www.aging-us.com AGING 2020, Vol. 12, No. 14 Research Paper Activation of C-reactive protein proinflammatory phenotype in the blood retinal barrier in vitro: implications for age-related macular degeneration Sara Romero-Vázquez1, Alfredo Adán1, Marc Figueras-Roca1, Victor Llorenç1, Mark Slevin2, Gemma Vilahur3, Lina Badimon3, Andrew D. Dick4,5, Blanca Molins1 1 Group of Ocular Inflammation, Clinical and Experimental Studies, Institut d’Investigacions Biomèdiques Agustí Pi i Sunyer (IDIBAPS), Hospital Clínic de Barcelona, Barcelona, Spain 2 Department of Life Sciences, Manchester Metropolitan University, Manchester, UK 3 Cardiovascular Research Center-ICCC, Hospital de la Santa Creu i Sant Pau, IIB-Sant Pau, CiberCV, Institute Carlos III, Barcelona, Spain 4 Academic Unit of Ophthalmology, School of Clinical Sciences and School of Cellular and Molecular Medicine, University of Bristol, Bristol, UK 5 National Institute for Health Research (NIHR) Biomedical Research Centre at Moorfields Eye Hospital and University College London Institute of Ophthalmology, London, UK Correspondence to: Blanca Molins; email: bmolins@clinic.cat Keywords: age-related macular degeneration, retinal pigment epithelium, inflammation, C-reactive protein Received: April 9, 2020 Accepted: June 20, 2020 Published: July 16, 2020 Copyright: Romero-Vázquez et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY 3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. ABSTRACT The retinal pigment epithelium (RPE) is considered one of the main targets of age-related macular degeneration (AMD), the leading cause of irreversible vision loss among the ageing population worldwide. Persistent low grade inflammation and oxidative stress eventually lead to RPE dysfunction and disruption of the outer blood- retinal barrier (oBRB). Increased levels of circulating pentameric C-reactive protein (pCRP) are associated with higher risk of AMD. The monomeric form (mCRP) has been detected in drusen, the hallmark deposits associated with AMD, and we have found that mCRP induces oBRB disruption. However, it is unknown how mCRP is generated in the subretinal space. Using a Transwell model we found that both pCRP and mCRP can cross choroidal endothelial cells and reach the RPE in vitro and that mCRP, but not pCRP, is able to cross the RPE monolayer in ARPE-19 cells. Alternatively, mCRP can originate from the dissociation of pCRP in the surface of lipopolysaccharide-damaged RPE in both ARPE-19 and primary porcine RPE lines. In addition, we found that the proinflammatory phenotype of mCRP in the RPE depends on its topological localization. Together, our findings further support mCRP contribution to AMD progression enhancing oBRB disruption. INTRODUCTION multiple genetic and environmental factors, age being the primary risk factor. Age-related macular degeneration (AMD) is the primary cause of irreversible vision loss among the The retinal pigment epithelium (RPE) monolayer is ageing population worldwide. The number of people believed to be among the initial targets of early disease. with AMD worldwide is expected to reach 196 million AMD presents RPE cell abnormalities, disruption of the in 2020, increasing to 288 million in 2040 [1]. AMD is outer blood-retinal-barrier (oBRB), and degeneration of a degenerative and progressive disease involving photoreceptors, which require a normally functioning www.aging-us.com 13905 AGING RPE to survive [2, 3]. Prior research has implicated variant can effectively bind to mCRP to dampen mCRP strong roles for inflammation, oxidative stress, lipid pro-inflammatory activity [27]. Notably, FH from AMD abnormalities, and RPE dysfunction in AMD patients carrying the risk polymorphism for AMD pathobiology, but their precise mechanisms and their shows an impaired binding to mCRP and, therefore, its relative contribution are unclear [4]. A multitude of proinflammatory effects remain unrestrained [28]. In systemic changes occur with ageing that contribute line with these findings, data demonstrates that mCRP to the initiation and development of inflammation. is the more abundant form of CRP in human RPE- Indeed, the immune system of elderly individuals is choroid [29], and that mCRP levels are elevated in characterized by a basal systemic inflammatory state [5]. individuals with the high-risk CFH genotype [29, 30]. Altered immune responses are thought to contribute to If mCRP pro-inflammatory capacity is unrestrained in the dry AMD phenotype. Parainflammation is a low- AMD and particularly in high risk patients, then we grade cytoprotective adaptation to local stress that is need to determine how mCRP is generated or intermediate between immune-mediated homeostasis accumulates in the subretinal space as there is no CRP and chronic inflammation that maintains cellular and transcription in the retinal tissue [30, 31]. In addition, it tissue function. Loss of parainflammation control is also unclear whether mCRP-induced barrier contributes to AMD by invoking a chronic, heightened disruption depends on its topological localization. immune response that causes tissue destruction [6–8]. Histochemical and proteomic analysis of ocular drusen, RESULTS the hallmark deposits of AMD, have shown that these deposits contain inflammatory proteins and complement Choroidal endothelial cells allow diffusion of CRP components that mediate local inflammation [9, 10]. isoforms Furthermore, the strongest genetic risk factor for AMD known to date is a common polymorphism in the We first interrogated whether circulating CRP could complement factor H (CFH) gene (c.1277T > C, reach the subretinal space using a Transwell model, in p.Tyr402His), a gene essential for the regulation of which confluent monolayers of primary porcine complement activation [11, 12]. choroidal endothelial cells (CECs) were grown on porous filters with their apical and basolateral surfaces C-reactive protein (CRP), a prototypical acute-phase exposed to separate chambers (Figure 1A). Addition of reactant, is an active regulator of the innate immune mCRP to the apical chamber that mimics blood vessel system. Among the multiple functions ascribed to CRP lumen (A to B red arrow in Figure 1A) resulted in CRP are activation of the classical complement pathway and diffusion into the basolateral chamber (tissue side) as inactivation of the alternative pathway [13]. CRP is Western blot (Figure 1B) and ELISA (Figure 1C) of the considered to be a serum biomarker for chronic culture media of the different compartments revealed inflammation, heart disease and, more recently, also the presence of mCRP in both chambers. Similarly, AMD [14, 15]. In plasma, CRP typically exists as a pCRP was able to reach the abluminal side of the CEC cyclic, disk-shaped pentamer (pentameric CRP, pCRP) monolayer, as seen by Western blot (Figure 1D). CRP composed of five noncovalently linked subunits of 23 isoforms were also able to reach the apical chamber kDa [16]. However, pCRP can undergo dissociation into when added in the abluminal compartment (B to A blue its subunits, acquiring distinct biological functions. arrow in Figure 1A), suggesting bidirectional diffusion Oxidative stress and bioactive lipids from activated or of the proteins. Immunofluorescence imaging showed damaged cells can dissociate pCRP into its 23-kDa that mCRP delivered to the apical compartment was subunits [17–19] through a mechanism that is dependent extensively bound to the CEC surface compared to on lysophosphatidylcholine (LPC) exposure after pCRP and to CRP (either mCRP or pCRP) delivered in phospholipase A2 activation [20]. This alternative the basolateral chamber (Figure 1E–1G). conformation of CRP, termed monomeric CRP (mCRP), has different antigenicity-expressing neoepitopes than Diffusion of CRP across the RPE pCRP and represents the tissue-based insoluble form of CRP. Unlike pCRP, mCRP displays a proinflammatory Given that CRP isoforms were able to cross the CEC phenotype in several cell types [21–23]. monolayer in our in vitro model, we next evaluated whether CRP isoforms could also reach the subretinal mCRP has been identified in ocular drusen and other space and cross the RPE, using the Transwell model. In subepithelial deposits [24, 25], as well as in the choroid, this scenario the basolateral side of the RPE monolayer and we have shown that mCRP, but not pCRP, represents the Bruch’s membrane/choriocapillaris side, contributes to oBRB disruption in vitro [26]. Moreover, whereas the apical side represents the subretinal space we also showed that the “non-risk” Factor H (FH) (Figure 2A). Western blot experiments revealed that www.aging-us.com 13906 AGING Figure 1. CRP isoforms are able to cross CECs. (A) Experimental setup. CRP (10 µg/ml) was added to either the apical or basolateral chamber of the Transwell for 48h, mimicking blood vessel lumen and RPE, respectively. The presence of CRP in the opposite chamber where it was added was determined by Western blot and ELISA, and CRP bound to the cell surface was determined by immunofluorescence. (B) Western blot of mCRP present in apical (Up) and basolateral (Down) chamber (N=4). (C) ELISA of mCRP (ng/ml) from apical (Up) and basolateral (Down) supernatants. Values are expressed as mean ± SD (N=3). (D) Western-blot of pCRP present in apical (Up) and basolateral (Down) supernatants (N=5). (E) Immunofluorescence of CRP (red) stained with monoclonal antibodies against mCRP (3H12) or pCRP (1C6). Nuclei stained with DAPI. Scale bar = 50 µm (N=6). (F) Quantification of CRP binding measured as stained area divided by the number of cells per image (µm2/cell). Results are expressed as mean area (µm2/cell) ± SD. Statistical analysis was performed by One-Way ANOVA and Tukey’s posthoc. **P<0.01 vs. all conditions. (G) Reconstruction of x-z sections with a 0.3 µm z axis step of immunofluorescence images. Images shown are representative of six independent experiments. www.aging-us.com 13907 AGING mCRP was able to diffuse across ARPE-19 cell Topological localization of mCRP determines the monolayer, as it was present in the apical chamber when impact on barrier disruption in RPE cells added in the basolateral chamber (B to A, red arrow in Figure 2A). Diffusion of mCRP was greater at 48 hours Next, we evaluated whether mCRP-induced barrier compared to 24 hours (Figure 2B). We also detected the disruption depended on the topological localization of presence of mCRP by ELISA 48 hours after treatment mCRP. For this purpose, ARPE-19 cells grown on in the opposite chamber where it was added (Figure inserts for at least 3 weeks were treated with CRP 2C). By contrast, pCRP did not seem to cross the isoforms (10 µg/mL) either in the apical or basolateral ARPE-19 monolayer. Western blot experiments showed compartment for 48h. As expected, mCRP delivered in that pCRP was not present in the supernatant of the the apical chamber, significantly decreased TEER opposite chamber where it was added, neither at 24h nor values (Figure 5A). This observation was accompanied at 48h after treatment (Figure 2D). Immunofluorescence by an increase in paracellular permeability (Figure 5B) imaging showed that mCRP delivered to the apical and an increased ZO-1 disorganization (Figure 5C, 5D, compartment was extensively bound to the ARPE-19 Supplementary Figure 3). Interestingly, abluminal cell monolayer compared to pCRP and to CRP (either treatment of mCRP also induced a significant decrease mCRP or pCRP) delivered in the basolateral chamber in TEER values. (Figure 2E–2G). We then aimed to replicate the experiments in primary We then tested whether mCRP was also able to cross porcine RPE cells, which are less permeable to mCRP porcine primary RPE cells. Primary RPE cells represent diffusion. As seen in Figure 6, apical treatment of a healthier and younger RPE than ARPE-19 cells, as mCRP, but not pCRP, induced barrier disruption also in they show more than 5 times higher TEER values primary RPE cells as seen by significant decrease in (Supplementary Figure 1D, 1E). Interestingly, unlike TEER (Figure 6A), increase in paracellular permeability ARPE-19 cells, primary porcine RPE cells did not allow (Figure 6B) and increased ZO-1 disorganization (Figure mCRP diffusion as mCRP was not detected in the 6C, 6D, Supplementary Figure 4). However, when opposite chamber where it was added, neither in mCRP was delivered into the abluminal compartment it Western blot (Figure 3A) nor ELISA (Figure 3B). As failed to induce barrier disruption, showing that mCRP- expected, pCRP was also unable to cross the RPE induced barrier disruption depends on its topological monolayer (Figure 3C). Immunofluorescence imaging localization. showed similar results to those with ARPE-19; mCRP delivered to the apical compartment was extensively DISCUSSION bound to the RPE cell monolayer compared to pCRP and to CRP (either mCRP or pCRP) delivered in the The present study aimed to understand the relative basolateral chamber (Figure 3D–3F). contribution of mCRP to the pathophysiology of AMD. Our in vitro work demonstrates that mCRP is capable of Damaged RPE dissociates pCRP into mCRP traversing through choroidal vascular endothelium and across RPE. Although no direct in vivo correlate, the Given that pCRP is able to cross the CEC monolayer and data suggest that mCRP can reach the subretinal space. reach the subepithelial space in vitro, we next studied Alternatively, mCRP may derive from the dissociation whether pCRP could dissociate into its monomeric of pCRP on the surface of damaged RPE. Moreover, we subunits within the RPE. It has been previously described found that the proinflammatory phenotype of mCRP in that LPS-induced inflammation induces CRP dissociation the RPE depends on its topological localization. in the cremaster muscle [20], and therefore we studied Together the data continues to build the evidence of whether LPS-induced inflammation could also lead to mCRP accentuation of AMD pathology and detriment CRP dissociation in RPE cells. RPE cells were treated to RPE health. with 100 µg/mL LPS for 24h before adding pCRP. After 24h, RPE cells were treated with pCRP for 48h and the CRP is mainly produced in the liver and, although presence of mCRP on the surface of RPE cells extrahepatic synthesis has been reported in some tissues, was measured by immunofluorescence. As observed in no evidence of CRP gene transcription has been detected Figure 4, LPS-induced inflammation triggered pCRP in the retinal tissue [30, 31]. This indicates that the dissociation into mCRP in both ARPE-19 (Figure 4A, systemic circulation is the main source of CRP in the 4B) and primary porcine RPE cells (Figure 4C, 4D). sub-RPE deposits. Indeed, we observed that both CRP Altogether these results show that mCRP present in isoforms, at clinically relevant concentrations, are able to drusen and in the subretinal space may either arrive from cross CECs from their apical side -simulating blood the choroidal circulation or it may originate from local side- and reach the basolateral side of the endothelium - dissociation of pCRP in damaged RPE. simulating the subepithelial side- in a Transwell model www.aging-us.com 13908 AGING Figure 2. Diffusion of CRP isoforms across ARPE-19 cells. (A) Experimental setup. CRP (10 µg/ml) was added to either the apical or basolateral chamber of Transwell for 48h, mimicking neural retina and choriocapillaris, respectively. The presence of CRP in the opposite chamber where it was added was determined by Western blot and ELISA, and CRP bound to the cell surface was determined by immunofluorescence. (B) Western blot of mCRP present in apical (Up) and basolateral (Down) supernatants after 24 and 48 hours of treatment (N=4). (C) ELISA of mCRP (ng/ml) from apical (Up) and basolateral (Down) supernatants 48 hours after treatment. Values are expressed as mean ± SD (N=3). (D) Western blot of pCRP present in apical (Up) and basolateral (Down) supernatants after 24 and 48 hours of treatment (N=4). (E) Immunofluorescence of CRP (red) stained with monoclonal antibodies against mCRP (3H12) or pCRP (1C6). Nuclei stained with DAPI. Scale bar = 50 µm (N=6). (F) Quantification of CRP binding measured as stained area divided by the number of cells per image (µm2/cell). Results are expressed as mean area (µm2/cell) ± SD. Statistical analysis was performed by One-Way ANOVA and Tukey’s posthoc. * P<0.05, ** P<0.01 vs. mCRP apical. (G) Reconstruction of x-z sections with a 0.3 µm z axis step of immunofluorescence images. Images shown are representative of six independent experiments. www.aging-us.com 13909 AGING Figure 3. Diffusion of CRP isoforms across primary porcine RPE cells. (A) Western blot of mCRP present in apical (Up) and basolateral (Down) supernatants 48 hours after addition of mCRP (N=4). (B) ELISA of mCRP (ng/ml) from apical (Up) and basolateral (Down) supernatants. Values are expressed as mean ± SD (N=5). (C) Western blot of pCRP present in apical (Up) and basolateral (Down) supernatants 48 hours after treatment (N=3). (D) Immunofluorescence of CRP (red) stained with monoclonal antibodies against mCRP (3H12) or pCRP (1C6). Nuclei stained with DAPI. Scale bar = 30 µm (N=3). (E) Quantification of CRP binding measured as stained area divided by the number of cells per image (µm2/cell). Results are expressed as mean area (µm2/cell) ± SD. Statistical analysis was performed by One-Way ANOVA and Tukey’s posthoc. **P<0.01 vs. all conditions. (F) Reconstruction of x-z sections with a 0.3 µm z axis step of immunofluorescence images. Images shown are representative of three independent experiments. www.aging-us.com 13910 AGING (Figure 1). The choroidal endothelium is fenestrated, especially in regions containing drusen-like deposits which allows the movement of macromolecules and [30]. A similar study compared differences in total CRP nutrients to nourish the RPE cells [32]. Although mCRP immunoreactivity in the retina based on AMD status and could extravasate the endothelium and reach the found that early and wet AMD eyes had higher levels of subepithelial space, we observed by immunofluorescence CRP compared to controls and that CRP was primarily that most mCRP was retained in the apical side of the detected into the BM [34]. In their work, Chirco et al. endothelium. Indeed, mCRP promiscuously interacts also showed that mCRP exerts an inflammatory effect on with a variety of immunoglobulins and other proteins CEC, as it increases CEC migration and paracellular [33]. Thus, we tested whether mCRP in the subepithelial permeability and upregulates inflammatory gene space could also originate from the dissociation of expression including ICAM1, suggesting a role for pCRP in the RPE. Using the approach of Thiele et al. mCRP in promoting inflammation in the choroid. [20], we found that LPS-induced inflammation lead to pCRP dissociation also in RPE cells (Figure 4). Besides the proinflammatory effect in the choroid, we Mechanistically, this process is dependent on exposure have previously demonstrated that clinically relevant of LPC, a bioactive lipid that is generated after concentrations of mCRP induce barrier disruption and phospholipase A2 activation on activated cell membranes have a proinflammatory effect in RPE in vitro and [20]. Chirco et al. found that mCRP is predominantly potential for driving angiogenesis [26, 27]. With respect localized in the choriocapillaris and Bruch´s membrane to angiogenesis and neovessels, in the context of [29] and a previous work by Johnson and colleagues, vascular disease, CRP inhibits VEGF production and looking at total CRP, showed that CRP was more angiogenesis [35, 36]. Conversely, others have shown abundant in donor eyes with the high-risk CFH CRP upregulates VEGF expression in adipose-derived polymorphism compared to age-matched controls, stem cells and in monocytes [37, 38]. mCRP has been Figure 4. LPS-induced inflammation promotes CRP dissociation in RPE cells. RPE cells were treated with 100 µg/mL LPS for 24h before adding pCRP. After 24h, RPE cells were treated with pCRP for 48h and the presence of mCRP on the surface of RPE cells was measured by immunofluorescence. mCRP immunostaining of ARPE-19 (A) and primary porcine RPE (B) cells treated with 10 µg/ml mCRP for 48h (I, IV), 25 µg/ml pCRP for 48h (II, V), or 100 µg/ml LPS 24h before treatment with 25 µg/ml pCRP for 48h (III, VI). Arrows point mCRP dissociated from pCRP on RPE surface. Nuclei stained with DAPI. Scale bar = 50 µm. Images shown are representative of three independent experiments. (C, D) Quantification of CRP dissociation measured as stained area with the monoclonal antibody 3H12 against mCRP (green) divided by the number of cells per image (µm2/cell). Results are expressed as mean area (µm2/cell) ± SD (N=3). Statistical analysis was performed by student t-test. *P<0.05 vs. pCRP. www.aging-us.com 13911 AGING localized around newly formed microvessels in carotid had no effect on barrier disruption when added from the artery plaques and in peri-infarct regions after an acute basolateral side, than in ARPE-19 cells. Indeed, we ischemic stroke [39, 40], promoting angiogenesis and observed by immunofluorescence a preferential binding inducing inflammation [41]. One notion is understanding of mCRP to the apical side (Figures 2, 3). The polarized of mCRP role is context dependent and in vivo studies proinflammatory effect of mCRP has been already are required to determine the contribution of mCRP to observed in endothelial cells using a similar approach neovessel formation in the context of AMD, either via [42]. These observations could be due to a polarized VEGF dependent or -independent mechanisms. On the distribution of the surface sensors for mCRP in the cell other hand, Lauer et al. showed that mCRP binds surface. necrotic RPE cells and that complement regulation at necrotic cell lesions is impaired by the FH His402 risk The receptors that mediate mCRP activities have not variant [17]. In these studies, RPE cells were stimulated been fully characterized. In human neutrophils, mCRP with mCRP from the apical side. However, given the binds FcγRIII (CD16) [21]. However, functional polarized nature of RPE cells it could be possible that blockade of CD16 showed only a slight attenuation of the proinflammatory effect of mCRP on RPE cells mCRP-induced activation in RPE and endothelial cells depends on its topological localization. Our results (ECs) [27, 43]. Instead, in ECs, lipid raft microdomains showed a polarized stimulation of mCRP on barrier seem to be the major sensors for mCRP [44]. Therefore, disruption in RPE cells, in both ARPE-19 and primary it could be speculated that mCRP interacts with RPE porcine RPE cells. The addition of mCRP to the apical cells through lipid raft microdomains. Nevertheless, side of RPE cells resulted in a significantly greater unlike many other surface receptors in epithelial tissues, barrier disruption -decreased TEER, increased caveolae seem to have a bipolar distribution in RPE permeability and disrupted membrane ZO-1- than the cells [45]. The increased immunoreactivity of mCRP addition of mCRP to the basolateral side. This effect was when added to the apical side could be also attributed to more pronounced in primary RPE cells, where mCRP the presence of the Transwell filter in the basolateral Figure 5. mCRP induces barrier disruption in ARPE-19 cells in a polarized manner. ARPE-19 cells were treated with CRP isoforms for 48h either from the apical side or the basolateral chamber and TEER (A) and paracellular permeability as determined by FITC-dextran diffusion rate (B) was determined. (C) Cells were then fixed and immunostained with anti ZO-1 (red) and DAPI (blue). Images shown are representative of four independent experiments. Scale bar = 20 μm. (D) Quantification of ZO-1 at the TJs expressed as relative (intercellular/cytoplasmic) ZO-1 distribution. Values are expressed as mean ± SD and statistical analysis was performed by one-way ANOVA and Dunnett´s posthoc analysis (N=4). * P<0.05, ** P<0.01 vs. control. www.aging-us.com 13912 AGING side that could hamper the binding of mCRP to the able to reach the apical side only in ARPE-19 cells, but basolateral side. This could also prevent mCRP to not in primary porcine RPE cells (Figures 2, 3). ARPE- induce barrier disruption when added from the 19, is a spontaneously arising RPE cell line that behaves basolateral side. However, we used filters with a pore in many ways like primary RPE cultures as they exhibit size of 0.4 μm, big enough to allow mCRP to reach the barrier functions mediated by tight junctions and secrete basolateral side of the RPE cells. It could be also cytokines. However, they exhibit reduced TEER [46]. possible that the receptors on the apical side of the cells Indeed, ARPE-19 cells are commonly used for studying have a greater affinity for mCRP than those in the oxidative stress and cell signaling in AMD because they basolateral side. However, we observed a similar exhibit features of aged RPE [47]. Thus, the fact that pattern with different cell types (CEC, ARPE-19, and ARPE-19 cells but not primary RPE cells allow mCRP primary RPE), and therefore it is likely that our diffusion to the apical side suggest that mCRP could observations are mainly the consequence of mCRP reach the subretinal space when the RPE is damaged. settling on the cell surface. These findings may explain why mCRP had some effect on barrier disruption on ARPE-19 cells but not in Given that mCRP seems to induce higher barrier primary RPE cells when added from the basolateral disruption when present in the apical side of RPE cells, side. As such, some mCRP may have crossed the we tested whether mCRP could cross the RPE and reach ARPE-19 monolayer reaching the apical side, thereby the apical side when added in the basolateral side. RPE inducing barrier disruption. cells are critical for oBRB function, enabling selective transport of molecules in and out of the retina to Our current work suggests a plausible mechanism by preserve its immune privilege [7]. We found that pCRP which mCRP may contribute to RPE dysfunction and was unable to reach the apical side -representing the AMD progression: the serum-associated isoform of subretinal space- of RPE cells. However, mCRP was CRP (pCRP), would reach the oBRB by diffusion Figure 6. mCRP induces barrier disruption in primary porcine RPE cells in a polarized manner. Primary porcine RPE cells were treated with CRP isoforms for 48h either from the apical side or the basolateral chamber and TEER (A) and paracellular permeability as determined by FITC-dextran diffusion rate (B) was determined. (C) Cells were then fixed and immunostained with anti ZO-1 (red) and DAPI (blue). Images shown are representative of four independent experiments. Arrows show disruption of ZO-1. Scale bar = 20 μm. (D) Quantification of ZO-1 at the TJs expressed as relative (intercellular/cytoplasmic) ZO-1 distribution. Values are expressed as mean ± SD and statistical analysis was performed by one-way ANOVA and Dunnett´s posthoc analysis (N=6). * P<0.05, ** P<0.01, *** P<0.0001 vs. control. www.aging-us.com 13913 AGING through CECs from the choroidal circulation. Once 4 °C. pCRP was also dialyzed with TBS to remove there, it would undergo dissociation into mCRP via sodium azide. LPC exposed in RPE surface in an inflammatory microenvironment. Alternatively, mCRP could also be Cell culture generated elsewhere -although in small amounts- or in the surface of CECs before reaching the oBRB. The ARPE-19, a spontaneously arising human retinal pigment mCRP in the subepithelial space could reach the apical epithelium cell line, was obtained from the American side of a damaged RPE and amplify the inflammation Type Culture Collection (ATCC®CRL-2302™). ARPE- further disrupting the RPE barrier integrity. 19 (passages 15-20) were cultured in a 50:50 mixture of Dulbecco modified Eagle medium (DMEM) and Ham’s The present work carries some limitations. Firstly, we F12 (Biowest) supplemented with 10% fetal bovine used a simple model that did not incorporate a proper serum (FBS, Biowest), 2 mM L-glutamine (Biowest), analogue of the Bruch’s membrane. We used 100 U/mL penicillin (PAA), 0.1 mg/mL streptomycin fibronectin to grow CECs, laminin to grow primary (Biowest), and 1mM sodium pyruvate (Sigma) in a porcine RPE cells, and ARPE-19 cells were grown humidified incubator at 37°C in 5% CO2. Cells were without any protein coating in the Transwell filters. passed every 4 to 6 days by trypsinization. ARPE-19 Secondly, we did not use RPE cells derived from cells were plated at confluence onto semi-permeable inducible pluripotent stem cells which would have polycarbonate Transwell® filters, 0.4 μm pore size. At added more translatability to our work. However, we day 3 FBS was reduced to 2 % and cells were maintained used two different models of RPE to understand how in a 37°C and 5% CO2 incubator for 2–4 weeks, mCRP contributes to AMD progression, albeit in an in changing media every 3–4 days. vitro setting. Primary porcine RPE cells were isolated and cultured In summary, our findings further support mCRP direct following the protocol described by [48] with some contribution to progression of AMD, at least at the modifications. Eyes were trimmed of excess tissue and RPE level. The topological experiments elicit that placed in 0.2% povidone iodine for 10 minutes on ice. mCRP is proinflammatory when present on the apical Eyes were rinsed with sterile distilled water and placed side of the RPE. However, mCRP is likely to only in 1000 U/mL Penicillin-Streptomycin on ice for a reach the apical side of the RPE in compromised RPE minimum of 5 minutes. Anterior segments were health and where barrier functions are compromised. removed with a scalpel at the ora serrata. Eyecups were Thus, a plausible scenario would infer that, in the filled with 1 mM EDTA and incubated at 37°C for 30 presence of an already aged/damaged RPE, mCRP minutes to loosen the neural retina from the RPE sheet. reaches the apical side of the RPE to amplify the The retina was gently pulled and detached from the proinflammatory microenvironment and enhance RPE sheet. RPE cells were collected after incubation of barrier disruption. With respect to previous findings, the eyecups with 0.05% trypsin with 0.67 mM EDTA at this pathologic mechanism will be more prevalent in 37°C. After trypsin inactivation with 10% FBS, RPE patients carrying the FH risk polymorphism for AMD, suspension was centrifuged and plated in DMEM High where mCRP proinflammatory effects remain Glucose (Capricorn Scientific), with L-glutamine and unrestrained [28]. sodium pyruvate, supplemented with 1% penicillin- streptomycin, 1% non-essential amino acids (Corning) MATERIALS AND METHODS and 10% FBS. At day 3 of culture, 5 µg/ml of ciprofloxacin (Sigma) was added to the medium and at CRP isoforms day 7, serum was decreased to 1%. Cell monolayers were pigmented and showed the characteristic High purity human pCRP (Calbiochem) was stored in cobblestone morphology. At day 14, RPE cells were 10 mM Tris, 140 mM NaCl buffer (pH 8.0) containing 2 trypsinized and plated at confluence onto semi- mM CaCl2 to prevent spontaneous formation of mCRP permeable polycarbonate Transwell® filters, 0.4 μm from the native pentamer. mCRP was obtained by urea pore size, previously coated with laminin. RPE cells chelation from purified human CRP as previously were maintained in a 37°C and 5% CO2 incubator for described [23]. Briefly, pCRP at 1 mg/mL was chelated 2–4 weeks and fed with 1% FBS growth medium every with 10 mM ethylene diaminetetraacetic acid (EDTA) 3–4 days. These RPE cells expressed RPE-specific and incubated in 8.0 M urea for 4 h at 37°C. Urea was markers and showed high levels of TEER (see removed via dialysis against low ionic strength TBS Supplementary Figure 1). (0.01 M Tris-HCl and 0.05 M NaCl, pH 7.3). Monomeric CRP concentration was determined by the Choroidal endothelial cells from porcine eyes were BCA protein assay. The filtered solution was stored at isolated and cultured as described by Browning et al. [49] www.aging-us.com 13914 AGING with few modifications. Briefly, after the retina and RPE was expressed as a percentage and calculated as follows: cell layer were removed with a cell scraper, the complex (amount of dextran lower chamber) x100 / (amount of choroid-Bruch’s membrane was peeled off from the dextran upper chamber). sclera, cut into small pieces and washed three times with Minimum Essential Medium (MEM, ThermoFisher). The Immunofluorescence pieces were incubated with 0.1% collagenase (Sigma) for 2 hours at 37°C. The collagenase was neutralized with The distribution of ZO-1 and RPE65 in RPE monolayers, MEM containing 10% FBS and the mixture was passed CD31 and VWF expression in CECs, and CRP binding to through a 20G syringe. After centrifugation and washing RPE and CECs was examined by immunofluorescence. with isolation medium, cells were resuspended in 0.1% Filters were cut out, washed with PBS and fixed with BSA-PBS, adjusted to 1 x 107 cells/ml, and incubated 3.8% paraformaldehyde in PBS for 15 min at room with rabbit anti-CD31 (Abcam) (20 µl per ml of cell temperature (RT). Cells were then washed with PBS, suspension) for 1 hour at room temperature (RT) with permeabilized with Triton X-100 (0.2% for ZO-1 and agitation. Cells were centrifuged, washed with PBS and CD31, and 0.5 % for RPE65) for 15 minutes and blocked incubated with Dynabeads® for 45 minutes at 4°C. twice with filtered 1% BSA. Cells were then incubated Endothelial cells positively selected were resuspended in with primary antibody anti-ZO-1 (clone 1A12, Thermo EGM-MV2 (PromoCell) without hydrocortisone and Scientific), anti-RPE65 (clone 401.8B11.3D9, Abcam), seeded onto 0.5% gelatin coated wells. CECs expressed anti-CD31 (Abcam), anti-mCRP (3H12 gently provided characteristic endothelial markers (CD31, and VWF) and by Dr LA Potempa) or anti-pCRP (1D6, gently provided showed the capacity of endothelial tube formation for up by Dr LA Potempa) overnight. After washing three times to passage 8 (see Supplementary Figure 2). CECs were with PBS, cells were incubated with secondary antibody plated at confluence onto semi-permeable polycarbonate Alexa Fluor anti-mouse 488 or 568 IgG or anti-rabbit 568 Transwell® filters, 0.4 μm pore size, previously coated IgG for 1h at RT. Nuclei were counterstained with DAPI. with 10 µg/mL fibronectin and maintained at 37°C and Controls were stained with secondary antibodies only. 5% CO2 incubator in EGM-MV2 media (Promocell). Stained cells were washed and covered with Prolong Gold antifade reagent (Life Technologies). Images of Measurement of transepithelial electrical resistance immunostained cells were recorded on the high-speed (TEER) spectral confocal microscope Leica TCS-SP5 and analyzed with ImageJ software. ZO-1 was intensity at the TEER was measured using a commercial electrical tight junctions (TJs) was measured as intensity at the resistance system (Millicell; Millipore) in ARPE-19 and intercellular junction divided by the intensity at the primary porcine RPE monolayers grown on Transwell cytoplasm [50]. filters as described above. TEER values were calculated by subtracting the value of a blank (transwell filter SDS-PAGE and immunoblotting without cells). Measurements were repeated at least three times for each filter, and each experiment was The presence of CRP isoforms on apical and basolateral repeated at least five times using 2 filters. compartments was detected by SDS-PAGE and immunoblotting. Supernatants were centrifuged at 1,000 Permeability assay g for 10 minutes, loaded onto 12.5% polyacrylamide gels and run at 30 mA for 60 minutes. In order to avoid The paracellular permeability of ARPE-19 and primary denaturalization of pCRP, samples were not heated and porcine RPE monolayers was assessed by measuring the the amount of SDS in the acrylamide gels and the passive permeation of FITC-dextran (40 kDa, Sigma- loading and electrophoresis buffers was reduced to 1/20 Aldrich) across confluent cells grown on filters for a [51]. Proteins were transferred to a 0.22 µm minimum of 3 weeks. Then, the RPE monolayers were nitrocellulose membrane performing a semi-dry transfer treated with CRP isoforms (10 μg/mL) for 48h. After protocol. Non-specific binding sites were blocked with 48h treatment, 500 μg/ml FITC-dextran were added to 5% non-fat dry milk in 0.1% PBS-Tween for 1 hour at the apical compartment of the chamber and samples (200 RT, before incubation with anti-mCRP-specific μl) from the basal medium (lower chamber) were monoclonal antibody 3H12 1:300 in blocking buffer collected 120 min after addition of FITC-dextran. The ON at 4°C. Membranes were incubated with the absorbance of basal and apical medium samples was secondary antibody linked to HRP (GAM-HRP, Bio- measured at 485 nm of excitation and 528 nm of Rad) for 1 hour at RT. Chemiluminescent signal was emission in a microplate reader (Infinite 200 PRO detected with the Amersham ECL™ Prime Western multimode, Tecan Group Ltd., Switzerland). Each blotting detection reagent (GE Healthcare) with condition was assayed in triplicate and repeated in at ImageQuant LAS4000 (GE Healthcare) and bands were least five independent experiments. The diffusion rate analyzed using ImageJ software. www.aging-us.com 13915 AGING Determination of mCRP in cell supernatants degeneration and disease burden projection for 2020 and 2040: a systematic review and meta-analysis. mCRP was detected in cellular supernatants (previously Lancet Glob Health. 2014; 2:e106–16. centrifuged at 1,000 g for 10 min) by an ELISA assay https://doi.org/10.1016/S2214-109X(13)70145-1 following the protocol recently described by Zhang et al. PMID:25104651 [52]. For this purpose, mouse anti-human CRP mAb 2. Zarbin MA. Current concepts in the pathogenesis of CRP-8 (Sigma-Aldrich, C1688) was immobilized as age-related macular degeneration. Arch Ophthalmol. capture antibody at 1:1,000 in coating buffer (10 mM 2004; 122:598–614. sodium carbonate/bicarbonate, pH 9.6) overnight at 4°C. https://doi.org/10.1001/archopht.122.4.598 After washing three times for 2 minutes each with TBS, PMID:15078679 non-specific binding sites were blocked with filtered 1% BSA-TBS for 1 hour at RT. Samples diluted 1:100 in 3. McLeod DS, Taomoto M, Otsuji T, Green WR, Sunness blocking buffer were added into wells for 1 hour at RT. JS, Lutty GA. Quantifying changes in RPE and choroidal Then, washing step was repeated and samples were vasculature in eyes with age-related macular incubated with sheep anti-human CRP polyclonal degeneration. Invest Ophthalmol Vis Sci. 2002; antibody (1:2,000 in blocking buffer) (BindingSite), prior 43:1986–93. incubation with a HRP-labeled donkey anti-sheep IgG PMID:12037009 (1:10,000 in blocking buffer) (Abcam). Signaling was 4. Ambati J, Atkinson JP, Gelfand BD. Immunology of age- detected with VersaMax Microplate Reader and The OD related macular degeneration. Nat Rev Immunol. 2013; value of each sample was calculated as OD450–OD570 nm. 13:438–51. https://doi.org/10.1038/nri3459 Statistical analysis PMID:23702979 Results were expressed as mean ± SD. Student’s t test or 5. Shaw AC, Goldstein DR, Montgomery RR. Age- ANOVA followed by Dunnett’s posthoc analysis were dependent dysregulation of innate immunity. Nat Rev used to determine statistical significance between Immunol. 2013; 13:875–87. treatments. A value of P<0.05 was considered significant. https://doi.org/10.1038/nri3547 All calculations were performed using GraphPad Prism PMID:24157572 (GraphPad Software, San Diego, CA, USA). 6. Donoso LA, Kim D, Frost A, Callahan A, Hageman G. The role of inflammation in the pathogenesis of age- CONFLICTS OF INTEREST related macular degeneration. Surv Ophthalmol. 2006; 51:137–52. Authors have no conflicts of interest to declare. https://doi.org/10.1016/j.survophthal.2005.12.001 PMID:16500214 FUNDING 7. Dick AD. Doyne lecture 2016: intraocular health and the many faces of inflammation. Eye (Lond). 2017; This work was supported by the Ministry of Science 31:87–96. and Innovation of Spain, ‘Instituto de Salud Carlos https://doi.org/10.1038/eye.2016.177 III’, ‘Fondo de Investigación Sanitaria’ (PI19/00265) PMID:27636226 to B.M. This work was supported by Plan Nacional de Salud (PNS) [PGC2018-094025-B-I00 to G.V and 8. Chen M, Luo C, Zhao J, Devarajan G, Xu H. Immune SAF2016-76819-R to L.B.] from the Spanish Ministry regulation in the aging retina. Prog Retin Eye Res. of Science and Innovation and funds FEDER “Una 2019; 69:159–72. Manera de Hacer Europa”; and CIBERCV (to L.B). https://doi.org/10.1016/j.preteyeres.2018.10.003 We thank the support of the Generalitat of Catalunya PMID:30352305 (Secretaria d’Universitats i Recerca del Departament 9. Crabb JW, Miyagi M, Gu X, Shadrach K, West KA, d’Economia i Coneixement de la Generalitat, 2017 Sakaguchi H, Kamei M, Hasan A, Yan L, Rayborn ME, SGR 0701 and 2017 SGR 1480 and the Fundación Salomon RG, Hollyfield JG. Drusen proteome analysis: Investigación Cardiovascular-Fundación Jesus Serra an approach to the etiology of age-related macular for their continuous support. degeneration. Proc Natl Acad Sci USA. 2002; 99:14682–87. REFERENCES https://doi.org/10.1073/pnas.222551899 PMID:12391305 1. Wong WL, Su X, Li X, Cheung CM, Klein R, Cheng CY, 10. Johnson LV, Leitner WP, Staples MK, Anderson DH. Wong TY. Global prevalence of age-related macular Complement activation and inflammatory processes in www.aging-us.com 13916 AGING drusen formation and age related macular structural intermediate: mCRP(m). FASEB J. 2007; degeneration. Exp Eye Res. 2001; 73:887–96. 21:284–94. https://doi.org/10.1006/exer.2001.1094 https://doi.org/10.1096/fj.06-6722com PMID:11846519 PMID:17116742 11. Haines JL, Hauser MA, Schmidt S, Scott WK, Olson LM, 19. Eisenhardt SU, Habersberger J, Murphy A, Chen YC, Gallins P, Spencer KL, Kwan SY, Noureddine M, Gilbert Woollard KJ, Bassler N, Qian H, von Zur Muhlen C, JR, Schnetz-Boutaud N, Agarwal A, Postel EA, Pericak- Hagemeyer CE, Ahrens I, Chin-Dusting J, Bobik A, Peter Vance MA. Complement factor H variant increases the K. Dissociation of pentameric to monomeric c-reactive risk of age-related macular degeneration. Science. protein on activated platelets localizes inflammation to 2005; 308:419–21. atherosclerotic plaques. Circ Res. 2009; 105:128–37. https://doi.org/10.1126/science.1110359 https://doi.org/10.1161/CIRCRESAHA.108.190611 PMID:15761120 PMID:19520972 12. Edwards AO, Ritter R 3rd, Abel KJ, Manning A, 20. Thiele JR, Habersberger J, Braig D, Schmidt Y, Goerendt Panhuysen C, Farrer LA. Complement factor H K, Maurer V, Bannasch H, Scheichl A, Woollard KJ, von polymorphism and age-related macular degeneration. Dobschütz E, Kolodgie F, Virmani R, Stark GB, et al. Science. 2005; 308:421–24. Dissociation of pentameric to monomeric c-reactive https://doi.org/10.1126/science.1110189 protein localizes and aggravates inflammation: in vivo PMID:15761121 proof of a powerful proinflammatory mechanism and a new anti-inflammatory strategy. Circulation. 2014; 13. Black S, Kushner I, Samols D. C-reactive protein. J Biol Chem. 2004; 279:48487–90. 130:35–50. https://doi.org/10.1161/CIRCULATIONAHA.113.007124 https://doi.org/10.1074/jbc.R400025200 PMID:24982116 PMID:15337754 21. Khreiss T, József L, Hossain S, Chan JS, Potempa LA, 14. Seddon JM, George S, Rosner B, Rifai N. Progression of Filep JG. Loss of pentameric symmetry of c-reactive age-related macular degeneration: prospective protein is associated with delayed apoptosis of human assessment of c-reactive protein, interleukin 6, and neutrophils. J Biol Chem. 2002; 277:40775–81. other cardiovascular biomarkers. Arch Ophthalmol. https://doi.org/10.1074/jbc.M205378200 2005; 123:774–82. PMID:12198121 https://doi.org/10.1001/archopht.123.6.774 PMID:15955978 22. Khreiss T, József L, Potempa LA, Filep JG. Conformational rearrangement in c-reactive protein is 15. Vine AK, Stader J, Branham K, Musch DC, Swaroop A. required for proinflammatory actions on human Biomarkers of cardiovascular disease as risk factors for endothelial cells. Circulation. 2004; 109:2016–22. age-related macular degeneration. Ophthalmology. https://doi.org/10.1161/01.CIR.0000125527.41598.68 2005; 112:2076–80. PMID:15051635 https://doi.org/10.1016/j.ophtha.2005.07.004 PMID:16225921 23. Molins B, Peña E, Vilahur G, Mendieta C, Slevin M, Badimon L. C-reactive protein isoforms differ in their 16. Volanakis JE. Human C-reactive protein: expression, effects on thrombus growth. Arterioscler Thromb Vasc structure, and function. Mol Immunol. 2001; Biol. 2008; 28:2239–46. 38:189–97. https://doi.org/10.1161/ATVBAHA.108.174359 https://doi.org/10.1016/s0161-5890(01)00042-6 PMID:18787187 PMID:11532280 24. Mullins RF, Russell SR, Anderson DH, Hageman GS. 17. Lauer N, Mihlan M, Hartmann A, Schlötzer-Schrehardt Drusen associated with aging and age-related macular U, Keilhauer C, Scholl HP, Charbel Issa P, Holz F, degeneration contain proteins common to Weber BH, Skerka C, Zipfel PF. Complement extracellular deposits associated with atherosclerosis, regulation at necrotic cell lesions is impaired by elastosis, amyloidosis, and dense deposit disease. the age-related macular degeneration-associated FASEB J. 2000; 14:835–46. factor-H His402 risk variant. J Immunol. 2011; PMID:10783137 187:4374–83. https://doi.org/10.4049/jimmunol.1002488 25. Anderson DH, Mullins RF, Hageman GS, Johnson LV. PMID:21930971 A role for local inflammation in the formation of drusen in the aging eye. Am J Ophthalmol. 2002; 18. Ji SR, Wu Y, Zhu L, Potempa LA, Sheng FL, Lu W, Zhao J. 134:411–31. Cell membranes and liposomes dissociate c-reactive https://doi.org/10.1016/s0002-9394(02)01624-0 protein (CRP) to form a new, biologically active PMID:12208254 www.aging-us.com 13917 AGING 26. Molins B, Pascual-Méndez A, Llorenç V, Zarranz- factor H in aged human eyes and eyes with age-related Ventura J, Mesquida M, Adán A, Martorell J. C-reactive macular degeneration. Br J Ophthalmol. 2011; protein isoforms differentially affect outer blood- 95:1323–30. retinal barrier integrity and function. Am J Physiol Cell https://doi.org/10.1136/bjo.2010.199216 Physiol. 2017; 312:C244–53. PMID:21633121 https://doi.org/10.1152/ajpcell.00057.2016 35. Schneeweis C, Gräfe M, Bungenstock A, Spencer- PMID:28003224 Hänsch C, Fleck E, Goetze S. Chronic CRP-exposure 27. Molins B, Fuentes-Prior P, Adán A, Antón R, Arostegui inhibits VEGF-induced endothelial cell migration. J JI, Yagüe J, Dick AD. Complement factor H binding Atheroscler Thromb. 2010; 17:203–12. of monomeric c-reactive protein downregulates https://doi.org/10.5551/jat.3004 proinflammatory activity and is impaired with at risk PMID:20173307 polymorphic CFH variants. Sci Rep. 2016; 6:22889. 36. Verma S, Wang CH, Li SH, Dumont AS, Fedak PW, https://doi.org/10.1038/srep22889 PMID:26961257 Badiwala MV, Dhillon B, Weisel RD, Li RK, Mickle DA, 28. Molins B, Romero-Vázquez S, Fuentes-Prior P, Adan A, Stewart DJ. A self-fulfilling prophecy: c-reactive protein Dick AD. C-reactive protein as a therapeutic target in attenuates nitric oxide production and inhibits age-related macular degeneration. Front Immunol. angiogenesis. Circulation. 2002; 106:913–19. 2018; 9:808. https://doi.org/10.1161/01.cir.0000029802.88087.5e https://doi.org/10.3389/fimmu.2018.00808 PMID:12186793 PMID:29725335 37. Bello G, Cailotto F, Hanriot D, Kolopp-Sarda MN, 29. Chirco KR, Whitmore SS, Wang K, Potempa LA, Halder Latger-Cannard V, Hess K, Zannad F, Longrois D, Ropars JA, Stone EM, Tucker BA, Mullins RF. Monomeric c- A. C-reactive protein (CRP) increases VEGF-a reactive protein and inflammation in age-related expression in monocytic cells via a PI3-kinase and ERK macular degeneration. J Pathol. 2016; 240:173–83. 1/2 signaling dependent pathway. Atherosclerosis. https://doi.org/10.1002/path.4766 PMID:27376713 2008; 200:286–93. https://doi.org/10.1016/j.atherosclerosis.2007.12.046 30. Johnson PT, Betts KE, Radeke MJ, Hageman GS, PMID:18280482 Anderson DH, Johnson LV. Individuals homozygous for the age-related macular degeneration risk-conferring 38. Chen J, Gu Z, Wu M, Yang Y, Zhang J, Ou J, Zuo Z, Wang variant of complement factor H have elevated levels of J, Chen Y. C-reactive protein can upregulate VEGF CRP in the choroid. Proc Natl Acad Sci USA. 2006; expression to promote ADSC-induced angiogenesis by 103:17456–61. activating HIF-1α via CD64/PI3k/Akt and MAPK/ERK https://doi.org/10.1073/pnas.0606234103 signaling pathways. Stem Cell Res Ther. 2016; 7:114. PMID:17079491 https://doi.org/10.1186/s13287-016-0377-1 PMID:27526687 31. Chirco KR, Flamme-Wiese MJ, Wiley JS, Potempa LA, Stone EM, Tucker BA, Mullins RF. Evaluation of serum 39. Slevin M, Matou-Nasri S, Turu M, Luque A, Rovira N, and ocular levels of membrane attack complex and c- Badimon L, Boluda S, Potempa L, Sanfeliu C, de Vera N, reactive protein in CFH-genotyped human donors. Eye Krupinski J. Modified c-reactive protein is expressed by (Lond). 2018; 32:1740–42. stroke neovessels and is a potent activator of https://doi.org/10.1038/s41433-018-0170-8 angiogenesis in vitro. Brain Pathol. 2010; 20:151–65. PMID:30013157 https://doi.org/10.1111/j.1750-3639.2008.00256.x PMID:19170684 32. Alizadeh E, Mammadzada P, André H. The different facades of retinal and choroidal endothelial cells in 40. Krupinski J, Turu MM, Martinez-Gonzalez J, Carvajal A, response to hypoxia. Int J Mol Sci. 2018; 19:3846. Juan-Babot JO, Iborra E, Slevin M, Rubio F, Badimon L. https://doi.org/10.3390/ijms19123846 Endogenous expression of c-reactive protein is PMID:30513885 increased in active (ulcerated noncomplicated) human carotid artery plaques. Stroke. 2006; 37:1200–04. 33. Boguslawski G, McGlynn PW, Potempa LA, Filep JG, https://doi.org/10.1161/01.STR.0000217386.37107.be Labarrere CA. Conduct unbecoming: c-reactive protein PMID:16601222 interactions with a broad range of protein molecules. J Heart Lung Transplant. 2007; 26:705–13. 41. Slevin M, Matou S, Zeinolabediny Y, Corpas R, Weston https://doi.org/10.1016/j.healun.2007.04.006 R, Liu D, Boras E, Di Napoli M, Petcu E, Sarroca S, Popa- PMID:17613401 Wagner A, Love S, Font MA, et al. Monomeric c- reactive protein—a key molecule driving development 34. Bhutto IA, Baba T, Merges C, Juriasinghani V, McLeod of Alzheimer’s disease associated with brain DS, Lutty GA. C-reactive protein and complement ischaemia? Sci Rep. 2015; 5:13281. www.aging-us.com 13918 AGING https://doi.org/10.1038/srep13281 ARPE-19 cells by targeting the silent mating type PMID:26335098 information regulation 2 homolog 1/p66shc pathway: 42. Li HY, Wang J, Wu YX, Zhang L, Liu ZP, Filep JG, implications for age-related macular degeneration. Front Aging Neurosci. 2019; 11:137. Potempa LA, Wu Y, Ji SR. Topological localization of monomeric c-reactive protein determines https://doi.org/10.3389/fnagi.2019.00137 PMID:31249522 proinflammatory endothelial cell responses. J Biol Chem. 2014; 289:14283–90. 48. Toops KA, Tan LX, Lakkaraju A. A detailed three-step https://doi.org/10.1074/jbc.M114.555318 protocol for live imaging of intracellular traffic in PMID:24711458 polarized primary porcine RPE monolayers. Exp Eye 43. Singh U, Devaraj S, Dasu MR, Ciobanu D, Reusch J, Jialal Res. 2014; 124:74–85. https://doi.org/10.1016/j.exer.2014.05.003 I. C-reactive protein decreases interleukin-10 secretion in activated human monocyte-derived macrophages PMID:24861273 via inhibition of cyclic AMP production. Arterioscler 49. Browning AC, Gray T, Amoaku WM. Isolation, culture, Thromb Vasc Biol. 2006; 26:2469–75. and characterisation of human macular inner choroidal https://doi.org/10.1161/01.ATV.0000241572.05292.fb microvascular endothelial cells. Br J Ophthalmol. 2005; PMID:16917108 89:1343–47. 44. Ji SR, Ma L, Bai CJ, Shi JM, Li HY, Potempa LA, Filep JG, https://doi.org/10.1136/bjo.2004.063602 PMID:16170129 Zhao J, Wu Y. Monomeric c-reactive protein activates endothelial cells via interaction with lipid raft 50. Molins B, Mora A, Romero-Vázquez S, Pascual-Méndez microdomains. FASEB J. 2009; 23:1806–16. A, Rovira S, Figueras-Roca M, Balcells M, Adán A, https://doi.org/10.1096/fj.08-116962 Martorell J. Shear stress modulates inner blood retinal PMID:19136614 barrier phenotype. Exp Eye Res. 2019; 187:107751. https://doi.org/10.1016/j.exer.2019.107751 45. Mora RC, Bonilha VL, Shin BC, Hu J, Cohen-Gould L, Bok D, Rodriguez-Boulan E. Bipolar assembly of caveolae in PMID:31394104 retinal pigment epithelium. Am J Physiol Cell Physiol. 51. Taylor KE, van den Berg CW. Structural and functional 2006; 290:C832–43. comparison of native pentameric, denatured https://doi.org/10.1152/ajpcell.00405.2005 monomeric and biotinylated c-reactive protein. PMID:16251472 Immunology. 2007; 120:404–11. 46. Ablonczy Z, Dahrouj M, Tang PH, Liu Y, Sambamurti K, https://doi.org/10.1111/j.1365-2567.2006.02516.x PMID:17163961 Marmorstein AD, Crosson CE. Human retinal pigment epithelium cells as functional models for the RPE in 52. Zhang L, Li HY, Li W, Shen ZY, Wang YD, Ji SR, Wu Y. An vivo. Invest Ophthalmol Vis Sci. 2011; 52:8614–20. ELISA assay for quantifying monomeric c-reactive https://doi.org/10.1167/iovs.11-8021 protein in plasma. Front Immunol. 2018; 9:511. PMID:21960553 https://doi.org/10.3389/fimmu.2018.00511 PMID:29593741 47. Tong N, Jin R, Zhou Z, Wu X. Involvement of microRNA- 34a in age-related susceptibility to oxidative stress in www.aging-us.com 13919 AGING SUPPLEMENTARY MATERIALS Supplementary Figures Supplementary Figure 1. Characterization of RPE cells. (A) Cells harvested for 1, 7 and 14 days after plating. Objective lens 10x. Primary porcine RPE cells cultured for 30 days were stained with antibodies to RPE65 (green) (B) and ZO-1 (red) (C). Scale bar = 20 µm. (D) TEER values of primary porcine RPE cells plated at 280,000 cells/cm2 on laminin coated Transwell™ filters. (E) TEER values of ARPE-19 cells plated at 250,000 cells/cm2 on Transwell™ filters for 35 days. www.aging-us.com 13920 AGING Supplementary Figure 2. Characterization of primary porcine CECs. (A) Primary porcine CECs were cultured into pure matrigel- coated wells and allowed to form capillary-like structures for 24 hours. Scale bar = 500 µm. Primary porcine CECs were stained with antibodies against CD31 (B) and VWF (C). Scale bar = 20 µm. (D) Negative control with cells stained without primary antibody. www.aging-us.com 13921 AGING Supplementary Figure 3. Effect of CRP isoforms on ZO-1 expression in ARPE-19 cells. Cells were fixed and immunostained with anti ZO-1 (red) and DAPI (blue). Scale bar = 30 μm. www.aging-us.com 13922 AGING Supplementary Figure 4. Effect of CRP isoforms on ZO-1 expression in primary porcine RPE cells. Cells were fixed and immunostained with anti ZO-1 (red) and DAPI (blue). Scale bar = 30 μm. www.aging-us.com 13923 AGING