*Article* **Expression of the Endoplasmic Reticulum Stress Marker GRP78 in the Normal Retina and Retinal Degeneration Induced by Blue LED Stimuli in Mice**

**Yong Soo Park 1,2 , Hong-Lim Kim <sup>3</sup> , Seung Hee Lee 1,2, Yan Zhang 1,4 and In-Beom Kim 1,2,3,4,5,\***


**Abstract:** Retinal degeneration is a leading cause of blindness. The unfolded protein response (UPR) is an endoplasmic reticulum (ER) stress response that affects cell survival and death and GRP78 forms a representative protective response. We aimed to determine the exact localization of GRP78 in an animal model of light-induced retinal degeneration. Dark-adapted mice were exposed to blue light-emitting diodes and retinas were obtained at 24 h and 72 h after exposure. In the normal retina, we found that GRP78 was rarely detected in the photoreceptor cells while it was expressed in the perinuclear space of the cell bodies in the inner nuclear and ganglion cell layers. After injury, the expression of GRP78 in the outer nuclear and inner plexiform layers increased in a time-dependent manner. However, an increased GRP78 expression was not observed in damaged photoreceptor cells in the outer nuclear layer. GRP78 was located in the perinuclear space and ER lumen of glial cells and the ER developed in glial cells during retinal degeneration. These findings suggest that GRP78 and the ER response are important for glial cell activation in the retina during photoreceptor degeneration.

**Keywords:** retinal degeneration; endoplasmic reticulum; stress response; unfolded protein response; GRP78; retinal glial cell

#### **1. Introduction**

Retinal degeneration (RD) is a leading cause of blindness and is characterized by the irreversible and progressive loss of photoreceptor cells. Age-related macular degeneration (AMD) is the most common RD with a multifactorial cause of progression. Various factors including a genetic contribution, light-induced stress, oxidative stress and inflammation affect photoreceptor cell death during AMD [1–3]. For the treatment of wet AMD, antivascular endothelial growth factor (VEGF) therapy is a unique, effective treatment option. However, long-term intravitreal injections of an anti-VEGF agent is needed, which can cause complications including ocular hypertension, inflammation, retinal detachment and hemorrhage [4,5]. Moreover, a therapeutic strategy for dry AMD has not yet been established [6]. Thus, studies on the pathogenesis of RD including AMD are needed.

The endoplasmic reticulum (ER) stress response is an important intracellular mechanism of neuronal cell death [7,8]. Stress conditions in the ER can trigger the unfolded protein response (UPR), which increases the number of endoplasmic chaperones to clear

**Citation:** Park, Y.S.; Kim, H.-L.; Lee, S.H.; Zhang, Y.; Kim, I.-B. Expression of the Endoplasmic Reticulum Stress Marker GRP78 in the Normal Retina and Retinal Degeneration Induced by Blue LED Stimuli in Mice. *Cells* **2021**, *10*, 995. https://doi.org/10.3390/ cells10050995

Academic Editor: Maurice Ptito

Received: 26 March 2021 Accepted: 22 April 2021 Published: 23 April 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/).

the accumulated unfolded proteins and to maintain homeostasis [7,9]. However, when ER stress exceeds the UPR, ER stress undergoes a cell death pathway via the activation of the C/EBP homologous protein (CHOP) [10]. There is increasing evidence that ER stress and ER responses are involved in the pathogenesis of AMD and its progression [11,12]. Previous animal studies revealed that ER stress markers were increased in the retina during photoreceptor degeneration [13,14]. In addition, a drug-induced ER stress model could promote photoreceptor loss in mice [15]. However, there is a contrary report that ER stress could protect photoreceptors in Drosophila [16]. Thus, the relationship between photoreceptor cell death and the ER stress of photoreceptors remains unclear.

Immunoglobulin heavy chain-binding protein (BiP), also known as 78-kDa glucoseregulated protein (GRP78), is a representative ER stress marker that belongs to the ER chaperone. GRP78 regulates protein folding to prevent the accumulation of misfolded proteins and maintains ER homeostasis and cell protection [17,18]. Previous studies have shown increased GRP78 expression in RD models induced by light injury [13,19] and retinal detachment [13,19] and in inner retinal degeneration induced by N-methyl-D-aspartate (NMDA) toxicity [20,21], suggesting the involvement of GRP78 in the pathogenesis of various types of retinal degeneration. However, its localization in the normal retina remains unclear. A few investigators have reported that GRP78 is expressed in the inner nuclear layer (INL) and ganglion cell layer (GCL) in the normal retina [20,21] while others have not detected GRP78 expression in cells in the outer nuclear layer (ONL), INL and GCL in a normal retina [19]. Therefore, detailed information on GRP78 expression at cellular and subcellular levels in normal and RD retinas needs to be elucidated to understand the role of ER stress in the pathogenesis of RD.

We aimed to determine the cellular and subcellular localization of GRP78 in normal retinas and to examine the changes in the GRP78 expression profile in RD induced by lightemitting diodes (LED) using immunohistochemistry and an advanced immuno-electron microscopic technique, correlative light and electron microscopy (CLEM).

#### **2. Materials and Methods**

#### *2.1. Animals*

A total of 30 male BALB/c mice at seven weeks of age were used in this study. The mice were kept in a climate-controlled condition with a 12 h light and dark cycle and divided into three groups: normal, 24 h and 72 h after LED exposure (*n* = 10 for each group). All procedures followed the regulations established by the Institutional Animal Care and Use Committee of the School of Medicine, The Catholic University of Korea (Approval number: CUMS-2017-0241-03), which acquired AAALAC International full accreditation in 2018.

#### *2.2. Exposure to a Blue LED*

The blue LED-induced RD model was described in detail in our previous studies [22–24]. Mice were kept in a dark room for 24 h before LED exposure and their pupils were dilated with Mydrin P (Santen Pharmaceutical Co., Osaka, Japan) under dim red-light conditions 30 min before LED exposure. Afterwards, mice were exposed to a 2000-lux blue LED for 2 h. After LED exposure, the mice were kept in a dark room for 1 h and then the mice were moved to a climate-controlled condition with a 12 h light and dark cycle.

#### *2.3. Tissue Preparation*

At 24 h and 72 h after LED exposure, the mice were anesthetized by an intraperitoneal injection of zolazepam (20 mg/kg) and xylazine (7.5 mg/kg) for tissue preparation. The anterior portion of the eye was dissected and eye cups were fixed in 4% paraformaldehyde for 2 h. After fixation, the eye cups were washed with a phosphate buffer (PB, 0.1 M) and then transferred to 30% sucrose in PB 0.1 M for one night. The tissues were embedded in an OCT compound to prepare the frozen tissue. Sections of the eye cups at the center of the retina of 8 µm thickness were obtained.

#### *2.4. Terminal Deoxynucleotidyl Transferase dUTP Nick End Labeling (TUNEL) Assay*

After washing, the retinal sections were incubated in a permeabilization solution containing 0.1% Triton X-100 and 0.1% sodium citrate in 0.01 M phosphate buffered saline (PBS) for 2 min. The TUNEL reaction mixture from the in situ cell death detection kit (Roche, Basel, Switzerland) was used to treat the permeabilized tissue sections for 1 h at 37 ◦C in a humidified atmosphere in the dark. After the end of the TUNEL reaction, the sections were counterstained with DAPI for 5 min.

#### *2.5. Immunohistochemistry*

We used 0.01 M PBS in every procedure. After washing with 0.01 M PBS, the retinal sections were blocked in 10% normal donkey serum for 1 h with 0.1% Triton X-100. Subsequently, they were incubated with primary antibodies for 18 h at 4 ◦C. The sections were washed in PBS and incubated with a secondary antibody for 2 h at room temperature. Cell nuclei were counterstained with DAPI for 5 min. Anti-GRP78 (1:1000; Abcam, Cambridge, UK), anti-ionized calcium binding adaptor molecule 1 (IBA1) (1:500; Wako Chemical, Osaka, Japan), anti-glutamine synthase (GS) (1:1000; Chemicon, Temecula, CA, USA), anti-glial fibrillary acidic protein (GFAP; 1:1000; Chemicon) and Cy3 (1:1500; Jackson Immunoresearch, West Grove, PA, USA) or Alexa 488-conjugated antibodies (1:1000; Molecular Probes, Eugene, OR, USA) were used as secondary antibodies. Images were obtained using a Zeiss LSM 800 confocal microscope (Carl Zeiss Co. Ltd., Oberkochen, Germany).

#### *2.6. Immuno-Electron Microscopy (EM) and CLEM*

For immuno-EM and CLEM, we followed the protocol described by Jin et al. [25] with modifications. The retinas were dissected from the eye cup and fixed with 4% PFA for 2 h. Fixed retinas were washed with 0.1 M PB and cryoprotected with 2.3 M sucrose in 0.1 M PB for 24 h. Retinas with 2.3 M sucrose were frozen in liquid nitrogen. Semi-thin cryosections (2 µm) of the frozen retina were obtained at −100 ◦C using a Leica EM UC7 ultramicrotome equipped with an FC7 cryochamber (Leica). For immunohistochemistry, sections were incubated with 10% normal donkey serum for 1 h without Triton X-100 at room temperature and then double-labeled with GRP78 and GS overnight at 4 ◦C. To detect GRP78 in the EM, FluoroNanogold-conjugated Alexa 488 goat anti-rabbit (1:100; Nanoprobes, Yaphank, NY, USA) was used as a secondary antibody for 2 h at room temperature. The sections were counterstained with DAPI for 3 min and confocal images were obtained. After confocal microscopy, the sections were washed with 0.1 M PB and post-fixed with 2.5% glutaraldehyde and 1% osmium tetroxide for 30 min. Following post-fixation, silver enhancement was performed using an HQ silver enhancement kit (Nanoprobes) and then sections were dehydrated in graded alcohols. After the completion of all procedures, the tissues were embedded in Epon 812. Interested areas were selected in the confocal image and selected areas were cut into ultrathin sections (80–90 nm) and observed under an electron microscope (JEM 1010, Tokyo, Japan).

#### *2.7. Image Analysis*

Confocal images were analyzed using ZEN 2.3 software (Carl Zeiss). The immunohistochemistry intensity was measured by a histogram function and co-localization ratio automatically calculated by ZEN's co-localization function.

#### *2.8. Statistical Analysis*

Quantified intensity values were statistically analyzed using Prism 8.0 software (GraphPad, San Diego, CA, USA). A one-way ANOVA was used to determine statistical significance with a *p*-value < 0.05.

#### **3. Results**

#### *3.1. GRP78 Expression in a Normal Retina*

First, we determined the GRP78 expression pattern and its localization in the normal retina at cellular and subcellular levels using immunohistochemistry and immuno-EM methods. In the normal retina, a strong immunoreactivity of GRP78 was observed in cell bodies in the INL and GCL and in the inner and outer segments (IS/OS) of photoreceptors and the outer plexiform layer (OPL) while weak punctate labeling was observed in the ONL and inner plexiform layer (IPL) (Figure 1A). To examine its localization at the subcellular level, we performed immuno-EM using a 1.4 nm gold-conjugated secondary antibody. In the ONL, gold particles were rarely detected in the cell bodies of the photoreceptor while they were mainly localized in thin cytoplasmic components within the intercellular spaces (Figure 1B), which might be the process of Müller glial cells, which are the main glial cells in the retina. In the INL and GCL, gold particles were distributed through the nuclear membrane (Figure 1C,D) and membranous structures near the nucleus (Figure 1D), which might be the ER.

#### *3.2. Changes in GRP78 Expression in Blue LED-Induced RD*

We subsequently examined the GRP78 expression in RD retinas at 24 h and 72 h after blue LED exposure (Figure 2) when photoreceptor degeneration started and peaked, respectively [22–24]. GRP78 immunoreactivity was increased in the ONL and IPL after LED exposure in a time-dependent manner (yellow rectangles in Figure 2B,F,J). In the RD retina 72 h after blue LED exposure, many GRP78-labeled lumps were observed in the ONL (yellow circles in Figure 2J). Considering that GRP78 is mainly localized in a cellular component of the intercellular spaces between photoreceptors in the normal retina (Figure 1B), which might be Müller cell processes, its cellular identity appeared to be Müller cells not photoreceptors. To confirm this point, we performed a double-label immunofluorescence experiment with anti-GRP78 and anti-GS, a representative Müller cell marker. In the normal retina, GS immunoreactivity was found in Müller cells whose bodies were situated in the middle of the INL and processes in the ONL, OPL, IPL and GCL from cell bodies to the outer and inner limiting membranes with a variety of lines in size (Figure 2C). In the RD retinas, GS immunoreactivity was remarkably increased in the ONL and thus it appeared as circles and outwardly extending thick irregular lines (Figure 2G,K,M), which suggested that Müller cell processes became hypertrophied and occupied the ONL where photoreceptors degenerated. In merged images (Figure 2D,H,L), the co-localization of GRP78 with GS in the ONL and INL was observed in approximately 23% and 41% of the normal retina, respectively, and significantly increased to 42% and 66%, respectively, in RD at 72 h after blue LED exposure (*p* < 0.05, *n* = 7 in Figure 2M,N).

However, GRP78-labeled lumps in the ONL in RD at 72 h after blue LED exposure did not show GS immunoreactivity (Figure 2L, yellow circles) suggesting that they belonged to different cellular profiles. As microglial cells occurred in similar locations in this RD model with the same morphology [22–24], we tested whether GRP78 was expressed in microglial cells in RD retinas by double-labeling with anti-GRP78 and anti-IBA1, a microglial cell marker. In the normal retina, IBA1-labeled microglial cells were mainly observed in the IPL and OPL (red in Figure 3A). However, in RD induced by blue LED exposure, IBA1 labeled microglial cells were mainly found in the ONL and inner and outer segment layers of the photoreceptors (red in Figure 3B,C). These microglial cells showed weak GRP78 immunoreactivity in their cell bodies and processes in normal retinas (Figure 3A) while it was increased in activated microglial cells, which migrated into the ONL where photoreceptors degenerated followed by the time course of RD (Figure 3B–C). In addition, GRP78-labeled lumps in the ONL observed in RD at 72 h after blue LED exposure were colabeled with IBA1 (Figure 3C). These results indicated that resting and activated microglial cells expressed GRP78 in the normal and RD retinas and GRP78 expression was increased as microglial cells were activated in RD.

**Figure 1.** GRP78 expression in normal retinas: (**A**) Representative normal retina labeled with DAPI (blue) and GRP78 (green). GRP78 was mainly labeled in the rod and cone layers and the cell bodies of the INL and GCL. In the ONL, the photoreceptor cell bodies were weakly surrounded by GRP78. (**B**–**D**) Representative EM images of the normal retina with immuno-gold-labeled GRP78. (**B**) Normal photoreceptor cells in ONL. Immuno-gold labeled-GRP78 was detected in the interspace of the photoreceptor cell bodies but not in the photoreceptor cells. (**C**) Cell bodies in INL. Immuno-goldlabeled GRP78 was detected in the perinuclear spaces of INL cells. (**D**) Representative EM images of GCL. Immuno-gold-labeled GRP78 was detected in the perinuclear spaces and ER lumen of the GC. INL, inner nuclear layer; GCL, ganglion cell layer; ONL, outer nuclear layer; EM, electron microscopy; ER, endoplasmic reticulum; GC, ganglion cell.

**Figure 2.** GRP78 expression in Müller glial cells in blue LED-induced RD retinas. (**A**,**E**,**I**) DAPI (blue) stained retina sections of the normal (**A**), 24 h (**E**) and 72 h (**I**) retinas after blue LED exposure. ONL thickness prominently decreased at 72 h. (**B**,**F**,**J**) GRP78 (green)-labeled sections of normal (**B**), 24 h (**F**) and 72 h (**J**) retinas after blue LED exposure. GRP78 was increased in the ONL and IPL in a time-dependent manner after LED exposure (yellow rectangles) and labeled in giant cells in the ONL (yellow circles). (**C**,**G**,**K**) GS (red)-labeled retina sections of the normal (**C**), 24 h (**G**) and 72 h (**K**) retinas after blue LED exposure. GS-labeled Müller cell processes enlarged and encircled photoreceptor cells in the ONL after injury. (**D**,**H**,**L**) Merged results of GRP78 and GS. GRP78 was increased in the Müller cell processes from IPL to ONL in a time-dependent manner while GRP78-labeled giant cells in the OPL showed no GS immunoreactivity (yellow circles). (**M**,**N**) Co-localization ratio of the GRP78/GS in the ONL (**M**) and IPL (**N**). In both the ONL and INL, the GRP78 co-localization ratio in GS-labeled Müller glial cells was significantly increased at 72 h after LED exposure (*n* = 7, *p* < 0.05 (\*), ANOVA).

**Figure 3.** GRP78 expression in microglial cells in blue LED-induced RD retinas. (**A**) A representative confocal image of the normal retina labeled by DAPI (blue), GRP78 (green) and IBA1 (red). IBA1-labeled microglial cells were weakly labeled by GRP78 in the normal retina. (**B**). After 24 h of light exposure, microglial cells were detected in the ONL with an enlarged shape and increased GRP78 compared with the normal retina. (**C**) After 72 h of light exposure, microglial cells were markedly increased in the ONL and GRP78 was prominently increased in microglial cells compared with those after 24 h.

#### *3.3. Subcellular Localization of GRP78 in Glial Cells of the Retina*

GRP78 is known to localize various subcellular components including the cell membrane, nuclear membrane and cytosol [26] according to cell type. To elucidate the functional role of GRP78 in two retinal glial cells, Müller cells and microglia, we employed CLEM, an advanced EM technique combined with immunohistochemistry to determine the subcellular localization (Figure 4). As it is difficult to delineate the Müller cell processes in the INL and IPL without markers, we performed double-labeling with an anti-GS/Cy3 conjugated secondary antibody and an anti-GRP78/Alexa488-FluoroNanogold-conjugated secondary antibody in semi-thin cryosections of the RD retina. Under confocal microscopy, GS/Cy3-labeled Müller cell processes that were apparently co-labeled with GRP78/Alexa 488-FluoroNanogold were identified and selected for EM observation.

**Figure 4.** Subcellular GRP78 localization in the injured retina 72 h after light exposure. Semi-thin sections of the retina stained by GRP78 (green, gold), GS (red) and DAPI (blue). (**A**) Microglia and migrated Müller glia in the ONL. Microglia (red asterisk) were more strongly labeled by GRP78 than Müller glia without GS while Müller glia (orange asterisk) were labeled by both GRP78 and GS. (**B**) EM images of microglia (red asterisk) and migrated Müller glia (orange asterisk) in the same region as (**A**). (**C**) Immuno-gold-labeled GRP78 was detected in the perinuclear space (orange arrow) and lumen of the ER structure (red arrows) of the microglia. (**D**) Immuno-gold-labeled GRP78 was detected in the perinuclear space of the Müller glial cell (orange arrows). (**E**) Confocal images in the INL of an injured retina with Müller glia processes. (**F**) EM images matched with the yellow box in (**E**). Processes of the Müller glia in the INL labeled by GRP78. (**G**) A higher magnification of the Müller glia process in (**F**). Membranous structures containing immuno-gold-labeled GRP78 located in Müller glia processes.

In the confocal (Figure 4A) and its correlated EM (Figure 4B) images taken from the ONL in a RD retina, two types of retinal glial cells were identified. One type was GRP78 single-labeled microglial cells without GS immunoreactivity (red asterisk in Figure 4A), which showed typical microglial cell morphology characterized by irregular nuclei with a characteristic peripheral heterochromatin and heterochromatin net (red asterisk in Figure 4B). The other was GRP78/GS double-labeled Müller glial cells (orange asterisk in Figure 4A), which had irregular nuclei in shape that were apparently distinguished from neighboring photoreceptors and microglial cells (orange asterisk in Figure 4B). In higher magnification EM views of the microglial cells (Figure 4C) and Müller cells (Figure 4D) in Figure 4B, the immuno-gold particles of GRP78 were localized to the nuclear membrane, membranous structures in the perinuclear region (orange arrows) and ER (red arrows) in the perinuclear region.

In the INL, GRP78/GS double-labeled Müller cell processes were frequently observed by confocal microscopy (Figure 4E). In correlative EM (Figure 4F,G), the immuno-gold particles of GRP78 were found in an elongated membranous structure suggesting that GRP78 was expressed in newly formed ER in hypertrophied Müller glial cells in the RD retina.

#### **4. Discussion**

We determined the cellular and subcellular localization of GRP78 in normal and blue LED-induced RD retinas. GRP78 is expressed in various types of retinal neurons and glial cells. In RD, GRP78 was not increased in the damaged photoreceptors, resulting in photoreceptor apoptosis while it was increased in Müller and microglial cells together with the development of a new ER.

UPR related to ER stress is expected to be a promising target for the treatment of various degenerative diseases because it is involved in cell survival and death and thus controls cell fate [27–29]. Among all UPR products, GRP78 belongs to the heat shock protein 70 family (HSP70), which deals with damaged or misfolded proteins to protect cells [17,18]. In previous studies [13,19–21], the protective role of GRP78 in response to various retinal injuries was proposed. For instance, that role against ganglion cell injury has been relatively well described [21] while GRP78′ s role in photoreceptor degeneration remains unclear. Moreover, it is not clear whether photoreceptors in the normal retina express GRP78. Previous studies have shown GRP78 expression in the normal 661 W cell line, which was derived from mouse cone cells [13] while other studies have shown that GRP78 is rarely expressed in the ONL in the retinas [19] of normal and P23H rats [30] and ER stress-activated indicator (ERAI) transgenic mice [20]. We therefore wanted to describe the exact cellular and subcellular localization of GRP78 in photoreceptors.

In this study, we confirmed GRP78 expression in photoreceptors in the normal retina, albeit at low levels. At the subcellular level, it was localized to membranous structures. This expression and localization pattern was observed in RD photoreceptors but did not change (Figure 1 and Supplementary Figure S1). These results appeared to be inconsistent with previous studies showing that GRP78 expression was increased in the photoreceptor cells in RD suggesting a protective role in RD by the inhibition of photoreceptor apoptosis [13,19]. First, this discrepancy might be caused by in vitro RD conditions [13,19] in which only a 661 W cell, a transformed cell line derived from the cone cell, exists and it cannot interact with the retinal pigment epithelium and/or rod and/or Müller glial cells. In addition, the finding of increased GRP78 in the ONL being due to photoreceptors might be falsely attributed [13,19]. Actually, increased GRP78 expression occurred in Müller cell processes that occupied a narrow interphotoreceptor space (Figures 2 and 4), not in the photoreceptors in RD (Supplementary Figure S1) as demonstrated in this study. These findings suggested that GRP78 was not directly involved in photoreceptor cell death. Instead of GRP78, apoptotic UPR pathways might be activated during photoreceptor apoptosis in RD. CHOP may be a strong candidate for cell death and is already known to

be expressed in photoreceptors and plays an essential role in photoreceptor apoptosis via ER stress in RD [31,32].

Glial cells play an important role in various types of brain injury. In particular, two glial cells, microglia and astrocytes, are involved in neuroinflammation. Each type is divided into two subpopulations (M1/M2 and A1/A2) depending on its function and with an opposing role: pro-inflammatory vs. anti-inflammatory [33,34], which eventually results in neurodegeneration or neuroprotection. In neuroinflammation related to various brain pathologies, ER stress contributes to determining the phenotypes of microglia [35,36] and astrocytes [37,38]. It has been reported that GRP78 induces phagocytic function and cytokine production in microglial cells [39] and activates astrocytes to promote neuroprotective cytokines [40]. Moreover, it protects astrocytes themselves under stress conditions [41].

In this study, we demonstrated that GRP78 expression was increased in two types of retinal glial cells, Müller cells and microglial cells, in a blue LED-induced RD model. These results were consistent with the increase in GRP78 expression in glial cells in response to several types of brain and spinal cord injuries [25,42–44]. Müller cells are the main glial cells in the retina, corresponding to astrocytes in the brain and spinal cord. As GRP78 could activate astrocytes to promote neuroprotective cytokines [40], increased GRP78 in Müller cells might have had a neuroprotective role in this RD model. Moreover, Müller cells are responsible for retinal metabolism through the UPR as demonstrated in diabetic retinopathy [45–48]. Considering that GRP78 belongs to the HSP70 family that deals with damaged or misfolded proteins to protect cells [17,18], GRP78 in Müller cells may be involved in retinal protection as a key component of UPR. In addition, a prominent GRP78 increase in the microglia occurred 72 h after blue LED exposure when microglial activation peaked [22,23]. This suggested that GRP78 was closely associated with microglia activation, which might phagocytose and clear degenerating photoreceptors [39]. However, the exact role of GRP78 and its mechanism of action in retinal glial cells should be further investigated in a study on GRP78 in Müller cells or microglia conditional knockout animals.

GRP78 is an ER chaperone and thus was used as a representative ER stress marker. At the subcellular level, it was localized to membranous structures including the nuclear membrane and putative ER in retinal neurons and glial cells (Figures 1 and 4, Supplementary Figure S1). In particular, activated Müller cell processes in blue LED-induced RD retinas showed an increased GRP78 expression with ER development (Figure 4). We confirmed this by double-labeling immunofluorescence with anti-GFAP, an activated Müller cell marker [49,50], and anti-calreticulin, a representative ER marker [51,52]. The expression of calreticulin became evident in GFAP-labeled Müller cell processes in the IPL and INL (Supplementary Figure S2). Based on the fact that Müller cells are responsible for metabolic modulation in the degenerative retina [45] and that it engulfs and clears the damaged photoreceptor [53,54], increased GRP78 within newly developed ER in Müller cells might therefore be essential to resolve the overloaded metabolic needs by preventing the accumulation of misfolded proteins. In addition, activated microglial cells showed higher GRP78 levels than inactivated ones (Figure 3) and calreticulin-immunoreactive lumps similar to GRP78-immunoreactive lumps in activated microglial cells in the ONL in RD retina were found at the same time and location (Supplementary Figure S2) suggesting the development of a new ER. This might be reasonable because microglia could be in an overloaded condition to produce the cytokines and phagocyte the damaged photoreceptors in RD.

Finally, we want to refer to GRP78 at the mitochondria and mitochondria-associated ER membrane. Recently, growing evidence has pointed towards GRP78 playing an important function in the mitochondria in association with co-chaperones known to be involved in calcium-mediated signaling between the ER and mitochondria, important for bioenergetics and cell survival [55]. As previous reports detected GRP78 in the mitochondria [56,57] and mitochondria-associated ER membrane [58], we tried to detect mitochondrial GRP78 localization in photoreceptors and glial cells by EM. However, immuno-gold-labeled GRP78 was rarely detected in the mitochondria either in photoreceptors or glial cells. There was

also no distinguishable ER region of a strong GRP78 expression close to the mitochondria (data not shown). Nevertheless, this issue needs to be further investigated in the near future.

#### **5. Conclusions**

In summary, we determined the cellular and subcellular localization of GRP78 in normal and blue LED-induced RD retinas. GRP78 was expressed in various types of retinal neurons and glial cells. In RD, GRP78 was not increased in the damaged photoreceptors, resulting in photoreceptor apoptosis while it was increased in Müller and microglial cells together with the development of a new ER. These results suggested that increased GRP78 played a role in glial cell activation and neuroprotective function by modulating the UPR under stress conditions. Further studies to reveal the relationship between GRP78 and ER stress and glial cell responses in RD and its mechanism are needed, considering that the findings in this study were limited to the histology.

**Supplementary Materials:** The following are available online at https://www.mdpi.com/article/10 .3390/cells10050995/s1, Figure S1: Subcellular localization of the GRP78 in the ONL of the normal retina and blue LED-induced RD retina. Figure S2: Expression of the calreticulin and GFAP in a blue LED-induced RD retina.

**Author Contributions:** Conceptualization, Y.S.P. and I.-B.K.; methodology, Y.S.P., H.-L.K., S.H.L. and Y.Z.; software, Y.S.P.; validation, Y.S.P. and I.-B.K.; formal analysis, Y.S.P. and I.-B.K.; investigation, Y.S.P., H.-L.K., S.H.L. and Y.Z. and I.-B.K.; resources, Y.S.P. and I.-B.K.; data curation, Y.S.P. and I.-B.K.; writing—original draft preparation, Y.S.P.; writing—review and editing, I.-B.K.; visualization, Y.S.P.; supervision, I.-B.K.; project administration, I.-B.K.; funding acquisition, I.-B.K. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by the Basic Science Research Program through the National Research Foundation (NRF) of Korea funded by the Ministry of Education, Science and Technology (grant number 2017R1A2B2005309).

**Institutional Review Board Statement:** The study was conducted in accordance with the guidelines of the Declaration of Helsinki and approved by the Institutional Animal Care and Use Committee (IACUC) at the College of Medicine, The Catholic University of Korea (Approval number: CUMS-2017-0241-03).

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** Data is contained within the article or supplementary material.

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

#### **References**


### *Article* **Long-Term Transplant Effects of iPSC-RPE Monolayer in Immunodeficient RCS Rats**

#### **Deepthi S. Rajendran Nair <sup>1</sup> , Danhong Zhu <sup>2</sup> , Ruchi Sharma <sup>3</sup> , Juan Carlos Martinez Camarillo 1,4, Kapil Bharti <sup>3</sup> , David R. Hinton <sup>2</sup> , Mark S. Humayun 1,4 and Biju B. Thomas 1,4,\***


**Abstract:** Retinal pigment epithelium (RPE) replacement therapy is evolving as a feasible approach to treat age-related macular degeneration (AMD). In many preclinical studies, RPE cells are transplanted as a cell suspension into immunosuppressed animal eyes and transplant effects have been monitored only short-term. We investigated the long-term effects of human Induced pluripotent stem-cellderived RPE (iPSC-RPE) transplants in an immunodeficient Royal College of Surgeons (RCS) rat model, in which RPE dysfunction led to photoreceptor degeneration. iPSC-RPE cultured as a polarized monolayer on a nanoengineered ultrathin parylene C scaffold was transplanted into the subretinal space of 28-day-old immunodeficient RCS rat pups and evaluated after 1, 4, and 11 months. Assessment at early time points showed good iPSC-RPE survival. The transplants remained as a monolayer, expressed RPE-specific markers, performed phagocytic function, and contributed to vision preservation. At 11-months post-implantation, RPE survival was observed in only 50% of the eyes that were concomitant with vision preservation. Loss of RPE monolayer characteristics at the 11-month time point was associated with peri-membrane fibrosis, immune reaction through the activation of macrophages (CD 68 expression), and the transition of cell fate (expression of mesenchymal markers). The overall study outcome supports the therapeutic potential of RPE grafts despite the loss of some transplant benefits during long-term observations.

**Keywords:** iPSC-RPE; retinal pigment epithelium; immunodeficient RCS rat; ultrathin parylene; retinal degeneration; retinal transplantation

### **1. Introduction**

Age-related macular degeneration (AMD), one of the most common causes of blindness in the developed world, is a degenerative disease of the retina often leading to progressive vision loss. Geographic atrophy, the advanced form of AMD, is characterized by dysfunction of retinal pigmented epithelium cells (RPEs) followed by degeneration of overlying photoreceptors leading to the loss of central vision. At present, no proven clinical treatments exist for the preservation or replacement of vulnerable RPE cells; however, RPE cell transplantation is perhaps the most obvious therapeutic option and has garnered significant interest. In the early stages of AMD, although the RPE cells are dysfunctional, surviving photoreceptors and the inner retina that transmit visual signals to the brain remain functional, rendering a realistic possibility that replacing the degenerating RPE with functional young RPE will restore vision.

**Citation:** Rajendran Nair, D.S.; Zhu, D.; Sharma, R.; Martinez Camarillo, J.C.; Bharti, K.; Hinton, D.R.; Humayun, M.S.; Thomas, B.B. Long-Term Transplant Effects of iPSC-RPE Monolayer in Immunodeficient RCS Rats. *Cells* **2021**, *10*, 2951. https://doi.org/ 10.3390/cells10112951

Academic Editors: Maurice Ptito and Joseph Bouskila

Received: 18 September 2021 Accepted: 25 October 2021 Published: 29 October 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/).

Potential sources of healthy RPEs are pluripotent cells derived from embryonic [1–4] or adult cell sources [5–9], which are differentiated into RPE cells by employing spontaneous or directed differentiation methods. Early-phase clinical trials by various research groups used embryonic stem-cell-derived RPE (ESC-RPE) for cell replacement, which has already shown early signals of safety and potential efficacy [10–14]. Our team has demonstrated that human embryonic stem-cell-derived RPE (hESC-RPE) grown as a polarized monolayer on ultrathin parylene substrates can remain functional after transplantation in athymic nude rats [15] and in Royal College of Surgeons (RCS) rats [16,17], a model for RPE dysfunction. The product, termed the California Project to Cure Blindness-RPE (CPCB-RPE1), is being assessed in an FDA-approved phase1/2a clinical trial (NCT 02590692) for advanced dry AMD and exhibits promising outcomes for improving visual activity [12].

The autologous induced pluripotent stem-cell-derived RPE (iPSC-RPE) transplantation is considered more advantageous as the chance of graft rejection issues can be minimized. Recent research focuses on the generation of iPSC lines from adult cell sources, such as skin fibroblasts or peripheral blood mononuclear cells [5–9]. The four-year report of iPSC-RPE sheet transplant surgery for CNV (choroidal neovascularization—the wet form of AMD) in one patient has been published recently [18]. Another major step forward is the allogeneic transplantation of off-the-shelf available iPSC-RPE. Due to concerns regarding possible oncogenic mutations in cell preparation, the attention is now focused on personalized screening for mutations and the development of autologous iPSC-RPE therapies including HLA matching [19,20]. A study by Sugita et al. [19] aimed at examining the safety of six-loci HLA-matched allogeneic iPSC-RPE transplantation under local steroids. RPE cells grafted as a suspension into the patient's subretinal space survived in all five cases for more than one year [19]. These observations suggest that it is possible to manage the survival of iPSC-RPE, under immunosuppression. However, regenerative medicine is still in its infancy and the cells may behave differently in each individual. In the first iPSC-RPE transplantation clinical study [14,21], three aberrations in the deoxyribonucleic acid (DNA) copy number (deletions) were observed in the cell preparation of the second patient, which caused the study to end due to possible adverse effects.

Existing evidence indicates that the delivery of cells as a suspension may not consistently develop into a monolayer of RPE and that their long-term survival rate will be low compared to RPE cells transplanted as a monolayer [15]. In many preclinical studies for geographic atrophy, the iPSC-RPEs were delivered into the subretinal space as a bolus injection [22–25] and the animals were monitored for survival under immunosuppression for only a short period. In other studies, iPSC-RPEs transplanted as a monolayer and maintained under immunosuppression regimes were followed for up to 5 months [9,18]. Published data suggest that a confluent polarized monolayer of iPSC-RPEs transplanted as a patch rather than as a cell suspension can perform several basic functions of RPEs including phagocytosis of photoreceptor outer segments, the renewal of visual pigment, and the transport of metabolites [9,14]

Although transplantation of iPSC-RPE cells to replace the diseased RPE has been tested by several groups through preclinical studies and there are preliminary reports of ongoing preclinical studies, no major attempt has been made to perform an in-depth analysis of the long-term viability and fate of the transplanted RPE. Based on previous reports from our group [17], the progressive deterioration of visual function after the transplantation of ESC- RPE was evident during long-term observations. Monitoring cell survival and assessing the long-term functional benefits of transplants are significant since the transplanted cells are exposed to a progressively degenerating environment and there may be immunological factors that can cause adverse effects.

In preclinical studies of RPE transplantation, one of the major factors that influence the long-term benefits is the immune reaction and associated xenograft rejection. Sharma et al. [9] showed 70% survival of subretinally transplanted human iPSC-RPE (hiPSC-RPE) cells up to 2.5 months post-implantation in immunosuppressed rodents. In the above study, systemic and resident innate immune responses in animal models

were suppressed by using prednisone, doxycycline, and minocycline whereas the adaptive immune responses were suppressed using tacrolimus and sirolimus [9]. Del Priore et al., in 2003 [26], demonstrated 'triple systemic' therapy with anti-inflammatory antibiotics to increase the survival of grafted RPE at four weeks post-implantation. However, based on a previous report, immunosuppressants can alter the visual function in RCS rats with depressed scores on behavioral and electrophysiological testing [27]. Hence, to minimize the complications associated with immunosuppressants, we used a newly developed immunodeficient RCS rat model characterized by an absence of T cells and a lack of natural cell-mediated cytotoxicity [28]. The iPSC-RPE cells grown as a polarized monolayer on ultrathin perylene substrates were transplanted into the subretinal space of immunodeficient RCS rats. The transplant effects were assessed at various post-implantation time points (1 to 11 months after transplantation).

#### **2. Materials and Methods**

#### *2.1. Human Pluripotent Stem Cells Generated from iPSCs*

iPSC-RPE (frozen, passage 2 cells) generated from iPSCs, reprogrammed from healthy adult fibroblasts, was obtained from Dr. Kapil Bharti, Unit on Ocular and Stem Cell Translational Research, National Eye Institute, NIH, Bethesda, USA [29]. Briefly, hiPSC were seeded at 20,000 cells per cm<sup>2</sup> on Matrigel and grown in mTeSR1 in a 10% CO2/5% O<sup>2</sup> incubator for 5 days. Afterward, they were transferred to a 5% CO2/20% O<sup>2</sup> incubator and cultured for 5 additional days. At this point, the culture medium was switched to a differentiation medium (DM) [29]. After 10–15 days, cells were maintained in DM for 3 more weeks and then switched to an RPE maintenance medium (RPEM) [29]. Differentiated cells were dissociated in Accumax (Sigma, Saint Louis, MO, USA), plated at 250–300,000 cells per cm<sup>2</sup> , and grown in RPEM. The passage 3 cells were used for transplantation experiments. These cells are extensively characterized for clinical applications in Dr. Bharti's lab [9,30,31].

#### *2.2. Preparation of Polarized hESC-RPE Implant on Parylene Membranes*

Ultrathin parylene membranes (0.3 µm thickness supported on a 6.0 µm-thick mesh frame) made from parylene C were specially designed for implantation on rat retinas (1.0 × 0.4 mm) [15,17] and used successfully for culturing iPSC-RPEs. These ultrathin membranes were coated with Matrigel (AMS Biotechnology, Frankfurt, Germany) and seeded with iPSC-RPE based on our previously established protocol [17]. The cells were grown to confluence for approximately 4 weeks before implantation. The final density of each implant was kept as approximately 2700 cells/membrane [15].

#### *2.3. Immunostaining of iPSC-RPE on Parylene Membrane*

iPSC-RPE cells grown on Matrigel-coated parylene were stained for RPE specific markers zonula occludens protein 1 (ZO-1), and RPE65, based on established protocols [17]. Stained cells were mounted with an anti-fading mounting medium (Invitrogen) and images were captured by confocal microscopy (FV1000 Confocal Microscope, Olympus, Centre Valley, PA, USA).

#### *2.4. Animals*

The Royal College of Surgeons (RCS) rat is an established model of retinal degeneration, which has been mainly used for studying photoreceptor rescue with treatment at the age of 3–4 weeks. These rats develop a fully functional visual system, which degenerates secondarily due to their dysfunctional RPE (MertK mutation), resulting in the loss of most photoreceptors at the age of 3 months. Immunodeficient RCS rats were produced from a cross between female homozygous RCS (RCS-p+/RCS-p+) and male athymic nude rats (Hsd: RH-Foxn1mu, a mutation in the foxn1 gene; no T cells) as described previously [28]. All rats were maintained in an aseptic and temperature-controlled environment. All animals were included in accordance with the Association for Research in Vision and

Ophthalmology (ARVO) statement for the use of animals in research, and the Institutional Animal Care and Use Committee (IACUC) of USC.

#### *2.5. Surgical Procedure*

Animals underwent surgery at postnatal day (P) 28. Anesthesia was induced by intraperitoneal injection of ketamine (37.5 mg/kg) and xylazine (5 mg/kg). Only the left eyes were used for transplantation surgeries. All surgeries were performed by the same surgeon. Topical anesthesia was administered with a 0.5% proparacaine hydrochloride ophthalmic solution (Akorn, Inc., Lake Forest, IL, USA). Pupils were dilated using ophthalmic solutions of 2.5% phenylephrine hydrochloride and 0.5% tropicamide (Akorn, Inc., Lake Forest, IL, USA). Once the conjunctiva is removed, a scleral incision was performed in the temporal superior quadrant followed by an anterior chamber paracentesis to reduce intraocular pressure. A small incision (approximately 0.8–1.0 mm) was cut transsclerally at the temporal equator of the eye until the choroid was exposed with the help of a 32-gauge needle, and a 5 µL balanced salt solution (Alcon Laboratories, Inc., Fort Worth, TX, USA) was injected to create a local retinal detachment. The implant held by forceps was introduced through the sub scleral space into the subretinal bleb. Clinical assessment as well as retinal imaging by optical coherence tomography (OCT) using a Spectralis HRA + OCT device (Heidelberg Engineering, Heildeberg, Germany) were performed to confirm the placement of the implant. The rats were then allowed to recover from anesthesia in a thermal care incubator. Animals with surgical complications such as excessive bleeding, perforation of the retina, and implant delivery into the vitreous were immediately excluded from the study. Based on OCT images, 15 animals were selected for short-term experiments (1-month and 4-month study group) and 15 animals were selected for the 11-month study group.

#### *2.6. Histopathology*

Cohorts of rats were euthanized by intracardiac injection of euthasol (Virbac AH, Inc., Fort Worth, TX, USA) at 1, 4-, and 11-months post-surgery, and eyes were processed for histology. Contralateral eyes were considered as controls. Whole eyes were fixed in Davidson's solution overnight, and the cornea and lens were removed. Finally, the eye cups were embedded in paraffin and processed for sectioning (5-µm sections). Groups of consecutive slides were stained with hematoxylin and eosin (HE) for light microscopy. The HE-stained slides were scanned and photographed using an Aperio Scanscope CS (Aperio Technologies, INL., Vista, CA, USA) microscope. Histological sections of cell-seeded membranes were evaluated to assess iPSC-RPE survival. The surgical placement was considered acceptable if more than 70% of the implant was located inside the subretinal area. Transplant survival was confirmed only if iPSC-RPE were observed in at least three consecutive sections based on light microscopy and immunostaining evaluations. Cell migration or dead cell aggregation was considered when pigmented cells or cell clumps were seen adjacent to the substrate and confirmed by immunohistochemistry. If no human/RPE marker was found, the specimen was considered as non-surviving RPE clumps. The outer nuclear layer (ONL) integrity was evaluated for photoreceptor preservation in the transplanted area. Cellular reaction around the implants, observed by light microscopy, was assessed for the presence of macrophages or the expression of glial cells. Adjacent sections of the implanted eye were processed for immunohistochemical analysis using the following antibodies as needed: Human-specific cell surface marker (anti-TRA-1-85), a marker of differentiated RPE cells (anti-RPE65), a macrophage marker (anti-CD68), an astrocyte/Müller cell marker (anti-GFAP), mesenchymal markers (α Smooth muscle actin and Vimentin), photoreceptor phagocytosis marker (Rhodopsin), and RPE binding protein (RBP1).

Details of the antibodies used are included in Table 1. For immunostaining, all slides were deparaffinized, rehydrated, and antigen retrieved (sodium citrate, pH 6.0). After staining, the slides were mounted with fluorescent-enhanced mounting medium with 4′ ,6-

diamidino-2-phenylindole (DAPI) (Vector Laboratory, Burlingame, CA, USA). Images were taken using the Ultra viewer ERS dual-spinning disk confocal microscope (PerkinElmer, Waltham MA, USA) equipped with a C-Apochromat (Carl Zeiss, Thornwood, NY, USA) ×10 high dry lens, a C-Apochromat ×40 water immersion lens NA 1.2, an electron multiplier charge-coupled device cooled digital camera (Hamamatsu Orce\_ERCC 12-bit camera]; PerkinElmer, Waltham, MA, USA) or by using a Keyence BZX-800 microscope. Images were captured and processed using PerkinElmer Velocity imaging software.


**Table 1.** List of antibodies used for immunostaining.

#### *2.7. Superior Colliculus Electrophysiology*

Electrophysiological mapping of the superior colliculus (SC) was performed at approximately 11-months post-surgery based on an established protocol followed in our laboratory [7,17,28]. Based on OCT screening, 8 rats (8/15) were selected for SC experiments. Rats dark-adapted overnight were anesthetized by an intraperitoneal injection of xylazine/ketamine. The gas-inhalant anesthetic (1–2.0% isoflurane) was administered via an anesthetic mask (Stoelting Company, Wood Dale, IL, USA). Rats were mounted in a stereotactic apparatus; a craniotomy was performed, and the SC was exposed. Multi-unit visual responses were recorded extracellularly from the superficial laminae of the SC using custom-made tungsten microelectrodes. For SC mapping, the responses were recorded from approximately 30 different SC locations. At each recording location, approximately 10 presentations of a full-field strobe flash (1300 cd m22, Grass model PS 33 Photic stimulator, W. Warwick, RI, USA), positioned 30 cm in front of the rat's eye, were delivered to the contralateral eye. An interstimulus interval of 5 s was used. The neural activities were recorded using a digital data acquisition system (Power lab; ADI Instruments, Mountain View, CA, USA) 100 milliseconds before and 500 milliseconds after the onset of the stimulus. All responses at each site were averaged. Blank trials, in which the illumination of the eye was blocked with an opaque filter, were also recorded at each site.

#### *2.8. Optokinetic Testing*

Optokinetic (OKN) testing was performed at 4 months and 11 months post transplantation using a previously described protocol [17]. To record OKN responses, two tablet screens were used to display the OKN stimuli consisting of high-contrast black and white stripes generated using 'OKN Stripes Visualization Web Application', a freely available software (http://mdds.nyc/okn-stripes-visualization, accessed on 12 April 2021). A clear

plexiglass holder was used to restrain the rat and keep its head continuously exposed to the tablet screen. A micro camera attached to the top of the rat holder recorded the headtracking responses during clockwise (1 min) and anticlockwise (1 min) stripe rotations. Visual acuity was tested by changing the stripe width at decrements of 0.5. Video recordings were evaluated to compute the head-tracking scores by two separate investigators who were both masked to the experimental condition. The OKN responses at various spatial frequencies were assessed based on the presence or absence of clear head-tracking and based on the duration of head-tracking.

#### *2.9. Statistical Analysis*

Statistical comparisons were made using GraphPad Prism software (GraphPad Software Inc., La Jolla, CA, USA). The Paired *t*-test was used for analyzing the OKN data. The remaining data were analyzed using Student's *t*-test or by the Analysis of Variance (ANOVA) followed by the appropriate post hoc test. For all comparisons, the significance level was determined at *p* < 0.05.

#### **3. Results**

#### *3.1. Human iPSC-RPE Cells Can Grow as a Polarized Monolayer over Ultrathin Parylene Membrane and Demonstrate High-Purity and RPE Marker Expression*

iPSC-RPE cells (Figure 1a,b) cultured on a Matrigel-coated ultrathin parylene substrate were expanded as a polarized confluent monolayer (Figure 1d) and expressed RPE 65 and ZO-1 as evidenced by immunocytochemistry (Figure 1e–h). This study demonstrated that iPSC-RPE can be grown as a polarized monolayer on ultrathin parylene, similar to our previous hESC-RPE implants [17].

#### *3.2. iPSC-RPE Implant Survival and Functionality Assessed by Short-Term in Vivo Experiments in Immunodeficient RCS Rats (1- and 4-Month Study)*

After the transplantation surgery, OCT imaging was performed to screen the animals for proper implant placement (Figure 2a–c). Animals with the implant placed as a flat sheet adjacent to the Bruch's membrane were selected for further analysis. Histological analysis at 1-month post-implantation showed the presence of a well-pigmented, intact iPSC-RPE cell layer attached to the parylene substrate in all transplanted eyes. No major signs of inflammation were observed in any of the implanted animals. A majority of the retinas (92.0%) maintained the basic retinal architecture without noticeable structural changes. The histological analysis revealed that transplanted cells survived very well, evidenced by TRA-1-85 (human specific marker) expression and retinol-binding protein expression (Figure 2e,f). Transplants in which iPSC-RPEs were present on the lower surface of the parylene membrane also showed good survival (Figure 2e). At the 1-month time point, the expression of CD68 (macrophage marker) and GFAP (retinal glial marker) was not observed in the transplant area (Supplementary Figure S1) or in the area outside the transplant (data not shown). The cells retained RPE 65 expression and human marker expression (Tra-1-85) without any evidence of mesenchymal marker expression (vimentin and αsmooth muscle actin, see Supplementary Figure S2). Good survival of iPSC-RPE implants was also observed at 4 months post-implantation (Figure 2g–i). Based on rhodopsin staining, the surviving cells performed a phagocytic function (Figure 2i). The absence of immunological markers comparable to the control group indicate the absence of detectable chronic inflammation induced by xenografts.

**Figure 1.** iPSC-RPE grown as polarized monolayer over parylene substrate. (**a**) iPSC-RPE polarized monolayer cultured on 24-well transwell insert, (**b**) enlarged view of iPSC-RPE monolayer, (**c**) ultrathin parylene membrane without cells, (**d**) iPSC-RPE grown as polarized monolayer on parylene membrane, low magnification (10×) image showing the whole implant stained for **(e)** RPE 65 (**f**) ZO-1 expression, enlarged view of expression of (**g**) RPE 65 and (**h**) ZO-1 on parylene membrane.

*3.3. In Vivo Assessment of Long-Term Transplant Effects in Immunodeficient RCS Rats (11 Month Study)*

Evaluation of histology images from serial sections at 11 months post-implantation showed the presence of transplanted RPE in seven eyes (7/15). Out of these seven eyes, four eyes retained an intact RPE monolayer structure. Immunostaining showed rhodopsincontaining phagosomes in the transplanted RPE. This was more prominent in eyes in which better preservation of the iPSC-RPE monolayer structure was observed (Figure 3c). In the remaining three eyes, the cells appeared as clumps (Figure 3e,f) out of which only two eyes retained RPE65 expression. There was no Ki 67 expression in the implanted areas, suggesting an absence of proliferative cells. Photoreceptor outer nuclear layer (ONL) preservation was evident in almost all eyes in which strong RPE65 expression was noticed (Figure 3c,f). A complete loss of transplanted cells was noticed in eight (8/15) eyes.

ONL preservation was not observed in the eyes in which iPSC-RPE survival was absent. The presence of fibrosis was noticed in the majority of the above eyes (Figure 4a,b). A summary of the histological result of the 11-month post-implantation study is given in Table 2.

**Figure 2.** Short-term assessment of iPSC-RPE implant survival and functionality, in immunodeficient RCS rats. (**a**) iPSC-RPE implants observed during fundus examination of immunodeficient RCS rats at 1 month post-implantation, (**b**) enlarged view, (**c**) vertical OCT b-scan image through the transplant area, (**d**) HE image showing subretinal implant placement. The choroidal layer that appears to be separated from the remaining retina is considered a histologic artifact. Yellow asterisk indicates iPSC-RPE cells; Red arrow heads indicate the parylene membrane (**e**) transplant is identified by TRA-1-85 (human specific marker, red) and RPE65 (green) expression; Red arrowhead indicates the parylene substrate; white rhombus represents endogenous rat RPE, yellow asterisk indicates RPE on Parylene membrane (**f**) Rhodopsin (red) and retinol-binding protein (RBP1, green) staining to demonstrate that implanted iPSC-RPE can phagocytose photoreceptor outer segments (white arrows). Inset is a higher magnification of the above area. Red arrowhead indicates the parylene substrate; white rhombus represents endogenous rat RPE; (**g**) HE image showing subretinal implant 4 months after transplantation; (**h**) transplant at 4 months is identified by TRA-1-85 (human specific marker, red) (**i**) Rhodopsin (red) and RPE 65 (green) staining were used to show that implanted iPSC-RPE can phagocytose photoreceptor outer segments at 4 months after transplantation. Inset is a higher magnification of the transplant area indicating phagocytosis (white arrow).

**Figure 3.** Representative HE and immunostaining images of immunodeficient RCS rats implanted with iPSC-RPE monolayer assessed at 11 months post implantation. Large white arrows (**a**,**d**) point to the parylene membrane. (**a**–**c**) Retina containing iPSC-RPE monolayer, (**d**–**f**) retina with iPSC-RPE appeared as multiple cell layers or cell clumps. TRA-1-85 (white triangle in (**b**,**e**) and RPE65 expression were used for identifying iPSC-RPE. Absence of Ki67 expression indicates absence of proliferative cells (**b**,**e**). Rhodopsin immunostaining is used to identify photoreceptor survival (yellow arrows in figure (**c**,**f**). Rhodopsin-containing phagosomes are found in the transplanted iPSC-RPE denoted by white arrows (**c**,**f**). Phagocytic activities were prominent in eyes in which monolayer structure was better preserved (**c**).

**Table 2.** Summary of the histological result for iPSC-RPE implantation in immunodeficient RCS rats (11-month post implantation).


The expression of CD68 and GFAP was used to analyze inflammatory and glial reactions to donor tissues. GFAP was strongly expressed in the ganglion cell layer, the inner nuclear layer, and the choroid area, but was absent in the transplant area (Figure 4c). CD68 positivity observed in some of the implanted eyes suggests inflammatory reactions associated with transplantation (Figure 4d). To validate the cell loss associated with the loss of tight junctions and consequent loss of cell–cell contact and cell–matrix contact, the tissue was tested for classic EMT markers—α smooth muscle actin (α SMA) and vimentin. Interestingly, in the implants in which RPE expression was absent or feeble, there was a strong expression of α smooth muscle actin (Figure 4e), and vimentin (Figure 4f) was also observed.

#### *3.4. Preservation of Low Light Level Visual Responses in the Superior Colliculus (SC) of iPSC-RPE-Implanted Rats at 11-Month Post-Implantation*

iPSC-RPE-implanted immunodeficient RCS rats were subjected to SC luminance threshold mapping (Figure 5). Electrophysiological mapping of the SC allowed for correlation of the response area in the SC with the location of the implant placement in the eye based on the established retinocollicular map properties [32]. Age-matched normal

Long Evans (LE) rats showed visual activity from all over the SC (Figure 5a). Among transplanted rats, visual preservation was observed in only five rats (5/8).

**Figure 4.** Representative HE and immunostaining images of immunodeficient RCS rat retinas implanted with iPSC-RPE monolayer assessed at 11 months post-implantation. Presence of fibrosis, immunoreactivity, and epithelial–mesenchymal transition (EMT) was assessed. (**a**) Retina with no surviving iPSC-RPE showing signs of inflammation and peri-membrane fibrosis indicated by white asterisk. (**b**) Absence of TRA-I-85 staining. (**c**) Retinas showing RPE65 expressing iPSC-RPE cells (white arrows) labelled for GFAP (glial cells), (**d**) CD68 (macrophages/microglia), (**e**) expression of classical mesenchymal markers α smooth muscle actin α SMA and vimentin. (**f**) Images in which iPSC-RPE monolayer appears to be present below the parylene membrane are either due to orientation difference in the implant placement or due to the survival of the RPE on the lower side of the parylene membrane.

Visually evoked activities in these rats were observed only in a small SC area corresponding to the implant placement in the eye. Visual activities were robust (higher light sensitivity) in two of the above rats (Figure 5b) whereas only weak visual activity (lower light sensitivity) was recorded in the remaining three rats (Figure 5c). No light-evoked visual activity was observed in the SC of age-matched non-transplanted rats (Figure 5d). All five rats that showed SC visual activity showed a presence of transplanted RPE in their eyes (Table 2). No correlation was observed between the light-sensitivity threshold and the degree of transplant survival.

**Figure 5.** Visual activities recorded from the SC of 11-month-old immunodeficient RCS rats. Map properties of SC-evoked responses from individual rats are represented by colored asterisks. Larger asterisks show higher light sensitivity in the SC. (**a**) Age-matched normal rat. (**b**) **\* Rat # 6005 and \* Rat # 6012**. (**c**) **\* Rat # 6001**, \* **Rat # 6006**, and **\* Rat # 6011**. Based on morphological examination, all these rats showed surviving iPSC-RPE in the retina (see Table 2). (**d**) No light-evoked visual activity was observed in the remaining transplanted rats and age-matched control RD rats.

#### *3.5. Optokinetic (OKN) Responses in iPSC-RPE-Implanted Rats*

Based on OKN data, visual improvement in the iPSC-RPE-implanted eyes was observed at 4 months post-transplantation (Figure 6). However, when tested at the 11-month time point, no measurable OKN responses were observed in any of the iPSC-RPE-implanted rats (data not shown).

**Figure 6.** OKN testing data based on the duration of head-tracking recorded from 4-month-old immunodeficient RCS rats (±SE). The data show improved head-tracking response in the iPSC-RPE transplanted left eyes (*n* = 12) compared to the non-transplanted eyes and age-matched control rats (*n* = 5).

#### **4. Discussion**

Ongoing multicentral clinical studies have proven that stem-cell-derived RPE transplantation is a practical option to restore failing vision in retinal dystrophies [12,13,19,24]. Previous animal studies and pilot data from our Phase l/ll clinical studies demonstrated the feasibility of using ultrathin parylene as a bio membrane for hESC-RPE growth and subretinal implantation [12,15,17,33]. Compared to hESC-RPE, using iPSC-derived RPE is considered more advantageous due to the possibility of generating a sufficient number of autologous RPE cells that strongly resemble primary human RPEs and its potential to minimize issues associated with immune rejection. Based on this, the present study evaluated the long-term benefits of polarized iPSC-RPE cells grown on an ultrathin parylene membrane that can act as an artificial Bruch's membrane. The ability of such implants to support transplant survival and viability of the host photoreceptors to preserve visual function is demonstrated in a new immunodeficient RCS rat model.

The majority of the previous iPSC-RPE transplantation studies were conducted in immunosuppressed animal models, and assessments were made for a short duration only, which may not be sufficient to extrapolate into long-term implications [9,34–36]. Hence, in the present study, transplant effects were analyzed in a new immunodeficient RCS rat model, and assessments were conducted up to one year after implantation. The results from this study demonstrated the safety and potential bioactivity of the iPSC-RPE implant both during the short-term (1–4 months) investigation and long-term investigation (11-month study). Based on histology assessments, good coverage of the implanted iPSC-RPE on the parylene membrane was observed in the majority of eyes up to 4 months post-implantation along with improvement in visual function confirmed by OKN testing. In our long-term studies (11 months post-transplantation), iPSC-RPE survival and phagocytic function were only observed in less than 50% of the transplanted rats (7/15). In another report, a loss of transplanted iPSC- RPE in RCS rats, immunosuppressed by oral administration of cyclosporin, was observed at 13 weeks post-transplantation [36]. Based on the available data, the loss of transplanted RPE over the course of time and alteration in the monolayer structure can be related to the immune reaction to xenografts [19,37,38].

Previous studies have shown that the survival and integration of transplanted ESC-RPEs in the pathologic environment of a diseased retina is challenging due to it being prone to attack by macrophages [15]. CD 68 expression has been reported in studies using hESC-RPE transplantation and hESC-RPE cell suspension injections in immunosuppressed RCS rats [15,38,39]. In the present investigation, we used immunodeficient RCS rats to reduce immunological issues. However, signs of inflammation and microglia activation have been previously reported in immunodeficient RCS rats [28]. Hence, we used CD68 as a marker for assessing reactive microglia in the transplanted eyes. In our long-term studies, some CD 68 expression was observed in the implants and in areas adjacent to them (Figure 5d). However, this phenomenon was not observed at the early timepoint (1-month post-implantation, see Supplementary Figure S1). This suggests that reactive microglia/macrophages can play a role in the loss of transplanted RPEs at a later time point. In contrast to the above reports, a recent study by Zhu et.al suggested that an iPSC-PRE cell suspension injection can lower the microglial activation (CD68 expression) in rd10 mice [34]. The discrepancies in the study outcomes may be related to the differences in the animal models used, the time points in which CD68 staining was conducted, and the cell types used for transplantation experiments.

Glial fibrillary acid protein (GFAP) expression, which is known to occur in response to retinal injuries [40], can be also suggested to play a role in transplant loss in the 11-month study group. However, GFAP expression in the above study group was found to be mostly adjacent to the inner nuclear region, choroid area, and ganglionic layer, which is far from the implant area. Since this GFAP expression pattern is comparable to that of the nontransplanted control eyes [41], the presence of GFAP cannot be correlated to the loss of transplanted iPSC-RPEs.

In some of our transplanted eyes, the iPSC-RPEs developed into cell clumps on the surface of the parylene membrane (Figure 3d). Previous studies suggested that when the RPE transplant fails to establish a monolayer and form cell clumps, its survival will be poor and the cells will not be capable of performing normal RPE functions [42,43]. Based on this finding, we suggest that implants may need to be microscopically examined for potential signs of clumping prior to subretinal implantation.

The cell clump formation observed in about 20% of the implanted eyes (long-term study group) might have occurred even after transplantation due to cell migration. Cell migration is mainly attributed to the loss of RPE tight junctions [44]. This can lead to a loss of cell-to-cell contact and anchorage dependence, which are critical for RPE survival and functionality [45]. Emerging evidence demonstrates that RPE cells can be less differentiated and undergo the epithelial–mesenchymal transition (EMT) and enhanced migration in retinal degenerative diseases, including macular degenerations and proliferative vitreoretinopathy [46–49]. Such a transition is also reported in higher-passage RPEs during in vitro observations [50,51]. Immunostaining of transplanted eyes from the 11-month study group revealed the presence of two classic mesenchymal cell markers, namely α SMA and vimentin, especially in areas of the parylene membrane where a loss of RPE expression was noticed. The transition to a mesenchymal fate may cause a loss of tight junctions and reduced cell adherence to the parylene membrane that can lead to a fibroblastic phenotype [44,45]. According to Zhou et al. [52], RPE cells retain the reprogramming capacity to move along a continuum between polarized epithelial cells and mesenchymal cells. This shift towards a mesenchymal phenotype can be defined as RPE dysfunction [52]. This change of RPE characteristics can cause senescence/fibrosis, eventually resulting in a loss of transplanted cells. In our transplanted rats, the expression of EMT markers was not evident at the earlier time point (1-month study group) when RPE survival was more robust (Supplementary Figure S2). Further studies are needed to identify the exact time point at which the EMT markers are expressed to determine whether changes in the iPSC-RPEs take place only in the long-term post-implantation period.

The RCS retina is widely known for its acute reactions to surgical interventions. Surgical trauma in rat eyes of a severe nature as a result of their small size can lead to increased tissue reactions in the implanted area. In support of this, mild inflammation and peri-membrane fibrosis were visible around the majority of the implants, in which the RPE monolayer was lost. It may be noted that, although the immunodeficient RCS rat is T-cell deficient, they possess bone-marrow-dependent B cells and natural killer (NK) cells. All of the above factors can contribute to the loss of transplanted iPSC-RPE cells.

In the long-term study group (11-month post-implantation), the visual functional preservation (based on SC electrophysiology) was correlated to the survival of the transplanted iPSC-RPE. However, no considerable OKN visual activities were observed in these rats at this time point. Previously, in hESC-RPE-implanted immunosuppressed RCS rats, a progressive loss of OKN responses has been reported [17]. Improved OKN visual activities observed at an earlier time point (4 month) may be attributed to residual photoreceptors present in the RCS retina. When the photoreceptor degeneration is more advanced, the transplant benefit can be limited to a very small area of the retina and its contribution may not be strong enough to evoke measurable head-tracking activities.

In conclusion, the present study demonstrated the survival and functionality of iPSC-RPE transplanted as a polarized monolayer on a non-degradable substrate containing similarities to an artificial Bruch's membrane. The transplant benefits are higher during the earlier post-implantation period. Progressive deterioration of the transplant benefits observed in this study was correlated with the loss of transplanted iPSC-RPEs. The immune reactions and subretinal fibrosis can be considered the major causes of the loss of transplanted iPSC-RPE. From a clinical perspective, many of these adverse effects can be less severe in humans due to the differences in the eye architecture, surgical procedures, and the nature of the disease microenvironment. Moreover, in human eyes, easy application of target-specific and effective immune suppressants can help to reduce potential immunological reactions.

**Supplementary Materials:** The following are available online at https://www.mdpi.com/article/ 10.3390/cells10112951/s1, Figure S1: Representative images of immunodeficient RCS rat retinas implanted with iPSC-RPE monolayer (shown by RPE65 expression) cultured on a parylene membrane assessed at 1-month post implantation, Additional labelling of the retina sections was performed for expression of (**a**) CD68 (macrophages/microglia) and (**b**) GFAP (glial cells), Figure S2: Representative images of immunodeficient RCS rat retinas implanted with iPSC-RPE monolayer (shown by RPE65 expression) cultured on a parylene membrane assessed at 1-month post-implantation. Additional labelling of the retina sections was performed for expression of classical mesenchymal markers (**a**) vimentin and (**b**) α smooth muscle actin (α SMA).

**Author Contributions:** Conceptualization, B.B.T.; methodology, D.Z., D.S.R.N., J.C.M.C. and B.B.T.; validation B.B.T., K.B., M.S.H., J.C.M.C. and D.S.R.N.; resources, B.B.T., K.B., M.S.H., D.R.H. and R.S.; writing—original draft preparation, D.S.R.N. and B.B.T., writing—review and editing, B.B.T., D.S.R.N., D.Z., J.C.M.C., K.B., D.R.H., R.S. and M.S.H.; supervision B.B.T., funding acquisition, B.B.T. All authors have read and agreed to the published version of the manuscript.

**Funding:** This study was funded by a grant from the Bright Focus Foundation (M2016186, Thomas, PI). Research reported in this publication was supported by the National Eye Institute of the National Institutes of Health under Award Number P30EY029220. CIRM (California Institute for Regenerative Medicine) grants (DISC1-09912 PI-Thomas, DR3-07438-PI-Humayun), Unrestricted Grant to the Department of Ophthalmology from Research to Prevent Blindness, New York, NY. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

**Institutional Review Board Statement:** All experiments were approved by the University of Southern California Animal Care and Use Committee and were performed in accordance with the National Institute of Health Guide for the Care and Use of Laboratory Animals and the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. Ethical approval codes: 21068, BUA-14-00059 and 2020-3.

**Informed Consent Statement:** Not Applicable.

**Data Availability Statement:** The data presented in this study are available on request from the corresponding author.

**Acknowledgments:** We thank Jane Lebkowski (Regenerative Patch Technologies, Portola Valley, CA) for critically reviewing the manuscript. The authors want to thank Xiaopeng Wang (USC) for histological processing of the tissue samples.

**Conflicts of Interest:** Regenerative Patch Technologies: MSH and DZ have proprietary interests in the culture of RPE on ultrathin parylene.

#### **References**

