(Pro)renin Receptor–Mediated Signal Transduction and Tissue Renin-Angiotensin System Contribute to Diabetes-Induced Retinal Inflammation

Long-Evans rats (SLC, Shizuoka, Japan), C57BL/6 mice (CLEA, Tokyo, Japan), and AT1-R–deficient mice (24) (based on the C57BL/6J strain and donated by Tanabe Seiyaku, Osaka, Japan) aged 6–8 weeks were used. The mutant mice undergo targeted replacement of AT1a-R (one of the two isoforms AT1a-R and AT1b-R in rodents) loci by lacZ gene (24). All animal experiments were conducted in accordance with the Association for Research in Vision and Ophthalmology Statement for the Use of Animals in Ophthalmic and Vision Research. Animals received an intraperitoneal injection of 80 mg/kg body wt streptozotocin (Sigma, St. Louis, MO) once for rats and for 3 consecutive days for mice. Animals with plasma glucose levels >250 mg/dl 7 days after injection were considered diabetic.

To cover the handle region (positions 11–15) of the prorenin molecule, which is the binding site of the (pro)renin receptor (25), we designed decoy peptides NH₂-RILLKKMPSV-COOH (6,9,10,22,26) and NH₂-IPLKKMPS-COOH (8,23) as rat and mouse (pro)renin receptor blockers (PRRBs), respectively, and purified them by high-performance liquid chromatography on a C-18 reverse-phase column as previously described. We applied rat PRRB to diabetic rats (Figs. 1 and 2) and mouse PRRB to diabetic mice (Figs. 3–7). The specific inhibitory action of PRRB against prorenin binding with the (pro)renin receptor and subsequent ERK activation has been confirmed in our present (supplemental Figs. 1 and 2, available in an online appendix at http://diabetes.diabetesjournals.org/cgi/content/full/db08-0254/DC1) and recent (27,28) in vitro data. The specific inhibitory action of PRRB against the tissue RAS or RAPS in vivo was also confirmed in our recent data (6,8,–10,22,23,26). As a negative control, rat PRRB was inactivated by heat denaturation at 100°C for 10 min and used as control peptide. We also generated a scramble peptide NH₂-KPMLISKP-COOH for mouse PRRB as another negative control.

Diabetic animals were intraperitoneally injected with vehicle (0.25% DMSO in PBS), 1.0 mg/kg control peptide, 1.0 mg/kg scramble peptide, 0.1 or 1.0 mg/kg PRRB, or 20 mg/kg of the AT1-R blocker losartan (Cayman Chemical, Ann Arbor, MI) daily for seven consecutive weeks immediately after the establishment of diabetes (7 days after streptozotocin injection) until the end of the study. Daily injection was based on our preliminary experiments showing the time course of serum concentration of PRRB injected into normal mice and measured by BIAcore 2000 (GE Healthcare Bio-Sciences KK, Tokyo, Japan) (supplemental Fig. 3). In our preliminary experiments, 20 mg/kg losartan proved to be the maximal-effect dose for inhibiting leukocyte adhesion to the retinal vessels of diabetic mice (data not shown).

The retinal vasculature and adherent leukocytes were imaged by perfusion labeling with fluorescein isothiocyanate (FITC)-coupled concanavalin A lectin (ConA) (Vector, Burlingame, CA), as previously described (29). After deep anesthesia, the chest cavity was opened and a 20 or 27 G cannula (for rats or mice, respectively) was introduced into the left ventricle. After injection of PBS to remove erythrocytes and nonadherent leukocytes, 20 or 2 ml (for rats or mice, respectively) of FITC-conjugated ConA was perfused. After the eyes were enucleated, the retinas were flat mounted. The flat mounts were imaged using an epifluorescence microscope (IX71; Olympus, Tokyo, Japan), and the total number of ConA-stained adherent leukocytes per retina was determined.

We isolated total RNA from the retina with an RNeasy Mini Kit (Qiagen, Tokyo, Japan) and performed a real-time quantitative RT-PCR with the TaqMan One-Step RT-PCR Master Mix Reagents Kit, an ABI Prism 7700 HT Detection System (Applied Biosystems, Foster City, CA), and probes and primers for the rat genes that encode prorenin, the (pro)renin receptor, and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) as previously described (9,10,26).

Rat eyes were fixed with 4% paraformaldehyde and embedded in paraffin. After deparaffinized sections were precoated with 4% skim milk to block nonspecific binding, a rabbit polyclonal antibody against the gate region of rat prorenin (1:100) or a goat polyclonal antibody against the rat (pro)renin receptor (1:100) was applied as the primary antibody. An antibody against the gate region binds only to nonproteolytically activated prorenin but not to inactive prorenin or proteolytically activated prorenin (i.e., renin) (6). The anti-(pro)renin receptor antibody was raised by using the previously established COS-7 cells producing rat (pro)renin receptor protein (9). The sections were then incubated with a biotin-conjugated IgG as the secondary antibody and TOTO-3 (1:500) (Molecular Probes, Eugene, OR) for nuclear staining. The immunohistochemical reactions were visualized by using a Vectastain ABC Standard Kit (Vector Laboratories, Burlingame, CA) and Tyramide Signal Amplification Fluorescein System (PerkinElmer, Boston, MA) according to the manufacturer's protocol. The samples were viewed with the scanning laser confocal microscope.

The isolated retina was placed into the lysis buffer. After blocking nonspecific binding with 5% skim milk, polyvinylidene fluoride membranes were incubated with a mouse monoclonal antibody against phosphorylated ERK1/2 (1:1,000; Cell Signaling Technology, Beverly, MA), total ERK1/2 (1:1,000; Cell Signaling Technology), or α-tubulin (1:1,000; Sigma). Membranes were then incubated with biotin-conjugated secondary antibody (Jackson ImmunoResearch Laboratories, West Grove, PA) followed by avidin-biotin complex (Vectastain ABC Elite Kit; Vector Laboratories). Finally, the signals were detected through enhanced chemiluminescence (ECL Blotting Analysis System; GE Healthcare).

The isolated retina was placed into the lysis buffer. The protein levels of VEGF and ICAM-1 were determined with enzyme-linked immunosorbent assay (ELISA) kits (R&D Systems). The tissue sample concentration was calculated from a standard curve and corrected for protein concentration.

Animals with streptozotocin-induced diabetes showed a significant decrease in body weight (P < 0.01) and a significant increase in blood glucose levels (P < 0.01) compared with those in age-matched nondiabetic controls (Table 1). Treatment with PRRB did not significantly change these metabolic parameters (P > 0.05) (Table 1).

To elucidate the involvement of prorenin and the (pro)renin receptor in the pathogenesis of diabetic retinopathy, we first performed quantitative RT-PCR analyses for prorenin and the (pro)renin receptor in the retina. Prorenin mRNA levels (ratio to GAPDH mRNA) were upregulated (P < 0.01) in the diabetic retina compared with those in age-matched normal control rats (Fig. 1,A). In contrast, mRNA levels of the (pro)renin receptor showed no significant differences (P > 0.05) between diabetic rats and normal controls (Fig. 1 B).

Immunohistochemistry for normal, vehicle-, or PRRB-treated diabetic rat eyes was performed to identify the tissue localization of nonproteolytically activated prorenin and the (pro)renin receptor. Immunoreactivity of activated prorenin in the retinal vessels was negligible in nondiabetic controls (Fig. 1,C), substantially enhanced in diabetes (Fig. 1,D), and abrogated by PRRB application to the background signal (Fig. 1,E). In contrast, immunoreactivity of the (pro)renin receptor (Fig. 1,F) was unaltered by induction of diabetes (Fig. 1,G) or treatment with PRRB (Fig. 1 H).

The retinal adherent leukocytes were imaged by perfusion labeling with FITC-coupled ConA. Leukocyte counts were evaluated in the whole retina from the posterior retina around the optic disc (Fig. 2,A–D) to the peripheral (anterior) retina next to the ora serrata. The total number of adherent leukocytes was significantly (P < 0.01 for both) higher in vehicle-treated (Fig. 2,B) and control peptide-treated (Fig. 2,C) diabetic rats than in normal age-matched controls (Fig. 2,A). PRRB-treated diabetic rats (Fig. 2,D) showed a significant decrease in the leukocyte count in a dose-dependent fashion compared with that in vehicle-treated (P < 0.05) or control peptide-treated (P < 0.01) diabetic rats (Fig. 2 E).

Retinal protein levels of VEGF (Fig. 3,A) and ICAM-1 (Fig. 3,B) were significantly (P < 0.01) higher in vehicle-treated diabetic mice than in age-matched normal controls. PRRB significantly suppressed retinal protein levels of these inflammatory molecules known as pathogenic in the diabetic retina (Fig. 3 A and B).

We examined retinal leukocyte adhesion for comparison among diabetic mice treated with vehicle, scramble peptide, the AT1-R blocker losartan, PRRB, and losartan plus PRRB. Compared with vehicle (Fig. 4,A) or scramble peptide treatment, the inhibitory effect of PRRB (71.3% inhibition) (Fig. 4,C) on diabetic retinal inflammation was more potent than that of losartan alone (32.4% inhibition) (Fig. 4,B) and was equivalent to that of losartan plus PRRB (Fig. 4,D), showing that PRRB in combination with losartan caused little or no additive effect compared with PRRB alone (Fig. 4 E). Thus, PRRB was suggested to inhibit both angiotensin II–dependent (32.4% contribution) and –independent (71.3 − 32.4 = 38.9% contribution) actions.

To clarify the role of RAS-independent intracellular signaling via the (pro)renin receptor, we used mice in which the RAS was deactivated by genetic ablation of AT1-R. The number of adherent leukocytes in the retinal vasculature was significantly (P < 0.01) higher in diabetic wild-type mice than in nondiabetic wild-type controls. Compared with diabetic wild-type animals, diabetic AT1-R–deficient mice exhibited a significant (P < 0.01; 27.3%) reduction of the number of adherent leukocytes. Compared with that in nondiabetic AT1-R–deficient mice (Fig. 5,A), induction of diabetes (Fig. 5,B) significantly (P < 0.01) increased the number of adherent leukocytes in the retinal vasculature. The number of retinal adherent leukocytes was significantly (P < 0.01; 46.6% of diabetic wild-type mice) smaller in PRRB-treated diabetic AT1-R–deficient mice (Fig. 5,C) at the dose of 1.0 mg/kg than in vehicle-treated diabetic AT1-R–deficient mice (Fig. 5,D). These data with pharmacological blockade (Fig. 4) and genetic ablation (Fig. 5) of AT1-R indicated dual contribution of the tissue RAS (32.4 and 27.3%, respectively) and RAS-independent (pro)renin receptor signaling (38.9 and 46.6%) to diabetes-induced retinal inflammation.

To further determine whether RAS-independent intracellular signaling via the (pro)renin receptor contributes to the activation of ERK1/2 in the diabetic retina, phosphorylated ERK1/2 was examined in the retina from AT1-R–deficient diabetic mice. PRRB treatment suppressed phosphorylated but not total ERK1/2 (Fig. 6,A). Relative phosphorylation of ERK1/2, increased by inducing diabetes, was significantly (P < 0.01) suppressed by PRRB application, whereas no significant (P > 0.05) differences were detected in total ERK1/2 protein levels (Fig. 6 B).

To examine whether RAS-independent intracellular signaling via the (pro)renin receptor contributes to the upregulation of the pathogenic molecules responsible for diabetes-induced retinal inflammation (Fig. 3), diabetes was induced in AT1-R–deficient mice to measure retinal levels of VEGF and ICAM-1 protein. Compared with that in nondiabetic AT1-R–deficient mice, retinal expression of VEGF (P < 0.01) (Fig. 7,A) and ICAM-1 (P < 0.05) (Fig. 7,B) was significantly enhanced by diabetes induction. PRRB application to AT1-R–deficient diabetic mice led to significant suppression of VEGF (P < 0.05) (Fig. 7,A) but not ICAM-1 (P > 0.05) (Fig. 7 B) compared with vehicle treatment in AT1-R–deficient diabetic mice.

The present study reveals for the first time, to our knowledge, several important findings concerning the role of the (pro)renin receptor in diabetes-induced retinal inflammation. First, induction of diabetes led to significant upregulation of prorenin expression in the retina, and diabetes-induced activation of prorenin colocalized with its receptor in the retinal vessels was substantially attenuated with PRRB (Fig. 1). Second, diabetes-induced leukocyte adhesion to the retinal vasculature was suppressed by PRRB treatment, indicating that prorenin binding with its receptor contributes to the pathogenesis in the diabetic retina (Fig. 2). Third, the molecular mechanisms in the suppression of retinal leukocyte adhesion proved to include the inhibitory effects of PRRB on the retinal expression of VEGF and ICAM-1, both of which are known as key factors responsible for diabetes-induced retinal inflammation (Fig. 3).

Recently, surgical sample data have revealed the elevated levels of angiotensin II in the vitreous of eyes with proliferative diabetic retinopathy (30) and diabetic macular edema (31). In accordance with the findings of these clinical laboratory studies, we have shown that induction of diabetes to mice led to a significant increase in retinal expression and production of the RAS components (12). Although these findings indicated the association of the tissue RAS with diabetic retinopathy, the precise mechanism for activating the tissue RAS in the diabetic retina has not been determined. In parallel, little has been known about the bioactivity of the prorenin molecule per se other than its function as the inactive precursor of renin. Bilateral nephrectomy resulted in significant decline of plasma renin but not prorenin to an undetectable level, suggesting that the proteolytic processing of prorenin to renin is confined to the kidney and that other organs including the eye are capable of producing only prorenin, but not renin (32,33). Elevated prorenin levels in the plasma of long-standing diabetes (2,34,35) have indicated that prorenin may be a hallmark of microvascular complications in diabetes. Indeed, the vitreous levels of prorenin were higher in eyes with proliferative diabetic retinopathy than in those with nonproliferative retinopathy or without retinopathy (36), suggesting the involvement of prorenin with the pathogenesis of diabetic retinopathy. Reasonably, retinal mRNA levels of prorenin were elevated following induction of diabetes (Fig. 1,A), whereas (pro)renin receptor levels were unaltered (Fig. 1,B). The observed ligand-receptor expression change is compatible with our recent data on diabetic nephropathy (9,10) and choroidal neovascularization (37). These findings, in concert with the immunohistochemical results (Fig. 1,C–H), suggest the significance of the ligand-based regulation in the activation of the RAPS. Importantly, nonproteolytically activated prorenin was localized in diabetic retinal vessels bearing the (pro)renin receptor (Fig. 1 D and G), showing that the initiating step of the tissue RAS is prepared in the diabetic retina.

The molecular and cellular mechanisms underlying the pathogenesis of diabetic retinopathy are not fully understood; however, increasing evidence has suggested the involvement of inflammatory processes including cytokine upregulation and leukocyte infiltration, causing diabetic retinopathy to be regarded as an inflammatory disease (14,17,29,38,39). Retinal vasculature in diabetes is accompanied by inflammatory cell adhesion (40), which triggers vascular hyperpermeability (14) and pathological neovascularization (39). ICAM-1, constitutively expressed on vascular endothelial cells at a low level, is swiftly upregulated during inflammation, resulting in an enhancement of leukocyte-endothelium interaction. Previous studies using donor eyes from diabetic subjects (13) and experimentally induced diabetes (14,29) demonstrated that retinal ICAM-1 expression was elevated together with leukocyte adhesion and infiltration. Antibody-based blockade or genetic ablation of ICAM-1 led to significant suppression of vascular hyperpermeability in early diabetes (14) or capillary loss in established diabetes (38). VEGF, a potent angiogenic and proinflammatory factor, plays a central role in the pathogenesis of diabetic retinopathy. In patients with diabetic retinopathy, VEGF levels in the intraocular fluid were increased not only during the proliferative stage (15) but also during the nonproliferative stage characterized by macular edema (16). Interestingly, angiotensin II levels are elevated and correlated with VEGF levels in the vitreous fluid of patients with diabetic macular edema (31). Angiotensin II has been shown to induce ICAM-1 (41) and VEGF (42) via AT1-R in previous in vivo and in vitro studies. Also in the murine model of streptozotocin-induced diabetes, we have recently shown that the tissue RAS enhanced retinal expression of these inflammatory molecules and subsequent leukocyte adhesion to the retinal vasculature, all of which were suppressed by AT1-R blockade (12). In concert with the previous data, the currently observed PRRB-induced suppression of diabetes-induced retinal inflammation (Figs. 2 and 3) indicates that the tissue RAS is activated in the diabetic retina by (pro)renin receptor–mediated nonproteolytic activation of prorenin, leading to AT1-R signaling–mediated VEGF and ICAM-1 upregulation and retinal leukocyte adhesion.

The present study further revealed the role of RAS-independent (pro)renin receptor signaling in diabetes-induced retinal inflammation. The use of two different methods of inactivating AT1-R (Figs. 4 and 5) confirmed that PRRB functioned to inhibit both RAS-dependent and -independent mechanisms underlying retinal leukocyte adhesion. These data are the first to show the involvement of the RAPS, i.e., (pro)renin receptor–mediated signal transduction and tissue RAS activation, in the pathogenesis of diabetic retinopathy. We have recently shown the contribution of (pro)renin receptor signaling to diabetic nephropathy using AT1-R–deficient mice. AT1-R–deficient mice with streptozotocin-induced diabetes exhibited reduced proteinuria and glomerulosclerosis in the early phase compared with wild-type mice with diabetes, indicating a significant role of the tissue RAS in diabetic nephropathy. Surprisingly, these renal events in AT1-R–deficient diabetic mice later progressed to the equivalent levels seen in wild-type diabetic mice (8). The glomerulosclerosis observed in AT1-R–deficient diabetic mice was associated with ERK activation, which was completely blocked together with the phenotype by sustained application of PRRB, suggesting that the redundant pathways of the RAPS were involved in the pathogenesis of diabetic nephropathy (8). In addition, we administered PRRB to AT1-R–deficient diabetic mice and confirmed the significant role of intracellular signaling via the (pro)renin receptor in the diabetes-induced retinal leukocyte adhesion (Fig. 5). Importantly, PRRB application to AT1-R–deficient diabetic mice led to significant suppression of ERK activation, a known key intracellular signaling via the (pro)renin receptor (Fig. 6). Of diabetes-related inflammatory molecules (the expression of which was inhibited by PRRB [Fig. 3]), VEGF but not ICAM-1 was also regulated by (pro)renin receptor signaling per se (Fig. 7). The data are supported by and consistent with our recent report on choroidal neovascularization showing that (pro)renin receptor signaling selectively induces VEGF and monocyte chemoattractant protein-1 out of several inflammatory and angiogenic molecules mediated by AT1-R (37). These new findings (Figs. 4,FIG. 5,FIG. 6–7) clarified molecular and cellular mechanisms mediated by RAS-independent intracellular signaling via the (pro)renin receptor in diabetes-induced retinal inflammation. In addition to our recent reports showing that the RAPS contributes to glomerulosclerosis in the kidney and fibrosis in the heart (6,,,–10,26), the present data are the first to show the association of the RAPS with inflammation in the eye (Fig. 8).

In the present data, there was a discrepancy between in vivo (Figs. 1,FIG. 2,FIG. 3,FIG. 4,FIG. 5,FIG. 6–7 and supplemental Fig. 3) and in vitro (supplemental Figs. 1–2) doses of PRRB. In vivo, intraperitoneal injection of PRRB at a dose of 1.0 mg · kg⁻¹ · day⁻¹ led to a serum concentration of ≤110 ng/ml (∼100 nmol/l), which was cleared within 24 h (supplemental Fig. 3). Because the maximal concentration of PRRB in the serum is 50–100 times higher than the physiological level of endogenous prorenin (1–2 nmol/l), the currently used dose of 1.0 mg · kg⁻¹ · day⁻¹ is thought to be sufficient to inhibit the in vivo binding of prorenin to the receptor. Indeed, a lower dose of 0.1 mg · kg⁻¹ · day⁻¹ (presumably 10 nmol/l in the serum [5–10 times higher than physiological prorenin]) was also effective (Fig. 2). In vitro, however, much higher doses of PRRB (10–100 μmol/l) were required to suppress prorenin-stimulated ERK activation (supplemental Fig. 2). This would be attributable at least in part to the in vitro property of cellular localization of the (pro)renin receptor. In cultured vascular smooth muscle cells (43), indeed, the (pro)renin receptor was shown to be localized mainly to the cytoplasm and sparsely to the cellular surface, suggesting less efficacy of PRRB exerted in vitro than in vivo.

Recently, several studies have failed to show the inhibitory effect of PRRB on in vivo (44,45) and in vitro (46,47) models. These negative data on PRRB were generated by using in vivo (44,45) models in which the tissue RAS does not originally play any role or in vitro (46,47) models applied with too low concentration (≤1 μmol/l) of PRRB. The glomerulosclerosis in the clipped kidney in the rat model of 2-kidney, 1-clip hypertension has already been shown to be ischemic lesion resistant to RAS inhibition (48). The dTG rats overexpressing human renin and angiotensinogen exhibited significant increase in plasma renin and angiotensin II (activation of the circulatory but not tissue RAS), causing hypertension-induced cardiac and renal damages (45). As expected, PRRB did not alter these pathological conditions as a result of tissue RAS-independent mechanisms (44,45).

Although hypertension is a known risk factor for the progression of diabetic retinopathy (49,50), there are indeed a large number of normotensive patients with diabetic retinopathy who have the potential risk of hypotension caused by the use of antihypertensive agents including AT1-R blockers and ACE inhibitors. In contrast, because the (pro)renin receptor is present in the major organs but not in the circulation, PRRB does not affect the circulatory RAS or systemic blood pressure (9,10). Interestingly, PRRB administration to the diabetic retina was shown to cause not only tissue RAS deactivation but also additional suppression of (pro)renin receptor signaling–mediated expression of VEGF—the major pathogenic factor responsible for the development of diabetic retinopathy (Fig. 8). Collectively, inhibition of the RAPS with PRRB may prove more useful as a novel therapeutic strategy for diabetic retinopathy than RAS suppression with conventional AT1-R blockers or ACE inhibitors.

This work was supported by the Japanese Ministry of Education, Culture, Sports, Science and Technology (grant-in-aid for scientific research 18791296 to S.S.).

No potential conflicts of interest relevant to this article were reported.