Oral Intake of Hydrogen Water Improves Retinal Blood Flow Dysregulation in Response to Flicker Stimulation and Systemic Hyperoxia in Diabetic Mice

Abstract

Purpose: To examine the effects of hydrogen water on retinal blood flow (RBF) dysregulation in diabetes, we evaluated changes in RBF in response to flicker stimulation and systemic hyperoxia in diabetic mice.

Methods: Twelve type 2 diabetic mice were divided into a group that received non-hydrogen water (n = 6, control group) and the other that received hydrogen-rich water (0.6–0.8 mM) (n = 6, HRW group) from six weeks of age. Body weight, blood glucose, intraocular pressure, and blood pressure were evaluated from eight to 14 weeks of age. RBF was measured in the vascular area of the optic disc as mean blur rate using laser speckle flowgraphy in the resting state and response to flicker stimulation and hyperoxia. We evaluated glial activation and oxidative stress based on immunofluorescence expression.

Results: At 14 weeks, blood glucose level was significantly lower in the HRW group, though still elevated. RBF changes improved significantly in the HRW group compared with the control group from eight weeks of age and persisted throughout the study. Immunofluorescent expression of glial fibrillary acidic protein, particularly in the outer plexiform layer, was significantly decreased in the HRW group. Among oxidative stress markers, 3-nitrotyrosine was significantly suppressed in the HRW group.

Conclusions: Hydrogen-rich water intake significantly improved RBF dysregulation in diabetic mice. Hydrogen may improve impaired neurovascular coupling function in diabetic mice by suppressing gliosis and oxidative stress in the retina.

Translational Relevance: This study highlights the potential of oral intake of hydrogen-rich water to mitigate retinal dysfunction in diabetic mice.

Introduction

Diabetic retinopathy (DR) is a visually threatening complication because of impairment of the retinal microvasculature caused by hyperglycemia.¹ Previous studies have reported that diabetes induces the dysfunction of Müller cells.^2,3 Gliosis (glial activation) of Müller cells, which can be detected by glial fibrillary acidic protein (GFAP) expression, occurs early in DR and may be involved in early microvascular dysfunction.⁴ We previously reported that impairment of neurovascular coupling, which is the link between retinal neurons, glial cells, and blood vessels, may be the earliest event in diabetes-related retinal dysfunction.⁵ We have also reported that improvement in Müller glial cell function appears to be responsible for restoring retinal blood flow (RBF) dysregulation in response to systemic hyperoxia and flicker stimulation because retinal glial cells play important roles in both responses.^6,7 

Given the well-established association between oxidative stress and the pathogenesis of diabetic retinopathy,⁸ antioxidative interventions are presumed to be effective in the therapeutic management of diabetic retinopathy. Oxidative stress induces glial dysfunction characterized by elevated GFAP expression in the diabetic retina, with thickening of GFAP-positive filaments in Müller cells compared with that in the non-diabetic retina.⁹ Treatment modalities, including eye drops of dipeptidyl peptidase-4 (DPP-4) inhibitors,¹⁰ fenofibrate,¹¹ and oral administration of tofogliflozin,¹² mitigate glial dysfunction. Nevertheless, whether oxidative stress is involved in glial and RBF dysfunction, which are detected in the early phases of diabetes in diabetic mice, remains unclear. 

Molecular hydrogen is the smallest molecule with a high ability to diffuse through membranes and affect organs inside the body. Hydrogen reacts with and reduces oxidants with high oxidative activity, such as hydroxyl radicals and peroxynitrite, and suppresses inflammation in various tissues. There are multiple ways to ingest hydrogen, and its effects are reported by drinking hydrogen-rich water (HRW), receiving hydrogen-rich saline solution as an injection, applying it as eye drops, or inhaling hydrogen gas.¹³ Recently, Igarashi et al.¹⁴ reported that long-term consumption of HRW mitigated the decrease in the thickness of the outer retinal layers, indicating cell death and degeneration. Kamimura et al.¹⁵ also reported that oral ingestion of HRW improved oxidative stress in the liver, obesity, and blood glucose levels in db/db mice. However, no study has examined the effects of HRW on oxidative stress and glial activation in diabetic mice. In this study, we evaluated the effects of oral HRW intake on RBF in mice with type 2 diabetes. 

Methods

This study was approved by the ethics committee of the Nihon University School of Medicine (October 7, 2022, AP22MED044-1). C57BL/KsJ-db/db (BKS.Cg-Dock7m+/+Leprdb/J; n = 12) male obesity-induced diabetic mice with a leptin receptor mutation and 13-week-old male db/m (non-diabetic congenic littermates, n = 5) control mice were purchased from Charles River Laboratories (Yokohama, Japan). The mice were classified into two groups: one was administered non-hydrogen water (n = 6), and the other was administered HRW (n = 6) from six weeks of age. 

HRW Administration

HRW (Melodian, Osaka, Japan) at a concentration of 1.2–1.6 ppm was administered from six weeks of age. Two packages of hydrogenated water were placed in each cage. The water was placed in an aluminum bag and replaced weekly. The hydrogen concentration was measured weekly using a hydrogen sensor needle (Unisense, Aarhus, Denmark) to confirm that it was maintained at a certain level. A metallic feed-water inlet specially designed to match the water valve was attached to prevent hydrogen release, and excess air was removed by compressing the water package when exchanged. The hydrogen concentration in the water was measured weekly when the package was exchanged. 

Measurements of Blood Glucose (BG), Blood Pressure (BP), Intraocular Pressure (IOP), and Body Weight (BW)

Casual BG was measured every two weeks from the tail vein using a glucose assay kit (Abbott Laboratories, Abbott Park, IL, USA). Systolic and diastolic BP were measured using a sphygmomanometer (THC-31; Softron, Tokyo, Japan) with the animals under anesthesia, followed by BW measurement. IOP was measured using a handheld tonometer (TonolabTV02; ME Technical, Tokyo, Japan) before RBF measurement after the mice were anesthetized with isoflurane. 

Inductions of Systemic Hyperoxia and Flicker Light Stimulation

Systemic hyperoxia was induced by inhaling 100% oxygen for 10 minutes, as described in our previous studies.^7,16 The baseline value was determined as the mean of three consecutive flow measurements, obtained at one-minute intervals over three minutes before initiating hyperoxia. RBF measurements were taken every minute for 20 minutes during hyperoxia (10-minute stimulation) and for 10 minutes after the termination of hyperoxia (10-minute recovery). 

Ficker light stimulation at 12 Hz was used, as this frequency elicits a maximal RBF response in mice. The ambient light was reduced to 1 lux or less prior to flicker stimulation. Mice were dark-adapted for two hours, and flicker stimulation was conducted at an intensity of 30 lux, suitable for the rod-dominant mouse retina.¹⁶ RBF was measured at 20-second intervals during both the three-minute flicker stimulation and the three-minute recovery. The baseline value was calculated as the mean of three consecutive flow measurements obtained at 20-second intervals for one minute before the initiation of flicker light stimulation. 

Measurement of RBF

RBF in the resting state, flicker stimulation, and hyperoxia were measured under anesthesia induced by the inhalation of 2% isoflurane (Pfizer, Tokyo, Japan). Resting-state RBF was measured before each stimulation with flicker light and hyperoxia. RBF under flicker stimulation was measured after more than six hours of dark adaptation. 

Blood flow in the vascular area of the optic disc was measured as the mean blur rate (MBR) using a laser speckle flowgraphy-micro device (Softcare Co., Ltd., Fukutsu, Japan) after the pupils were dilated with 0.5% tropicamide (Santen Pharmaceutical Co., Osaka, Japan). The MBR values obtained from the vascular area around and at the optic nerve head reflect overall retinal circulation and serve as an index of RBF.¹⁷ The average MBR of the vessels was analyzed using LSFG analyzer software (version 3.2.19.0; Softcare Co., Ltd.). Body temperature was maintained at 37°C to 38°C with a heating mat during measurement. 

Study Protocol

Measurements were conducted every two weeks from eight to 14 weeks of age. Two days were set for the measurement. RBF under conditions of flicker stimulation after dark adaptation and BG measurements were performed on day 1. RBF under hyperoxic conditions and BP, IOP, and BW measurements were performed on day 2. After 14 weeks of data collection, the mice were euthanized, and the organs were stored for further analysis. We have previously reported that electroretinography abnormalities appear at 14 weeks of age, whereas deficiencies in flow regulation and neurovascular coupling in the retina can be detected at eight weeks of age.⁵ Because the goal of this study is to determine whether HRW can prevent neural dysfunction in a mouse model of type 2 diabetes, longitudinal monitoring was conducted from eight to 14 weeks of age in db/db mice. 

Immunohistological Analysis

After measurement at 14 weeks of age, all mice were anesthetized by peritoneal injection of three mixed anesthetic agents: medetomidine, midazolam, and butorphanol. A sternotomy was performed, and normal saline solution was perfused into the left ventricle to wash out the circulating blood, followed by perfusion with 4% paraformaldehyde (PFA). After PFA fixation, the eyes were enucleated and stored at 4°C. 

Immunohistological analysis was performed based on the fluorescence intensity of GFAP (IR 52461, Ready to use; Agilent, Santa Clara, CA, USA) as an indicator of glial activation, 8-hydroxy-2′-deoxyguanosine (8-OHdG) (1:50) (MOG-100P; JaICA, Shizuoka, Japan), and 3-nitrotyrosine (1:200) (AB5411; Chemicon International, Temecula, CA, USA) as oxidative stress markers. For 8-OHdG, a secondary antibody of goat anti-mouse IgG (H+L) Highly Cross-Adsorbed, Alexa Fluor 488 (A-11029; Thermo Fisher Scientific, Waltham, MA, USA) was used at a 1:500 dilution. For GFAP and 3-nitrotyrosine, a secondary antibody of goat anti-rabbit IgG (H+L) Cross-Adsorbed, Alexa Fluor 488 (A-11008) was used at a 1:500 dilution. Microscopic imaging was performed using an All-in-One Fluorescence Microscope BZ-X710 (Keyence, Itasca, IL, USA). Immunohistological expression was measured using ImageJ Version 1.54 (National Institutes of Health, Bethesda, MD, USA). The outline of the outer plexiform layer (OPL) was traced by a line in each retinal sample to analyze GFAP expression, and the mean gray value was measured. Intravascular fluorescence intensity was measured to analyze oxidative stress markers. The retinal layers were divided into three layers depending on the location of the retinal blood vessels: the outer layer included vessels in the OPL, the middle layer included vessels at the bottom end of the inner plexiform layer, and the inner layer included vessels in the ganglion cell layer. The perforating vessels were divided into two parts at the vascular plexus's midpoint. The intensity of each marker in the retinal vessels was measured and divided by the area of the vessels in each sample. 

Statistical Analysis

Statistical analyses were performed using Prism 9 (GraphPad Software, San Diego, CA, USA) and EZR Software Version 4.1.2 (Saitama Medical Center, Jichi Medical University). Data are presented as mean ± SD. Statistical significance was set at P <0.05. 

Results

Monitoring of Hydrogen Concentration in HRW

The hydrogen concentration in HRW was measured using a Clark-type needle sensor at 20°C to 25°C. Six db/db mice (six weeks old) were placed in a cage with two aluminum vapor-deposited packages containing 300 mL of HRW (0.6–0.8 mM of hydrogen) and drank HRW as desired. The water solubility of hydrogen was 0.806 mM at 20°C and 1013 hPa. The water packages were replaced weekly. Each time, the hydrogen concentration of the leftover HRW was measured, and the concentration ranged from approximately 0.34 to 0.8 mM. 

Effects of Drinking HRW on Systemic and Ocular Parameters

There were no significant differences in systolic BP, diastolic BP, or BW between the control and HRW groups from eight to 14 weeks of age (Figs. 1A–D). BG levels were significantly lower in the HRW group at 12 weeks of age (P < 0.05), although the median BG level in the HRW group at 14 weeks of age remained an elevated at 427 mg/dL (Fig. 1E). There were also no significant differences in ocular parameters between the control and HRW groups regarding IOP, ocular perfusion pressure, and RBF in the resting state (Figs. 2A–C). 

Figure 1.

 

Change in measurement data of systemic parameters after drinking hydrogen or non-hydrogen water (n = 6 mice per group). Systemic parameters: (A) body weight, (B) systolic BP, (C) diastolic BP, (D) mean BP, (E) BG. All measurements were conducted every two weeks from eight to 14 weeks of age. (A)–(D) were compared between db/db mice fed by non-hydrogen water and db/db mice fed by HW by repeated-measures ANOVA. (E) was compared by the Wilcoxon signed-rank test. BG was significantly lower in the HW group from 12 weeks of age (P = 0.03). NS; not significant.

Figure 2.

 

Change in measurement data of ocular parameters after drinking HW or non-hydrogen water. Data are presented for six mice per group. Ocular parameters: (A) IOP; (B) OPP; (C) RBF in the resting state. All measurements were conducted every two weeks from eight to 14 weeks of age. (A)–(C) were compared between db/db mice fed by non-hydrogen water and db/db mice fed by HW by repeated-measures ANOVA. NS, not significant.

Suppressive Effects of Drinking HRW on Longitudinal Changes in RBF

Two weeks after the start of HRW administration, the RBF responses to systemic hyperoxia (Fig. 3A) and flicker stimulation (Fig. 4A) were both significantly improved in the HRW group compared with those in the control group from eight weeks of age (P < 0.05, two-way repeated measures analyses of variance [ANOVA]). Improvement in RBF responses to both stimulations in the HRW group persisted throughout the study period (Figs. 3B, 3C, 4B, 4C). 

Figure 3.

 

The RBF response to systemic hyperoxia was significantly improved in the HW group from eight to 14 weeks of age: (A) eight weeks of age; (B) 10 weeks of age; (C) 12 weeks of age; (D) 14 weeks of age. * P < 0.05 between groups analyzed by repeated-measures ANOVA; solid bar = period of hyperoxia.

Figure 4.

 

The RBF response to flicker stimulation was significantly improved in the HW group from eight to 14 weeks of age: (A) eight weeks of age; (B) 10 weeks of age; (C) 12 weeks of age; (D) 14 weeks of age. * P < 0.05 between groups analyzed by a repeated-measures ANOVA; solid bar = period of flicker stimulation.

Maximum RBF Change in Response to Hyperoxia and Flicker Stimulation at 14 Weeks in Diabetic Mice and db/m Nondiabetic Control Mice

At 14 weeks of age, systemic hyperoxia and flicker stimulation were performed using the same protocol for both diabetic mice and db/m non-diabetic control mice (Fig. 5). Because we previously confirmed that the RBF response to both stimuli remained stable in db/m non-diabetic control mice from eight to 20 weeks of age, RBF measurements were conducted only once at 14 weeks, in accordance with our institution's animal welfare criteria. Figure 5 illustrates that the RBF changes observed in db/m control mice in response to systemic hyperoxia (A) and flicker stimulation (B) were abolished in db/db mice treated with non-hydrogenic drinking water. However, these changes were restored to normal in db/db mice treated with HRW, compared to the db/m control group. 

Figure 5.

 

Maximal changes in RBF from baseline in response to systemic hyperoxia (A) and flicker stimulation (B) in 14-week-old mice. Data are shown for db/m (n = 5), untreated db/db (n = 6), and HRW-treated db/db mice (n = 6). Under flicker stimulation, the maximum rate of RBF change was 51.0% ± 6.6% in db/m, −38.2% ± 4.8% in untreated db/db, and 28.7% ± 11.2% in HRW-treated db/db mice. Under hyperoxic conditions, the maximum rate of RBF change was −39.4% ± 3.4% in db/m, 42.7% ± 6.4% in untreated db/db, and −25.6% ± 5.5% in HRW-treated db/db mice. * P < 0.05, ** P < 0.01 compared with db/m, as determined by one-way ANOVA followed by the Holm-Sidak test.

Suppressive Effects of Drinking HRW on Gliosis and Oxidative Stress in the Retina

The upregulation of GFAP in Müller cells in diabetic db/db mice has been reported. We observed increased immunofluorescence staining for GFAP in the OPL of the control group at 14 weeks of age. The immunofluorescence intensity of GFAP was significantly suppressed in the HRW group (P < 0.05, Mann–Whitney U test) (Fig. 6). 

Figure 6.

 

Comparison of GFAP immunofluorescence (green) in OPL between non-hydrogen and hydrogen-rich water groups. Immunofluorescence images of the (A) non-hydrogen water group and (B) HW group. (C) Comparison of fluorescein intensities of GFAP in OPL as mean gray value by analyzing with Image J (n = 5 retinal images per group). There were significant differences in the fluorescein intensities of GFAP (P = 0.03) between non-hydrogen group (left box) and HW group (right box) by Mann–Whitney U-test. Nuclei were labeled with DAPI (blue), and blood vessels were labeled with lectin (red). *P < 0.05

Hyperglycemia-induced oxidative stress results in glial dysfunction, as indicated by increased GFAP expression in the retina. Regarding oxidative stress, there were no significant differences in the accumulation of 8-OHdG between the two groups in all three layers (inner layer; P = 0.14, middle layer; P = 0.851, outer layer; P = 0.624) at 14 weeks of age (Fig. 7). In contrast, 3-nitrotyrosine accumulation was significantly suppressed in the HRW group in all three layers (P < 0.05) (Fig. 8). Nitrotyrosine is a biomarker of oxidative stress formed by the nitration of protein-bound and free tyrosine residues by peroxynitrite molecules. 

Figure 7.

 

Comparison of intravascular immunofluorescence intensity of 8-OHdG between non-hydrogen and hydrogen-rich water groups. Analysis includes seven or eight retinal images per group. Intensity was compared between vessels of retinal layers in different depths. (A) Inner layer, (B) middle layer, and (C) outer layer. (D) Intravascular intensity of 8-OHdG; nuclei were labeled with DAPI (blue), and intravascular intensity of 8-OHdG are labeled with green. There was no significant difference in intensity between the two groups in all layers. Mann–Whitney U test.

Figure 8.

 

Comparison of intravascular immunofluorescence intensity of 3-nitrotyrosine between non-hydrogen and hydrogen-rich water groups Data are based on eight to 12 retinal images analyzed per group. Intensity was compared between different depths. (A) Inner layer, (B) middle layer, and (C) outer layer. (D) Intravascular intensity of 3-nitrotyrosine; nuclei were labeled with DAPI (blue), and intravascular intensity of 3-nitrotyrosine are labeled with green. Fluorescence intensities were suppressed in all three layers (inner layer, P = 0.04; middle layer, P = 0.021; outer layer, P = 0.029). Mann–Whitney U test. *P < 0.05.

Discussion

This longitudinal interventional study demonstrated that long-term administration of hydrogen water ameliorated the dysregulation of RBF, intravascular oxidative stress within the retina, and dysfunction of retinal glial cells. There were no significant differences in the changes in blood flow in response to systemic hyperoxia and flicker stimulation between non-diabetic control mice (db/m) and db/db mice treated with HRW. Consequently, retinal neurovascular coupling in diabetic mice was nearly restored to normal levels following HRW treatment. 

The hydrogen concentration was maintained at a certain level throughout the study period, as previously reported.⁴ We found that long-term drinking of HRW improved blood flow dysregulation in response to systemic hyperoxia and flicker stimulation in diabetic mice from eight to 14 weeks of age but did not affect RBF in the resting state. We also observed a significant reduction in GFAP in the OPL of db/db mice administered HRW. This finding suggests suppression of glial activation, particularly in the deep retinal layer. The deep vascular plexus in the retina includes Müller cells as fluid transporters, and disruption of this system is presumed to cause macular edema.¹⁸ Therefore, based on GFAP expression, long-term intake of hydrogen water may be a treatment option for early-stage diabetic retinopathy, possibly preventing the progression of macular edema. 

Our longitudinal study also supports the findings of a previous study by Oharazawa et al.,¹⁹ which reported the suppression of GFAP expression by hydrogen water eye drops in an acute ischemia/reperfusion model rats. However, they examined only the acute effects of hydrogen water eye drops on the ocular surface during the ischemia/reperfusion period. Igarashi et al.¹⁴ recently reported that long-term drinking of hydrogen water has neuroprotective effects and inhibits photoreceptor death in mice with retinitis pigmentosa (RP), suggesting the potential of drinking hydrogen water for treating RP. However, there was no significant difference in GFAP expression between the hydrogen and normal water–drinking groups. This may be because GFAP expression was not strong in the retinas of the RP model mice. Therefore we first confirmed that longitudinal oral intake of hydrogen improves RBF dysregulation and glial activation, as detected by GFAP expression, in type 2 diabetes. 

In this study, there was no significant difference in the intravascular expression of 8-OHdG in the retina; however, the formation of 3-nitrotyrosine was significantly suppressed in the hydrogen water group compared with the control group. According to a report by Ramos et al.,²⁰ two weeks of administration of DPP-4 inhibitor eye drops in db/db mice prevented oxidative stress, as shown by the decrease in 8-OHdG and 3-nitrotyrosine levels in the nuclear layers, which was increased in the vehicle group. In our study, hydrogen improved oxidative stress in retinal vessels, indicating that hydrogen is antioxidative during transportation in the bloodstream after oral intake. Moreover, our result that intravascular oxidative stress especially improved in the OPL, in which GFAP immunofluorescence also decreased by the administration of hydrogen, may indicate that long-term consumption of hydrogen water may have the therapeutic potential to prevent the development of DR and macular edema via the reduction of oxidative stress in the retinal vessels. However, further clinical studies are required to confirm this hypothesis. 

This study revealed a slight but significant decrease in blood glucose levels in the hydrogen water–drinking group compared with the control group at 14 weeks of age. Among the multiple routes of administration, hydrogen water consumption is a convenient way to intake hydrogen compared with inhaling gas or administering injections, owing to its low invasiveness and the lack of additional equipment. In a clinical study on patients with impaired glucose tolerance, hydrogen water consumption normalized the oral glucose tolerance test, demonstrating its preventive effect against type 2 diabetes.²¹ There is a possibility that hydrogen exerts beneficial insulinogenic effects on the pancreatic β cells via the neuronal network and regulates the blood glucose level²² or is directly diffused from the liver, which is thought to accumulate hydrogen with glycogen,¹⁵ resulting in improved insulin resistance. In the current study, HRW did show a slight decrease in BG levels compared to the control group; however, BG levels remained elevated, with a median of 427 mg/dL in the HRW group at 14 weeks of age. We have previously found that blood flow dysregulation occurs at approximately 400 mg/dL in db/db mice.⁵ Therefore we hypothesize that the observed improvement in RBF response was due to a reduction in oxidative stress rather than a decrease in BG levels. 

Diabetes induces enodothelial nitric oxide synthase uncoupling and peroxynitrite formation, leading to endothelial dysfunction.²³ Our study showed the suppression of 3-nitrotyrosine in the area of retinal vessels in the hydrogen water group, indicating that drinking hydrogen water provides an antioxidant effect as a peroxynitrite scavenger. Further studies are needed to examine whether suppressing nitrosative stress in retinal vessels leads to improvements in neurovascular coupling function. 

In the current study, we induced two types of provocations—flicker stimulation and systemic hyperoxia—every two weeks from eight to 14 weeks of age. The mechanisms regulating blood flow in response to these provocations differ significantly. Flicker stimulation increases RBF through nitric oxide production by retinal neurons²⁴ and glial cells.⁶ In contrast, systemic hyperoxia decreases blood flow because of endothelin-1 (ET-1) production²⁵ from vascular endothelium, pericytes,²⁶ and glial cells.⁷ By evaluating RBF under both conditions, we aim to elucidate the detailed pathology of the retina in diabetes. 

In our previous research, we observed that retinal circulation became unresponsive to both systemic hyperoxia and flicker stimulation, with flow responses tending to reverse in the later stages of diabetes.⁵ This phenomenon was also noted in the untreated db/db mice in the current study. In addition, flicker-evoked negative RBF responses in db/db mice without HRW treatment align with previous findings, which reported flicker-induced reductions in vessel density and relative blood flow across the three capillary plexuses in STZ-induced diabetic mice using a functional optical coherence tomography angiography.²⁷ Although we could not investigate the exact mechanisms behind these reversed flow responses, the observed impairment in response to systemic hyperoxia might be attributed to diminished vascular responsiveness to ET-1,²⁸ potentially related to endothelin A receptor desensitization²⁹ in vascular smooth muscle cells or pericytes exposed to elevated ET-1 levels in diabetes. 

This study has some limitations. First, the sample size was small owing to the decision of the ethics committee of our institution. Second, we could not measure the actual hydrogen concentration in the eyes of the mice. We should investigate how much hydrogen reaches the eye through experiments with larger animals to measure the hydrogen concentration in the vitreous or retina after consumption of hydrogen water. Third, the mechanism of the beneficial effect of hydrogen water in chronic conditions such as diabetes may differ from the direct antioxidant activities of hydrogen water eye drops, as reported in a previous study using an acute ischemic/reperfusion mouse model.¹⁸ Further investigation is needed using other administration methods, including eye drops, that may directly increase the intraocular hydrogen concentration. Finally, because non-diabetic mice consuming HRW were not included in the current study, we cannot exclude the possibility that HRW itself might alter RBF regulation in response to systemic hyperoxia and flicker stimulation in normal mice. Further study is needed to exmine this possibility. 

In conclusion, the long-term effects of hydrogen water consumption improved RBF dysregulation in type 2 diabetic mice via the suppression of oxidative stress. Hydrogen water may be a therapeutic agent against early-stage diabetic retinopathy. 

Acknowledgments

The authors thank Akiko Tomioka for her contributions to the animal experiments. 

Supported by a Grant-in-Aid for Scientific Research (C) 26861430 from the Ministry of Education, Science, and Culture, Tokyo, Japan (to TN). 

Disclosure: R. Sugiyama, None; J. Hanaguri, None; H. Yokota, None; A. Kushiyama, None; S. Kushiyama, None; T. Kikuchi, None; T. Igarashi, None; M. Iketani, None; I. Ohsawa, None; S. Harino, None; H. Nakashizuka, None; S. Yamagami, None; T. Nagaoka, None