*Review* **Mesenchymal Stem Cell-Based Therapy for Retinal Degenerative Diseases: Experimental Models and Clinical Trials**

**Vladimir Holan 1,2,\*, Katerina Palacka 1,2 and Barbora Hermankova <sup>1</sup>**


**Abstract:** Retinal degenerative diseases, such as age-related macular degeneration, retinitis pigmentosa, diabetic retinopathy or glaucoma, represent the main causes of a decreased quality of vision or even blindness worldwide. However, despite considerable efforts, the treatment possibilities for these disorders remain very limited. A perspective is offered by cell therapy using mesenchymal stem cells (MSCs). These cells can be obtained from the bone marrow or adipose tissue of a particular patient, expanded in vitro and used as the autologous cells. MSCs possess potent immunoregulatory properties and can inhibit a harmful inflammatory reaction in the diseased retina. By the production of numerous growth and neurotrophic factors, they support the survival and growth of retinal cells. In addition, MSCs can protect retinal cells by antiapoptotic properties and could contribute to the regeneration of the diseased retina by their ability to differentiate into various cell types, including the cells of the retina. All of these properties indicate the potential of MSCs for the therapy of diseased retinas. This view is supported by the recent results of numerous experimental studies in different preclinical models. Here we provide an overview of the therapeutic properties of MSCs, and their use in experimental models of retinal diseases and in clinical trials.

**Keywords:** retinal degenerative diseases; mesenchymal stem cells; stem cell therapy; experimental models; clinical trials

#### **1. Introduction**

The retina is a highly specialized structure composed of several layers of morphologically and functionally different cell types. The individual layers are mutually interconnected and their primary function is to capture a light signal via the photoreceptors and to convert it into electrical impulses. These impulses are relayed to ganglion cells and then pass through the optic nerve into the visual cortex of the brain.

Individual retinal layers have an irreplaceable role in the capture and transduction of light signals. A disease or damage to any particular cell layer has a negative impact on the surrounding cell types and is reflected by the impairment of the vision. The progression of the retinal damage results in the development of retinal degenerative disorders. Although the exact etiology, causes and starting mechanisms of these diseases are mostly unknown, many factors, such as oxidative stress, light-induced damage, chemical insults, vascular defects, cytokine imbalance, damage of blood–retinal barrier and infiltration with immune cells or aging, have been suggested to contribute to the development of retinal degeneration [1–3]. Irrespectively of the different etiologies and various causes of retinal disorders, cumulative damage and loss of retinal cells, chronic inflammation, immune cell infiltration and enhanced cytokine secretion by immune and retinal cells represent the main pathological signs of retinal degenerative diseases, which represent the leading cause of blindness worldwide.

**Citation:** Holan, V.; Palacka, K.; Hermankova, B. Mesenchymal Stem Cell-Based Therapy for Retinal Degenerative Diseases: Experimental Models and Clinical Trials. *Cells* **2021**, *10*, 588. https://doi.org/10.3390/ cells10030588

Academic Editor: Maurice Ptito

Received: 3 February 2021 Accepted: 2 March 2021 Published: 7 March 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

#### **2. Retinal Degenerative Diseases**

Retinal diseases are a heterogenous and multifactorial group of light-threatening disorders, which include age-related macular degeneration (AMD), retinitis pigmentosa (RP), diabetic retinopathy (DR), pediatric Stargardt's macular dystrophy, glaucoma and many other similar forms. Although retinal diseases have various causes and different etiologies, a common characteristic is the death or dying of the specialized retinal cells and the loss of integrity of the retina or the degeneration of the photoreceptors; and this process then results in a visual impairment and ultimately in blindness.

In several of these disorders, including AMD, the earliest changes observed are caused by a loss of the cells of the retinal pigment epithelium (RPE), which play a major role in photoreceptor nutrition and in the maintenance of homeostasis. The degeneration of RPE and photoreceptors is the main cause of AMD which may have its onset in choroidal neovascularization or in accumulation of amorphous deposits. Both these causes lead to alterations in the retina and to the impairment of its functions. Early AMD is characterized by the appearance of soft drusen and pigmentary changes in the RPE, which can progress into two forms of advanced AMD—dry and wet AMD. Both of these forms result in the loss of central vision [1]. AMD is a leading cause of vision loss, affecting tens of millions of elderly people worldwide. Similarly, in DR, which is primarily caused by hyperglycemia in diabetes mellitus, the reduction in the number of pericytes at the vascular level, and the decreased number of retinal neurons and glial cells result in the interruption of retinal integrity and a progressive loss of vision. Nearly all patients with type I diabetes and over 60% of patients with type II diabetes have some degree of retinopathy after 20 years duration. DR is a leading cause of blindness in developed countries. RP is a genetic disorder of the eye which is caused by a progressive loss of the rod photoreceptor cells on the back of the eye [4]. Similarly, Stargardt's macular dystrophy is also an inherited retinal disease that begins in childhood or adolescence and that affects the macula. On the contrary, the main risk for a group of eye diseases called glaucoma is increasing age and high pressure in the eye. The mechanism of glaucoma is believed to be a slow exit of the aqueous humor through the trabecular meshwork, which results in damage to the optic nerve and causes vision loss. In addition to these degenerative processes associated with the loss of the specialized retinal cells, local inflammation significantly contributes to the triggering and development of retinal diseases. Among the main contributors to inflammation belong the various types of infiltrating immune cells and activated microglia. Increased numbers of glial cells have been observed in the retina with the degeneration of the photoreceptors [5]. Activated microglia can contribute to the production of proinflammatory factors and to the damage of the hematoretinal barrier [6]. Furthermore, different populations of glial cells expressing genes associated with AMD (such as *VEGFA* and *HTRA1*), have been identified in the retina with this type of disease [7]. Alternatively, it has been shown that the inhibition of the microglia delays retinal degeneration in the experimental retinal vein occlusion in mice [8].

To date, the treatment options for retinal diseases have been very limited. In the advanced stages, laser photocoagulation remains the main method of treatment for DR. Other therapeutic approaches are represented by a vitrectomy or different microsurgery interventions, which involve complicated surgery and are highly invasive procedures. Recently, less invasive treatment of some forms of retinal degenerative diseases has been based on the administration of inhibitors of vascular endothelial growth factor (VEGF) or other drugs. However, these inhibitors only induce short-term effects and just slow down the progression of the disease. Therefore, the need for a safe and less invasive approach to prevent development and to treat these sight-threatening manifestations of retinal diseases is vital.

#### **3. Perspectives of Cell Therapy for Retinal Diseases**

Since the loss of specialized retinal cells and local inflammatory reactions are the main causes contributing to the progression of retinal degenerative diseases, the inhibition of inflammation and a support for the surviving retinal cells appear to be prospective approaches to manage these diseases. Recent studies have indicated that various types of stem cells could contribute, by paracrine effects, to the support of the survival of the residual retinal cells, and to

the inhibition of inflammation [9]. A therapeutic possibility is offered by embryonic stem cells (ESCs), which can be isolated from blastocysts and which have a high differentiation potential. Another possibility is represented by the induced pluripotent stem cells (iPSCs), prepared by the reprogramming of normal adult fibroblasts or other cells. Both ESCs and iPSCs have the potential for differentiation into various retinal cell types [10,11]. However, the use of ESCs or iPSCs is limited by the possibility of immune rejection, teratogenicity and ethical restrictions in the case of ESCs. For these reasons, mesenchymal stem cells (MSCs) show great potential and could be a prospective tool for the treatment of retinal diseases. MSCs can be obtained from the bone marrow or adipose tissue of a particular patient and after separation and culturing in vitro could be used as autologous cells without the danger of immune rejection. It has been shown that after an injection of MSCs into the vitreous body, the cells can survive for a long period of time and can protect retinal ganglion cell survival or stimulate axon regeneration after optic nerve crush [12,13].

MSCs are multipotent stem cells which can be obtained relatively easily in a sufficient amount from various types of tissues and expanded in vitro for autologous application. It has been shown that MSCs retain their differentiation potential during their in vitro expansion, and that they can be differentiated into different cell types including cells expressing RPE or photoreceptor cell markers [14–16]. Similarly, the anti-inflammatory properties of MSCs [17,18] and their ability to support ocular surface healing [19–23] have been well documented. An advantage of MSCs is also their safety in use. The experiments in animal models confirmed that the subcutaneous administration of MSCs did not induce tumor growth during several months of observation [24]. Similarly, an extensive meta-analysis of studies using MSCs in over 1000 patients did not reveal a significant association between MSC treatment and the toxicity of infusions, internal organ infection, cancer or death [25].

#### **4. Mesenchymal Stem Cells**

#### *4.1. Characteristics of MSCs*

MSCs currently represent the most frequently studied type of adult stem cells. Originally, these cells were described as a population of bone marrow-derived cells that adhere to plastic and form fibrocyte-like colonies [26]. They have differentiation potential, which they retain during their in vitro expansion, as was demonstrated by their differentiation into other cell types of the mesenchymal cell line [27,28]. For therapeutic purposes, MSCs are mainly isolated from the bone marrow or adipose tissue. However, no specific marker that could characterize these cells has been identified. According to the International Society of Cellular Therapy, human MSCs are characterized by the ability to adhere to plastic surfaces in standard culture conditions, by being positive for the surface markers CD105, CD73 and CD90 and negative for hematopoietic markers CD45, CD34, CD14, CD19 and CD11b, and by their ability to differentiate into adipocytes, chondroblasts and osteoblasts [29]. It has been shown that MSCs possess potent immunmodulatory and anti-inflammatory properties, produce a number of cytokines and growth factors, and contribute to tissue healing and regeneration. The great advantage of these cells is their relatively easy isolation from the bone marrow or adipose tissue, good growth properties during their propagation in vitro and the possibility to use them as autologous (patient´s own) cells. It has also been demonstrated that MSCs from different sources (bone marrow, adipose tissue, umbilical cord blood, etc.) have similar function properties [30–32]. All these characteristics make them a promising candidate for the cell therapy of inflammatory and degenerative diseases.

#### *4.2. Immunoregulatory and Anti-Inflammatory Properties of MSCs*

The immunomodulatory properties of MSCs are mediated by multiple mechanisms including regulation by direct cell-to-cell contact, the production of various immunomodulatory molecules, the negative effects on antigen-presenting cells or the activation of regulatory T cells (Tregs). The complexity of the immunoregulatory effects of MSCs is also evident from the observation that MSCs inhibited lymphocyte proliferation induced by

mitogens and alloantigens by different mechanisms [33]. In general, MSCs have potent immunosuppressive properties. It has been shown that MSCs inhibit T and B cell proliferation, the production of cytokines and activity of cytotoxic T and NK cells [17,34]. In in vivo experimental models, the administration of MSCs prolonged the survival of skin allografts in baboons [35] and mice [36], prevented the rejection of corneal allografts [37,38], decreased the incidence of graft-versus-host disease in mice and humans [39,40] attenuated septic complications [41] and suppressed the incidence and severity of autoimmune diseases [42,43]. These suppressive effects of MSCs can be mediated by multiple mechanisms. It has been shown that MSCs express numerous molecules contributing to the immunosuppression, such as indoleamine 2,3-deoxygenase (IDO), cyclooxygenase-2 (Cox-2), TNF-α stimulated gene 6 protein (TSG-6), programmed death-ligand 1 (PDL-1) or Fas-L molecule [38,44–46]. Furthermore, MSCs produce a number of cytokines which can negatively influence in immune reaction. It has been shown that MSCs produce transforming growth factor-β (TGF-β) and interleukin-6 (IL-6) which are the principal cytokines regulating the development of anti-inflammatory Tregs and proinflammatory Th17 cells [47,48]. The spectrum of cytokines produced by MSCs depends on the state of their activation. We have demonstrated that a cytokine environment, where MSCs reside, considerably influences their secretory and immunoregulatory potential [49]. The beneficial effects of MSCs after their systemic application in vivo are supported by the demonstration of their ability to migrate to the site of injury or inflammation and to contribute to tissue healing and regeneration [50–52]. In this respect we showed that mouse bone marrow-derived MSCs (BM-MSCs) administered intravenously migrated in a significantly higher number to the injured eye than into the contralateral healthy eye [53], and that adipose tissuederived MSCs (A-MSCs) delivered intraperitoneally into transplanted mice were detected in a significantly higher amount in skin allografts than in healthy skin [54]. It has been suggested that the cytokines and chemokines produced by immune cells in the site of an injury attract MSCs to migrate to the damage site, where they participate in the attenuation of inflammation [55,56].

#### *4.3. Antiapoptotic Properties of MSCs*

Degenerative and inflammatory reactions in the diseased retina are regularly associated with a locally enhanced production of a variety of cytokines. These molecules can be produced either by inflammatory immune cells or by the activated cells of the retina [57]. It has been shown in vitro and in vivo that increased levels of proinflammatory cytokines can induce apoptosis of the surrounding cells [58,59]. Moreover, chronic inflammation is associated with endoplasmic reticulum stress, which also promotes the induction of apoptosis [60]. Furthermore, proinflammatory cytokines induce changes in the expression of various genes (such as *Bcl-2*, *Bax*, *p53*) associated with apoptosis. Any damage in the retina attracts the cells of the immune system which produce chemokines and cytokines, and thus potentiate inflammatory and apoptotic reactions. Therefore, the inhibition of a local inflammatory reaction and attenuation of apoptosis might be promising approaches to alleviate and inhibit the development of retinal injury. In this respect, MSCs by their immunoregulatory, anti-inflammatory and antiapoptotic properties could also be a promising therapeutic tool for developing retinal disorders [18,61]. We have recently shown that MSCs inhibit the expression of proapoptotic genes and decrease the number of apoptotic cells in the corneal explants cultured in the presence of apoptosis-inducing proinflammatory cytokines [62].

#### *4.4. The Production of Growth Factors by MSCs*

MSCs are potent producers of various growth and trophic factors. Some of these factors are produced by MSCs constitutively, while others are only secreted after activation with proinflammatory cytokines, mitogens or other signals. It has been suggested that the production of growth factors and their paracrine action are the main mechanisms of the therapeutic action of MSCs. Among the growth factors which are produced by

MSCs and that could contribute to retinal regeneration are hepatocyte growth factor (HGF), nerve growth factor (NGF), glial cell-derived neurotrophic factor (GDNF), insulin-like growth factor-1 (IGF-1), pigment epithelium growth factor (PEGF), fibrocyte growth factor (FGF), platelet-derived growth factor (PDGF), epidermal growth factor (EGF), angiopoietin-1, erythropoietin, VEGF and TGF-β [16,57,63–66]. Some of these factors are secreted spontaneously by untreated MSCs, and their production is enhanced after stimulation with proinflammatory cytokines [16]. On the contrary, the production of some other cytokines which are produced spontaneously (such as TGF-β, HGF) is significantly decreased in the presence of proinflammatory cytokines [16]. We also showed that higher levels of some growth factors are produced by MSCs after their differentiation into cells expressing retinal cell markers [16]. The expression of genes *Ngf*, *Gdnf* and *Il-6* was enhanced in differentiated MSCs, which suggests a higher potential of differentiated MSCs for the regeneration of diseased retinal tissue. It was demonstrated that the supernatants from light-injured retina significantly promote the secretion of neurotrophic factors by MSCs and slow down the process of apoptosis in damaged retinal cells [67]. Another study showed that the secretion of neurotrophic factors by MSCs promoted the viability of photoreceptors in vitro, and supported their survival after the subretinal transplantation of MSCs in a retinal degeneration model [68]. All these observations indicate that MSCs differentiated into cells with characteristics of retinal cells have a higher secretory activity than untreated MSCs, and could have a better regenerative potential than primary MSCs.

#### *4.5. The Ability of MSCs to Differentiate into Cells with Retinal Cell Characteristics*

One of the characteristics of stem cells is the ability to differentiate or even transdifferentiate into different cell types. With regards to differentiation into ocular cells, relatively extensive data exist about the differentiation of MSCs and other stem cells into cornea-like cells [69–72], but the data are less abundant on the differentiation of MSCs into neurons [73] or various types of retinal cells [74,75].

The ability of different types of stem cells and MSCs to differentiate into retinal cells has been reviewed by Salehi et al. [76]. For example, MSCs isolated from rat conjunctiva and cultured in the presence of taurine expressed markers characteristic of photoreceptors and bipolar cells [75]. Taurine, together with activin A and EGF, has been used in other studies to differentiate MSCs to photoreceptors. The cells cultured in differentiation conditions for 8–10 days expressed the *Rho* and *Rlbp* genes [74]. The same authors also showed that MSCs injected into the subretinal space are able to integrate into the retina and express markers specific for photoreceptors. Other studies demonstrated that the transplantation of MSCs into the damaged retina induced the expression of markers typical for photoreceptors, bipolar and amacrine cells in grafted MSCs [77–79]. Several other studies also showed the differentiation of MSCs into RPE cells [79,80], which play an important role in the nourishment of photoreceptors.

In our study, to simulate the environment of the damaged retina, we cultured mouse BM-MSCs with the retinal cell extract and with supernatant from Concanavalin A-stimulated mouse spleen cells. MSCs cultured for 7 days in such conditions differentiated to cells expressing retinal cell markers such as rhodopsin, S antigen, retinaldehyde binding protein, calbindin 2, recoverin and retinal pigment epithelium 65 [16]. Interferon-γ, present in the supernatant from activated spleen cells was identified as the main factor supporting the retinal differentiation of MSCs. In addition, the differentiated MSCs produced a number of neurotrophic factors which are important for retinal regeneration. This study, and the results of other authors [78–80], indicate that the signals from the damaged retina induce the differentiation of MSCs into cells expressing retinal cell markers, and that the MSC differentiation is supported by cytokines produced by activated immune cells [81,82].

#### *4.6. Additional Mechanisms Contributing to the Therapeutic Action of MSCs*

In addition to the ability of MSCs to produce several growth, immunoregulatory or neurotrophic factors, MSCs release various types of extracellular vesicles (EVs). These particles encapsulate different functional molecules which could support the survival of cells [83,84]. For example, it has been shown that intravitreally injected EVs were as effective as MSCs in improving vision in experimental model of retinal laser injury [85]. Similarly, Mead and Tomarev [86] showed that MSC-derived exosomes protected retinal ganglion cell function in a rat optic nerve crush model.

Furthermore, mitochondrial transfer has been described as additional mechanism which MSCs can use to support anti-inflammatory conditions and cell survival [87,88]. Since mitochondrial disfunction has been proved in many retinal diseases, the mitochondrial transfer therapy might have an impact on the treatment of retinal diseases [89].

Finally, the ability of MSCs to fuse with other cell types has been documented in various models [90,91]. Therefore, the possibility of the fusion of MSCs and the cells of diseased retina cannot be excluded, and should be considered as another mechanism contributing to the therapeutic action of intraocularly administered MSCs.

#### **5. The Potential of MSCs for the Treatment of Retinal Diseases**

Abundant experimental data demonstrate the beneficial therapeutic effects of MSCs on retinal diseases [92–94]. It has been shown that MSC transplantation significantly delays retinal degeneration, supports the regeneration of RPE, cone cells and axons, and improves the survival of retinal ganglion cells. On the basis of these encouraging results, the potential to use MSCs for the treatment of retinal diseases has been proposed and tested [95–99]. The main mechanisms of the therapeutic action of MSCs for the treatment of retinal diseases are shown in Figure 1.

**Figure 1.** The main mechanisms of the therapeutic effect of mesenchymal stem cells (MSCs) for retinal diseases. MSCs contribute to treatment of retinal disorders by multiple mechanisms involving the production of growth and neurotrophic factors, immunomodulatory actions, by antiapoptotic effect and by direct cell differentiation.

> Although several questions about the clinical use of MSCs still remain unanswered, for the purpose of great interest to use stem cells for the treatment of currently incurable retinal diseases, the first clinical trials using MSCs have been initiated [100,101]. However, before the introduction of the stem cell-based therapies into clinical practice, extensive research is needed to optimize the therapeutic procedures. For this reason, experimental animal models using pharmacologically induced degeneration of the individual retinal cell types or using animals with genetically induced retinal diseases, have been introduced. The pharmacological models use the application of sodium iodate (NaIO3) for the destruction of RPE cells that mimic progression of macular degeneration [102,103], or the application of *N*methyl D-aspartic acid (NMDA) or *N*-methyl-*N*-nitrosourea (MNU) that induces apoptosis and the selective degeneration of ganglion cells and photoreceptors that are processes

resembling hereditary RP or glaucoma [104,105]. The studies using these preclinical experimental models support the idea that the treatment of retinal diseases with stem cells could represent the most modern and prospective approach for the treatment of currently incurable severe retinal disorders, and to improve the patient´s quality of life [106–108].

#### **6. Possible Problems and Limitations Associated with MSC-Based Therapy**

Although stem cell therapy is safe, as shown in both animal studies and clinical trials [24,25], there are still several issues that have to be taken into account before final translation of MSC-therapy from preclinical models into clinical therapy.

First at all, there is a heterogeneity of individual MSC samples, based on differences in the cell source, isolation and culture procedure. MSCs are used at different time intervals after their isolation and different doses of cells are used. It has been documented that a longer cell culture duration has an impact on MSC morphology, secretory potential and migratory properties [109,110]. Therefore, there should be an agreement about the preparation of MSCs for individual types of application.

There is still controversy about in vivo survival of in vitro cultivated MSCs. Although MSCs are considered immune privileged cells which do not express costimulatory molecules and MHC Class II molecules, in the presence of some cytokines they can express these molecules and become a target for immune cells. Without respect to these observations, numerous studies suggested a long-term survival of allogeneic or even xenogeneic MSCs in immunocompetent recipients. In contrast to these studies, Eggenhofer et al. [111] claimed that in vitro cultured MSCs are extremely short-lived and do not survive in vivo. However, there is a possibility that in immunologically privileged sites, such as those in the eye, MSCs could survive.

Another unresolved issue is the fate and immunological functions of MSCs after their transfer into the inflammatory environment of the diseased retina. There is a possibility that immunosuppressive MSCs, transferred into an environment where there are proinflammatory cytokines present, can turn out into a cell population supporting the development of aggressive proinflammatory Th17 cells.

The route of MSC administration is also very important. After the intravenous injection of MSCs, only a small proportion of the delivered cells can be found in the eye (our preliminary observations). Therefore, for the treatment of the retina, the intraocular delivery of cells appears more effective. Using experimental models, especially based on small animals, the intravitreal injection of MSCs is the most common way. However, in the healthy eye, only a few cells can be detected in the vitreous cavity, and the occurrence of side effects after such delivery of MSCs was reported [112]. Therefore, other routes of MSC application have been tested. These approaches include subretinal application [113], suprachoroidal delivery [114] and subtenon injection [115].

#### **7. The Use of MSCs for the Treatment of Retinal Diseases in Experimental Models**

To study the mechanisms of retinal diseases and to validate new therapeutic approaches for these diseases, various experimental models resembling different types of retinal damage have been established and tested. These models are based on pharmacological interventions which induce the degeneration of specialized retinal cells, or utilize mutant animal strains, genetically modified recipients or various mechanical damages or injuries [116]. We review here the selected experimental models that have been used to test the therapeutic potential of the various types of MSCs.

#### *7.1. Experimental Models of AMD*

AMD is characterized by a progressive degeneration of the RPE and photoreceptors, and this process represents the major cause of visual impairment and irreversible blindness in the elderly population. Numerous experimental models have been established to study the individual steps of AMD progression. Transgenic experimental animal models provide systems to explore the cellular and molecular mechanisms of this disease. Some advantages

are offered by laser-induced models. Other approaches are based on the application of pharmacological agents inducing pathological changes in the retina. One of the wellestablished pharmacological models is based on the systemic or a local administration of NaIO3. NaIO<sup>3</sup> is a chemical which selectively induces the degeneration and death of RPE cells. It was shown in vitro that the exposure of human RPE cell line ARPE-19 to NaIO<sup>3</sup> induces the activation of inflammasome, changes the expression of molecules involved in the apoptosis, induces cell dysfunctions resembling conditions in AMD and finally causes RPE cell death [117,118]. In this model, human A-MSCs decreased the levels of mRNA for proapoptotic molecules and provided a rescue effect for ARPE-19 cells cultivated in the presence of NaIO<sup>3</sup> [118]. In our recent study we have observed that NaIO<sup>3</sup> increases the expression of genes for proinflammatory cytokines IL-1α and IL-6 or for proapoptotic Bax and p53 molecules in cultured mouse retinal explants. This increase was inhibited in the presence of mouse BM-MSCs, or by using a supernatant obtained after the cultivation of MSCs (Palacka et al., preliminary observations).

The intravenous or intraperitoneal application of NaIO<sup>3</sup> in vivo causes a rapid degeneration of the RPE cells and consequent damage to the outer nuclear layer. Increased levels of mRNA for Htra-1 a C3, the genes associated with the development of AMD, were detected in the retina of the NaIO3-treated recipients [103,119,120]. The intravitreal or subretinal application of NaIO<sup>3</sup> thus provides a suitable experimental model for study of the late phase of nonvascular AMD called geographic atrophy. The subretinal delivery of NaIO<sup>3</sup> in rats causes the formation of an atrophic area characterized by the degeneration of RPE cells and photoreceptors [121]. The intravitreal administration of NaIO<sup>3</sup> in rabbits after vitrectomy induced retinal atrophy and diffused outer retinal degeneration [122]. In these in vivo models, MSCs provided protection of the retinal cells from degeneration. The intravitreal injection of human A-MSCs in mice treated with NaIO<sup>3</sup> protected the RPE layer, photoreceptors and other nuclear cells from the damage [123]. Gong et al. [124] showed that rat BM-MSCs transplanted into the subretinal space can differentiate into cells expressing retinal markers, and can protect the retina in the experimental models of NaIO3-induced retinal damage. Since the RPE cells are the first damaged cell type in the progression of AMD, the protective effect of MSCs may be a promising option for the treatment of this condition.

#### *7.2. Experimental Models of DR*

DR represents a common complication of diabetes which is caused by hyperglycemia and by injury in retinal microvasculature and neurons. This disorder represents one of the leading causes of blindness globally. Despite the high prevalence of DR and extensive research, the treatment options for this disease are still strongly limited. Various experimental animal models have been established for the study of treatment possibilities. These models have been generated by a selective inbreeding or genetic modifications, the feeding of a galactose diet, or by a pharmacological induction using streptozotocin. This chemical selectively damages the β cells of the pancreas, increases blood glucose level and decreases the number of ganglion cells. To date, various animal models have been used to test the possibilities of treating DR by the application of stem cells [125]. In the majority of these models, the beneficial effects of a systemic or local application of MSCs were observed.

In models of streptozotocin-induced diabetes and DR, the application of MSCs had a positive effect on the retinal architecture. For example, Ezquer et al. [126] showed that the local application of mouse A-MSCs prevented the loss of retinal ganglion cells in diabetic mice. Levels of neurotrophic factors, such as NGF, GDNF and bFGF were increased in the eyes treated with A-MSCs. Although donor A-MSCs were found integrated into the host retina, these authors did not observe the differentiation of MSCs into retinal cells. In other studies, the intravitreal administration of MSCs obtained from the human umbilical cord attenuated capillary damage in streptozotocin-induced DR and increased levels of BDNF and NGF in the treated eyes. Donor MSCs also restored the visual function measured by ERG [94,127,128]. Yang et al. [92] showed that the administration of human A-MSCs

improved the integrity of the blood–retinal barrier and ameliorated DR in streptozotocin diabetic rats. Slightly enhanced levels of BDNF in the retina were also obtained after the transplantation of neural stem cells differentiated from umbilical cord MSCs, thus suggesting that this type of cells originated from MSCs may represent another suitable option for neuroprotection in DR [127]. Since MSCs isolated from mice with DR have lower proliferative abilities and higher levels of apoptosis compared to cells from healthy individuals [129,130], attempts were made to improve their therapeutic properties with the aim of using these cells for autologous transplantation in patients with DR. It has been shown that the treatment of BM-MSCs from mice with streptozotocin-induced diabetes with Wharton's jelly extract (containing a number of growth factors and other cytokines) significantly improved their proliferative abilities and therapeutic potential [130]. It suggests that the preconditioning of diabetic MSCs could improve their therapeutic properties.

In addition to models of pharmacologically induced diabetes, several studies used models of spontaneously or genetically induced diabetes. These models were described in detail by Robinson et al. [131] and Lai and Lo [132]. For example, the Akita (*Ins2Akita*) mice were created by a point mutation in the *insulin-2* gene and represent a spontaneous type-l diabetes model. It was shown that hyperglycemia in these mice causes neurodegenerative effects in the retina resulting in retinal thickness [133]. In addition, elevated levels of VEGF, PEGF and placental growth factor (PlGF) and an increased expression of Iba-1 (activated glial marker) and monocyte chemoattractant protein-1 (MCP-1) were observed in the neural retina and RPE layer in Ins2Akita mice during the progression of the diseases [133]. The therapeutic administration of human A-MSCs into the vitreal cavity of Ins2Akita mice improved vascular permeability and vision in this model of nonproliferative DR. Similar results were also obtained after the application of the conditioned medium from human A-MSCs which were pretreated with TNF-α and IFN-γ. The conditioned medium from A-MSCs also reduced the retinal expression of GFAP, the gene associated with neuroinflammation [134]. The experimental mouse model of proliferative DR was created by the mutation causing an overexpression of VEGFa (Akimba mice). In the retina of the Akimba mice, hemorrhage and neovascularization, the degeneration of photoreceptors, the activation of microglial cells and infiltration with monocytes and macrophages were detected. The inflammatory environment is manifested by a local increase in the expression of genes for IL-1β and IL-6 and by the upregulated activation of the NLRP3 inflammasome in the retina [135,136]. Locally administered A-MSCs obtained from mice without mutation in the Insulin 2 gene were mainly found in the perivascular space and improved the vascular density in the retina [129].

#### *7.3. Experimental Models for RP*

RP is a group of inherited neurodegenerative diseases characterized by a loss of photoreceptor cells, leading to visual impairment and eventually to blindness. The experimental models (natural and transgenic) of this disease are based on the use of spontaneous or genetically induced degeneration of the photoreceptors, and on the administration of chemicals inducing degeneration of the retinal cells [137–139]. A frequently used model for RP is the rd mouse with a mutation causing the early loss of the photoreceptors. For example, the rd1 mouse is characterized by mutation in the *PDE6b* gene which is, under physiological conditions, important for the signal transmission. In addition, it was shown that activated microglia with proinflammatory polarization occur in the rd1 retina [140]. It is also possible to use an rd10 mouse which has a spontaneous mutation in the *PDE* gene for rod-phosphodiesterase. This mutation causes the degeneration of photoreceptors and other retinal cell types [141]. Moreover, it has been shown that (as with rd1 mice) activated microglia can play a role in the development of a pathological condition [142]. Another example is an rd6 mouse carrying a mutation in the *Mrfp* gene, which is expressed in the RPE layer of cells [143]. In addition to mouse experimental models, the Royal College of Surgeons (RCS) rat is often used to study RP. The RCS rat carries a mutation in the *Merkt* gene, causing photoreceptor damage and an increase in microglial activation in the retina, resulting in inherited retinal degeneration [144,145]. Some of these models have been used to study the therapeutic effects of MSCs. Treatment with MSCs has supported the survival of photoreceptors and showed therapeutic benefits. For example, the application of MSCs to the eyes of rd1 and rd10 mice provided a rescue effect for retinal cells [146]. The administration of genetically modified MSCs with an overexpression of BDNF resulted in increased antiapoptotic signaling in the retina, and in a reduction in cell damage in the rd6 mouse [147]. Moreover, the donor cells preferentially integrated into the outer retinal layers. In addition, the combined transplantation of the human retinal progenitor cells and BM-MSCs into the subretinal space provided an effective immunomodulation in the eye of RCS rats and prevented pathological changes more effectively than with a single therapy [148]. Decreased levels of TNF-α and IL-1β and an increased expression of growth factors, such as BDNF and NGF, were observed in the treated eye.

Another approach to imitate the retinal degeneration observed in RP is based on the administration of *N*-methyl-*N*-nitrosourea (MNU). A single systemic administration of MNU causes retinal degeneration in various species [139]. Deng et al. [149] showed that the treatment of mice with MNU induces retinal degeneration that can be attenuated by the administration of MSCs, and that the therapeutic effect was decreased if MSCs were prepared from the aging mice with bone progeria.

Thus, as in the case of other experimental models of retinal degeneration, the positive therapeutic effects of MSC therapy were also demonstrated in RP models.

#### *7.4. Experimental Models for Glaucoma*

Glaucoma is a heterogenous group of eye diseases mainly caused by increased intraocular pressure and characterized by the progressive loss of retinal ganglion cells. So far, numerous experimental models have been established to study this disease. They include the intracameral injection of microbeads, laser photocoagulation, episcleral vein cauterization, the injection of hyaluronic acid and various models based on genetically modified rodents [150–153]. These models lead to increased intraocular pressure, the degeneration of retinal ganglion cells, the activation of glial cells in the retina and increased levels of inflammatory factors in the retina [154,155]. All these models, having their advantages and limitations, were used to study new therapeutic approaches involving MSC-based therapy. Various types of MSCs have been tested with a positive impact on the decrease of intraocular pressure and on the protection of the retina. For example, Mead et al. [156] showed that an intravitreal administration of A-MSCs, BM-MSCs or dental pulp stem cells decreased ocular pressure and offered a neuroprotective effect.

A trabecular meshwork regeneration observed after intraocular administration of BM-MSCs was described in a model of a laser-induced retinal damage model [157,158]. The enhanced neuroprotective effects were observed in these models using MSCs with an increased secretion of BDNF. The neuroprotective effect of A-MSCs was also described in a model of hyaluronic acid-induced glaucoma in rats [159]. Another therapeutic approach is provided by the application of BM-MSC-derived exosomes. It has been shown that the administration of the exosomes secreted by BM-MSCs promoted the survival of retinal ganglion cells and improved the retinal structure in the eye of rats after optic nerve crush injury, and in glaucoma models with ocular pressure induced by an intracameral injection of microbeads [86]. Similarly, the application of umbilical cord MSC-derived exosomes in a glaucoma model induced by an optic nerve crush injury in rats promoted retinal ganglion cell survival and glial cell activation [160]. Thus, MSC-derived exosomes injected into the vitreous provide a significant therapeutic benefit for glaucomatous eyes and for other types of retinal degenerative diseases.

Altogether, the animal models indicated positive therapeutic effects of MSC-based therapies in various types of retinal diseases. Selected experimental models where MSCs were used are summarized in Table 1.


**Table 1.** Selected experimental models of MSC-based therapy for retinal degenerative disorders.

#### **8. Clinical Trials Using MSCs for Retinal Diseases**

Currently, several clinical trials are in progress to test the potential of MSCs for the treatment of retinal degenerative diseases. Most of these studies are in the phase 1 or 2 focused on the safety of MSC application. The first finished studies showed that the administration of MSCs is not associated with serious complications [164,165]. Moreover, some studies have also showed an improvement in visual function based on the examination of in visual acuity, visual field and electroretinography. The effects of the treatment have been studied in patients with variety types of retinal diseases, such as AMD, DR, RP, glaucoma, inherited retinal dystrophy, optic nerve diseases or macular holes. MSCs obtained from various sources have been used for the treatments and BM-MSC or A-MSC were tested as an option for autologous transplantation. In addition, the effects of the administration

of the conditioned medium obtained from cultured MSCs or the application of exosomes prepared by the ultracentrifugation of a conditioned medium are also examined.

The application of MSCs appears safe and no serious treatment-related problems were observed in the eyes of patients with AMD, RP or retinal vascular occlusion six months after the administration of autologous BM-MSCs in the phase 1 testing. An improvement in visual function was also noted, but as the studies were designed to assess the safety of the treatment, it was not possible to definitely confirm whether this improvement was caused by the MSC application [164]. Similar results were observed in clinical trials after the application of autologous BM-MSCs in patients with RP [165]. There were no severe complications associated with cell transplantation in the treated eye. This conclusion is also supported by the study of Gu et al. [166] who showed that autologous BM-MSCs were beneficial in DR subjects with correction in macular thickness, and the improvement in visual acuity was also observed. In addition, the administration of autologous BM-MSCs improved visual acuity in patients with RP [167]. A similar improvement in visual function after the injection of autologous BM-MSC was observed in patients with optic nerve diseases and nonarteritic ischemic optic neuropathy [168,169]. In addition to the application of autologous BM-MSC, Wharton's jelly-derived MSC transplantation improved outer retinal thickness and visual acuity in patients with RP in phase 3 clinical study [115].

However, further clinical trials with a higher number of patients and a longer followup are still needed to evaluate the efficacy of MSC therapy. It will also be necessary to evaluate the benefits and advantages of autologous MSCs and the transplantation of stem cells obtained from another source, such as Wharton's jelly or the umbilical cord, or the use of a conditioned medium or exosomes. Selected clinical trials with MSCs for retinal degenerative disorders and their results are shown in Table 2.


**Table 2.** Selected examples of clinical trials using MSCs for retinal degenerative diseases.

#### **9. Conclusions**

Sight-threatening retinal degenerative diseases represent the main cause of visual impairment or even blindness worldwide. Despite great endeavors, there is still a lack of effective therapeutic approaches to stop or even cure these disorders. A prospective option has recently been offered by stem cell therapy. Experimental data from numerous animal models and using different types of stem cells offer promising results. Although the data from the experimental models are encouraging, numerous questions about the use of stem cell therapy have to be resolved to make this therapy more effective and safe [170]. Nevertheless, abundant clinical trials on the use of stem cells for retinal diseases have already been initiated [101,171,172]. These trials are focused on the study of the safety of the therapy, the selection of the optimal stem cells and their activation or modification prior to application, the optimalization of the dose of cells, the routes of application and the possibility of replacing the cells with their paracrine products.

One of the most important issues associated with MSC-based therapy, the safety of MSC administration, has been tested in numerous preclinical studies and clinical trials. Various experimental models and clinical studies using an intravitreal administration of MSCs demonstrated the safety of this therapy without any undesirable side effects [100,101]. Another issue that deserves attention is the mechanism of MSC therapy. Although some authors showed a long-time survival of therapeutically applied MSCs and demonstrated their presence in the eye even a few months after application, other studies suggested that MSCs are short-lived, do not survive in the recipients and can be only detected for a few days [111]. Moreover, in some experimental animal studies human MSCs were administered intravitreally and their biocompatibility, long survival and positive therapeutic effects were observed [173–175]. There arises a question about the mechanisms of this therapeutic effect across the interspecies barrier. Namely, Lohan et al. [176] demonstrated that human MSCs injected into rats do not have the same therapeutic effect as rat MSCs have, and that the immunoregulatory action of human MSCs is strongly limited by the interspecies barrier. These differences in the therapeutic effect between autologous/syngeneic and xenogeneic MSCs have to be taken into consideration, when human MSCs are applied therapeutically in rodent recipients and the knowledge from such experimental studies is translated to the clinical situation.

Finally, the immunoregulatory action of MSCs could strongly depend on the cytokine environment [49]. The inflammatory conditions in the diseased retina can significantly change the immunoregulatory properties of MSCs. We showed that unstimulated MSCs are immunosuppressive and spontaneously produce high levels of TGF-β, but not IL-6 [48]. TGF-β is a negative regulator of immunity and is also a factor determining the development of suppressive Tregs. However, in the presence of proinflammatory cytokines MSCs secrete, in addition to TGF-β, high levels of IL-6 [48,177]. The combination of TGF-β and IL-6 determines the development of proinflammatory Th17 cells [178]. Therefore, there is a danger that the application of MSCs into the inflammatory environment of the diseased retina could result in the inhibition of the immunosuppressive action of MSCs and in the preferential activation of proinflammatory Th17 cell population.

Although there are still many issues to be addressed before the final approval of MSC therapy for retinal degenerative diseases, the results obtained so far in preclinical animal models and in clinical trials are promising and encouraging. Therefore, the stem cellbased therapy offers a prospective option, especially for the patients without alternative therapeutic options. However, before the definitive expansion of the clinical use of MSCbased therapies, several questions, such as the sources of MSCs, the conditions of the in vitro propagation of MSCs, the routes of the cell applications or the possibility of the use of MSC products have to be answered and carefully verified.

**Author Contributions:** All authors participated equally in the preparation and writhing of the manuscript. All authors have also read and agreed to the published version of the manuscript.

**Funding:** This work was supported the grant No. 19-02290S from the Grant Agency of the Czech Republic and by the Charles University programs SVV 260435 and 20604315 PROGRES Q43.

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** Not Applicable.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


**Rahmeh Othman 1,2,\* , Gael Cagnone <sup>3</sup> , Jean-Sébastien Joyal <sup>3</sup> , Elvire Vaucher 1,\* and Réjean Couture 2,\***


**Abstract:** The kallikrein-kinin system (KKS) contributes to retinal inflammation and neovascularization, notably in diabetic retinopathy (DR) and neovascular age-related macular degeneration (AMD). Bradykinin type 1 (B1R) and type 2 (B2R) receptors are G-protein-coupled receptors that sense and mediate the effects of kinins. While B2R is constitutively expressed and regulates a plethora of physiological processes, B1R is almost undetectable under physiological conditions and contributes to pathological inflammation. Several KKS components (kininogens, tissue and plasma kallikreins, and kinin receptors) are overexpressed in human and animal models of retinal diseases, and their inhibition, particularly B1R, reduces inflammation and pathological neovascularization. In this review, we provide an overview of the KKS with emphasis on kinin receptors in the healthy retina and their detrimental roles in DR and AMD. We highlight the crosstalk between the KKS and the renin–angiotensin system (RAS), which is known to be detrimental in ocular pathologies. Targeting the KKS, particularly the B1R, is a promising therapy in retinal diseases, and B1R may represent an effector of the detrimental effects of RAS (Ang II-AT1R).

**Keywords:** kallikrein-kinin system; kinin receptors; diabetic retinopathy; age-related macular degeneration

### **1. Preface**

Ocular pathologies involving chronic inflammation of the retina are particularly devastating in terms of visual acuity. Among these, age-related macular degeneration (AMD) and diabetic retinopathy (DR) are the leading cause of severe vision loss in the elderly and active population of industrialized countries, respectively. In addition to the chronic inflammation, vascular dysfunction and neovascularization, which correspond to the formation of new pathological branches from pre-existing retinal or choroidal vessels, occur. The inflammatory process includes a breakdown of the blood–retinal barrier, leukocyte adhesion on the blood vessel wall, macrophage and microglial activation, and cytokine and chemokine production. Current treatments of these diseases are only compensatory and consist commonly of invasive treatments such as quarterly intravitreal (ITV) injections of anti-angiogenesis agents (anti-VEGF antibodies) or laser coagulation to prevent loss of sight due to aberrant neovascularization. Moreover, a large population of patients does not respond to anti-VEGF therapy. To offer alternative and comfortable treatment to nonresponders, such as a topical approach, our team's ongoing research effort has shown that the kallikrein-kinin system (KKS)—involved in inflammation—is overexpressed in the human AMD and DR retina and contributes to the development of pathological events in animal models of these diseases. Moreover, we were able to specifically target the KKS via topical ocular kinin B1 receptor (B1R) antagonist administration, which decreased

**Citation:** Othman, R.; Cagnone, G.; Joyal, J.-S.; Vaucher, E.; Couture, R. Kinins and Their Receptors as Potential Therapeutic Targets in Retinal Pathologies. *Cells* **2021**, *10*, 1913. https://doi.org/10.3390/ cells10081913

Academic Editors: Maurice Ptito and Joseph Bouskila

Received: 14 June 2021 Accepted: 24 July 2021 Published: 28 July 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

neovascularization and retinal inflammatory responses. The purpose of this review is thus to better describe the possible involvement of the KKS in retinal diseases and therapeutical approaches that can prevent deleterious events that lead to blindness.

#### **2. The kallikrein-kinin system**

The kallikrein-kinin system (KKS) is a complex multi-enzymatic and peptidergic system known to play a critical role in human physiology, but also in pain and inflammation [1–3]. Its physiological functions encompass nociception, cardiovascular and renal functions, vasomotricity, and host defense to infectious diseases [2,4]. The KKS is constituted by a panel of vasoactive peptides (kinins), synthesized and metabolized by different enzymes (kallikreins and kininases), and two G-protein-coupled receptors (GPCR) (Figure 1).

**Figure 1.** Biosynthesis and metabolism of kinins. Low- and high-molecular-weight kininogens are cleaved by tissue kallikrein and plasma kallikrein, respectively, into kallidin (in humans) and kallidin-like peptide (in rodents) [5,6], and bradykinin. Bradykinin, kallidin, and kallidin-like peptide are then either converted by the action of kininase I to des-Arg<sup>9</sup> -bradykinin, des-Arg10-kallidin, and des-Arg10-kallidin-like peptide, respectively, or inactivated by kininase II, neutral endopeptidase 24.11 (neprilysin, NEP), and the endothelin-converting enzyme [4,7–11]. ACE, angiotensin-1- converting enzyme (also known as kininase II); APP, aminopeptidase P; B1R, bradykinin type 1 receptor; B2R, bradykinin type 2 receptor; BK, bradykinin; ECE, endothelin-converting enzyme; KD, kallidin; KLP, kallidin-like peptide, which is Arg(1)-kallidin (Arg(0)-bradykinin); NEP, neutral endopeptidase; PK, plasma kallikrein; TK, tissue kallikrein.

#### *2.1. Kinins Generation*

Kinins are small peptides of 9–11 amino acids, including bradykinin (BK), kallidin (KD or Lys-BK), kallidin-like peptide (Arg(1)-kallidin (Arg(0)-bradykinin)), and T-kinin (Ile-Ser-BK; expressed exclusively in rats), which are generated from high-molecularweight kininogen (88 to 120 kDa) (HK) and low-molecular-weight kininogen (50 to 68 kDa) (LK) under the action of plasma kallikrein (PK) and tissue kallikrein (TK) [4–6]. BK generation in plasma takes part in the intrinsic coagulation pathway activation, involving the interaction of Factor XII (Hageman factor), prekallikrein (PPK), and Factor XI with HK on negatively charged surfaces, such as components of the extracellular matrix or other negatively charged particles (cholesterol sulfate, urate, or phospholipid acid), leading to prothrombotic and inflammatory effects [2,12]. Aminopeptidase P transforms KD and KLP into BK, while kininase I that includes carboxypeptidases N (CPN) and M (CPM) transforms BK, KD, and KLP into des-Arg<sup>9</sup> -BK, lys-des-Arg<sup>9</sup> -BK (or des-Arg10-KD), and des-Arg<sup>10</sup> - KLP, respectively. Alternatively, kininase II (also called angiotensin-1-converting enzyme (ACE)), neutral endopeptidase 24.11 (neprilysin, NEP), and the endothelin-converting enzyme (ECE) degrade BK, KD, and KLP into inactive fragments on the canonical B1 and B2 receptors [4,7–11]. Moreover, ACE and NEP can metabolize des-Arg<sup>9</sup> -BK, des-Arg10-KD, and des-Arg10-KLP into inactive metabolites. It is worth noting that the enzymes involved in the catabolism of kinins are also involved in the metabolism of other peptides belonging to other systems such as angiotensin, endothelin, anaphylatoxins C3a and C5a, substance P, neurotensin, enkephalins, atrial natriuretic peptides, and chemotactic peptide [5,8,13–20].

#### *2.2. Kinin Receptors*

The KKS operates through the activation of two GPCR, bradykinin type 1 (B1R), and type 2 (B2R) receptors. While BK, KD, and KLP are the endogenous agonists of B2R, their kininase I metabolites (deprived of C-terminal Arginine) are the preferred agonists of B1R [3,21]. The agonist selectivity of mouse B1R differs from human and rabbit B1R; des-Arg<sup>9</sup> -BK is the preferred B1R agonist in mice, while des-Arg10-KD displays much higher selectivity for human and rabbit B1R [22]. B2R can activate a plethora of signaling pathways either indirectly by interacting with guanine nucleotide-binding proteins, mainly with Gαq and less commonly with Gαs, Gαi, and Gα12/13, as reviewed in [23]; or directly by interacting with endothelial nitric oxide synthase (eNOS), phospholipase A 2 (PLA2), and tyrosine phosphatase (SHP2) [4]. B1R interacts with the same guanine nucleotide-binding proteins as those of B2R, but preferentially with Gαq to activate the phosphatidyl inositol-mitogen-activated protein kinase (MAPK) pathway, and with Gαi to activate the extracellular signal-regulated kinase (ERK)-nducible nitric oxide synthase (iNOS) pathway [4,23,24].

Most physiological effects of kinins are mediated by the constitutive B2R, since B1R is virtually absent in healthy tissues. BK is a potent endothelium-dependent vasodilator that has important cardiovascular and renal functions via the B2R [25]. Moreover, B2R contributes to the therapeutic effects of angiotensin-1-converting enzyme inhibitors (ACEI) and angiotensin AT1 receptor blockers [26]. These benefits derive primarily from its vasodilatory, antiproliferative, antihypertrophic, antifibrotic, antithrombic, and antioxidant properties [4,26–36]. However, it is worth noting that B2R can also contribute to inflammation. Indeed, uncontrolled production of kinins and excessive activation of B2R may lead to unwanted pro-inflammatory side effects as observed in angioedema, septic shock, stroke, hypertension, and Chagas vasculopathy, in which B2R antagonism is salutary [4,26,37–41].

B1R, however, is induced and upregulated during tissue injury involving the cytokine pathway, oxidative stress, and the transcriptional nuclear factor NF-κB [2,38,39,42–44]. The highly inducible character of B1R is often symptomatic of the occurrence of autoimmune, infectious, cardiac, kidney, and bowel inflammatory diseases [2,45–49]. However, B1R may play a compensatory role for the lack of B2R, and its upregulation during tissue damage may be a useful mechanism of host defense [25,50–52].

#### *2.3. Kinin Receptors in Inflammation and Neovascularization*

B1R antagonism or deletion plays a protective role in inflammation, organ damage, and lethal thrombosis in septic shock in diabetes [53]; lipopolysaccharide (LPS) mediated

acute renal inflammation [54]; renal ischemia-reperfusion injury [55]; and in cardiovascular [56] and retinal [57–60] inflammatory diseases. B1R inhibition reversed vascular [61] and retinal [58,60] inflammation induced by diabetes mellitus. Moreover, genetic deletion of B1R or administration of B1R antagonist in mice reduced pro-inflammatory mediators' expression and increased anti-inflammatory mediators [55]. Besides the well-described pro-inflammatory roles of both kinin receptors, an anti-inflammatory effect has been attributed to B2R. For instance, intramyocardial injections of tissue kallikrein reduced the expression of many inflammatory mediators through B2R activation [62]. Moreover, BK can counteract the inflammation in the brain [63]. Indeed, BK reduced LPS-induced TNF-α release from microglia activated by B1R [63]. Recently, a neuroprotective role for B2R was highlighted, and the use of B2R agonists was proposed as a possible therapeutic option for patients diagnosed with Alzheimer's disease [64]. Altogether, these findings support a dual role of B2R in inflammation, whereas B1R is mainly involved in the inflammatory responses, especially those triggered by cytokines or pathogens [65–67]. Because B1R is a potent activator of iNOS and NADPH oxidase, it is associated with vascular inflammation, increased vascular permeability, insulin resistance, endothelial dysfunction, and diabetic complications [24,43,44,68–70].

The contribution of kinin receptors to neovascularization has been widely studied in various models and diseases. In some vascular diseases, drugs are used to inhibit neovascularization (i.e., cancer, neovascular retinal pathologies, etc.), while in others such as ischemia, treatments aim to stimulate neovascularization. Therefore, both activation and inhibition of kinin receptors are important drug targets of vascular diseases. For instance, the activation of B1 and/or B2 receptors may be beneficial, notably in neovascularization and angiogenesis in diabetic mice, renal ischemia/reperfusion injury, diabetic nephropathy, and cerebral and heart ischemia [38,71–78]. B1R deletion or antagonism was shown to impair neovascularization, while B1R agonist had a positive outcome in a model of hindlimb ischemia in diabetic mice [77]. In the same model, B1R or B2R agonists administration induced revascularization by stimulating the mobilization of monocytes and proangiogenic CD34/VEGFR-2 mononuclear cells, and the infiltration of macrophages [76]. Moreover, B1R inhibition prevented the revascularization, as well as VEGF, eNOS, and basic fibroblast growth factor (FGF2) upregulation, induced by ACE inhibitor [79]. While the proangiogenic effect of ACE inhibitor was attributed to an increase in BK generation (Figure 1) and the activation of B2R in diabetic ischemia [80], B1R was more implicated than the B2R in ACE inhibitor mediated angiogenesis in Ang II type 1a receptor knockout (AT1aKO) mice after hindlimb ischemia [81]. Indeed, the B1R antagonist reversed the neovascularization and reduced VEGF-A and VEGFR-2 expression, while the B2R antagonist had less impact [81].

Cancer is among the diseases for which inhibiting kinin receptors would be beneficial. Indeed, the role of kinin receptors in promoting angiogenesis was supported by many experimental studies using cancer cells/tissues. For instance, B1R activation was shown to increase IL-4 and VEGF generation from human keratinocytes and to stimulate endothelial cell migration, thus promoting neovascularization [82]. Furthermore, when human endothelial cells were co-cultured with neuroblastoma cells, B1R and B2R expression was observed at the sites of interaction between these two cell types, regulating angiogenesis and tumorigenesis [83]. Interestingly, blockade of either B1R or B2R reduced tumor vascularization in vivo and significantly inhibited proliferation and migration of colorectal cancer cells in vitro [84]. In studies of mice bearing sarcoma 180 cells, it was suggested that BK promotes angiogenesis in the early phase of tumor development by increasing vascular permeability via B2R, expressed in the endothelial cells and not via B1R, and in the late phase by stimulating the upregulation of VEGF via B2R in the stromal fibroblasts [85–87]. BK was also found to increase VEGF expression in human prostate cancer cells and further promote tumor angiogenesis. Interestingly, B2R blockade using antagonists or genetic deletion reduced VEGF expression and abolished prostate cancer cell conditional mediummediated angiogenesis [88]. Altogether, these studies suggest that kinins play a pivotal role in angiogenesis through B1R, B2R, or both.

The dual beneficial and deleterious effects of kinin receptors raise questions about the therapeutic value of B1R/B2R agonists or antagonists in various diseases. Hence, the Janus face of kinin receptors needs to be seriously addressed in each pathological setting. The discovery of the expression of kinin receptors and other KKS components in the eye led many investigators to address their physiological and pathological roles, particularly in the retina.

#### **3. kallikrein-kinin system in the Eye**

Similarly to other organs, the KKS in the eye is a double-edged sword, as it contributes to many physiological processes including blood-flow regulation and vascular tone control, but also partakes in the complex processes of inflammation [4,57,89]. It was reported that the KKS underlies a number of ocular pathologies (DR, AMD, choroidal neovascularization, macular edema) associated with inflammation and pathological neovascularization, particularly in the human and rat retina [57–60,69,90–93]. For instance, PK and HK, by binding to the vascular endothelium, release BK and subsequently activate B2R, which plays a key role in the control of vascular tone [4]. However, in diabetic rats, an increase in PK mediates retinal vascular dysfunction and induces retinal thickening [91]. Moreover, tissue kallikrein (TK) was expressed in the human retina, cornea, and ciliary body [94]. TK does not seem to be implicated in retinal pathologies, particularly in diabetic retinopathy, as it was slightly detectable in vitreous fluids of patients with severe proliferative DR [95]. An expression of TK, B1R, and B2R was also reported at multiple tissue sites in the anterior portion of the human eye [96]. Nevertheless, B2R but not B1R was expressed in the control human retinae [93]. BK produces B2R-mediated vasodilatation of retinal vessels in control rats [97]. This response involves the COX-2 pathway, including prostacyclin [97]. Hence, B2R contributes to retinal blood flow control. On the other hand, the vasodilatation mediated by kinins is associated with B2R and B1R in streptozotocin (STZ)-diabetic rats and involves both NO and prostacyclin [97]. A protective compensatory role on retinal microcirculation was attributed to B1R at day 4 but not at 6 weeks following diabetes induction [98]. Likewise, both B1R and B2R contribute to the increased retinal vascular permeability in STZ-diabetic rats [58,60,99]. Collectively, these studies support the presence of the KKS throughout the eye and its ability to influence ocular function in health and disease.

It is still unclear whether the KKS expression is generated locally in the eye, or if it is a result of a systemic infiltration of KKS components. While the observation of some KKS components in the healthy eye [94] suggests a local production of these components, Phipps and Feener have suggested an infiltration of these components from the systemic circulation that could happen in DR [100]. This was explained by the increase of KKS components expression in the plasma of diabetic patients, and their infiltration in the retinal interstitium and vitreous that may occur following the increase in vascular permeability and hemorrhages in the retinal vessels [100]. Nonetheless, whether the origin of the KKS expression is local or a result of its infiltration from the systemic circulation in the eye, all these studies support an implication of the KKS in the pathogenesis and development of retinal diseases, such as DR and age-related macular degeneration (AMD).

#### **4. kallikrein-kinin system in Diabetic Retinopathy**

DR is one of the most common microvascular complications of diabetes, observed in up to 90% of patients with type 1 diabetes and 50 to 60% of patients with type 2 diabetes, despite a tight glycemic control [101–103]. If left untreated, DR can cause severe vision loss. Current therapeutic strategies target the advanced stages of the disease and aim to slow its progression without really reversing its outcome [104]. Among the current treatments for proliferative DR and macular edema are laser photocoagulation, vitrectomy, and intravitreal injections of corticosteroids or anti-VEGF that could prevent further vision loss [105]. However, the curative activity of these treatments is limited by side effects. For instance, pan-retinal photocoagulation can cause a loss of peripheral vision, color vision, and night vision [106]; intravitreal injection of anti-VEGF has a short effect duration and

can cause a tractional retinal detachment and endophthalmitis [107], and many patients are refractory to it [108–110]. Importantly, there is no effective treatment for the highly widespread early stages of the disease [111]. Thus, there is an urgent need for less-invasive and more-effective therapeutic strategies.

#### *Kallikreins and Kinin Receptors in Diabetic Retinopathy*

A decrease in the concentration of kallikrein-binding protein (KBP), a serine protease that binds to tissue kallikrein and inhibits kallikrein activity, was reported in the vitreous humor of patients with proliferative DR [112]. Parallel to this study, the levels of KBP were reported to be decreased by 60% for at least 4 months in the retina of STZ-diabetic rats [113]. Moreover, tissue kallikrein was significantly elevated in vitreous fluid in proliferative DR patients when compared with control patients [114]. Interestingly, intravitreal injection of kallistatin, a tissue kallikrein inhibitor, in STZ-diabetic rats reduced retinal neovascularization; however, these effects have been attributed to the tissue kallikrein effects on the VEGF system [115]. Other components of the plasma KKS, including PK, FXII, and HK were also found in the vitreous fluid of patients with advanced DR [116,117]. Increased levels of PK and PK activity were observed in the retina of diabetic rats compared with nondiabetic controls [90,91]. PK injection increased vascular permeability in the healthy retina, and further in the diabetic retina, yet these effects were reversed by the inhibition of PK [91]. Furthermore, the retinal thickening, as well as the increase in vascular permeability caused by intravitreal injections of VEGF, were reduced (by 47% and 68%, respectively) in plasma prekallikrein knockout mice [118]. In phase I.B of a recent clinical trial, PK inhibition by one-time intravitreal injection of KVD001 was shown to be effective in treating macular edema without creating a safety concern. The injection improved visual acuity and central retinal thickness, and no exacerbation of the severity of DR was observed [119]. PK contribution to DR pathogenesis was, however, attributed to B2R activation. Indeed, C1 inhibitor-deficient mice caused vasogenic edema due to increases in PK expression, BK synthesis, and activation of B2R [120]. Given the fact that PK is a constitutive enzyme involved in other systems, including thrombosis and blood hemostasis, its inhibition may risk interfering with its physiological role [57].

Alternatively, B1R expression was shown to be significantly increased in retinae of rats and humans affected by type 1 and type 2 diabetes [57,58,60,69,92]. B1R expression was enhanced on the 4th day of STZ-diabetic retina [97], and it remained upregulated even 6 months after the induction of type 1 diabetes [58]. B1R upregulation in STZ-diabetic rats leads to retinal microvessel vasodilation [97], vascular hyperpermeability, and inflammation [60]. Importantly, these responses were reversed by eye-drop application of B1R antagonists (LF22-0542 and R-954) [58,60]. B1R was strongly expressed in vascular endothelial cells and in the retinal pigment epithelium of human and rats' retinae, suggesting its implication in altering the integrity of the internal and external blood–retinal barrier (BRB) in DR and AMD [58,59,69,92,93]. B1R might disrupt the BRB [59,60,69] either by the suppression of tight junction components (occludin, claudin, and zonula occludens-1), or by a rearrangement of the cytoskeleton filamentous actin [121]. In human cerebral microvascular endothelial cells, B1R agonist (des-Arg<sup>9</sup> -BK) was shown to decrease the expression of zonula occludens-1 and occludin in vitro [122]. Altogether, these data support a strong implication of B1R in DR.

One additional mechanism by which B1R contributes to the pathogenesis of DR has been recently suggested that involves the activation of the iNOS pathway [69]. In HEK293 cells, it was shown that B1R associated with Gαi can activate iNOS through ERK [24,123], thereby producing sustained amounts of NO. Interestingly, iNOS inhibition in the retina of diabetic mice caused a decrease in occludin and zonula occludens-1 expression, thus protecting the dissociation of BRB [124]. The elevated concentrations of NO, nitration of proteins, prostaglandin E2, superoxide, leukostasis, and retinal thickness induced by diabetes were significantly inhibited in diabetic iNOS (−/−) mice [125]. In addition, diabetes-induced acellular capillaries and pericyte ghosts were significantly inhibited in

diabetic iNos (−/−) mice [125]. Given that B1R can also enhance the production of superoxide anion through PLC and the activation of NADPH oxidase [44], NO produced by iNOS upon B1R activation can react with superoxide anion to yield peroxynitrite, a highly toxic molecule [126–128], which causes endothelial and neuronal cell apoptosis, neuronal degeneration, and BRB breakdown in DR [125,126,129–133]. Peroxynitrite can also activate NF-κB, and thereby can increase the expression of several pro-inflammatory mediators, including B1R [4,43]. Hence, B1R activation can further amplify and perpetuate the inflammatory response, as well as the oxidative stress, through a positive feedback loop [43,69] In resonance with this hypothesis, pharmacological iNOS inhibition in the retina of STZ-diabetic rats reversed peroxynitrite formation, the upregulation of inflammatory mediators (notably B1R), and the enhanced vascular hyperpermeability induced by B1R agonist [69]. Collectively, these data support a robust implication of the B1R in DR, mainly by increasing and perpetuating inflammation and oxidative stress. Hence, targeting B1R represents a promising therapeutic approach in DR, and deserves further investigation.

In the retina of 2-week STZ-diabetic rats, B2R mRNA and protein expression did not change when compared to the control retina [69], yet a significant increase in B2R mRNA was observed at 24 weeks in the retina of diabetic rats [58]. B2R contributes to the increased retinal vascular permeability in STZ-diabetic rats [99]. Indeed, BK induces vascular endothelial cadherin phosphorylation and a subsequent rapid internalization and ubiquitination, leading to an opening of endothelial cell junctions and plasma leakage [134] (Figure 2). However, more studies are needed in DR using recently developed selective and stable B2R antagonists or biostable kinin analogs [41,46,135].

**Figure 2.** kallikrein-kinin system in diabetic retinopathy. Schematic proposal of the signaling pathways activated by B1R and B2R in diabetic retinopathy. PGs, prostaglandins; PGI2, prostacyclin; PLA2, phospholipase A2; Src, kinase proto-oncogene tyrosine-protein kinase; VEC, vascular endothelial cadherin; VEC-P, phosphorylated vascular endothelial cadherin. The human eye anatomy diagram was acquired from Shutterstock (http://www.shutterstock.com, accessed on 16 July 2021).

#### **5. kallikrein-kinin system in Age-Related Macular Degeneration**

AMD is a multifactorial disorder, highly heritable, and caused by an interplay of many factors including age, genetic, and environmental risk factors. The prevalence of AMD is

α

β α

rising worldwide, and it is expected to increase from 196 million in 2020 to 288 million by 2040 [136,137]. In its early stages, AMD is characterized by pigmentary abnormalities and deposits of lipoproteinaceous debris (soft drusen) between the basal lamina of the retinal pigment epithelium (RPE) and the inner collagenous layer of Bruch's membrane (BM) of the central retina [138]. Early and late forms of AMD include wet (exudative) AMD and dry (nonexudative) AMD. The late form of dry AMD is also called geographic atrophy AMD. Exudative/wet AMD is mainly characterized by neovascularization that arises from the choroid, but in about 10–15% of the cases, originates from the retinal vasculature in the subretinal space [139,140]. Dry AMD is more prevalent, affecting 85–90% of patients suffering from AMD [136,137], and is characterized by an extending lesion of the RPE and photoreceptors [141]. Current treatments target only the neovascular AMD mainly by anti-angiogenic therapy (anti-VEGF), which aims to decrease the vascular permeability and to inhibit the formation of new vessels without treating the degenerative processes and the vision loss in 30% of patients that occur in the long term [142]. On the other hand, no effective treatment options are available for dry AMD, besides lifestyle modification and nutrient supplementation [143].

Similarly to DR, the pathogenesis of AMD is driven by both inflammation and microvascular alterations leading to BRB dysfunction and pathological neovascularization. Indeed, an increase of diverse transcriptional factors (NF-kB, HIF-1α) and pro-inflammatory mediators (cyclooxygenase-2 (COX-2) products, IL-1β, TNF-α, iNOS, NO) has been reported in different models of DR and AMD [58–60,69,144]. Consistently with the roles of kinin receptors in both inflammation and neovascularization, we showed that most upregulated inflammatory mediators were blocked by B1R inhibition in DR and AMD [58–60]. B1R was shown to be expressed on Müller cells and astrocytes in these retinal pathologies in rat and post-mortem human retina [58,59,93], and on microglia in post-mortem human wet AMD retina [93]. Macroglia play a primary role in vascular function and neuronal integrity of the retina [145]. These results deserve closer scrutiny and encourage further investigations to assess the impact of an ocular treatment with a B1R antagonist on macroand microglial reactivity in DR and AMD.

B1R expression was upregulated in a rat model of choroidal neovascularization (CNV), and B1R blockade reduced the size of the neovascularization [59]. B1R contribution to retinal neovascularization in humans was also suggested in a recent study in post-mortem human wet AMD retinae. In these retinae, B1R was strongly expressed in endothelial/vascular smooth muscle cells, and co-localized with iNOS and fibrosis markers. Its presence on vascular smooth muscle cells can induce prolonged vessel constriction and consequently contribute to retinal ischemia, a main trigger of neovascularization, mainly by activating the VEGF-A pathway [146]. Altogether, these data highlight a contribution of B1R to retinal pathologies associated with neovascularization. By analogy with another ocular pathology, B1R agonist administration in the rabbit eye induced corneal neovascularization, an effect that was reversed by B1R inhibition with the same efficacy as VEGF-A inhibition [147]. The implication of B2R in ocular neovascular pathologies has also been suggested. For instance, in an ischemic retinopathy model, B2R antagonist (Fasitibant) significantly decreased the expression of VEGF and FGF2, as well as pathological retinal neovascularization [148]. In a mice model of CNV, B2R blockade with Icatibant had a limited effect, yet concomitant inhibition of B2R and kininase II had additive suppression of the CNV size [149]. We reported no significant modification of B2R mRNA and protein expression in human neovascular AMD retinae [93].

In addition to KKS gene expression in the ocular pathologies reviewed above, we also mined a recent public single-cell transcriptomics database of post-mortem choroid tissues from neovascular AMD human patients [150], using previously described analyses [151,152]. KKS genes were detected in fibroblasts and immune, RPE, and endothelial cells (Figure 3, unpublished original results). Choroidal endothelial cell specifically expressed *KLKB1*, *BDKRB1,* and *BDKRB2* (genes for prekallikrein, B1R, and B2R, respectively), albeit at low expression levels. Subclustering of the heterogenous choroidal endothelial cell popula-

tion identified four subtypes (see legend of Figure 3), including vein clusters 1 and 2, discriminated by the higher expression of *SELECTIN E* and *VCAM1* (Figure 3f), a pattern reminiscent of post-capillary venous identity [153]. Interestingly, vein cluster 2 showed greater expression of KKS genes, notably *BDKRB1*, *BDKRB2,* and *MME* (genes for B1R, B2R, and neprilysin, respectively) in choroid endothelial cells from AMD patients (Figure 3g). Although the relatively low detection levels for these three genes (less than 10%) requires cautious interpretation, their specific expression in post-capillary venous endothelial cells of neovascular AMD patients is intriguing and warrants further investigation of kinin receptors in AMD.

**Figure 3.** Transcriptomic impact of age-related macular degeneration on the kallikrein-kinin system

in human choroid tissues by single-cell RNA seq. (**a**) Dimensionality reduction and cluster visualization with UMAP plot. Color-coded clusters represent the different choroid cell types (see legend in bottom right corner) identified by single-cell RNAseq analysis of post-mortem control (left panel) and neovascular AMD (right panel) choroids (public data deposited on GEO with reference number GSE135922). (**b**) Dotplot of the expression of the gene markers used to identify choroid cell types. (**c**) Dotplot of the expression of the genes involved in the KSS pathway across choroidal cell types from control (black legend) and neovascular (nv)-AMD (red legend) samples. As shown in (**a**–**c**), KLKB1, BDKRB1, and BDKRB2 are mainly expressed in choroidal endothelial cells, albeit at low expression levels; some expression was also detected in fibroblasts. (**d**) Dimensionality reduction and cluster visualization with UMAP plot of the subpopulations of choroid endothelial cells of postmortem control (left panel) and neovascular AMD (right panel) choroids. Choroid cells clustered into four distinct endothelial cell subtypes: two venous subtypes, one choriocapillaris subtype, and one arterial subtype (see legend on right side). (**e**) Dotplot of the expression of the specific gene markers in these endothelial subcluster, as annotated by Voigt et al., [150]. (**f**) Visualization with UMAP plot of E SELECTIN (SELE) and VCAM1 expression co-localizing to vein 2 subcluster, a signature reminiscent of post-capillary venous identity. (**g**) Dotplot of the expression of genes involved in the KSS pathway across choroidal endothelial cell subtypes from control (black legend) and neovascular (nv)-AMD (red legend) samples. Vein cluster 2 showed greater expression of KKS genes, notably BDKRB1, BDKRB2, and MME, across all choroidal endothelial cells of control and nv-AMD choroid samples. In all the dotplots, the size of the dots encodes the percentage of cells within a class, and the color scale encodes the average expression level across all cells within a class (red being the strongest value). Av. Exp., average gene expression across all cells within each cluster; % Exp., percentage of cells with detectable gene expression within each cluster.

Although recent studies support the implication of the KKS in wet neovascular AMD, it is still not clear if the KKS is implicated in the dry form. In the retina of aged rats, an increase of KKS components was demonstrated, where 4-month-old rats showed a significant decrease in KBP, and consequently an increase in tissue kallikrein compared to 2-week-old rats [113]. Recent data using post-mortem human retinal sections showed only a weak expression of B1R and no changes of B2R in dry AMD [93].

#### **6. kallikrein-kinin system in Other Retinal Damage**

This review highlights the implication of the KKS in retinal pathologies associated with inflammation and neovascularization. However, KKS can also be implicated in ocular pathologies such as glaucoma and ocular ischemia. For instance, BK alters the shape of cells in both bovine and human trabecular meshwork [154–156]. Moreover, FR-190997, a B2R agonist, was shown to lower the intraocular pressure by promoting uveoslceral outflow in monkeys [157]. Taken together, these results suggest that the KKS can also be implicated in ocular diseases with elevated intraocular pressure. Intravenous administration of TK protected against retinal ischemic damage in a retinal ischemia/reperfusion model in mice [158]. In this model, TK administration inhibited retinal ganglion cell death, counteracted the retinal permeability induced by ischemia, and improved the visual function [158]. However, these protective effects seem to be independent of blood flow and might be mediated by eNOS activation and subsequent NF-κB silencing.

#### **7. Crosstalk between the kallikrein-kinin system and the Renin–Angiotensin System (RAS) in Ocular Pathologies**

There is compelling evidence for a local renin–angiotensin system (RAS) within the human eye that is activated in ocular disorders and DR [159–163]. Multiple interactions (crosstalk) exist between the RAS and the KKS [4,25,164–166] (Figure 4). In addition to the implication of ACE (kininase II) in the degradation of kinins (acting on B1R and B2R) and the formation of angiotensin II (Ang II) from angiotensin I (Ang I) [165], the activation of the angiotensin II type 2 receptor (AT2R) leads to BK generation, which promotes vasodilation through the NO/cGMP system [167]. Under the action of angiotensin-converting enzyme

2 (ACE 2), Ang I is cleaved into angiotensin-(1-9) this is corrected (Ang-(1-9)), a peptide that elicits vasodilation and anti-inflammatory effects through activation of AT2R [168,169]. ACE 2 can also cleave Ang II to Ang-(1-7), an agonist of AT2R and Mas-receptor (MasR) that elicits the release of BK, vasodilatory, antiproliferative, anticoagulation, anti-inflammatory, and antifibrotic activity, thus counterbalancing the adverse effects of Ang II mediated by AT1R [170–172]. Importantly, ACE 2 hydrolyses B1R agonists (des-Arg<sup>9</sup> -BK and Lys-des-Arg<sup>9</sup> -BK) into inactive metabolites and therefore impairment of ACE 2 (as under COVID-19 infection) is expected to enhance the pro-inflammatory effects of the des-Arg<sup>9</sup> -BK/B1R axis [164,173] (Figure 4). Moreover, the pro-inflammatory effects of Ang II was attributed to AT1R and B1R activation [174]. Following AT1R activation, Ang II enhances B1R expression in vitro [175,176] and in vivo [174,177] by activating NADPH oxidase, IL-1β, IL-6, TNFα, and NF-κB [174,176]. Besides ACE2, neutral endopeptidase 24.11 (NEP) was described to be biochemically capable of producing Ang-(1-7) from Ang I and Ang-(1-9) [178]. NEP can also hydrolyze Ang-(1-7) to form angiotensin-(1-4) (Ang-(1-4)), an inactive metabolite [179]. Hence the reciprocal interaction between the RAS and the KKS must be considered in the development of novel therapeutic approaches in the treatment of retinal diseases.

**Figure 4.** Crosstalk between the kallikrein-kinin system (KKS) and the renin–angiotensin system (RAS). ACE, angiotensin -1 -converting enzyme (kininase II); ACE2, angiotensin-converting enzyme 2; Ang I, angiotensin I; Ang II, angiotensin II; AT1R, angiotensin type 1 receptor; AT2R, angiotensin type 2 receptor; BK, bradykinin; MasR, Mas receptor; NEP, neutral endopeptidase 24.11 (neprilysin, NEP).

#### *7.1. Renin–Angiotensin System in Diabetic Retinopathy*

The RAS is implicated in inflammation, vascular alterations, neovascularization, and edema in retinal pathologies, notably in DR and retinopathy of prematurity [163]. An increase in prorenin level was reported in the vitreous fluid of patients with proliferative DR [161]. Ang II induces pericyte apoptosis in the retina in vivo and in vitro in hypertensive rats by increasing the expression of RAGE receptor for advanced glycation end products (AGEs); these effects were reversed by an Ang II-AT1R blocker [180]. An AngII-AT1R blocker (Candesartan) inhibits the development of DR by reducing the accumulation of AGEs and the expression of VEGF in the retina in a rat model of type 2 diabetes [181]. This AT1R blocker reduces retinal vascular permeability induced by diabetes and Ang II in rats [182]. Importantly, the DIRECT study based on more than 1400 patients found that Candesartan reduces the progression of microaneurysms in both type 1 and type 2 diabetic patients, yet no effects were observed on the DR regression and progression, or on the prevention of diabetic macular edema risk [183]. Another multicenter study of 285 patients with type 1 diabetes reported that Losartan, another AT1R blocker, slows the progression of DR [184]. Together, these studies suggest that angiotensin AT1R blockers may be effective against DR independently of their anti-hypertensive action.

β

ACE inhibition was also shown to lower the risk and prevent the development and the evolution of DR in humans [185,186]. ACE inhibition reduces retinal VEGF overexpression and hyperpermeability in experimental diabetes [187] and vitreous VEGF concentrations in patients with proliferative DR [188]. Interestingly, changes in circulating VEGF do not account for the beneficial effect of ACE inhibition on retinopathy in patients with type 1 diabetes [189]. Previous clinical trials have associated the decrease in DR progression in type 1 or type 2 diabetic patients with a reduction of hypertension [190,191]. The United Kingdom Prospective Diabetes Study (UKPDS) with more than 1000 patients reported a reduction in the progression of DR with ACE inhibitor and β1-adrenergic receptor blocker, suggesting that the beneficial effect may be related to the anti-hypertensive and not to the ACE-inhibition-specific effect [190]. Nevertheless, other studies have reported a slowdown in DR progression in normotensive diabetic patients taking an ACE inhibitor, suggesting a possible therapeutic effect of ACE inhibitors not related to the anti-hypertensive effect [184,191]. In resonance with this, a meta-analysis of 21 clinical trials with more than 13,000 patients disclosed no effects of RAS inhibitors on DR progression in hypertensive patients, but a reduced risk of DR, and increased possibility of DR regression in normotensive patients [186]. In rank order of anti-hypertensive drug classes, the association with risk of DR progression was lowest with ACE inhibitors, followed by Ang II-AT1R blockers, β-blockers, and finally with calcium-channel blockers [186].

While ACE inhibitors show promising results against DR and diabetic macular edema, several safety questions related to increased kinin levels can be raised, such as hypotension, angioedema, and pain associated with inflammation, which are B2R-mediated [192–195]. Increased kinin levels are also associated with retinal vascular permeability, inflammation, and neovascularization (Figure 2). A decrease in the degradation of endogenous B1R agonist (des-Arg<sup>9</sup> -BK) was also observed in the plasma of patients treated with an ACE inhibitor [196]. In theory, the use of kinin receptor antagonists can overcome the side effects of ACE inhibitors in the retina.

Furthermore, Ang II-AT1R is a potent enhancer of the pro-inflammatory B1R [174–177] and ACE inhibition ablated B1R expression in diabetic vessels [197], suggesting that B1R acts as an effector of the RAS (Figure 4). Therefore, targeting the RAS (AT1R and ACE) in DR may be a promising approach to prevent the induction and deleterious effects of B1R. Nonetheless, further studies are needed to unveil the exact mechanism(s) and crosstalk with other components of the RAS/KKS (ACE2, AT2R, and MasR) to address the beneficial versus the detrimental effects of the dual pro- and anti-inflammatory role of B2R in retinal disorders. Until these questions are fully answered, targeting B1R in retinal pathologies associated with inflammation and/or vascular alterations remains by far the best asset, with less possible interaction with other axes involved in physiological signaling pathways.

#### *7.2. Renin–Angiotensin System in Age-Related Macular Degeneration*

The implication of RAS was also reported to contribute to CNV pathogenesis. Indeed, our single-cell RNA seq showed a high expression of ACE in the neovascular AMD arteries and choriocapillaries (Figure 3). Furthermore, prorenin receptor blockade in a murine model of laser-induced CNV exhibited a significant reduction of CNV, macrophage infiltration, and the upregulation of ICAM-1, monocyte chemotactic protein-1, (MCP-1), VEGF, VEGFR1, and VEGFR2 [198]. Moreover, AT1R inhibition pharmacologically or genetically inhibited CNV and macrophage infiltration [198]. VEGF, ICAM-1, and MCP-1 levels, elevated by CNV induction, were significantly suppressed by ACE inhibition, which led to significant suppression of CNV development to the level seen in AT1R-deficient mice [149]. Despite these significant beneficial effects in rodents, antihypertensive drugs (ACE inhibitors and angiotensin receptor blockers) failed to show any positive effects on AMD in humans [199–201].

#### **8. Conclusions**

Inflammatory and neovascular retinal diseases, including DR and AMD, can lead to severe vision loss if left untreated. Current treatments for these pathologies are invasive and can sometimes worsen the pathology. Besides these side effects, many patients do not respond well or become refractory to these treatments, thus there is an urgent need to identify new therapeutic targets and new treatment strategies. Interestingly, the pro-angiogenic, pro-inflammatory, and vasoactive effects of the KKS make it a promising therapeutic target for treating retinal pathologies associated with inflammation and neovascularization. However, KKS targeting needs to be carefully documented before clinical application, as this system is also involved in physiological functions (such as organ blood-flow perfusion and blood coagulation) [4]. To minimize as much as possible the side effects of a complete shutdown of this system that may lead to ischemia and thrombotic events, it is advisable to use a more selective approach by targeting directly kinin receptors in retinal pathologies. Conflicting data are available regarding the implication of B2R in retinal pathologies. This may be related to its important physiological role on the vasculature and the regulation of blood flow. Thus, the inhibition of this receptor may cause unwanted side effects, notably ischemia, and its role in retinal pathology warrants further investigation. In contrast, currently available data strongly support the contribution of B1R in inflammatory and neovascular retinal diseases. Inhibiting the inducible B1R, by topical eye-drop treatment represents a promising noninvasive therapeutic approach in retinal diseases. This is keeping with the finding that B1R acts as an effector of the RAS (Ang II-AT1R) and may subserve its deleterious effects in ocular diseases.

**Author Contributions:** Writing—original draft and designing figures, R.O.; analysis of scRNAseq (Figure 3), G.C. and J.-S.J.; editing the manuscript, R.O., G.C., J.-S.J., E.V., and R.C. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by the Canadian Institutes of Health Research (PJT-175061) and the FRQS Vision Health Research Network, in partnership with the Antoine-Turmel Foundation, to E.V. and R.C. R.O. received PhD Studentship Awards from the Graduate Program of Physiology, the Faculty of Graduate and Postdoctoral Studies, and the Faculty of Medicine, Université de Montréal.

**Conflicts of Interest:** The authors declare no conflict of interest. The funders had no role in the writing of the manuscript or in the decision to publish the results.

#### **References**

