morbidity and mortality rates. Studies indicate a relatively high incidence of rehemorrhage after ICH surgery, closely tied to the overall prognosis of patients (40, 41). Rehemorrhage not only prolongs hospitalization and increases treatment complexity but also significantly elevates the risk of mortality, imposing a substantial psychological and economic burden on patients and their families.

Despite the widespread application of DFO in the treatment of ICH, its expected clinical effectiveness has not been fully realized (42). One possible reason is the systemic administration of DFO in clinical practice, making it challenging to precisely control the dosage reaching the target site (43, 44). This systemic delivery method may limit the local concentration of DFO in brain tissues, affecting its effectiveness in effectively suppressing rehemorrhage. Moreover, the occurrence of rehemorrhage is challenging to detect promptly in clinical settings, making it difficult for health care professionals to determine the optimal timing for administering DFO treatment. The lack of accurate monitoring tools for rehemorrhage may lead to uncontrolled treatment timing, thereby diminishing the efficacy of DFO. Therefore, to better harness the therapeutic potential of DFO, exploring more precise administration methods and improving monitoring tools are crucial to ensure timely adjustments to treatment plans when the risk rehemorrhage increases. This improvement holds promise for enhancing the clinical efficacy of DFO.

Further research reveals that the direct use of DFO may induce iron imbalance in patients, leading to a series of adverse reactions (45). This underscores the urgent need to find a safer and more effective treatment approach. In this study, the introduction of controlled-release hydrogel offers a feasible solution to this challenge. It can release drugs locally based on rehemorrhage signals (excessive Hb), reducing the systemic drug concentration burden on patients, effectively alleviating the risks of rehemorrhage, and avoiding side effects such as iron imbalance.

It is noteworthy that similar challenges exist not only in ICH surgery but also in surgeries for other brain diseases. Therefore, researching the application of the DNA hydrogel holds the potential to address rehemorrhage issues in patients with ICH and provides unique insights for the treatment of other brain surgeries. In-depth research and practical applications of this topic are expected to offer more effective and safer methods for the postoperative treatment of various brain diseases, positively affecting patient recovery and survival rates.

However, ICH treatment is complex and requires a combination of multiple drugs. DFO in this study is a model drug, and the potential of the DNA hydrogel as a drug delivery carrier for other drugs, including cell transplantation, requires further research. In addition, drug loading capacity and detailed release kinetics of the hydrogel warrant further investigation.

In conclusion, this study introduces an Hb-responsive in situ implantable DNA hydrogel for postoperative rehemorrhage therapy. The hydrogel selectively captures excess Hb, detects rehemorrhage, and releases DFO to counter ferroptosis. This innovative hydrogel offers a diagnostic and therapeutic intervention for postoperative rehemorrhage, potentially improving outcomes in clinical settings.

MATERIALS AND METHODS

Study design

The objective of this study was to develop implantable DNA hydrogel for postoperative rehemorrhage. First, we prepared the DNA hydrogel that is equipped with excellent injectability and suitable modulus properties. Second, the stimuli responsiveness and drug release of the DNA hydrogel were evaluated in vitro. Then, transwell model was used to study the therapeutic effect of the hydrogel in counteracting ferroptosis. After that, we investigated the stability, specificity, and biosafety of the DNA hydrogel in vivo. Last, collagenase-induced mouse model and autologous blood-induced mouse model were carried out to inspect the self-diagnosis and self-treatments of DNA hydrogel.

Animals

Male C57BL/6 mice (8 weeks old, 20 to 22 g) were purchased from Si Pei Fu Biotechnology Co. Ltd. (Beijing, China) (license no. 110324220100896834). All mice were housed in 55% humidity at 25°C with a 12-hour light/dark cycle. All animal experiments were performed in accordance with the Life Sciences Ethical Review Committee of Zhengzhou University. The animal laboratory’s accreditation number is SCXK (YU) 2017-0001.

Preparation of aptamer–cross-linked hydrogel

Desalted 5′ end acrydite modified strand A (S-A) and strand B (S-B) were added separately in centrifuge tubes filled with tris-HCl buffer containing 4% acrylamide and 200 mM NaCl. They were removed oxygen using deaerator for 15 min. Then, adding 1.4% Ammonium Persulfate (APS) and N,N,N’,N’-Tetramethylethylenediamine (TEMED) fresh aqueous solutions, and the centrifuge tubes were quickly put in the deaerator to allow the polymerization in vacuum at 37°C. Linear-chain polyacrylamide S-A (PS-A) and polyacrylamide S-B (PS-B) were produced after 18 min. The resulting mixtures were stored at 4°C for 12 hours for further polymerization.

To generate hydrogel, PS-A (1 mM) and PS-B (1 mM) were mixed with aptamer linker (0.3, 0.7, and 1 mM) with addition of DFO solution. The mixture was shaken vigorously to guarantee the homogeneity of the solution before incubating in PCR amplifier at 65°C for 5 min and then allowed to slowly cool to 25°C to produce aptamer cross-linked hydrogel with DFO trapped inside.

Characterization of the hydrogel

S-A, S-B, PS-A, and PS-B were analyzed by agarose gel electrophoresis on a 3% gel for 50 min at 120 V. Effect of hybridizing region on aptamer-Hb interaction was explored by magnetics beads (Beaver Biosciences Inc., China). A scanning electron microscope (SEM SU8020, Japan) was used to visualize the morphology of the samples. The rheological properties of the hydrogel were characterized by a Rheometer (MCR92, Austria). Fluorescently labeled hydrogel was obtained by mixing hydrogel-labeled Cy5 with rhodamine B (1 mg/ml). Fluorescence field images of the hydrogel were taken by confocal laser scanning microscope (Leica).

The loading efficiency of DFO

Initially, a standard absorbance curve was established using the UV-vis spectrophotometer from a series of standard DFO solution with Ferric chloride (FeCl₃; 3 mM) at the wavelength of 485 nm. The content of drug in supernatant was determined with the aid of the standard absorbance curve. The loading efficiency of DFO was calculated in the following equation

$$\text{Loading efficiency}=(M_0-M_{\text{sup}})/M_0$$

where M₀ is the initial amount of DFO and M_sup is the amount of DFO in the supernatant.

Swelling Test

Hydrogels (400 μl) were prepared, followed by lyophilization to ensure the complete removal of water content from the hydrogels. Subsequently, the lyophilized hydrogels were immersed in 2 ml of deionized water, and their weights were measured at specific time points. Calculate the swelling ratio using the formula

$$\text{Swelling ratio} (\%)=(W_{\mathrm{w}}-W_{\mathrm{d}})/W_{\mathrm{d}}\times 100\%$$

where W_w is the weight of the wet hydrogel and W_d is the weight of the dry hydrogel.

In vitro release of the DFO

UV-vis spectroscopy was used to determine the cumulative release of DFO from hydrogel complex in vitro. In brief, 30 μl of drug-loaded hydrogel was evenly dispersed in 10 μl of PBS (pH 7.4, 10 mM) and 2 mM Hb incubated at 37°C to simulate the physiological environment in vitro. A volume of 5 μl of supernatant was used for the determination of drug concentration, and, then, one more 5 μl of PBS was added in hydrogel. FeCl₃ (3 mM) was incubated with the supernatant at 37°C for 20 min, and the DFO content in the supernatant was measured at 485 nm (DFO-Iron III chelate complex, maximum absorption) using a spectrophotometer. Meanwhile, a standard absorbance curve was established using the UV-vis spectrophotometer from a series of standard DFO solution with FeCl₃ (3 mM) at the wavelength of 485 nm. The drug concentration at each time point was calculated according to the pretested standard curve, and the following formula was used to calculate the cumulative release

$$\text{Cumulative release} (\%)=100\times (C_t{V}_0+{\displaystyle \sum _{n=1}^{\mathrm{t}-1}}{C}_i{V}_i)/W$$

where C_t and C_i are the t and i sampling concentration, respectively; V₀ and V_i represent initial volume and the volume obtained at i; and W is the total quality of the drugs.

For the controllability of hydrogel drug release, briefly, 30 μl of drug-loaded hydrogel was evenly dispersed in 10 μl of Hb (2 mM) incubated at 37°C to simulate the physiological environment in vitro. A volume of 10 μl supernatant was taken out at a certain time to detect the drug concentration, and, then, one more 10 μl of PBS was added in hydrogel. The PBS and Hb were orderly added in hydrogel repeatedly. The drug concentration at each time point was calculated according to the pretested standard curve.

Confront ferroptosis

HT-22 cells were first seeded in 24-well plates at a density of 5 × 10⁵ cells well and were then cultured overnight. Ferroptosis model was established by FAC with transwell. First, Hb, FAC, and different group hydrogels were incubated in the apical chamber, and the HT-22 cells were put in the basal chamber. The apical chamber was withdrawn after incubating 3 hours, and the basal chamber continued to incubate 24 hours to induce ferroptosis.

FerroOrange probes were used to detect intracellular ferrous irons. Cells were stained with Hoechst 33342 (10 μM) and FerroOrange (1 μM) in PBS at 37°C for 20 min. After washing three times with PBS, the fresh culture medium was supplemented for fluorescence microscopy imaging (DMI6000, Leica).

Intracellular ferrous iron was also determined by flow cytometry analysis using Ca-AM. Cells were stained with Ca-AM at 10 μM at 37°C for 30 min. Last, after washing three times with PBS, cells were resuspended in PBS for flow cytometry analysis on a NovoCyte flow cytometer. The ROS level in HT-22 cells was determined by fluorescence microscopy imaging due to the oxidation of DCFH-DA activated by ROS.

Cell viability was measured using the CCK-8. The hydrogel was pretreated with Hb and centrifuge at 1000 rpm/min (10 min) to obtain the hydrogel compound. HT-22 cells (1 × 10⁴) were seeded in a 96-well plate and first replenished with 150 μM FAC for 24 hours to induce cellular iron accumulation prior to compound treatment.

Iron and MDA assay

The iron and MDA concentrations were measured using the Iron Assay Kit and the MDA Assay Kit, respectively. The collected cells were disrupted by ultrasound, and the supernatant was collected after high-speed centrifugation (12,000 rpm/min, 30 min) and placed on ice for detection. A sufficient volume of the working solution was prepared and transferred with the collected supernatant to a 96-well plate sequentially. The Iron Assay Kit detects the optical density at 510 nm, and the MDA Assay Kit detects the optical density at 530 nm. The concentrations of iron and MDA were further calculated according to the instructions.

4-HNE assay

The 4-HNE concentration was measured using a commercial kit from Nanjing Jiancheng Bioengineering Inst. (Nanjing, China). The collected cells were disrupted by ultrasound, and the supernatant was collected after high-speed centrifugation (12,000 rpm/min, 30 min). Next, 4-HNE was measured according to the manufacturer’s instructions. The absorbance of 4-HNE was measured by a multimode microplate reader at 450 nm. All tests were performed three times.

Western blot analysis

Then, the cells were treated with FAC transwell model for a further 24 hours. The cells were harvested and lysed with radioimmunoprecipitation assay buffer containing 1 mM phenylmethanesulfonyl fluoride to collect the total protein. The total protein concentration was measured using the bicinchoninic acid (BCA) protein assay kit (Solarbio Science & Technology Co. Ltd.) according to the manufacturer’s instructions. Then, 80 μg of total protein was separated by 10% SDS-PAGE and transferred to polyvinylidene fluoride (PVDF) membranes, followed by incubating with 5% skimmed milk powder for 1 hour. After, the PVDF membranes were incubated with anti–β-tubulin (Cell Signaling Technology, MA, USA; 1:4000) and GPX4 (Proteintech, Wuhan, China; 1:2000) antibodies at 4°C overnight, followed by incubating with horseradish peroxidase–conjugated secondary antibodies for 2 hours at room temperature. The final result was harvested by BeyoECL Star (Beyotime Biotechnology Co. Ltd.) using Image Lab (Bio-Rad).

Neuroprotective effects of the hydrogel

HT-22 cells were seeded into 24-well plates at a density of 5 × 10⁵ cells per well and cultured in 1 ml of fresh medium containing 10% fetal bovine serum until adherence. Then, the cells were treated with fresh medium containing Hb and different groups for another 24 hours. To investigate the neuroprotective effect of the hydrogel, cell apoptosis was measured using the Ca-AM/PI Double Stain Kit and analyzed with a fluorescence microscope.

In vivo imaging

Mice with ICH were injected with no-res. gel and gel that both labeled with the same concentration of Cy5. The stability and specificity effect of hydrogel was investigated using a small-animal imager, with which images were taken at different time points after hydrogel injection.

Brain tissue harvesting, immunohistochemistry, and immunofluorescence staining

Mice with ICH subjected to different treatments were anesthetized with sodium pentobarbital and euthanized at 72 hours after surgery. The brains were removed immediately for subsequent measurements. H&E and Prussian blue staining were used to investigate the protective effect of the hydrogel in mice with ICH. The brain sections were processed for immunofluorescence staining using the following primary antibodies: anti-NeuN antibody for neurons, anti-GPX4 antibody for GPX4 protein, and DCFA-DA for ROS.

MDA measurement

Brain tissue samples were homogenized at 4°C in precooled tris buffer [1:4 (w/v), 50 mM, pH 7.4] and centrifuged (4°C, 10,000g, 10 min). The supernatant was harvested to measure MDA content using the MDA assay kit (Nanjing Jiancheng Bioengineering Institute, China). Meanwhile, the total protein of the supernatant was measured using the BCA protein assay kit (Solarbio Science & Technology Co. Ltd.) according to the manufacturer’s instructions. The MDA content was expressed as nanomoles of MDA/milligram of total protein.

Hematoma size

Mouse brains were removed after perfusing with precooled PBS. Then, the brains were sliced into five equally spaced brain sections at a thickness of 1 mm. The hemorrhage area and brain section area were calculated using ImageJ. The hematoma size was calculated using the following formula: hematoma size (%) = hemorrhage area ÷ brain section area × 100%.

Measurement of brain cytokines

Healthy and hydrogel brain tissue samples were homogenized at 4°C in precooled tris buffer [1:4 (w/v), 50 mM, pH 7.4] and centrifuged (4°C, 10,000g, 10 min). The supernatant was harvested to measure the concentrations of IL-6 and TNF-α in the brain using the mouse IL-6 ELISA kit and the mouse TNF-α ELISA kit (Multisciences Biotech Co. Ltd.), respectively.

Supplementary Materials

This PDF file includes:

Supplementary Text

Figs. S1 to S24

Table S1