Homocysteine: A Potential Biomarker for Diabetic Retinopathy

Diabetic retinopathy (DR) is the most common cause of blindness in people under the age of 65. Unfortunately, the current screening process for DR restricts the population that can be evaluated and the disease goes undetected until irreversible damage occurs. Herein, we aimed to evaluate homocysteine (Hcy) as a biomarker for DR screening. Hcy levels were measured by enzyme-linked immuno sorbent assay (ELISA) and immunolocalization methods in the serum, vitreous and retina of diabetic patients as well as in serum and retina of different animal models of DM representing type 1 diabetes (streptozotocin (STZ) mice, Akita mice and STZ rats) and db/db mice which exhibit features of human type 2 diabetes. Our results revealed increased Hcy levels in the serum, vitreous and retina of diabetic patients and experimental animal models of diabetes. Moreover, optical coherence tomography (OCT) and fluorescein angiography (FA) were used to evaluate the retinal changes in mice eyes after Hcy-intravitreal injection into normal wild-type (WT) and diabetic (STZ) mice. Hcy induced changes in mice retina which were aggravated under diabetic conditions. In conclusion, our data reported Hcy as a strong candidate for use as a biomarker in DR screening. Targeting the clearance of Hcy could also be a future therapeutic target for DR.


Introduction
Diabetic retinopathy (DR) is one of the most substantial microvascular complications of diabetes mellitus (DM), and is the most common cause of blindness in people under the age of 65 [1]. Unfortunately, DR can go undetected and not even noticed until irreversible damage and blindness has occurred [2]. DR is caused by damage to the blood vessels, resulting in retinal ischemia and increased permeability. New blood vessel formation (neovascularization) and diabetic macular edema (DME) are common characteristics for the disease [3]. Currently, retinopathy can only be diagnosed by a qualified specialist, either via direct proper examination of the eye or the examination of images The concentration of CBS in the serum from human patients with and without diabetes and from diabetic mice was determined using a CBS ELISA assay kit from My BioSource (MBS700623) (San Diego, CA, USA). Blood samples were allowed to clot in serum separator tubes (SST) for a minimum of 2 h at room temperature prior to centrifugation at 1000× g for 15 min. The serum was collected and immediately assayed according to the protocol provided with the kit.

Optical Coherence Tomography (OCT) and Fluorescein Angiography (FA)
OCT and FA imaging were used to evaluate the retinal vasculature in living mice according to our published methods [28,31]. Briefly, 2% isoflurane was used to anesthetize mice and 1% tropicamide eye drop was used to dilate their pupils. The anesthetized mice were then individually placed on the imaging platform of the Phoenix Micron III retinal imaging microscope accompanied with an OCT imaging device (Phoenix Research Laboratories, Pleasanton, CA, USA). Then, Genteal gel was applied liberally to keep the eye moist during imaging. For FA, 10% fluorescein sodium (Apollo Ophthalmics, Newport Beach, CA, USA) was injected to the mice (IP, 10 to 20 µL), and fluorescent images were rapidly captured for~5 min. Indistinct vascular borders progressing to diffusely hazy fluorescence was considered as fluorescein leakage.

Immunofluorescent Assessment of Hcy Level
Retinal cryosections were prepared according to our previously published method [28,30]. First, the retinal sections were fixed with 4% paraformaldehyde. The fixation was followed by washing with PBS-Triton X-100, blocking with Power Block (BioGenex, Fremont, CA, USA), and then incubation with an anti-homocysteine antibody (Catalog number; ab5512 rabbit polyclonal, Chemicon International Inc., Temecula, CA, USA) at 37 • C for 3 h or at 4 • C overnight. The sections were subsequently washed with PBS-Triton X-100 three times. Then, the sections were incubated with the appropriate secondary antibody at 37 • C for 1 h. Sections were then washed with PBS-Triton X-100 and Fluoroshield with DAPI (4'6-diamidino-2-phenylindole) was applied, followed by the placement of a coverslip (Sigma Aldrich F6057, Saint Louis, MO, USA) to label the nuclei. Thereafter, sections were examined and images were captured using fluorescent microscopy (Carl Zeiss, Göttingen, Germany).

Statistical Analysis
The results are expressed as mean ± standard deviation SD. The assessment of differences among experimental groups was performed using the two-tailed t test or one-way analysis of variance (ANOVA). Detection of statistical differences by ANOVA was followed by a post hoc Tukey's test to determine which groups differed. Statistical significance was considered at a confidence level of p < 0.05.

Elevated Hcy Levels in Serum, Vitreous and Retina of Diabetic Patients and Experimental Diabetic Animals
To evaluate Hcy as a potential biomarker for the development of retinal complications in diabetes, we tested changes in Hcy level in the blood and retina of experimental models of type1and type 2 diabetes (STZ mice and rats and db/db mice, respectively). Our results showed increased Hcy levels in the blood and retina of diabetic animals compared to corresponding non-diabetic controls. In humans, the serum and vitreous levels of Hcy were significantly increased in diabetic patients compared to non-diabetic controls, suggesting that elevated Hcy levels might represent a risk factor for development of DR ( Figure 1). Chemicon International Inc., Temecula CA,USA) at 37°C for 3 h or at 4°C overnight. The sections were subsequently washed with PBS-Triton X-100 three times. Then, the sections were incubated with the appropriate secondary antibody at 37°C for 1 h. Sections were then washed with PBS-Triton X-100 and Fluoroshield with DAPI (4'6-diamidino-2-phenylindole) was applied, followed by the placement of a coverslip (Sigma Aldrich F6057, Saint Louis, MO, USA) to label the nuclei. Thereafter, sections were examined and images were captured using fluorescent microscopy (Carl Zeiss, Göttingen, Germany).

Statistical Analysis
The results are expressed as mean ± standard deviation SD. The assessment of differences among experimental groups was performed using the two-tailed t test or one-way analysis of variance (ANOVA). Detection of statistical differences by ANOVA was followed by a post hoc Tukey's test to determine which groups differed. Statistical significance was considered at a confidence level of p < 0.05.

Elevated Hcy Levelsin Serum, Vitreous and Retina of Diabetic Patients and Experimental Diabetic Animals
To evaluate Hcy as a potential biomarker for the development of retinal complications in diabetes, we tested changes in Hcy level in the blood and retina of experimental models of type1and type 2 diabetes (STZ mice and rats and db/db mice, respectively). Our results showed increased Hcy levels in the blood and retina of diabetic animals compared to corresponding non-diabetic controls. In humans, the serum and vitreous levels of Hcy were significantly increased in diabetic patients compared to non-diabetic controls, suggesting that elevated Hcy levels might represent a risk factor for development of DR ( Figure 1).

Hcy Immunolocalization in the Retina of Diabetic Patients and Experimental Diabetic Animals
In order to identify Hcy expression in the retina of diabetic patients, we performed Hcy immunostaining in frozen sections from human diabetic retina. Our results showed both marked immunoflourescent and immunohistochemical staining of Hcy in diabetic human retinal sections ( Figure 2a). Further, we performed Hcy immunofluorescence in various experimental models of type1and type 2 diabetes. STZ-injected mice and rats as well as Akita mice were used as experimental models of type 1 diabetes. However, db/db mice were used as an experimental model of type 2 diabetes. Immunoflourescent staining showed an increased expression of Hcy in retinal sections from various diabetic animals compared to the corresponding WT non-diabetic controls (Figure 2b). immunostaining in frozen sections from human diabetic retina. Our results showed both marked immunoflourescent and immunohistochemical staining of Hcy in diabetic human retinal sections ( Figure 2A). Further, we performed Hcy immunofluorescence in various experimental models of type1and type 2 diabetes. STZ-injected mice and rats as well as Akita mice were used as experimental models of type 1 diabetes. However, db/db mice were used as an experimental model of type 2 diabetes. Immunoflourescent staining showed an increased expression of Hcy in retinal sections from various diabetic animals compared to the corresponding WT non-diabetic controls ( Figure 2B).

Downregulation of CBS in the Serum of Diabetic Patients and Experimental Diabetic Animals
Cystathionine beta-synthase (CBS) catalyzes the first step in Hcy clearance through the transsulfuration pathway by converting Hcy and serine to cystathionine and water. To determine whether changes in CBS levels are associated with altered Hcy levels in diabetes, we tested the CBS

Down Regulation of CBS in the Serum of Diabetic Patients and Experimental Diabetic Animals
Cystathionine beta-synthase (CBS) catalyzes the first step in Hcy clearance through the transsulfuration pathway by converting Hcy and serine to cystathionine and water. To determine whether changes in CBS levels are associated with altered Hcy levels in diabetes, we tested the CBS level in the serum of diabetic patients and experimental diabetic animals. Our data demonstrated significantly lower levels of CBS in diabetic subjects compared to non-diabetic control (Figure 3).

3.4.Optical Coherence Tomography (OCT) and Fluorescein Angiography (FA)
To confirm the link between Hcy and microvascular dysfunction in diabetic retina, OCT and FA were performed in WT and diabetic mice injected intravitreally with Hcy. Consistent with our previous publications[26 -28,33],in the current study, OCT examination showed alterations in retinal vasculature with neovascularization in both inner (white arrows) and outer retina (yellow arrows),

Optical Coherence Tomography (OCT) and Fluorescein Angiography (FA)
To confirm the link between Hcy and microvascular dysfunction in diabetic retina, OCT and FA were performed in WT and diabetic mice injected intravitreally with Hcy. Consistent with our previous publications [26][27][28]31], in the current study, OCT examination showed alterations in retinal vasculature with neovascularization in both inner (white arrows) and outer retina (yellow arrows), and disrupted retinal morphology, such as sub-retinal fluid accumulation, separation of the retinal pigmented epithelial layer, and thickening of the basal laminar membrane and the choroid, suggesting neovascularization. Also, STZ-injected mice demonstrated decreased vessel integrity and an impaired blood-retinal barrier (BRB), indicated by increased fluorescein leakage as well as disrupted retinal morphology, when compared to the WT mice. Furthermore, to evaluate the changes of the combined effect of elevated Hcy and diabetes on the retina, STZ-treated mice given intravitreal injections of Hcy were subjected to FA and OCT. Hcy-injected diabetic mice showed more deleterious effects on the retinal architecture, decreased vessel integrity and more impairment of the BRB.FA showed increased fluorescein leakage with focal spots of hyperfluorescence, and OCT results showed more distortion in retinal morphology and BRB integrity and neovascularization compared to Hcy-injected and diabetic mice alone (Figure 4).

3.4.Optical Coherence Tomography (OCT) and Fluorescein Angiography (FA)
To confirm the link between Hcy and microvascular dysfunction in diabetic retina, OCT and FA were performed in WT and diabetic mice injected intravitreally with Hcy. Consistent with our previous publications [26][27][28]33],in the current study, OCT examination showed alterations in retinal vasculature with neovascularization in both inner (white arrows) and outer retina (yellow arrows), and disrupted retinal morphology, such as sub-retinal fluid accumulation, separation of the retinal pigmented epithelial layer, and thickening of the basal laminar membrane and the choroid, suggesting neovascularization. Also, STZ-injected mice demonstrated decreased vessel integrity and an impaired blood-retinal barrier (BRB), indicated by increased fluorescein leakage as well as disrupted retinal morphology, when compared to the WT mice. Furthermore, to evaluate the changes of the combined effect of elevated Hcy and diabetes on the retina, STZ-treated mice given intravitreal injections of Hcy were subjected to FA and OCT. Hcy-injected diabetic mice showed more deleterious effects on the retinal architecture, decreased vessel integrity and more impairment of the BRB.FA showed increased fluorescein leakage with focal spots of hyperfluorescence, and OCT results showed more distortion in retinal morphology and BRB integrity and neovascularization compared to Hcyinjected and diabetic mice alone (Figure 4).

Discussion
This study was conducted to highlight Hcy as a marker and may be a future therapeutic target for DR. Recently, the association between Hcy and (DM) has gained increasing attention. Many studies reported increased levels of Hcy in the plasma of type 1 diabetes mellitus (T1DM) patients compared to non-diabetic controls [34,35]. It was also reported in a meta-analysis study that plasma Hcy concentrations in T1DM patients without any complications were normal compared with healthy people. However, plasma Hcy concentrations showed significant elevations only in T1DM patients with DR compared with T1DM patients without any complications [12], suggesting that increased Hcy levels during DM contributes to the development of retinopathy. In contrast, some studies reported no significant difference in Hcy levels between T1DM and non-diabetic patients [36], yet patients with proliferative retinopathy display significantly higher values of Hcy than those

Discussion
This study was conducted to highlight Hcy as a marker and may be a future therapeutic target for DR. Recently, the association between Hcy and (DM) has gained increasing attention. Many studies reported increased levels of Hcy in the plasma of type 1 diabetes mellitus (T1DM) patients compared to non-diabetic controls [33,34]. It was also reported in a meta-analysis study that plasma Hcy concentrations in T1DM patients without any complications were normal compared with healthy people. However, plasma Hcy concentrations showed significant elevations only in T1DM patients with DR compared with T1DM patients without any complications [12], suggesting that increased Hcy levels during DM contributes to the development of retinopathy. In contrast, some studies reported no significant difference in Hcy levels between T1DM and non-diabetic patients [35], yet patients with proliferative retinopathy display significantly higher values of Hcy than those without [21]. Hcy level was also reported to be more elevated in type 2 diabetes mellitus (T2 DM) patients than controls [36][37][38][39][40]. On the other hand, some studies refuted the link between HHcy and incidence of retinopathy in diabetic patients [40,41]. Indeed, it was reported that T1DM patients had lower levels of Hcy and global DNA methylation [42]. A recent study reported that elevated Hcy levels were associated with an increased risk of DR, especially in T2DM patients [43].
Therefore, the current study aimed to solve this controversy by confirming what has been reported in human studies and also measured Hcy levels in different animal models of DM representing T1DM (STZ mice, Akita mice and STZ rats) and T2 DM (db/db mice). Our study showed a significant increase in Hcy levels in both human and different animal models of diabetes. There are many causes that have been suggested for increased Hcy levels, such as the deficiency of vitamin cofactors needed for Hcy metabolism, such as folic acid and vitamin B 12 [44][45][46][47], or the deficiency of any of the enzymes involved the remethylation [48] or transsulfuration [49] pathways of the Hcy metabolism. Our study showed that the differences in the serum and retinal levels of Hcy were statistically significant between the control and diabetic groups in both humans and animals. In addition, our study showed a significant reduction of the CBS enzyme level, which is a key enzyme needed for Hcy clearance via the transsulfuration pathway, in both human and animal models of DR. Ratnam et al.reported that insulin plays a role in regulating Hcy metabolism, and impaired insulin levels in diseases such as diabetes may influence Hcy metabolism by regulating the hepatic transsulfuration pathway [50]. Diabetic nephropathy is a common microvascular complication of diabetes and could play a role in elevated Hcy levels due to chronic kidney insufficiency and impaired Hcy clearance. However, a recent clinical study was conducted on 163 normo-albuminuric patients with T1DM and normal renal function to examine whether there is an independent relationship between plasma total homocysteine (tHcy) and retinopathy in normo-albuminuric T1DM patients with normal estimated glomerular filtration rate (eGFR). The study suggested that tHcy is independently associated with retinopathy in normo-albuminuric T1DM with normal eGFR [17]. Furthermore, another large study of European type 1 diabetic patients stated that increased concentrations of tHcy were independently related to macro-albuminuria, renal function and hypertension [51].
Moreover, various studies reported the mechanism of action of Hcy leading to retinal neurodegeneration using different animal models. Hcy has been reported to induce apoptosis in retinal ganglion cells and induced ganglion cell loss via the dysregulation of mitochondrial dynamics in vivo and in vitro [52][53][54]. The activation of N-methyl D-aspartate (NMDA) receptors has been also suggested as a possible mechanism of HHcy-induced retinal ganglion cell death during DR in several studies [54][55][56][57][58][59]. Other studies suggested that HHcy exerts its toxic effect via the activation of inflammatory and oxidative stress mechanisms leading to the activation of mitogen-activated protein kinases (MAPK), macrophage infiltration and enhanced pro-inflammatory cytokines production [60]. Moreover, HHcy elicits oxidative stress and decreases nitric oxide's bioactivity, leading ultimately to vascular dysfunction [61]. Our previous work demonstrated that HHcy caused dysfunction of the BRB, disrupts retinal pigment epithelial structure and function, activates oxidative and endoplasmic reticulum stresses and induces retinal neovascularization and epigenetic modifications [27][28][29][30][31]. The current study found that HHcy caused similar structural changes to what has been reported in our previous publications [27,28,31] when injected intravitreally in mice retina, and these changes were more deleterious when Hcy was injected in diabetic mice, suggesting the involvement of Hcy in the pathogenesis of diabetes-induced retinal damage. This was consistent with what has been reported by Srivastav et al.-That Hcy level was correlated with the decrease in the thickness of the retinal nerve fiber layer in diabetic patients-and suggested a correlation between increased serum levels of Hcy and an increased severity of retinopathy [11].

Conclusions
The current study is innovative in suggesting a correlation between elevated levels of Hcy in serum and in retina and DR in both diabetic human and animal models of diabetes. Furthermore, our results suggest an association between increased serum levels of Hcy and an increased severity of retinopathy. Therefore, Hcy could be a useful diagnostic marker for screening to predict the incidence and severity of retinal damage in diabetic patients. In addition, enhancing Hcy clearance via pharmacological or genetic manipulations could be a future preventive/therapeutic strategy in targeting diabetic retinopathy.