Contributions of Sodium-Hydrogen Exchanger 1 and Mitogen-Activated Protein Kinases to Enhanced Retinal Venular Constriction to Endothelin-1 in Diabetes

The retinal microcirculation supplies oxygen and nutrients to and removes metabolic wastes from the inner two-thirds of the metabolically active retinal tissue. Sufficient retinal perfusion is essential for maintaining normal vision (1). Several studies have shown that alterations in retinal perfusion are associated with diabetic retinopathy (2–4), the most common cause of vision impairment or blindness among working-age adults. In general, decreased retinal blood flow has been reported in the early stage of diabetes in humans (5) and in rodent models of type 1 diabetes (6,7). Although changes in vasomotor tone of retinal arterioles play a major role in regulation of resistance to blood flow in the retina (1), recent studies have demonstrated the ability of retinal venules to constrict in response to endogenous factors (8–11). The constriction of venules in the retinal microcirculation could increase retinal capillary pressure and reduce retinal perfusion when compensatory or defense mechanisms for blood flow regulation are exhausted during the onset of diabetes before development of retinopathy. However, there is a limited understanding of the microvascular vasomotor dysregulation in early diabetes.

Intravenous administration of endothelin-1 (ET-1), a potent vasoconstrictor peptide, has been shown to reduce retinal blood flow in healthy subjects (12). Elevated level of ET-1, detected in the vitreous humor and plasma of patients with diabetic retinopathy (13) and in the vitreous of pigs after 2 weeks of hyperglycemia (9), is sufficient to increase retinal vasomotor tone (9–11) and thus possibly compromises retinal perfusion. Interestingly, the reduction of retinal blood flow in mice subjected to 4 weeks of hyperglycemia was prevented by administration of an ET-1 subtype A receptor (ET_AR) antagonist (6), suggesting an adverse role of ET-1/ET_AR in altering retinal vasomotor tone in the early stages of diabetes. Recent studies in the porcine retinal microcirculation indicated that the constriction of isolated retinal arterioles to ET-1 was not altered in early diabetes or by a short-term exposure to hyperglycemia (14). However, we recently found that the vasoconstrictor responses to norepinephrine, thromboxane analog U46619, and ET-1 were enhanced in retinal venules isolated from pigs subjected to hyperglycemia, and the most pronounced increase in venular constriction was elicited by ET-1, via ET_AR activation (9). This enhanced responsiveness to ET-1 was recently observed in isolated human retinal venules after exposure to hyperglycemic challenge (11). Although in our recent study we demonstrated the involvement of ET_AR and reverse-mode sodium-calcium exchanger (NCX) in enhanced retinal venular constriction to ET-1 during hyperglycemic insult (9), the contribution of ET-1 subtype B receptor (ET_BR) remains to be determined. Moreover, there has been no evaluation of whether this ET-1 signaling pathway is linked to activation of sodium-hydrogen exchanger 1 (NHE1), a potential signaling protein upstream of NCX (15).

ET-1 has been shown to elevate intracellular calcium (Ca²⁺) (16) and to activate protein kinase C (PKC) (17) and mitogen-activated protein kinases (MAPKs), which include c-Jun N-terminal kinase (JNK) (18), p38 kinase (19), and extracellular signal–regulated kinase (ERK) (20). These protein kinases have been reported to influence vasomotor responses of some arterial vascular beds (21–23). However, it is unclear whether extracellular Ca²⁺ and these signaling molecules contribute to or modulate the venular constriction to ET-1 in the retina. In the current study, we addressed the roles of ET_BR, extracellular Ca²⁺, NHE1, PKC, and MAPKs in ET-1–induced constriction of retinal venules and investigated their contribution to enhanced constriction to ET-1 in pigs with early type 1 diabetes.

All animal procedures were performed in accordance with the Association for Research in Vision and Ophthalmology Statement for the Use of Animals in Ophthalmic and Vision Research and were approved by the Baylor Scott & White Health Institutional Animal Care and Use Committee. Domestic (Yorkshire) male pigs (age range 6–10 weeks and weight range 9–20 kg) were purchased from Real Farms (San Antonio, TX). Type 1 diabetes was induced by selective ablation of pancreatic β-cells with intravenous injection of streptozocin (STZ) (200 mg/kg in saline; Zanosar) via an ear vein (39 pigs) as we previously described in detail (9,14,24). The control group was intravenously injected with saline (43 pigs). Fasting blood glucose levels were obtained every other day with a Bayer Contour glucometer (Bayer, Pittsburgh, PA). Pigs were treated with insulin (Humulin 70/30, 2–8 units; Lilly, Indianapolis, IN) if blood glucose was sustained >600 mg/dL to keep the level between 250 and 600 mg/dL. After 2 weeks, pigs were sedated with Telazol (4–8 mg/kg, intramuscularly), anesthetized with 2–5% isoflurane, heparinized with heparin (1000 units/kg, intravenously via marginal ear vein), and intubated for eye harvesting as previously described (25).

The techniques used for visualization, identification, isolation, cannulation, and pressurization of the retinal vasculature have previously been described (9,10). In brief, the isolated second-order retinal venules were cannulated on each end with glass micropipettes containing a physiological saline solution (PSS)-albumin (1%) (9,10,25) and pressurized to 5 cmH₂O intraluminal pressure without flow by two independent pressure reservoir systems. Vasomotor activity of isolated venules was recorded using videomicroscopic techniques throughout the experiments (9,10,25).

Cannulated and pressurized retinal venules were bathed in PSS-albumin at 36–37°C to allow development of basal tone (stable within 60–90 min) (10). For evaluation of the effect of diabetes on vasomotor function, diameter changes in response to cumulative administration of ET-1 (1 pmol/L to 10 nmol/L; Bachem, Torrance, CA) (10) and selective ET_BR agonist sarafotoxin S6c (10 pmol/L to 0.1 μmol/L; Tocris Bioscience/Bio-Techne, Minneapolis, MN) (10) were recorded and compared in vessels isolated from diabetic and control pigs. Retinal venules were exposed to each concentration of agonist for 10 min until a stable diameter was attained. Because of potential ET-1–induced tachyphylaxis, as noted in retinal arterioles (26), and sustained vasoconstriction to ET-1 after washing it out, only one concentration-response curve was evaluated in each vessel. The contributions of NHE1 and PKC in vasoconstriction to ET-1 were evaluated after incubation with NHE1 inhibitor cariporide (1 μmol/L; Tocris Bioscience/Bio-Techne) (27) and PKC inhibitor Gö 6983 (10 μmol/L; EMD Millipore, Billerica, MA) (10,28). Concentration-dependent responses to PKC activator phorbol-12,13-dibutyrate (PDBu) (0.1–10.0 μmol/L; Tocris Bioscience/Bio-Techne) were also evaluated (28). The involvement of MAPKs in the vasomotor responses of retinal venules to ET-1 was examined after treatment with p38 kinase inhibitor SB203580 (0.1 μmol/L; EMD Millipore) (29), ERK inhibitor PD98059 (10 μmol/L; EMD Millipore) (21), or JNK inhibitor SP600125 (10 μmol/L; EMD Millipore) (22). In some vessels, the ET-1–induced response was examined after treatment with both cariporide and NCX inhibitor KB-R7943 (10 μmol/L; Tocris Bioscience/Bio-Techne) (9). All vessels were pretreated with pharmacologic inhibitors extraluminally for at least 30 min. For assessment of the contribution of extracellular Ca²⁺ entry to ET-1–induced vasoconstriction, vasomotor activity was examined in vessels exposed to Ca²⁺-free PSS-albumin (with 1 mmol/L EDTA) (10).

ET-1, sarafotoxin S6c, and PDBu were dissolved in water; Gö 6983, KB-R7943, cariporide, SB203580, PD98059, and SP600125 were dissolved in DMSO. Subsequent concentrations of these drugs were diluted in PSS. The final concentration of DMSO in the vessel bath did not exceed 0.1% by volume and had no significant effect on vessel viability, vasoconstrictor responses, or maintenance of basal tone (data not shown).

Total RNA was isolated from retinal venules (sample pooled from both eyes) and neural retina tissue with an RNeasy Mini Kit (QIAGEN, Crawley, U.K.) as previously described (9,10). For performance of real-time PCR experiments, specific primer sets for NHE1 (SLC9A1) (forward primer, 5′-CATGCTGTCAGGCAGATGAA-3′; reverse primer, 5′'-CGCAGCTGGAAGATACTGGA-3′) and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) (forward primer, 5′-CCACCCACGGCAAGTTCCACGGCA-3′; reverse primer, 5′-GGTGGTGCAGGAGGCATTGCTGAC-3′) were used. The same amount of total RNA for each sample was used to synthesize cDNA followed by real-time PCR with the Applied Biosystems QuantStudio 6 Flex Real-Time PCR System (Life Technologies, Carlsbad, CA), as previously described (10).

Retinal venules (sample pooled from both eyes, ∼5 µg per pig) and neural retina tissue were isolated and sonicated in lysis buffer. The protein content of each lysate was determined with the BCA protein assay kit (Thermo Fisher Scientific, Rockford, IL). Protein was separated by Tris-glycine SDS-PAGE (4–15% Tris-HCl Ready Gels; Bio-Rad Laboratories, Hercules, CA), transferred onto a nitrocellulose membrane, and incubated with rabbit anti-JNK or anti-phosphorylated (p)-JNK polyclonal antibody (1:1,000 dilution, cat. nos. 9252 and 4668; Cell Signaling Technology, Danvers, MA) or with rabbit anti-p38 or anti-p-p38 polyclonal antibody (1:1,000 dilution, cat. nos. 9212 and 4511; Cell Signaling Technology). Membranes were stripped and reprobed with mouse anti-vinculin monoclonal antibody (1:1,000, cat. no. V9264; Sigma-Aldrich, St. Louis, MO). After incubation with an appropriate secondary antibody (anti-rabbit or anti-mouse IgG, 1:1,000, cat. nos. 7074s and 7076s; Cell Signaling Technology), the membranes were washed and developed by enhanced chemiluminescence (Thermo Fisher Scientific). Densitometric analyses of immunoblots were performed with National Institutes of Health ImageJ software. Results for total p38 kinase in venule samples were normalized to total vinculin.

In pigs under isoflurane anesthesia, after intravenous administration of heparin (1,000 units/kg) and a left thoracotomy, blood samples were obtained by cardiac puncture followed by heart removal as the terminal procedure. Collected blood samples were centrifuged at 1,000 rpm for 10 min at 4°C, and the supernatants considered as plasma were then stored at −80°C for subsequent analysis within 3 months. Levels of ET-1 in the plasma were measured with an ET-1 ELISA kit (ADI-900-020A; Enzo Life Sciences, Farmingdale, NY) according to the manufacturer’s instructions.

At the end of each functional experiment, the vessel was relaxed with 0.1 mmol/L sodium nitroprusside in Ca²⁺-free PSS with 1 mmol/L EDTA to obtain its maximal diameter at 5 cmH₂O intraluminal pressure (9,10). Diameter changes in response to agonists were normalized to the resting diameter and expressed as percentage changes in diameter (9,10,30). All data are reported as mean ± SEM, and n represents the number of animals (1–2 vessels per pig per treatment group for functional studies). Student t test or two-way ANOVA followed by Bonferroni multiple range test was used to determine the significance of experimental interventions, as appropriate (GraphPad Prism, version 6.0; GraphPad Software, La Jolla, CA). P < 0.05 was considered statistically significant.

Retinal venules isolated from the pigs with diabetes for 2 weeks (in vivo hyperglycemia 460 ± 17 mg/dL, within range of 250–600 mg/dL) and age-matched control pigs (in vivo euglycemia 93 ± 3 mg/dL) had similar average maximum diameters (∼130 µm) and developed comparable levels of basal tone with constriction to ∼92% of maximum diameter. Administration of ET-1 elicited concentration-dependent vasoconstriction, which was greater in venules from diabetic pigs than from control pigs (Fig. 1A). By contrast, retinal venules from both control and diabetic pigs did not respond to ET_BR agonist sarafotoxin S6c (Fig. 1B). In STZ-treated pigs with a plasma glucose level <160 mg/dL, retinal venules constricted to ET-1 in a manner comparable with that of vessels isolated from control pigs (Fig. 1C). During exposure to NHE1 inhibitor cariporide, retinal venules from either control or diabetic animals lost ∼50% of basal tone. The cariporide treatment also prevented the diabetes-enhanced constrictions of retinal venules to ET-1 (Fig. 1A) but had no effect on ET-1–induced constriction of normal vessels (Fig. 1A). Similar attenuation of diabetes-enhanced constrictions to ET-1 occurred in the presence of both cariporide and NCX inhibitor KB-R7943 (Fig. 1A). The NHE1 mRNA expression level was greater in retinal venules than in neural retina tissue but was unaffected by diabetes (Fig. 1D).

In the presence of a Ca²⁺-free solution, the ET-1–induced constrictions of venules from control and diabetic pigs were abolished (Fig. 2). In another cohort, the PKC inhibitor Gö 6983 did not affect basal tone or constrictions to ET-1 in vessels from either control or diabetic animals (Fig. 3A). Also, PKC activator PDBu did not significantly alter resting diameter of control or diabetic vessels (Fig. 3B).

In the presence of JNK inhibitor SP600125, retinal venules from control and diabetic pigs lost ∼70% of basal tone, and the vasoconstrictions at lower concentrations of ET-1 were abolished (Fig. 4A). However, the ET-1–induced constriction at concentrations ≥1 nmol/L, after SP600125 treatment, remained significantly greater for vessels from diabetic than control pigs (Fig. 4A). The ERK inhibitor PD98059 did not affect basal tone or the constrictions to ET-1 in vessels from control or diabetic animals (Fig. 4B). The p38 kinase inhibitor SB203580 also did not influence basal tone (Fig. 4C) but prevented the diabetes-enhanced vasoconstriction to ET-1. In contrast, SB203580 had no impact on the venular responses to ET-1 in control vessels (Fig. 4C). The p38 kinase protein level in retinal venules isolated from control and diabetic pigs was comparable (Fig. 5A) when normalized to loading control protein vinculin (p38/vinculin: control 0.50 ± 0.02 vs. diabetes 0.39 ± 0.05, P = 0.10). Although p-p38, JNK, and p-JNK were undetected in retinal venules (data not shown), these proteins along with p38 were expressed in neural retina tissue from control pigs (Fig. 5B).

The ET-1 level was greater in the plasma isolated from diabetic (21.7 ± 3.9 pg/mL [≈8.7 ± 4.4 pmol/L]) compared with control (11.0 ± 1.6 pg/mL [≈4.4 ± 1.8 pmol/L]) pigs (Fig. 6).

A reduction in retinal blood flow is associated with the early stages of type 1 diabetes (5–7). Although the perfusion and distribution of retinal blood flow are dictated by vasomotor activity and resistance of arterioles and venules in the retina, there is a poor understanding about the microvascular dysregulation that occurs before the development of pathologic changes in the diabetic retina. In our previous studies we have provided some insight into this issue by demonstrating the increase of vitreous levels of ET-1 in pigs with early diabetes (9) and enhanced constriction of porcine (9) and human (11) retinal venules to ET-1 during acute hyperglycemia. The augmented constriction of retinal venules could elevate capillary pressure and reduce blood flow in the retina, as shown previously in other vascular beds (31,32). However, there is a paucity of information on the molecular mechanisms involved in constriction of retinal venules to ET-1 under normal or diabetes conditions. The salient findings of the current study are that the ET-1–induced constriction of normal retinal venules is mediated by extracellular Ca²⁺ influx independent of ET_BR and partly by JNK MAPK activation. The selective activation of NHE1 and p38 MAPK augments the JNK-dependent ET-1–induced constriction of retinal venules in early diabetes.

In the current study, the data for enhanced constriction of diabetic retinal venules were analyzed in vessels from diabetic pigs (plasma glucose within 250–600 mg/dL) with an average plasma glucose of ∼460 mg/dL (plasma glucose from control pigs was ∼90 mg/dL). However, four pigs administered STZ for induction of diabetes had blood glucose levels that did not exceed 160 mg/dL. The retinal venules isolated from these STZ-administered pigs constricted to ET-1 in a manner comparable with that of vessels isolated from saline control pigs (Fig. 1C). Therefore, it seems that STZ did not exert toxicity on the retinal venules to enhance ET-1–induced constriction. On the other hand, these results provided evidence that the observed enhancement of venular constriction to ET-1 is directly related to the elevated level of blood glucose. We have previously shown that exposure of retinal venules to high glucose for 2 h enhances ET-1–induced constriction of porcine (9) and human (11) retinal venules, supporting the ability of acute hyperglycemia to directly impact this vasomotor response. Furthermore, in vitro exposure to high d-glucose but not high l-glucose enhanced ET-1–induced constriction of retinal venules (9), indicating that the venular reactivity was specifically influenced by the biological form of hyperglycemia.

The ability of an ET_AR but not ET_BR antagonist to prevent the diabetes-enhanced constriction of retinal venules to ET-1 (9) supports the role of ET_AR activation in altered venular function in early diabetes. To further corroborate this notion, we examined the responsiveness of retinal venules to ET_BR agonist sarafotoxin S6c. Although sarafotoxin S6c can cause constriction of isolated retinal arterioles (30), it failed to alter tone of retinal venules from control or diabetic pigs (Fig. 1B). The ET_BRs are commonly localized in endothelial cells, so this finding suggests little role of the endothelium in mediating the ET-1–induced response of retinal venules. However, we were not able to directly confirm this via endothelial denudation because of the thin venular wall that posed difficulty in removing the endothelium without damage to the underlying vascular smooth muscle in our pilot studies. Collectively, our results indicate differences in vasomotor characteristics of retinal venules and arterioles with a dominant role of venular smooth muscle ET_AR in mediating constriction to ET-1. Because ET_AR blockade reversed the diminished resting blood flow in retinal venules after 4 weeks of STZ-induced diabetes (6) and our previous (9) and current findings demonstrated an enhanced ability of these vessels to constrict to ET-1, activation of ET-1/ET_AR and subsequent signaling pathways (see schematic diagram in Fig. 7) appears to contribute to this early complication of diabetes.

In our previous studies we provided the first evidence for the prominent activation of ET_ARs and subsequent signaling mechanisms of extracellular Ca²⁺ entry and activation of Rho kinase (ROCK) in constriction of retinal venules to ET-1 (9–11). Additional proteins that may contribute to the ET-1–induced constriction of retinal venules are the MAPKs (18–20). This highly conserved family of serine/threonine protein kinases that respond to extracellular stimuli consists of three conventional isoforms: JNK, p38 kinase, and ERK (33). We assessed for the first time whether MAPKs contribute to or modulate the constriction of retinal venules to ET-1. In the current study, we found that only the JNK inhibitor diminished the ET-1–induced constriction of control retinal venules (Fig. 4), suggesting the critical role of this MAPK in mediating venular constriction to ET-1. Also, the maintenance of basal tone of retinal venules appears dependent on JNK because the JNK inhibitor increased the resting diameter of retinal venules. The overlap of JNK-mediated vasoconstriction between maintenance of basal tone and pharmacologic activation by ET-1 in the retinal venule deserves further investigation. The current findings from retinal venules extend previous evidence that ET-1 and other vasoconstrictors such as norepinephrine and local anesthetic levobuvicaine cause JNK-dependent contraction of the rat aorta (22,34). Interestingly, a previous report in rat aortic vascular smooth muscle cells suggested that ROCK can activate downstream target JNK for cell migration (35). Our recent finding on the involvement of ROCK in retinal venular constriction to ET-1 (10) appears to support the upstream link of ROCK to JNK in mediating venular constriction to ET-1 (Fig. 7). Moreover, the requisite role of extracellular Ca²⁺ entry in retinal venular constriction to ET-1 is corroborated by our present findings (Fig. 2). Because ROCK activation can be Ca²⁺ dependent (36), the potential link of the Ca²⁺ entry–ROCK–JNK axis for retinal venular constriction to ET-1 deserves further investigation.

In our recent study we showed that pharmacologic blockade of reverse-mode NCX with KB-R7943 prevented the enhanced ET-1–induced constriction of diabetic retinal venules without affecting the vasoconstrictor response of control vessels (9). Because activation of reverse-mode NCX causes efflux of sodium (Na⁺) and influx of Ca²⁺ (37), it is likely that the enhanced extracellular Ca²⁺ entry in diabetic venules augments the venular constriction to ET-1 (Fig. 7). However, the events that contribute to reverse-mode NCX activation remain unclear. Based on evidence from diabetic cardiomyocytes, possible signaling events triggered by hyperglycemia upstream of NCX include activation of NHE1 or PKC (15,38). In the current study, we demonstrated that pharmacologic blockade of PKC did not affect the enhanced ET-1–induced constriction of retinal venules in pigs with early diabetes. Furthermore, both control and diabetic vessels were unresponsive to PKC activator PDBu, suggesting that PKC activation is not involved in constriction of retinal venules to ET-1 under normal or diabetes conditions (Fig. 3). On the other hand, we found that the selective NHE1 inhibitor cariporide abolished the diabetes-enhanced constriction of retinal venules to ET-1 (Fig. 1A), in a similar manner to that of KB-R7943, as we reported in our previous study (9). The activation of NHE1 causes transmembrane exchange of a proton for an extracellular Na⁺, which elevates the intracellular Na⁺ level that can drive reverse-mode NCX to increase Ca²⁺ entry (39). This pathway, linkage of NHE1 to reverse-mode NCX, seems to be activated in retinal venules by diabetes because the combined treatment with cariporide and KB-R7943 did not further inhibit ET-1–induced vasoconstriction (Fig. 1A). Collectively, these data suggest the series connection of NHE1 to NCX. Since NHE1 mRNA expression was unaltered during 2 weeks of diabetes (Fig. 1D), our molecular and functional data support the increased NHE1 activity in mediating enhanced venular constriction to ET-1 via reverse-mode NCX activation (Fig. 7). This notion is consistent with and supported by the prevention of diabetes-induced increase in retinal microvascular resistance and improvement of retinal blood flow velocity after oral administration of cariporide in diabetic rats (27).

The NHE1 protein is composed of a membrane domain for ion transport and a cytosolic regulatory domain. Ion transport activity of NHE1 via the membrane domain can be regulated by phosphorylation of the cytosolic COOH terminus tail region of this protein (40). There are several lines of evidence suggesting the involvement of two conventional MAPKs, ERK and p38, in regulation of NHE1 activity in different tissues (41–43). Both ERK and p38 kinase can increase activity of NHE1 by binding and phosphorylating specific COOH terminus domain regions (40,44). High glucose exposure can increase NHE1 activity in kidney cells by enhancing p38 kinase signaling (41). Moreover, the ability of ET-1/ET_AR to promote p38 MAPK-dependent activation of NHE1 was suggested in Chinese hamster ovary cells (42), and ERK-dependent phosphorylation and activation of NHE1 was reported in rat cardiomyocytes (43). Herein, we assessed the roles of ERK, p38, and JNK MAPKs in the enhanced constriction of retinal venules to ET-1 in diabetes. Our current data show that p38 kinase inhibitor SB203580, but not ERK inhibitor PD98059, prevented the enhanced constriction of diabetic retinal venules to ET-1 (Fig. 4). Interestingly, JNK blockade not only reduced the basal tone but also influenced the vasoconstriction to ET-1 in diabetic venules similarly to what was observed in control vessels, i.e., a parallel shift of ET-1 concentration–response curves (Fig. 4A). It appears that JNK inhibition has a general impact on vasoconstriction to ET-1 in retinal venules. The failure of JNK to completely block ET-1–induced vasoconstriction may be due to the insufficient concentration of SP600125 (10 μmol/L) used in the current study. This contention is supported by a previous report that 100 μmol/L SP600125 was required to nearly abolish the constriction of the rat aorta to norepinephrine (22). A limitation of our study is the inability to use SP600125 concentrations >10 μmol/L because the required amount of DMSO solvent compromises constriction of retinal venules and confounds our data interpretation. Nonetheless, the MAPK data from the current study suggest that p38 kinase may act as a sensor signaling molecule for hyperglycemic insult and mediate the enhanced ET-1–induced constriction of retinal venules in early type 1 diabetes. The JNK may act as an effector signaling molecule to exert vasoconstriction to ET-1 (Fig. 7).

One of the limitations of the current study is the inability to reliably assess the expression/activity of some signaling proteins in the venular tissue. We are only able to isolate at most ∼5 µg protein of retinal venule tissue from a single pig. The results of our previous study showed much greater protein expression of p38 in neural retina than retinal venules for 5 µg samples (10). Since the neural retina tissue is more available, we used it for the antibody optimization and efficacy test. In the neural retina tissue, p38, p-p38, JNK, and p-JNK were expressed at 15–20 µg of sample (Fig. 5B), which supports the reliability of these antibodies for porcine tissue. The p38 protein expression was also detected in retinal venule samples (Fig. 5A), but p-p38, JNK, and p-JNK were undetected. Although these proteins were not detected, this result does not preclude the expression of these proteins in the retinal venules. Because the enhanced ET-1–induced constriction of diabetic retinal venules was sensitive to p38 kinase inhibitor SB203580 and the p38 protein expression was comparable for retinal venule samples from control and diabetic pigs (Fig. 5A), we speculate that diabetes-elevated p38 kinase activity in retinal venules contributes to the enhanced constriction to ET-1 independent of changes in the p38 protein level.

Because there was not sufficient protein from retinal venules to measure all indicated potential signaling molecules, the pharmacological approach was used in the current study to characterize the altered vasomotor activity. Our data interpretation relies on the selectivity and specificity of pharmacological blockers, which were based on evidence for specific target efficacy from our and other previous studies as discussed below. The concentration of SB203580 (0.1 μmol/L) used in the current study has been shown to be in the IC₅₀ range for specifically blocking p38 kinase activity without altering JNK, ERK, and several other protein kinases (45) and was the concentration that nearly abolished p38 phosphorylation of monocytes treated with inflammatory stimuli (46) and that blocked p38 kinase activation for superoxide production in isolated coronary arterioles (47). The JNK inhibitor SP600125 at 3–5 μmol/L was shown to reverse diabetes-induced vasomotor dysfunction of coronary (48) and retinal (24) arterioles in our previous studies. In addition, the 10 μmol/L concentration used in the current study can nearly abolish JNK activity and reduce JNK phosphorylation without altering p38 and ERK phosphorylation (49). We previously showed that the specific ERK inhibitor PD98059 at 0.3–1.0 μmol/L effectively blocked retinal arteriolar dilations in response to nitric oxide synthase activation (50,51). This inhibitor at 10 μmol/L has been shown to abolish ERK activity in vitro without affecting p38 and JNK activities (52), but this concentration did not alter the constriction of retinal venules to ET-1 in the current study. Interestingly, we found that retinal arterioles constrict to the PKC activator PDBu in a manner sensitive to the PKC inhibitor Gö 6983 (10 μmol/L) (28), but this inhibitor did not alter resting tone or ET-1–induced constriction of retinal arterioles (28) and venules (10) and had no impact on enhanced vasomotor activity under diabetic insult (Fig. 3). These findings suggest that PKC does not play a role in vasoconstriction to ET-1 in retinal microvessels. The NHE1 inhibitor cariporide was used in our current study at a concentration of 1 μmol/L, which is in the IC₅₀ range (0.03–3.4 μmol/L) for selectively blocking NHE1 activity (53). Overall, the potential limitations of the pharmacological approach were noted in the current study and the results were interpreted with caution.

The direct interaction of p38 and NHE1 was also not established in the current study. However, because p38 kinase activation can be regulated by Ca²⁺ signaling (54), we surmise that the initial increase in intracellular Ca²⁺ in response to ET-1 triggers the p38 pathway and subsequently augments further Ca²⁺ entry via downstream NHE1/NCX signaling, which is specifically activated by hyperglycemia or early diabetes (Fig. 7). The contribution of increased ET-1–stimulated intracellular Ca²⁺ is supported by the current finding that removal of extracellular Ca²⁺ abolished the ET-1–induced constriction of retinal venules from diabetic pigs (Fig. 2). Interestingly, NHE1-dependent activation of reverse-mode NCX by phenylephrine can lead to JNK signaling in rat ventricular myocytes (55). Therefore, it is possible that the heightened intracellular Ca²⁺, via NHE1/NCX signaling, in early diabetes augments JNK activity, leading to enhanced retinal venular constriction to ET-1 (Fig. 7).

The increased plasma ET-1 level in diabetic pigs (Fig. 6) is consistent with our previous finding of elevated ET-1 in the vitreous within 2 weeks of diabetes induction (9), further supporting activation of the endothelin system in functional abnormality of the retinal microcirculation. The plasma level of ET-1 in the diabetic pigs was in the 8–10 pmol/L range, while the constriction to 10 pmol/L ET-1 was significantly greater in retinal venules from diabetic pigs than from control pigs. A similar elevation of plasma ET-1 (∼10 pmol/L) has been reported in human patients with diabetes (56,57). In addition, the estimated concentration of ET-1 at the local microvasculature is in the nanomolar range (58). Therefore, the concentrations of ET-1 eliciting retinal venular constriction in the current study are within the physiological and pathophysiological ranges.

Structural changes in the retinal microvessels, such as the composition of the endothelial glycocalyx (59), may contribute to the alteration in vascular function during diabetes. Interestingly, ET_AR blockade has been shown to prevent degradation of the endothelial glycocalyx of the renal microcirculation of diabetic mice (60). Although the impact of the glycocalyx on ET-1–induced constriction is unclear, a significant reduction in the thickness of the glycocalyx of retinal arterioles but not of retinal venules was noted in mice with type 1 diabetes (Ins2 Akita) (61). This finding does not rule out the potential impact of diabetes on the glycocalyx of porcine retinal venules and future evaluations of structural integrity and endothelial glycocalyx in relation to ET-1–mediated vasomotor dysregulation are factors to consider in the retinal microcirculation during diabetic insult. The elevated level of ET-1 in the vitreous and plasma during early diabetes may exert a dual impact on retinal perfusion by directly promoting smooth muscle–dependent constriction of retinal venules and indirectly impairing endothelium-dependent nitric oxide–mediated dilation of retinal arterioles (14) via glycocalyx degradation (62).

In summary, we found that activation of NHE1 and JNK contributes in part to the development and maintenance of basal tone of isolated porcine retinal venules. The constriction of normal retinal venules to ET-1 is mediated by Ca²⁺ extracellular entry and JNK activation. The enhanced ET-1–induced constriction of retinal venules, observed in early type 1 diabetes, is mediated by augmented extracellular Ca²⁺ entry via activations of p38 kinase, NHE1, and NCX but not by PKC or ERK. Molecular targeting of these signaling molecules linked to enhanced Ca²⁺ entry in retinal venules might provide effective therapeutic strategies to alleviate microvascular dysregulation and retinal complications in early diabetes.

Acknowledgments. The authors thank Wenjuan Xu, Dr. Xin Xu, and Dr. Shu-Huai Tsai (Department of Medical Physiology at the Texas A&M University Health Science Center) for technical assistance.

Funding. This work was supported by National Eye Institute, NIH, grants R01EY023335 and R01EY024624 (T.H.) and the Retina Research Foundation (T.H. and L.K.).

Duality of Interest. There are no potential conflicts of interest relevant to this article.

Author Contributions. Y.-L.C. designed experiments, conducted experiments, analyzed data, and wrote the manuscript. Y.R. conducted experiments. R.H.R. reviewed and edited the manuscript. L.K. reviewed and edited the manuscript. T.W.H. designed experiments and reviewed and edited the manuscript. T.W.H. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.