Cyclic Stretch–Induced Reactive Oxygen Species Generation Enhances Apoptosis in Retinal Pericytes Through c-Jun NH2-Terminal Kinase Activation
Hypertension is known to exacerbate diabetic complications, such as retinopathy and nephropathy. Apoptosis of retinal vascular pericytes has been well established as the earliest conceivable change in diabetic retinopathy. In this study, we investigated the contribution of cyclic stretch, which mimics a hypertensive state to pericyte apoptosis. A 48-hour cyclic stretch induced DNA fragmentation in porcine retinal pericytes and increased the number of TUNEL+ cells at a pathophysiologically relevant extension level (10%/60 cycles per minute). Stretch also increased intracellular reactive oxygen species generation and increased c-Jun NH2-terminal kinase phosphorylation in a time- and magnitude-dependent manner, which were reduced by the nicotinamide-adenine dinucleotide phosphate oxidase inhibitor diphenylene iodonium or dominant-negative protein kinase C-δ. Stretch activated protein kinase C-δ and increased its association with p47phox. Stretch induced cleavage of caspase-9 and -3 and increased caspase-3 activity. Protein kinase C-δ or c-Jun NH2-terminal kinase inhibition normalized stretch-induced caspase-3 activity and prevented stretch-induced apoptosis. These data indicate that cyclic stretch induces apoptosis in porcine retinal pericytes by activation of the reactive oxygen species–c-Jun NH2-terminal kinase–caspase cascades, suggesting a novel molecular mechanism to explain the exacerbation of early diabetic retinopathy by concomitant hypertension.
Etiologic studies demonstrate that hypertension increases the risk of diabetic retinopathy progression.1 In a hypertensive state, hemodynamic overload increases vascular wall stress to cause various responses in vascular cells. Recent advances have elucidated the molecules that affect vascular changes, such as nicotinamide-adenine dinucleotide phosphate [NAD(P)H] oxidase and mitogen-activated protein kinases, including c-Jun NH2-terminal kinase (JNK).2–4
Intramural retinal pericytes, which seem to be like smooth muscle cells, fibroblasts or mesenchymal cells surround a single layer of endothelial cells.5 Pericytes exhibit a 1:1 ratio with endothelial cells in human retinal capillaries and might control the retinal vascular tone directly or indirectly. In addition, pathology reports have definitely shown that one of the earliest changes in diabetic retinopathy is the loss of retinal pericytes.6 The studies of coculture of endothelial cells and pericytes have suggested that pericytes have both antiapoptotic and antiproliferative effects on endothelial cells.7 This evidence supports that pericytes play an important role in the prevention of diabetic retinopathy. However, it remains to be investigated how hypertension exacerbates pericyte loss in diabetic retinopathy.
In this study, we elucidated that cyclic stretch, which mimics systemic hypertension, induced the activation of protein kinase C (PKC)-δ–reactive oxygen species (ROS)–JNK–caspase cascades, resulting in apoptotic changes in porcine retinal pericytes (PRPCs). This observation suggests that coexistent systemic hypertension may accelerate the development of diabetic retinopathy by promoting pericyte loss via generation of ROS.
All of the experiments were repeated ≥4 times using a different preparation of PRPCs. Results are expressed as mean±SD. Statistical analysis used Student’s t test or ANOVA to compare quantitative data populations with normal distributions and equal variance. Data were analyzed using the Mann–Whitney rank sum test or the Kruskal–Wallis test for populations with nonnormal distributions or unequal variance. A P value <0.05 was considered statistically significant. Reagents, cells and cell culture, recombinant adenoviruses, mechanical stretch, immunoprecipitation and immunoblotting, caspase-3 activity, TUNEL, analysis of DNA fragmentation, and measurement of ROS are in the expanded materials and methods in the online data supplement available at http://hyper.ahajournals.org.
Stretch-Induced Apoptosis in PRPCs
We first investigated whether cyclic stretch affects cell viability in PRPCs. DNA fragmentation analysis showed that 10%/60 cycles per minute pulsatile cyclic stretch for 48 hours, but not 2% cyclic stretch, induced accumulation of extranuclear DNA (Figure 1A). In addition, TUNEL analysis revealed that the cells subjected to 10% cyclic stretch showed significant increases in apoptotic cells in a time-dependent manner (P<0.01 versus control; Figure 1B), whereas cells under 2% cyclic stretch did not (Figure 1C). These data suggest that 10% cyclic stretch, which had been shown to be pathophysiologically relevant,8,9 can induce apoptotic changes in PRPCs.
Cyclic Stretch Activates JNK in PRPCs
To clarify the potential molecular mechanisms involved in stretch-induced cell death, JNK phosphorylation was analyzed. As shown in Figure 2A and 2B, mimicking pathophysiological amounts of cyclic stretch caused an increased phosphorylation of JNK and SAPK/extracellular signal regulated kinase kinase (an upstream regulator of JNK) in a time-dependent manner. The JNK activation depended on the magnitude of stretch applied to the cells (Figure 2C).
Stretch-Induced PRPC Apoptosis Is Reduced by JNK Inhibition
To discover whether activation of JNK was essential for stretch-induced apoptosis, we performed experiments of JNK inhibition. Both JNK inhibitor I and dominant negative (dn) JNK decreased stretch-induced apoptosis significantly (P<0.01 versus Ad-green fluorescent protein with stretch), whereas wild type (wt) JNK increased the apoptosis (Figure 3A and 3B).
Recently, it has been shown that JNK activation increases ROS generation,10 and we evaluated the increase in our experiments. PRPCs were incubated with 2′-7′-dichlorofluorescein diacetate and exposed to 10%/60 cycles per minute of cyclic stretch for 30 minutes. In PRPCs transfected with control adenoviruses, stretch increased the ROS level significantly (Figure 3C). However, wt JNK or dn JNK did not change the ROS level significantly in PRPCs subjected to stretch (Figure 3C).
Stretch-Induced JNK Phosphorylation Is Mediated by ROS and PKC-δ
On the other hand, JNK activation is induced by ROS in various cell types, and we investigated whether ROS increases JNK phosphorylation in PRPCs subjected to stretch. We found that N-acetyl-l-cysteine and diphenylene iodonium (DPI), both of which interfere with ROS generation, decreased JNK phosphorylation levels (Figure 4A). In addition, GF109203X was shown to attenuate JNK activation. Because GF109203X inhibits both classic and novel PKCs, we attempted to identify which PKC isoforms were required for the stretch-induced JNK activation and found that GÖ6796, a classical PKC inhibitor, had no effect on JNK phosphorylation.
These results prompted us to investigate the role of PKC-δ, which is one of the novel PKCs, using adenoviruses encoding PKC-δ mutants (Figure 4B). In stretched cells, wt PKC-δ augmented JNK activation, whereas dn PKC-δ abolished phosphorylation to the control level. These data suggest that stretch-induced JNK phosphorylation is mediated by ROS and PKC-δ.
Next we tested whether ROS and PKC-δ are involved in stretch-induced PRPC apoptosis. As Figure 5A shows, DPI reduced stretch-induced PRPC apoptosis significantly (P<0.01 versus control with stretch). Furthermore, wt PKC-δ overexpression augmented stretch-induced apoptosis by 58%, and the inhibition of PKC-δ with dn PKC-δ overexpression reduced apoptotic cells to the control level (Figure 5B).
Stretch Increased ROS Generation via NADPH Oxidase and PKC-δ
We investigated whether NAD(P)H oxidase and PKC-δ affect ROS generation induced by stretch and found that stretch-induced ROS generation was significantly decreased by NAD(P)H oxidase inhibitors DPI and apocynin (P<0.01 versus control with stretch, in both; Figure 6A). In addition, wt PKC-δ increased stretch-induced ROS generation by 46%, whereas dn PKC-δ decreased ROS generation by 104% (Figure 6B).
PKC-δ was reported recently to be required for the phosphorylation of p47phox and subsequent assembly and activation of NAD(P)H oxidase.11 To elucidate how PKC-δ is involved in NAD(P)H oxidase activation in PRPCs, the cells were exposed to cyclic stretch for 15 minutes, and immunoprecipitation and immunoblot analysis were performed using anti-PKC-δ or anti-p47phox antibody (Figure 7A). p47phox and PKC-δ did not show assembly in the static state.11 However, stretch induced rapid PKC-δ and p47phox association, which suggests that stretch-induced NAD(P)H oxidase activation accompanies increasing association between PKC-δ and p47phox.
Stretch-Induced PKC-δ Activation
Among the PKCs, the PKC-δ isoform is thought to participate in apoptosis in several cell systems.12 Then, we assessed stretch-induced PKC-δ phosphorylation at Thr505. As Figure 7B shows, stretch caused rapid and sustained phosphorylation of PKC-δ as quickly as 2 minutes after stretch began. Phosphorylation increased gradually to 2.7-fold after 60 minutes.
Mechanical Stretch Activates Caspase Cascades Through PKC-δ and ROS-Dependent JNK Activation
Among many caspases, the downstream effector, caspase-3, has been shown to play a pivotal role in the terminal, execution phase of apoptosis induced by a variety of stimuli. As Figure 8A shows, cyclic stretch treatment induced cleavage of procaspase-3 in a time-dependent manner. To examine whether caspase-3 activation is involved in stretch-induced PRPC apoptosis, PRPCs were pretreated with a caspase-3–specific inhibitor (DEVD-CHO) and exposed to cyclic stretch in a similar manner. DEVD-CHO decreased the number of stretch-induced TUNEL-positive cells (P<0.01 versus control with stretch), confirming that stretch-induced PRPC death is caspase–cascade dependent (Figure 8B). Next, the activity of caspase-3 was evaluated. Cyclic stretch significantly increased the activity of caspase-3 by 548% (Figure 8C).
Furthermore, we investigated whether stretch-induced activity of caspase-3 is mediated by ROS, PKC-δ, or JNK. The activation was significantly suppressed by pretreatment with N-acetyl-l-cysteine and DPI (Figure 8C). Figure 8D shows that, under the stretch, wt PKC-δ increased the caspase-3 activity, whereas dn PKC-δ decreased it significantly (Figure 8D). Additional experiments using JNK inhibitor I and dn JNK elucidated that the inhibition of JNK decreased caspase-3 activity induced by cyclic stretch (Figure 8E and 8F).
Although stretch has been reported to induce cell death in several cell systems, the underlying molecular mechanisms remain largely unknown. The present study has explored the intracellular pathways involved in cell death induced by cyclic stretch, which is one of several mechanical stresses in the hypertensive state.
JNK, p38, and extracellular-signal regulated protein kinase 1/2 have all been shown to activate by mechanical stretch.3,4,13 Investigations using cardiac myocytes and smooth muscle cells suggest that activated JNK is involved in hypertrophic or apoptotic changes contributing to pathological alterations in cardiovascular systems.14 In the present study, we observed that JNK is rapidly activated by cyclic stretch, resulting in the sustained activation that may lead to the apoptotic process. A similar long-lasting activation of the kinase has been reported in different forms of stress-induced apoptosis.15
Cellular oxidative stress is another important regulator of JNK.16 Previous studies indicate that ROS-induced JNK activation is involved in hypertrophic response.4 We show here that DPI and N-acetyl-l-cysteine, both counteracting with ROS, not only block JNK activation but also provide protection against stretch-induced apoptotic death. This result demonstrates a link among oxidative imbalance, JNK activation, and apoptosis. In addition, recent reports have shown that JNK activation results in ROS generation through NAD(P)H oxidase or mitochondria,10 which might lead to a positive feedback loop. However, stretch-induced ROS generation was not mediated via JNK activation in PRPCs. It suggests that the ROS is one of the upstream regulators of JNK in this system. In addition, the experiments with long-term stretch remain to be performed.
Proapoptotic roles for PKCs have been described for several cell systems and stimuli.12 We have found that PKC-δ is involved in executing the apoptotic process induced by stretch. We have also observed that PKC-δ inhibition completely blocks JNK activation resulting in caspase-3 activation. This dependence on PKC-δ activity for full activation of caspase-3 has been reported in other cellular systems.12
Interestingly, PKC-δ phosphorylation at Thr505 was gradual, in contrast to the rapid activation seen with JNK. Extensive studies suggest that PKC-δ is activated late in the apoptotic process and that it is involved in translocating bcl-2 into mitochondria.17 Activation of caspase-3 is also known to be PKC-δ dependent,12 and PKC-δ is, in turn, activated by caspase-3.18 PKC-δ is reported to be involved in the cell cycle as well.19 Our observation of gradual PKC-δ phosphorylation suggests that PKC-δ activation may occur through multiple steps. As Garcia-Fernandez et al12 suggested, a positive feedback loop may exist whereby caspase-mediated activation of PKC-δ results in amplification of the caspase cascade and subsequent apoptosis. We found that DPI inhibition of ROS generation successfully rescued retinal pericytes from stretch-induced apoptosis. The same result was also seen with PKC-δ inhibition using adenoviruses. We could not exclude a possible late effect of PKC-δ on a stretch-induced apoptotic event. However, successful inhibition of ROS generation and JNK activation by dn PKC-δ overexpression indicates that stretch-induced PKC-δ activation is an early step of the apoptotic event. This upstream activation of PKC-δ likely plays an important role in stretch-induced apoptosis. In addition, a recent report showed that the PKC-δ phosphorylation at Phe500/Phe527 in or near the activation loop also affects its activation.20 It remains to be investigated whether stretch induces the phosphorylation of these residues.
Vascular cells produce ROS through activation of NAD(P)H oxidase, and NAD(P)H oxidase is a more important source of ROS in intact arteries than arachidonic acid-metabolizing enzymes, xanthine oxidase sources, or mitochondrial sources.21 In smooth muscle cells, NAD(P)H oxidase consists of the proteins p47phox, p22phox, Nox1 or 4, and rac.22 Once stimulated, these protein subunits assemble and act as a functioning kinase. Translocation of the cytoplasmic components and activation of the oxidase are controlled by phosphorylation of residues in the C terminus of p47phox.23 Depletion of p47phox eliminates NAD(P)H oxidase activation in smooth muscle cells.24 Very recently, Brown et al11 elegantly demonstrated that it is PKC-δ that induces this assembly and activation of NAD(P)H oxidase by phosphorylating p47phox. Our observation that stretch treatment induced rapid association of PKC-δ and p47phox fits this scenario, which also explains how PKC-δ inhibition was able to decrease stretch-induced ROS generation and suppress the stretch-induced apoptotic response.
As mentioned, caspase activities are heavily involved in inducing programmed cell death. Our results indicate that stretch induces procaspase-9 and -3 activation. Caspase-3 inhibition reduced stretch-induced apoptosis, which supports the conclusion that stretch-induced cell death is an apoptotic event and that the activation of caspase cascades is required in stretch-induced cell apoptosis. The inhibition of ROS, JNK, or PKC-δ all prevented stretch-induced caspase-3 activation, which implies that ROS, JNK, and PKC-δ are indispensable.
Caspase-8 and caspase-9 play a role in the upstream events of caspase-3 activation. Although caspase-8 is involved in death receptor–dependent apoptosis, caspase-9 is involved in the mitochondrial pathway.25 As Brownlee26 suggested, in diabetes, hyperglycemia might induce mutations in mitochondrial DNA, which could eventually cause increased superoxide production at physiological concentrations of glucose, even in the later absence of hyperglycemia. Ellis et al27 reported that NAD(P)H oxidase activity is increased in the diabetic rodent retina. Our observation that stretch is capable of activating caspase-9 suggests that the hypertensive state may also participate in this mitochondrial pathway, which is modified by hyperglycemia, and promote retinopathy progression.
In summary, we have shown that cyclic stretch increases the intracellular generation of ROS in retinal pericytes via NAD(P)H oxidase activation. This activation occurs through PKC-δ. The increased ROS levels from stretch treatment cause PRPC apoptosis through JNK-dependent caspase-3 activation. Coexistent systemic hypertension may accelerate the development of diabetic retinopathy by promoting pericyte loss via generation of ROS, thus augmenting the mechanisms of diabetes-induced complications.
Sources of Funding
This work was supported by Grants-in-Aid for Scientific Research from the Ministry of Education, Science, and Culture and the Ministry of Health and Welfare of the Japanese Government (70283596 and 80335265), by the Japan National Society for the Prevention of Blindness, and by the Takeda Science Foundation.
The first 2 authors contributed equally to this work.
- Received May 31, 2006.
- Revision received June 19, 2006.
- Accepted November 10, 2006.
Hishikawa K, Luscher TF. Pulsatile stretch stimulates superoxide production in human aortic endothelial cells. Circulation. 1997; 96: 3610–3616.
Komuro I, Kudo S, Yamazaki T, Zou Y, Shiojima I, Yazaki Y. Mechanical stretch activates the stress-activated protein kinases in cardiac myocytes. FASEB J. 1996; 10: 631–636.
Mayr M, Hu Y, Hainaut H, Xu Q. Mechanical stress-induced DNA damage and rac-p38MAPK signal pathways mediate p53-dependent apoptosis in vascular smooth muscle cells. FASEB J. 2002; 16: 1423–1425.
Betsholtz C, Lindblom P, Gerhardt H. Role of pericytes in vascular morphogenesis. Exs. 2005: 115–125.
Safar ME, Peronneau PA, Levenson JA, Toto-Moukouo JA, Simon AC. Pulsed Doppler: diameter, blood flow velocity and volumic flow of the brachial artery in sustained essential hypertension. Circulation. 1981; 63: 393–400.
Garcia-Fernandez LF, Losada A, Alcaide V, Alvarez AM, Cuadrado A, Gonzalez L, Nakayama K, Nakayama KI, Fernandez-Sousa JM, Munoz A, Sanchez-Puelles JM. Aplidin induces the mitochondrial apoptotic pathway via oxidative stress-mediated JNK and p38 activation and protein kinase C delta. Oncogene. 2002; 21: 7533–7544.
Wang Y, Su B, Sah VP, Brown JH, Han J, Chien KR. Cardiac hypertrophy induced by mitogen-activated protein kinase kinase 7, a specific activator for c-Jun NH2-terminal kinase in ventricular muscle cells. J Biol Chem. 1998; 273: 5423–5426.
Tobiume K, Matsuzawa A, Takahashi T, Nishitoh H, Morita K, Takeda K, Minowa O, Miyazono K, Noda T, Ichijo H. ASK1 is required for sustained activations of JNK/p38 MAP kinases and apoptosis. EMBO Rep. 2001; 2: 222–228.
Benhar M, Dalyot I, Engelberg D, Levitzki A. Enhanced ROS production in oncogenically transformed cells potentiates c-Jun N-terminal kinase and p38 mitogen-activated protein kinase activation and sensitization to genotoxic stress. Mol Cell Biol. 2001; 21: 6913–6926.
Pongracz J, Webb P, Wang K, Deacon E, Lunn OJ, Lord JM. Spontaneous neutrophil apoptosis involves caspase 3-mediated activation of protein kinase C-delta. J Biol Chem. 1999; 274: 37329–37334.
Fukumoto S, Nishizawa Y, Hosoi M, Koyama H, Yamakawa K, Ohno S, Morii H. Protein kinase C delta inhibits the proliferation of vascular smooth muscle cells by suppressing G1 cyclin expression. J Biol Chem. 1997; 272: 13816–13822.
Liu Y, Belkina NV, Graham C, Shaw S. Independence of protein kinase C-delta activity from activation loop phosphorylation: structural basis and altered functions in cells. J Biol Chem. 2006; 281: 12102–12111.
Griendling KK, Sorescu D, Ushio-Fukai M. NAD(P)H oxidase: role in cardiovascular biology and disease. Circ Res. 2000; 86: 494–501.
Lassegue B, Sorescu D, Szocs K, Yin Q, Akers M, Zhang Y, Grant SL, Lambeth JD, Griendling KK. Novel gp91(phox) homologues in vascular smooth muscle cells: nox1 mediates angiotensin II-induced superoxide formation and redox-sensitive signaling pathways. Circ Res. 2001; 88: 888–894.
Nauseef WM, Volpp BD, McCormick S, Leidal KG, Clark RA. Assembly of the neutrophil respiratory burst oxidase. Protein kinase C promotes cytoskeletal and membrane association of cytosolic oxidase components. J Biol Chem. 1991; 266: 5911–5917.
Joza N, Susin SA, Daugas E, Stanford WL, Cho SK, Li CY, Sasaki T, Elia AJ, Cheng HY, Ravagnan L, Ferri KF, Zamzami N, Wakeham A, Hakem R, Yoshida H, Kong YY, Mak TW, Zuniga-Pflucker JC, Kroemer G, Penninger JM. Essential role of the mitochondrial apoptosis-inducing factor in programmed cell death. Nature. 2001; 410: 549–554.