H2 inhalation does no disrupt retinal physiologic vascularization during postnatal development

To investigate the effect of H2 on the formation and maturation of the developmental retinal vascular network, we investigated developmental angiogenesis in the retina in H2-treated WT mice and untreated littermate controls from P1 to P17.

As shown in Fig. 1B and C, after birth, the retinal vascular system starts to develop as a sprout from the optic disc and initially forms a primitive vascular plexus that is rapidly remodeled into large and small vessels.By around P8, retinal vessels extend radially over the superficial layer of the retina to form a two-dimensional vascular structure and subsequently sprout into the retina, establishing a secondary, deeper plexus. To assess retinal vascularization, we measured the ratio of the retinal vascular area (area of retinal blood vessels within the white dotted line) to the area of the entire retina [22]. The mice inhaling H2 (concentration: 3–4%) displayed normal primary vascular plexus and vascular density, similar to littermate controls, indicating that H2 inhalation does not interfere with the normal extension of retinal blood vessels and the density of the vascular network (Fig. 1F, E).

Endothelial tip cells are leading cells at the tips of the area undergoing vascular sprouting during angiogenesis. Additionally, retinal astrocytes are found in species that possess a retinal vasculature, providing the foundation for retinal angiogenesis and guiding the migration of endothelial tip cells. We examined the retinal distribution of astrocytes and the number of tip cells at P4 and P7, as shown in Fig. 1D and H. Mice that underwent H2 inhalation and controls showed no differences in tip cells at P4 and P7 (Fig. 1G). Moreover, normal distribution, morphology, and densities of glial fibrillary acidic protein (GFAP) expression in astrocytes were observed in both the central and peripheral retinas at P7 and P17 (Fig. 1H). Notably, the complete formation of a three-layered vascular system was detected in both groups at P17 through 3D image reconstruction by confocal laser scanning microscopy (Figure S1). Altogether, the results indicate that H2 does not interfere with retinal angiogenesis during postnatal development.

H2 attenuates retinal vaso-obliteration and neovascularization in OIR mice

We utilized a mouse model of oxygen-induced retinopathy (OIR), where mice were exposed to 75% oxygen from P7 to P12 and returned to room air until P17 [4]. The hyperoxic phase (P7–P12) results in vaso-obliteration of the central retina. This phase is followed by a second phase (P12–P17) of vascular regeneration in this central avascular retina (Fig. 1A). To gain insights into the role of H2 role in regulating vascularization in the retina, we examined the changes in OIR mice retinas with H2 inhalation on P12 and P17. As shown in Fig. 2A, OIR mice displayed a 37.44% ± 5.56% higher avascular area compared to the control group during the hyperoxic period P7–P12). After five days of H2 inhalation from P7 to P12, the avascular area in OIR mouse retinas decreased to 23.96% ± 3.97% (Fig. 2B). Subsequently, we extended the H2 treatment for the entire period. At P17, when the OIR mice were in a hypoxic state, a neovascular tuft (NVT) formed due to retinal vessel occlusion, resulting in an avascular area of 19.98% ± 6.30% and an NVT area of 18.34% ± 6.56%. However, H2 dramatically reduced the avascular area to 9.77% ± 4.49% after the complete treatment, while simultaneously suppressing NVT area formation by 10.63% ± 2.23% (Fig. 2C, D). It is noteworthy that the use of H2 led to an reduction in vaso-obliteration during both hyperoxic and hypoxic phases, whether it was administered solely during those phases or throughout the entire treatment (Fig. 2A–C). However, inhalation of H2 only during the hypo or hyperoxic phase did not yield as favorable results as full treatment with H2 on neovascularization, as the latter significantly decreased neovascularization (Fig. 2D).

Fig. 2

H2 attenuates retinal vaso-obliteration and neovascularization in OIR mice. A. Whole-mount retinas from the Control, OIR and OIR+H2 groups of wild type mice were harvested on P12 or 17 and examined by immunostaining with isolectin B4 (bar=500μm). H2 treatment reduces retinal vaso-obliteration at P12 and pathologic neovascularization at P17 in OIR. Areas of vaso-obliteration (B, C) and neovascularization (D) were quantified and analyzed. Error bars were the mean ± SEM. *P < 0.05 n=10-15/group. E-G. H2 resulted in a decrease in endothelial cell nuclei numbers anterior to the inner limiting membrane in OIR mice. F. The presence of PCNA-positive cells was confined to the retinal vessel area after H2 inhalation

To evaluate the degree of retinal neovascularization accurately, we opted for the evaluation of neovascular cell nuclei located anterior to the internal limiting membrane (ILM) as an indicator, due to the imprecise nature of assessing retinal neovascularization using the retinal vessel ratio.At P17,retinas from the OIRP17 group contained multiple NVTs extending into the vitreous. In contrast, inhalation of H2 resulted in a decrease in endothelial cell nuclei numbers anterior to the inner limiting membrane, with inhibitory effect of 64.4%, 77.17% and 88.67% reduction in the OIRP17 + H2P7–P12, OIRP17 + H2P12–P17 group and OIRP17 + H2P7–P17 group,respectively (Fig. 2E, G). These findings demonstrate that H2 inhalation suppressed pathological angiogenesis in the retina, preventing it from extending into the vitreous through the ILM.

In addition, proliferating cell nuclear antigen (PCNA), which is crucial for DNA replication and acts as a processivity factor for DNA polymerase δ in eukaryotic cells, is closely linked to the proliferation of vascular endothelial cells [25]. As depicted in Fig. 2F, the control group exhibited lower levels of PCNA expression, while PCNA-positive cells were mainly found in the retinal ganglion cell layer (GCL), inner nuclear layer (INL), and the vascular distribution region of the retina in the OIRP17 group, regardless of whether it was the neovascularization zone or the normal vessel area. Within the neovascular area of the OIRP17 group, the number of PCNA cells increased significantly, surpassing the control group by more than 20-fold. Following inhalation of H2, the presence of PCNA-positive cells was confined to the retinal vessel area. This led to reductions of 85.91%, 60.81%, and 92.94% in the OIRP17+H2P7–P17, OIRP17+H2P7-P12, and OIRP17+H2P12–P17 groups, respectively (Fig. 2F, H).These results indicated that in response to increased DNA damage and breaks caused by hypoxia and ischemia, the expression of PCNA is elevated to adapt to these changes.H2 demonstrated its efficacy in reducing abnormal blood vessel growth by decreasing the proliferation of vessels positive for PCNA during the hyperoxic or hypoxic phases, or throughout the entire treatment period.

H2 preserves the astrocytic template and inhibits microgliosis in the retina of OIR mice

To understand the impact of H2 on glial cells in OIR mouse retinas [19], we examined the changes in astrocytes and microglia after H2 inhalation on P12 and P17. We investigated the distribution, cell density, and morphology of astrocytes labeled by GFAP immunostaining to assess the evolution of astrocyte changes. Normally, astrocytes are located between the vasculature and retinal neurons and are mostly found in the retinal nerve fiber layer (RNFL). Figure 3A illustrates the distribution and cell density of astrocytes in the avascular area of the retinas on P12. In the OIR mouse retinas, there was a 43.92% reduction in astrocytes compared to the control group, although the right morphology was maintained. However, H2 treatment improved the density of astrocytes to 56.02% of the control in the avascular area (Fig. 3A, E). These findings suggest that H2 preserved the astrocytic template for the retinal vasculature in the avascular area during the hyperoxic period.

Fig. 3

H2 preserves the astrocytic template and inhibits microgliosis in the retina of OIR mice. A.E H2 preserved the astrocytic template for the retinal vasculature in the avascular area during the hyperoxic period. B.F.G H2 treatment maintained the integrity and arrangement of astrocytes, effectively minimizing the disturbance of the astrocyte template in both areas of retinal avascularity and abnormal blood vessel growth in OIR retinasduring the hypoxic phase. C.D H-J. In OIR mice, microglia with ameboid and ramified forms were observed throughout the entire retina at P12 and P17; inhalation of H2 in OIR mice resulted in a reduction in the proportion of ameboid microglial cells compared to the OIR group at P12, with minimal changes observed at P17 in the avascular area. Error bars were the mean ± SEM. *P < 0.05

During the hypoxic phase, astrocytes in the avascular areas of OIRP17 mouse retinas exhibited reduced ramification, hypertrophy, and disorganization compared to the control groups. The number of astrocytes was also significantly decreased, representing only 38.49% of the control group (Fig. 3B, F). In the neovascular area, the density of astrocytes in the OIR group dropped even further, reaching only 29.85% of the control group. The morphology and distribution of astrocytes were particularly disrupted in areas with neovascular tufts (Fig. 3B, G). However, in the avascular areas of OIRP17 + H2P7–P17 mouse retinas, the morphology of astrocytes was similar to that of OIRP17 mice, but the cell density increased to 71.62% of the control group after the entire treatment period. Furthermore, in contrast to the poorly distinguishable and irregularly arranged astrocytes mixed with neovascular tufts in OIRP17 mice, the astrocytes in the neovascular area of the OIRP17 + H2 P7–P17 group showed a reticular arrangement surrounding vessels, and the density increased to 47.83% of the normal control group (Fig. 3B, F, G). Additionally, in both the OIRP17 and OIRP17 + H2 P7–P17 groups, astrocytes were normally distributed in the area with normal vascular distribution (Fig. 3B). The findings demonstrated that H2 treatment maintained the integrity and arrangement of astrocytes, effectively minimizing the disturbance of the astrocyte template in both areas of retinal avascularity and abnormal blood vessel growth in OIR retinas.

We then examined the relationship between astrocytes and tip cells, which are leading cells in the sprouting of blood vessels. Immunostaining was performed to determine the colocalization of astrocytes and tip cells using GFAP and IB4. In the avascular and neovascularized regions of OIRP17 mouse retinas, there was either no colocalization or disordered colocalization between IB4 and GFAP, as shown in Fig. 3B. In contrast, a significant degree of colocalization was observed between anti-GFAP staining and IB-4 fluorescence in the avascular and neovascularized zones of OIRP17 + H2 P7–P17 mouse retinas. This suggests that astrocytes played a crucial role in guiding the tip cells in the OIRP17 + H2 P7–P17 group, as evidenced by the attachment of tip cells to astrocytes and the extension of their filopodia along the astrocytes (Fig. 3E–G). These findings indicate that H2 treatment maintained the quantity, morphology, and distribution of astrocytes, preserving the astrocyte template in OIR mouse retinas. Furthermore, our data suggest that H2 treatment improved the formation of tip cells, guided by the astrocyte network.

Furthermore, multiple studies have demonstrated the involvement of microglia in normal retinal vascular development [21]. Immunohistochemical analysis of the Iba-1 antigen on postnatal day 12 (P12) and P17 whole-mount retinas of normal mice revealed that microglia exhibited small cell bodies and long/thin cellular processes (Fig. 3C and D). In OIR mice, isolectin-immunopositive microglia with ameboid and ramified forms were observed throughout the entire retina at P12 and P17. The density of activated ameboid microglia in both the avascular and neovascular areas of OIR retinas was significantly higher than that in the control group, while the density of the nonactivated dendritic phenotype remained relatively unchanged. In contrast to OIRP12 mice, microglial cells in the peripheral neovascular area of OIRP17 mice tended to cluster near the neovascular tufts (NVTs) and exhibited enlarged cell bodies with shorter and fewer branched processes (Fig. 3D, H–J). Interestingly, inhalation of H2 in OIR mice resulted in a 59.1% reduction in the proportion of ameboid microglial cells compared to the OIR group at P12, with minimal changes observed at P17 in the avascular area, while the density of the dendritic phenotype remained unaffected (Fig. 3H–J). Moreover, treatment with H2 in OIR mice led to a reduction of approximately 52.33% in the proportion of ameboid microglial cells in the neovascular area, accompanied by a decrease in NVT size (Fig. 3J). These findings clearly show that H2 treatment inhibits microglial activation in the retinas of OIR mice, demonstrating the protective effect of H2 on neuroglial dysfunction in the OIR model.

H2 promotes Nrf2 activation in OIR mice

Nrf2 has been found to regulate retinal angiogenesis [20]. To gain insights into the role of Nrf2 and the relationship between H2 and Nrf2 in the regulation of vascular regeneration in ischemic neuronal tissue, we first investigated the expression of Nrf2 using quantitative PCR, immunoblotting, and immunofluorescence staining. Nrf2 expression could be observed in the INL and GCL of the normal retina, where vessels and neurons are mainly located. We found that the expression of Nrf2 in the OIRP12 group was the same as that in the control group in the avascular and normal blood vessel areas. However, in the OIRP17 group, Nrf2 expression in the neovascular and normal vascularized areas was higher than in the control group.

Western blot results indicated that the level of Nrf2 expression did not differ between the OIRP12 group and control group but increased after H2 inhalation (Fig. 4A, B). These results were consistent with previous studies showing that Nrf2 was activated after the onset of ischemia. Interestingly, H2 application induced early Nrf2 activation on P12, which revealed that H2 promoted Nrf2 activation during the hyperoxic phase. At P17, higher Nrf2 levels were observed in both the OIRP17 group and the OIRP17 + H2P7–P17 group than in the control group (Fig. 4C, D). There was a trend of higher Nrf2 expression in the H2 treatment group than in the OIR group. Additionally, H2 treatment notably enhanced the expression of the main target genes of Nrf2, such as HO-1 and NQO1, in the retinas of OIR mice (Fig. 4E; 2.11-fold increase in NQO1; 1.87-fold increase in HO-1). Thus, we found that the upregulation of Nrf2 promoted by H2 inhalation was sustained throughout the whole procedure. (Fig. 4A, C). These results therefore confirmed that H2 might exert its protective effect on retinopathy in the OIR model by promoting the activation of the Nrf2 pathway.

Fig. 4

H2 promotes Nrf2 activation in OIR mice. A-B the level of Nrf2 expression did not differ between the OIRP12 group and control group but increased after H2 inhalation. C-D At P17, higher Nrf2 levels were observed in both the OIRP17 group and the OIRP17+H2P7-P17 group than in the control group. E. H2 treatment enhanced the expression of the main target genes of Nrf2 (HO-1 and NQO1)

Genetic ablation of Nrf2 impeded the protection provided by H2 against retinopathy in OIR mice

To determine whether Nrf2 is required for the protective effects of H2 on retinal angiogenesis in OIR mice, we investigated retinal angiogenesis in Nrf2−/− mice treated with H2. According to Koichi and Wei’s study [22, 26], avascular areas in Nrf2−/− mice on P9 were significantly larger but similar on P12 to those in Nrf2 +/+ mice. Hyperoxia treatment revealed a similar window of time where Nrf2-regulated antioxidant production was beneficial and contributed to endothelial survival.

Thus, we chose OIR mice on P9 to estimate vaso-obliteration of hyperoxia-exposed retinas in Nrf2−/− and WT mice. The extent of vaso-obliteration and NVT can be determined by the ratio of the avascular/NVT area accounting for the whole-mount retina at P9 and P17. The avascular areas accounted for 32.38% ± 0.81% and 31.71% ± 1.18% of the total retinal area in the Nrf2−/− OIRP9 and Nrf2−/− OIRP9 + H2 P7-P9 groups, respectively, with no statistically significant difference during the hyperoxia period (Fig. 5A). During the hypoxia stage, neovascular areas and avascular areas were observed in retinas from the Nrf2−/− OIR P17 and Nrf2−/− OIR P17 + H2 groups. No changes in avascular or neovascular areas were found in the Nrf2−/− OIRP17 and Nrf2−/− OIR P17 + H2 groups. As mentioned above, Nrf2−/− OIR mice exhibited a remarkably higher extent of avascular retina than OIR-WT mice at P17, which is consistent with previous reports [22]. In addition, H2 could not ameliorate retinal pathological neovascularization and vaso-obliteration in Nrf2−/− OIR mice, indicating that H2 treatment participates in vascular remodeling through Nrf2 activation in OIR mice. (Fig. 5A, E–G).

Fig. 5

Genetic ablation of Nrf2 impeded the protection provided by H2 against retinopathy in OIR mice A． E-G H2 could not ameliorate retinal pathological neovascularization and vaso-obliteration in Nrf2−/− OIR mice. B.H-I The number of tip cells at the junction of the avascular and neovascular regions and the density of blood vessels in the peripheral region in each group. D.M-N Double labeling of isolectin B4 and GFAP showed that tip cells in the retina of the OIR+H2 group were attached to astrocytes, and their filopodia were stretched along them. They were not present in Nrf2-KO mice. C,L. H2 could not exert its antiangiogenic effect without Nrf2 participation. There were no differences in the expression of PCNA between the Nrf2-/-OIR and Nrf2-/- OIR+H2 groups

Next, we investigated the number of tip cells at the junction of the avascular and neovascular regions and the density of blood vessels in the peripheral region in each group. As shown in Fig. 5B, the number of tip cells and vascular density was higher by 2.08 times and 1.19 times, respectively, in the OIRP9 + H2 P7–P9 group than in the OIRP9 group (Fig. 5B). The number of tip cells and vascular density was higher by 1.77 times and 1.28 times in the OIRP17 + H2 P7–P17 group than in the OIRP17 group (Fig. 5H–K). Nevertheless, there was no significant difference in the number of tip cells or vascular density between the OIR (Nrf2−/−)P9 and OIR (Nrf2−/−)P9 + H2P7–P9 groups, and no differences were observed in the number of tip cells or vascular density between OIR (Nrf2−/−)P17 mice and the OIR (Nrf2−/−)P17 + H2P7–P17 groups (Fig. 5B, H, I). Moreover, double labeling of isolectin B4 and GFAP showed that tip cells in the retina of the OIR + H2 group were attached to astrocytes and that their filopodia were stretched along them,which were not present in Nrf2-KO mice (Fig. 5D). These results suggested that H2 inhalation improved revascularization and the formation of tip cells, which is guided by the astrocyte network. Nrf2 ablation weakened the protective effect of H2 against retinopathy in OIR mice.

We further explored the expression of PCNA, which closely correlates with the proliferation activity of vessel endothelial cells. The density of PCNA/IB4-positive cells within the neovascular area in the retinas of the Nrf2−/− OIR and Nrf2−/− OIR + H2 groups was higher than that in the controls and even showed an increase in the normal vasculature area. The results showed that the level of PCNA in the Nrf2−/− OIR and Nrf2−/− OIR + H2 groups was higher by 9.54-fold and 9.44-fold, respectively, the levels in the controls (Nrf2−/− mice) under normoxia. In addition, as shown in Fig. 6K, there were no differences in the expression of PCNA between the Nrf2−/− OIR and Nrf2−/− OIR + H2 groups. These results revealed that H2 could not exert its antiangiogenic effect without Nrf2 participation(Fig. 5C, L).

Fig. 6

H2 promotes retinal vascular regeneration and ameliorates neovascularization of OIR mice Via Nrf2-notch axis. A. H2 upregulated Nrf2 target genes (NQO1 and HO-1) and downregulated Notch/Dll4 mRNA expression. Nrf2 KO abolished the effect of H2 on Notch/Dll4 expression at the transcriptional level. B. Dll4 and Notch expression was suppressed in the OIRP17 + H2 P7–P17 group. C. H2 halation suppressed the elevation in the Notch1 and Dll4 protein level of OIR mice at P17. D.F Nrf gene ablation abolished the suppression of Notch/Dll4 in H2-treated OIR mice. E.Notch1 and Dll4 expression have an increasing trend after H2 treatment, although with no significant difference

Moreover, due to the vital function of astrocytes and microglia, we examined whether Nrf2 gene deletion blocked the effect of H2 on neuroglia. Surviving retinal astrocytes were labeled by GFAP immunostaining in Nrf2−/− OIR mice on P17 and exhibited finer processes and a decrease in astrocyte density in the avascular area of retinas. However, H2 administration preserved the distribution of astrocytes and reduced the disruption of the astrocyte template in Nrf2−/− OIR mice (Fig. 5D, N). In addition, microglial cells in Nrf2−/− OIR mice showed enlarged cell bodies and shorter processes with fewer branches. After H2 treatment in Nrf2−/− OIR mice, the proportion of amoeboid microglial cells was reduced by almost half (Fig. 5D, M). These data indicated that H2 could preserve the astrocyte template and inhibited the activation of microglia in the retinas of OIR mice without Nrf2 signal pathway participation.

H2 promotes retinal vascular regeneration and ameliorates neovascularization of OIR mice Via Nrf2-notch axis

The aforementioned results encouraged us to explore the underlying mechanism of the effects of H2 on the OIR mouse model. As mentioned above, increased vascular regeneration and alleviated pathological neovascularization by H2 were eliminated after Nrf2 knockout. It is known that the transcription factor Nrf2, well known for regulating the stress response in multiple pathologic settings, is a critical intracellular regulator in ECs for sprouting angiogenesis in vascular development through delta-like 4 (Dll4)/Notch signaling [20]. Given the critical role of Notch ligand Dll4 in negatively regulating angiogenic sprouting, we hypothesized that Dll4/Notch signaling might be modulated by Nrf2 in H2-induced vascular remodeling.

To confirm that inhibition of vascular Notch1 / Dll4 is a key mechanism for hydrogen-induced vascular remodeling of OIR,we observed the mRNA and protein expression levels of different groups of Notch1 and Dll4. H2 upregulated Nrf2 target genes (NQO1 and HO-1) and downregulated Notch/Dll4 mRNA expression. Nrf2 KO abolished the effect of H2 on Notch/Dll4 expression at the transcriptional level (Fig. 6A). Compared with the OIRP17 group, Nrf2 target genes (NQO1 and HO-1) expression was higher in the OIRP17 + H2 P7–P17 group. Accordingly, Dll4 and Notch expression was suppressed in the OIRP17 + H2 P7–P17 group (P < 0.05).

Contrary to inhibiting the Notch1/Dll4 pathway in OIRP17 + H2 P7–P17 mice,Notch and Dll4 target gene (Hes1, and Hey1) expression was higher in Nrf2−/− retinas with H2 inhalation than in those with OIRP17(WT) + H2 P7–P17 (P < 0.05 Fig. 6D). In consistent with mRNA level expression,western blot results showed that H2 halation suppressed the elevation in the Notch1 and Dll4 protein level of OIR mice at P17 (Fig. 6C). In addition, compared with controlP17 (Nrf2−/−) mice, Notch1 and Dll4 protein expression was upregulated in the OIRP17 (Nrf2−/−) group,even higher in OIRP17 (Nrf2−/−) + H2 group,which indicated Nrf gene ablation abolished the supression of Notch/Dll4 in H2-treated OIR mice (Fig. 6D, F). Next,we, examined the Notch/Dll4 protein level of OIR mice at P12. Western blot results showed that H2 reduce the elevation in the Notch1 protein level of OIR mice,but not Dll4. In Nrf2−/− OIR mice,Notch1 and Dll4 expression have a increasing trend after H2 treatment,although with no significant difference (Fig. 6B, E). These results suggest that H2 inhalation inhibited the Notch1/Dll4 pathway through Nrf2 activation. Therefore, H2 negatively regulates Dll4/Notch signaling to participate in retinal angiogenesis through Nrf2 activation in OIR mice.

H2 suppressed notch/Dll4 and ROS via Nrf 2 activation in vitro

To verify whether the interaction between Nrf 2 and Notch/Dll4 pathway and the H2 in regulating the Dll4/Notch axis via Nrf 2, we conducted an in vitro study using human umbilical vein endothelial cells (HUVECs) cultured under hypoxic conditions. We observed an increase in Nrf2 expression and its target genes, as well as Dll4/Notch expression, in the hypoxic group compared to the normal group. Treatment with H2 further increased Nrf2 expression and decreased Dll4/Notch expression in the hypoxia + H2 group. We determined that the maximum inhibitory concentration of the Nrf2 inhibitor ML385 was 5 μM (Figure S2) [27]. In the presence of ML385, the combination of hypoxia and hydrogen gas resulted in decreased Nrf2 expression and increased Dll4/Notch expression(Fig. 7A–E). These findings suggest that Nrf2 activation negatively regulates the expression of the Dll4/Notch signaling pathway, consistent with in vivo results. Additionally, under hypoxic conditions, the presence of H2 decreased the intensity of reactive oxygen species (ROS) fluorescence in cells, while Nrf2 inhibition led to an increase in ROS levels once again(Fig. 7F, G).

Fig. 7

H2 suppressed notch/Dll4 and ROS via Nrf 2 activation in vitro. A-E an increase in Nrf2 expression and its target genes, as well as Dll4/Notch expression, in the hypoxic group compared to the normal group. Treatment with H2 further increased Nrf2 expression and decreased Dll4/Notch expression in the hypoxia + H2 group. In the presence of ML385, the combination of hypoxia and hydrogen gas resulted in decreased Nrf2 expression and increased Dll4/Notch expression. F.G Under hypoxic conditions, the presence of H2 decreased the intensity of reactive oxygen species (ROS) fluorescence in cells, while Nrf2 inhibition led to an increase in ROS levels

Based on this, it can be inferred that hydrogen protect endothelial cell function.The MTT results showed that hydrogen enhanced the inhibitory effect of hypoxia on endothelial cell growth, but this effect was reduced when Nrf2 was inhibited. The Transwell data revealed that hydrogen significantly improved HUVEC migration, while hypoxia inhibited it. However, this effect was significantly diminished after Nrf2 inhibition(Fig. 7G, H). This suggests that hydrogen gas promotes the proliferation of normal endothelial cells by regulating negatively the Dll4/Notch pathway through Nrf2, and it also protects these cells from damage caused by oxidative stress by activating Nrf2 under hypoxia treatment.

H2 protected retina against retinopathy of OIR mice via dual-directional regulation of HIF-1α-VEGF pathway both in hyperoxic and hypoxic phases.

In addition to Notch signaling, numerous studies showed that the HIF-1α-VEGF pathway was involved in retinal neovascularization [28]. Hypoxia-inducible factor 1 alpha subunit (HIF-1α) is a key transcription factor in cellular responses to hypoxia and plays a critical role in angiogenesis by activating the transcription of genes encoding angiogenic growth factors. Therefore, we also explored whether H2 could regulate the HIF-1α-VEGF pathway in the retina during different phases. The protein levels of HIF-1α and VEGF in the retinas of different groups were examined by Western blot and immunofluorescence staining.

Our immunofluorescence staining results showed that HIF-1α was expressed in both the INL and GCL in the control littermate on P12, but the expression of HIF-1α in the avascular area was suppressed in mice in the OIRP12 group. In contrast, H2 inhalation increased the expression of HIF-1α in OIR mice at P12 (Fig. 8A). There was an evident decrease in HIF-1α and VEGF protein expression in the OIR group and an increase in HIF-1α and VEGF protein expression after H2 treatment(Fig. 8B, C).

Fig. 8

H2 protected retina against retinopathy of OIR mice via dual-directional regulation of HIF-1α-VEGF pathway both in hyperoxic and hypoxic phases. A．HIF-1α was expressed in both the INL and GCL in the control littermate on P12, but the expression of HIF-1α in the avascular area was suppressed in mice in the OIRP12 group. In contrast, H2 inhalation increased the expression of HIF-1α in OIR mice at P12；HIF-1α was slightly expressed in the control group at P17. Unlike the increased HIF-1α expression in the OIR group, HIF-1α expression was decreased in OIR mice on P17 after H2 treatment. B-C an evident decrease in HIF-1α and VEGF protein expression in the OIR group and an increase in HIF-1α and VEGF protein expression after H2 treatment. D-E H2 inhalation reversed the trend of increased HIF-1α and VEGF protein expression in OIR mice to the level in control littermates on P17

HIF-1α was slightly expressed in the control group at P17. Unlike the increased HIF-1α expression in the OIR group, HIF-1α expression was decreased in OIR mice on P17 after H2 treatment (Fig. 8A). In agreement with these results, Western blot analysis revealed that H2 inhalation reversed the trend of increased HIF-1α and VEGF protein expression in OIR mice to the level in control littermates on P17 (Fig. 8D, E). These data suggest that in the OIR mice, hypoxia reduced the level of retinal HIF-1α/VEGF expression, whereas hyperoxia increased the activation of the retinal HIF-1α/VEGF pathway. H2 therapy exerts bidirectional regulation to dynamically control appropriate HIF-1α/VEGF expression. Therefore, we could further infer that the HIF-1α-VEGF pathway was also involved in the protective effects of H2 on the retinas of OIR mice.