Effects of Intravitreal Injection of Hypoxia-Induced Umbilical Cord Mesenchymal Stem Cell Exosomes on Diabetic Retinopathy

BACKGROUND: This study aims to investigate the effect of intravitreal injection of hypoxia-induced human umbilical cord mesenchymal stem cell exosomes (hypo-hucMSC-Exs) on retinal microvasculature in rats with diabetic retinopathy (DR) and to determine whether hypo-hucMSC-Exs exhibit greater effects on DR than human umbilical cord mesenchymal stem cell exosomes (hucMSC-Exs) METHODS: Exosomes were isolated from normal hucMSCs culture media and hypoxic preconditioning culture media. transmission electron microscope, nanoparticle tracking analysis, and western blot were applied to identication of hucMSC-Exs. Streptozotocin-induced diabetic rats were used as a model for DR. Fundus uorescein angiography (FFA) was conducted to evaluate retinal microvasculature changes in vivo 4 week, 8 weeks and 12 weeks following intravitreal injection. RESULTS: The control group did not experience any signicant change throughout the study. After STZ injection, all rats developed hyperglycemia. The development of retinal microvascular changes include tortuous vessels, massive microaneurysms and late leakage of uorescein dye was visualized using FFA after induction of DR. This response was ameliorated in diabetic rats treated with hucMSC-Exs. To further investigate the effects of hypo-hucMSC-Exs we found that injection with hypo-hucMSC-Exs remarkably reduced the microvasculature lesions compared with the hucMSC-Exs group. CONCLUSION: These ndings suggest that intravitreal injection of hucMSC-Exs prevented diabetes-induced microvasculature lesions. hypo-hucMSC-Exs could enhance this effects and could be considered as a potential target for DR prevention and treatment. microscope; WB, western blotting


Background
Diabetic retinopathy (DR) is a common microvascular complication of diabetes mellitus and the main cause of blindness in adults. Blindness caused by DR is preventable but irreversible and represents a huge economic burden on families and society. Microvascular dysfunction is a major characteristics of DR, which is diagnosed on the basis microvascular lesions [1]. Laser treatment, anti-vascular endothelial growth factor therapy, and vitrectomy are commonly used to treat DR; however, not all patients respond well to these approaches.
In recent years, the study of exosomes has gradually attracted increasing attention, opening up potential new avenues for treatment of retinal diseases. Speci cally, research has shown that exosomes are involved in the occurrence and progression of many retinal diseases including DR, and they have potential value for application in the diagnosis and targeted treatment of DR.
Mesenchymal stem cells (MSCs) are derived from a wide range of tissues, are easy to and culture in vitro, and secrete relatively large amounts of exosomes, making them the best source cells for basic and clinical studies in exosome research. Vitreous MSC injection can exert a protective effect on the retina in DR by secreting various neurotrophic factors [2]. Oxygen concentration has an important role in the processes of MSC proliferation, differentiation, and self-renewal [3] [4]. Under hypoxia, MSCs can exhibit enhanced biological function and improved therapeutic e cacy [3,5]. Furthermore, MSC-derived exosomes have great potential for clinical application. For example, MSCs grown in a medium with conditions similar to those in peripheral artery disease (0% fetal bovine serum [FBS], 1% O 2 ) generate exosomes containing numerous pro-angiogenic factors that could be bene cial to ischemic tissue [6].
Another study using a model of heart infarction showed that hypoxia-induced MSC-derived exosomes increased vascularization, reduced myocardial apoptosis rate, and elevated numbers of cardiac progenitor cells [7]. Hence, hypoxic preconditioning of MSCs can signi cantly enhance their biological function, thereby improving their curative effects [8,9]. Since exosome content is greatly affected by the microenvironment in which the host cells reside [10], exosome levels can be improved by hypoxic preconditioning of source of MSCs. Hypoxic preconditioning can enhance MSC-derived exosome function in models of myocardial infarction [11], spinal cord injury [12], and bone fracture [13]; however, it remains unclear whether exosomes from MSCs exposed to hypoxic conditions exert better therapeutic effects in DR. Since the hypoxic microenvironment is a prominent feature of DR, we questioned whether hypoxiainduced human umbilical cord mesenchymal stem cell exosomes (hypo-hucMSC-Exs) would ameliorate DR-induced retinal vascular injury.
In this study, we established a streptozotocin (STZ)-induced DR model in rats to test the hypothesis that exosomes from MSCs grown under hypoxia would have better therapeutic e cacy than exosomes from MSCs cultured under normoxic conditions. The DR model was used to determine whether intravitreal injection of hypo-hucMSC-Exs had better protective effects on the retina than that of hucMSC-Exs not exposed to hypoxic conditions.

Ethics statement
All experiments involving animals were performed in accordance with the Animals for Scienti c Procedures Act 1986 and the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research, and the study protocols were approved by the Medicine Ethics Review Committee for Animal Experiments of our hospital.

Characterization of hucMSC-Exs
HucMSC-Ex morphology was characterized using an FEI Tecnai G2 transmission electron microscope (TEM) (Philips, Amsterdam, Netherlands). Nanoparticle tracking analysis (NTA) was used to analyze the size distribution of exosomes according to vesicle diameter by ZetaView (Particle Metrix, Meerbusch, Germany). A bicinchoninic acid protein assay kit was employed to determine the concentration of protein extracted from exosomes. Western blotting (WB) was used to analyze the expression of speci c exosome surface markers, including CD9 and CD63.

Experimental animals
Male Sprague-Dawley rats (6-8 weeks old, 180-220 g, n = 24) were obtained from Beijing Vital River Laboratory Animal Technology Co., Ltd. Rats were housed under standard laboratory conditions at a room temperature of 23°C ± 3°C and a humidity of 40-65% on a 12-h light/dark cycle and were provided food and water ad libitum.

Diabetic model establishment and intravitreal injections
A rat model of diabetes was constructed by administering a single intraperitoneal injection of STZ (60 mg/kg body weight) (Sigma, St. Louis, MO, USA). Nonfasting blood glucose level was measured using blood drawn from the tail vein one week after the injection of STZ. Rats with hyperglycemia (glucose levels > 16.7 mmol/L) on three consecutive days were identi ed as having diabetes mellitus and selected for use in subsequent experiments. Eligible DR model rats (n = 24) were arbitrarily assigned to diabetic, hucMSC-Exs, and hypoxia preconditioned hucMSC-Exs (hypo-hucMSC-Exs) groups (n = 8 rats per group). In addition, 8 healthy age-matched rats were selected as the normal control group and fed a standard diet. Intravitreal injections were administered 60 days after diabetic model establishment. All injections were administered into the right pars plana under iso urane (2-3% in oxygen) anesthesia. After instilling 5% betadine solution into the fornix, a sterile 33-G needle attached to a Hamilton syringe was used to deliver 1 µL solution per eye of either hypo-hucMSC-Exs, hucMSC-Exs, or PBS. Hypo-hucMSC-Exs group rats received intravitreal injection of hypo-hucMSC-Exs, the hucMSC-Exs group rats received intravitreal injection of hucMSC-Exs, and rats in the normal control and diabetic groups were administered the same volume of PBS.

Fundus uorescein angiography (FFA)
FFA was performed to assess potential changes in the retinal vasculature in vivo 4, 8, and 12 weeks after intravitreal injection. Fluorescein sodium (10%, 0.2 ml) (Akorn Inc., Lake Forest, IL, USA) was administered via intravenous injection. Angiography was performed (Heidelberg HRT2, Heidelberg, Germany) to investigate for signs of vascular leakage or changes in vessel tortuosity. SPSS 21.0 (IBM Corp, Armonk, NY, USA) was used for data analysis. Data are presented as the mean ± standard deviation. Comparisons among multiple groups were performed using one-way analysis of variance (ANOVA). Following one-way ANOVA, pairwise comparisons were conducted using the least signi cant difference method. P < 0.05 was considered statistically signi cant.

Characterization of hucMSC-Exs and hypo-hucMSC-Exs
We rst isolated exosomes from the culture supernatants of MSCs grown in normoxic or hypoxic environments. Exosomes were then analyzed using TEM, NTA, and WB. TEM revealed nanoparticles with a typical spheroid morphology, ranging in size from 50 to 150 nm in diameter ( Figure 1). There were no apparent differences in morphological parameters (size, shape, and electron density) between hucMSC-Exs and hypo-hucMSC-Exs. NTA was used to quantitatively assess the size and concentration of exosomes, which exhibited a similar size distribution, with means of 124.8 and 121.6 nm for hucMSC-Exs and hypo-hucMSC-Exs, respectively. Hypoxic preconditioning led to generation of more hucMSC-Exs than normoxic conditions. WB con rmed the expression of characteristic hucMSC-Exs surface markers including CD9 and CD63. Notably, the protein concentration in hypo-hucMSC-Exs was signi cantly higher than that in hucMSC-Exs. Hence, relative to the normoxic conditions, hucMSC-Exs were produced at higher levels following MSC exposure to hypoxic conditions.

Intravitreal injection of hucMSC-Exs did not alter blood glucose levels
We established a rat model of diabetes using a single intraperitoneal injection of STZ. Body weight and blood glucose levels were monitored routinely. There were no signi cant differences in blood glucose or body weight among all groups at baseline. At 72 hours after STZ injection, blood glucose levels of the treated rats increased signi cantly, and levels in selected rats were all > 16.7 mmol/L. Blood glucose levels were markedly increased in the diabetic, hucMSC-Exs, and hypo-hucMSC-Exs groups compared with the normal control group. Intravitreal injection of hucMSC-Exs and hypo-hucMSC-Exs did not in uence blood glucose levels, and there was no signi cant difference in blood glucose levels among the diabetic, hucMSC-Exs, and hypo-hucMSC-Exs groups at 4, 8, and 12 weeks after intravitreal injection (P > 0.05) ( Table 1).

Effect of hypo-hucMSC-Exs on retinal vascular abnormalities
We performed FFA at 4, 8, and 12 weeks after intravitreal injection to assess potential changes in retinal microvasculature and permeability in vivo. The normal control group had even caliber vessels radiating from a well-de ned optic nerve head. No signi cant changes in vascular permeability or vessel tortuosity were detected over the course of the experiment in the control group. In contrast, FFA revealed intensely hyper uorescent lesions associated with DR. DR control group rats had tortuous vessels and massive microaneurysms in the central and peripheral retina 4 weeks after PBS administration. Representative images showing pathological features associated with DR, including microaneurysm, venous abnormalities, hemorrhage and vessel leakage, are summarized in Figure 2. Fluorescein leakage was visualized as indistinct vascular borders progressing to diffusely hazy uorescence. Hence, longitudinal FFA revealed that rats in the DR group had extensive vascular changes manifested as heavy uorescein leakage that showed characteristic changes over time.
As shown in Figure 3, DR control group rats had the highest increase in uorescein leakage among the three groups of STZ-induced DR rats compared with the normal group at 12 weeks after treatment. DR group rats also presented with extensive areas of capillary non-perfusion and preretinal hemorrhage. Changes in the hucMSC-Exs group remained relatively stable, with some increases in the size and numbers of hyper uorescent spots on follow-up examination, indicating that administration of hucMSC-Exs signi cantly reduced retinal vascular leakage. At 4 weeks post-administration, rats in the hucMSC-Exs group showed few microaneurysms, focally dilated vessels, and isolated leaky lesions. At 8 weeks after administration, vascular leakage remained relatively stable, although slight progression was observed in individual lesions. Most retinal capillaries were relatively well preserved, with only a few microaneurysms. At 12 weeks after administration, there were increased numbers of readily identi able and isolated leaky lesions in hucMSC-Exs group rats, with retention of large areas of apparently normal retinal capillaries. There were no notable alterations 4 weeks after administration in hypo-hucMSC-Exs group rats, including no vascular leakage, venous dilatation, or tortuosity. Eight weeks post-administration, retinal vessels showed regular course and caliber, without microvascular changes or vessel leakage; however, at 12 weeks post-administration, there was an increase in vascular tortuosity, with evidence of some microvascular changes (e.g., microaneurysms).
The FFA data indicate that intravitreal injection of hucMSC-Exs and hypo-hucMSC-Exs decreased retinal damage. Leakage from the uorescein spot and vascular tortuosity were less extensive in the hypo-hucMSC-Exs group than in the hucMSC-Exs group, indicating that DR progression was reduced in rats treated with hypo-hucMSC-Exs. Our ndings suggest that hypo-hucMSC-Ex injection has a stronger potential to protect against DR progression than that of hucMSC-Exs.

Discussion
With the development of regenerative medicine, MSC transplantation is now considered a viable option for the treatment of a variety of refractory clinical diseases. MSCs have great therapeutic potential as they can self-replicate, differentiate, and regulate hematopoietic and immune cells [14,15]. There is evidence that MSCs act via a paracrine mechanism in the treatment of various diseases, and exosomes are important in this process [16,17]. The characteristics of exosomes and their application in a variety of disease models have been widely recognized by researchers. Further, technologies for exosome extraction, gene modi cation induction, and transplantation are under continued development. Compared with traditional MSC transplantation therapy, cell-free therapy based on MSC-Exs transplantation has clear advantages but also some disadvantages. There remain many de ciencies in treatment using MSC-Ex transplantation [18,19]. (1) Transplanted MSC-Exs cannot self-proliferate or differentiate, so the effective duration and degree of e cacy after transplantation are dose dependent, and targeting ability is weak when the transplantation dose is insu cient. (2) The steps involved in separation and extraction of high-purity MSC-Exs are complex, and the exosomes extracted in each round are far from su cient to meet the requirements for an experiment.
Our TEM and NTA data indicated that there were no morphological differences between hucMSC-Exs and hypo-hucMSC-Exs in terms of size, shape, and electron density; however, further experiments demonstrated that hypoxic conditions promoted exosome release from MSCs.
After STZ injection, blood glucose levels were signi cantly higher in the DR, hucMSC-Exs, and hypo-hucMSC-Exs groups than those in the normal control group, while body weight was signi cantly lower in the DR, hucMSC-Exs, and hypo-hucMSC-Exs groups than normal controls. The differences in blood sugar and body weight among the DR, hucMSC-Exs, and hypo-hucMSC-Exs groups were not statistically signi cant. These ndings re ect that the number of exosomes delivered by intravitreal injection of hucMSC-Exs and hypo-hucMSC-Exs is small, and the number that can reach systemic circulation through the intraocular barrier is even smaller and insu cient to lower blood sugar.
We conducted a series of in vivo FFA assessments to test our hypothesis that exosomes from cells subject to hypoxia would be superior for treatment of DE. Intravitreal injection of hucMSC-Exs and hypo-hucMSC-Exs were used to treat DR model rats, and our results demonstrate that both hucMSC-Exs and hypo-hucMSC-Exs protected against retinal vascular in this context, and that these bene cial therapeutic effects were more apparent after treatment with hypo-hucMSC-Exs.
DR is a common microvascular complication of diabetes mellitus. Exosomes can play a pathogenic role in DR occurrence and development by promoting endothelial proliferation and angiogenesis. Monocyte-/platelet-derived exosomes are positively correlated with DR progression, particularly in patients with retinal vascular occlusion [20]. Exosomes can stimulate the coagulation cascade in retinal blood vessels, thereby aggravating microvascular damage. In addition, vitreous exosomes can accelerate proliferative DR progression by stimulating endothelial cell proliferation and promoting angiogenesis.
Exosomes secreted by different tissues and cells have various effects on angiogenesis, and exosomes play a key role in the balance between angiogenesis stimulation and inhibition. Intravitreal injection of hucMSC-Exs in rats can effectively reduce the expression of in ammatory markers and adhesion molecules, thereby reducing the in ammatory response in DR [21]. Similar to our ndings, Safwat et al. injected fat-derived MSC-Exs into STZ-induced diabetic rabbits via different routes (intravenous, subconjunctival, and intravitreal), and the results showed that both subconjunctival and intravitreal injection of adipose MSC-derived exosomes could protect from retinal tissue and structure damage, con rming the role of rabbit adipose MSC exosomes in retinal repair [22].
Although the protective effect of hypo-hucMSC-Exs in DR is considered a potential therapeutic strategy, its regulation and mechanism of action are not well elucidated. It remains unknown how long exosomes can survive after administration into the vitreous space. In a glaucoma model study, exosomes were delivered by weekly and monthly intravitreal injection, and the results con rmed that weekly administration of bone marrow-derived MSC-Exs could promote retinal ganglion cell survival through microRNA-dependent mechanisms with a better protective effect than monthly injections [23]. In our study, only a single injection was administered. At 4 and 8 weeks after injection, no obvious microvascular lesions were observed in the hypo-hucMSC-Exs group. After 12 weeks of intravitreal injection, the retinas in the hypo-hucMSC-Exs group showed changes characteristic of DR. In the hucMSC-Exs group, vascular lesions remained stable at 4 and 8 weeks after intravitreal injection, with only a small number of microaneurysms, local vascular dilatation, and isolated leaky lesions. Vascular lesions became worse 12 weeks after intravitreal injection, indicating that the protective effect of hypo-hucMSC-Exs on vascular injury was signi cantly reduced 2 months after intravitreal injection. Mead et al. used a glaucoma model to demonstrate that the neuroprotective effects of bone marrow-derived MSC-Exs were completely absent without a second injection after 1 month [23]. Whether repeated administration can enhance the protective effect of exosome hypo-hucMSC-Exs against microvascular injury in DR is unclear.
In conclusion, we established an STZ-induced DR model and used FFA to analyze the protective effects of hypo-hucMSC-Exs and hucMSC-Exs against microvascular injury. Our results suggest that hypoxia preconditioning was effective in improving the effect of hucMSC-Exs for the treatment of DR.