From the Department of Physiology (Z.B., W.L., T.H.N., A.K., G.K.), New York Medical College, Valhalla, NY; the Department of Pathophysiology (A.K.), Semmelweis University, Budapest, Hungary; and the Division of Clinical Physiology Z.B., N.E., A.T.), Institute of Cardiology, University of Debrecen, Debrecen, Hungary.
Correspondence to Zsolt Bagi, MD, PhD, Division of Clinical Physiology, Institute of Cardiology, University of Debrecen, 4004 Debrecen, PO Box 1, Hungary. E-mail email@example.com
Objective— Type 2 diabetes mellitus (T2-DM) is frequently associated with vascular dysfunction and elevated blood pressure, yet the underlying mechanisms are not completely understood. We hypothesized that in T2-DM, the regulation of peripheral vascular resistance is altered because of changes in local vasomotor mechanisms.
Methods and Results— In mice with T2-DM (C57BL/KsJ-db–/db–), systolic and mean arterial pressures measured by the tail cuff method were significantly elevated compared with those of control (db+/db–) animals (db/db, 146±5 and 106±2 mm Hg versus control, 133±4 and 98±4 mm Hg, respectively; P<0.05). Total peripheral resistance, calculated from cardiac output values (measured by echocardiography) and mean arterial pressure were significantly elevated in db/db mice (db/db, 25±6 versus control, 15±1 mm Hg[middot]mL–1[middot]min–1). In isolated, pressurized gracilis muscle arterioles (diameter 80 μm) from db/db mice, stepwise increases in intraluminal pressure (from 20 to 120 mm Hg) elicited a greater reduction in diameter than in control vessels at each pressure step (at 80 mm Hg, db/db, 66±4% versus control, 79±3%). The passive diameters of arterioles (obtained in Ca2+-free solution) and the calculated myogenic index were not significantly different in the 2 groups. The presence of the prostaglandin H2/thromboxane A2 receptor antagonist SQ29548 did not affect arteriolar diameters of control mice but reduced the enhanced arteriolar tone of db/db mice back to control levels (at 80 mm Hg, 80±4%). The inhibitor of cyclooxygenase-1 (COX-1), SC-560, did not affect the basal tone of arterioles, whereas NS-398, an inhibitor of COX-2, caused a significant shift in the arteriolar pressure–diameter curve of vessels from db/db mice (at 80 mm Hg, 76±3%) but not in those of control mice. Also, in aortas of db/db mice, expression of COX-2 was enhanced compared with controls.
Conclusions— Collectively, these findings suggest that in mice with T2-DM, the basal tone of skeletal muscle arterioles is increased because of an enhanced COX-2–dependent production of constrictor prostaglandins. These alterations in microvascular prostaglandin synthesis may contribute to the increase in peripheral resistance and blood pressure in T2-DM.
Here we report that mice with type 2 diabetes mellitus have elevated systolic blood pressures and increased peripheral vascular resistance. In type 2 diabetic mice, these alterations are associated with enhanced skeletal muscle arteriolar tone, which is likely attributable to increased release of COX-2–derived constrictor prostaglandins within the arteriolar wall.
Key Words: type 2 diabetes mellitus ? microvessels ? basal arteriolar tone ? cyclooxygenase-2
Type 2 diabetes mellitus (T2-DM), which has reached epidemic proportions in Western countries, is associated with a markedly increased incidence of cardiovascular diseases, accounting for 70% of deaths in the diabetic population.1 However, the exact relations among T2-DM, obesity, and cardiovascular disease are not completely understood and have been the subject of some dispute. T2-DM is often part of an array of complex abnormalities referred to as the metabolic syndrome, which is frequently accompanied by elevated blood pressure.2
Several studies have demonstrated that vasomotor dysfunction of microvessels is an early manifestation of the vascular complications in T2-DM.3,4 Alterations in local vasoregulatory mechanisms intrinsic to the vascular wall, such as enhanced pressure-induced arteriolar tone5–7 and reduced endothelium-dependent dilation,6,8–11 have been reported previously as characteristic of T2-DM. Changes in the local vasoregulatory mechanisms of peripheral microvessels may significantly influence vascular resistance in T2-DM; however, the possible underlying mechanisms are still open to question.
Recently, a key role for low-grade vascular inflammation has received great attention in the development of diabetic vascular complications.12,13 Among other factors, prostaglandins (PGs) are important mediators of several inflammatory mechanisms14; however, it is also known that many PG derivatives have specific vasoactive properties, thereby contributing to the local regulation of arteriolar diameter.15,16 Early reports have already proposed a key role for altered vascular PG metabolism in diabetes-related changes in local vasoregulatory mechanisms.17 It was found that in mesenteric arteries of T1-DM dogs, exogenous arachidonic acid elicited thromboxane A2 (TxA2)–mediated constriction, whereas in control animals, it caused prostacyclin (PGI2)-dependent dilation.18 A recent study found that in aortas of T2-DM mice, phenylephrine-induced contraction was reduced and acetylcholine-induced relaxation was enhanced by the nonselective cyclooxygenase (COX) inhibitor indomethacin, indicating agonist-induced release of constrictor PGs.19 Although these studies suggested that the vascular synthesis of constrictor PGs is enhanced, the role of different COX isoforms and the functional consequences of altered PG synthesis affecting vascular resistance in T2-DM remain unclear.
It is known that COX-2 expression and activity are readily upregulated by inflammatory and physical stimuli.20 Recent biochemical studies have proposed a possible role for enhanced COX-2 expression in high glucose–induced alterations in constrictor prostanoid production in cultured endothelial cells.21 Also, it has been demonstrated that upregulation of COX isoforms is associated with a significant elevation of vascular PG synthesis.22 However, there are only a limited number of studies investigating alterations in vascular COX-2–dependent mechanisms in T2-DM, and little is known about the functional consequences of altered microvascular prostanoid synthesis.23 The aforementioned prompted us to investigate the cellular sources and the specific role of altered vascular prostanoid synthesis in the regulation of microvascular resistance in normal mice and mice with T2-DM.
We used 12- to 14-week-old, male db/db (C57BL/KsJ-db–/db–) and control heterozygous (C57BL/KsJ-db+/db–) mice in our experiments.11,24 Animals were fed standard chow and drank tap water ad libitum. All protocols were approved by the Institutional Animal Care and Use Committee at New York Medical College.
Determination of Blood Pressure and Calculation of Total Peripheral Resistance
In conscious mice, systolic and diastolic blood pressures were measured by the tail-cuff method and mean arterial pressure was calculated. Total peripheral vascular resistance was also calculated from mean arterial pressure and cardiac output data, which were obtained by echocardiography in awake animals.25
Isolation of Gracilis Muscle Arterioles
Mice were anesthetized with an intraperitoneal injection of pentobarbital sodium (50 mg/kg). Using microsurgery instruments and an operating microscope, we isolated a gracilis muscle arteriole (0.5 mm long) running intramuscularly and transferred it into an organ chamber containing 2 glass micropipettes filled with physiologic salt solution (PSS) composed of the following (in mmol/L): 110 NaCl, 5.0 KCl, 2.5 CaCl2, 1.0 MgSO4, 1.0 KH2PO4, 5.0 glucose, and 24.0 NaHCO3 equilibrated with a gas mixture of 10% O2 and 5% CO2, balanced with nitrogen, at pH 7.4. Vessels were cannulated on both ends, and micropipettes were connected with silicone tubing to an adjustable PSS reservoir. Inflow and outflow pressures were set to 80 mm Hg and continuously measured by a pressure servocontrol system (Living Systems Instrumentation). Temperature was set at 37°C by a controller. The internal arteriolar diameter at the midpoint of the arteriolar segment was measured by videomicroscopy with a microangiometer (Texas Instruments). Changes in arteriolar diameter and intraluminal pressure were continuously recorded with the Biopac-MP100 system connected to a computer and analyzed with AcqKnowledge data acquisition software (Biopac Systems, Inc).11,24
Pressure-Induced Arteriolar Response
After a 1-hour incubation period, spontaneous basal arteriolar tone developed in response to 80 mm Hg intraluminal pressure, without the use of any constrictor agent. Then, the change in diameter of arterioles was measured in response to stepwise increases in intraluminal pressure from 20 to 120 mm Hg. To obtain the passive arteriolar characteristics, pressure-induced arteriolar responses were reassessed in the presence of Ca2+-free PSS. Active arteriolar tone (in Ca2+-containing PSS) was expressed as a percentage of passive diameters (in Ca2+-free PSS). Myogenic index was also calculated, as described previously.26
Arteriolar Response to Arachidonic Acid
In separate experiments, at an intraluminal pressure of 80 mm Hg, arachidonic acid (10–9 to 10–7 mol/L, Cayman Chemicals) was applied to the superfusion solution. Each dose was incubated with the vessel for 10 minutes, and steady-state diameters were recorded. Acetylcholine (10–9 to 10–7 mol/L) and the NO donor sodium nitroprusside (SNP, 10–8 and 10–7 mol/L) were used to test the function of the endothelium and smooth muscle of arterioles. In separate experiments, the arteriolar endothelium was removed by air, and arachidonic acid–induced responses were obtained again. Endothelium denudation was ascertained by the loss of dilation to acetylcholine and the maintained dilation to the NO donor SNP.
Selective Inhibition of the PGH2/TxA2 Receptor and COX-1 and COX-2 Enzymes
After inhibition of PGH2/TxA2 receptors with SQ29548 (10–6 mol/L for 15 minutes, Cayman Chemicals), pressure- and arachidonic acid–induced arteriolar responses were obtained once more. For specific inhibition of COX isoenzymes, arterioles were incubated with the selective COX-1 inhibitor SC-560 (10–6 mol/L for 30 minutes, Cayman Chemicals) or with the selective COX-2 inhibitor NS-398 (10–5 mol/L for 30 minutes, Cayman Chemicals), and pressure- and arachidonic acid–induced arteriolar responses were reassessed.
Mouse aortas were dissected from control and db/db mice, cleared of connective tissue, and briefly rinsed in ice-cold PSS. After the addition of 100 μL of sample buffer (from Sigma Inc), tissues were homogenized. Aliquots were separated by electrophoresis on a 10% polyacrylamide gel at 125 V for 1 hour and transferred onto a polyvinyl difluoride membrane. Immunoblot analysis was performed as described before.24 Antibodies used for detection (anti–COX-1 IgG, anti–COX-2 IgG) were obtained from Cayman Chemicals. Anti–?-actin IgG from Abcam Ltd was used for loading control. Signals were revealed with chemiluminescence and visualized autoradiographically. Optical density of bands was measured and quantified by Image J software.
Data are expressed as mean±SEM. To obtain the passive diameter, arterioles were exposed to a Ca2+-free solution containing EGTA (10–3 mol/L) and 10–4 mol/L SNP. Statistical analyses were performed by a 2-way ANOVA for repeated measures, followed by the Tukey post hoc test or Student’s t test, as appropriate. P<0.05 was considered statistically significant.
Previously, we had found that at 12 weeks of age, body weight, serum glucose, and serum insulin values of db/db mice were significantly elevated compared with age-matched wild-type animals, resembling data obtained from patients with obesity and T2-DM.11,24
Blood Pressure and Calculated Peripheral Resistance
Systolic and mean arterial pressures were significantly elevated in conscious db/db mice compared with wild-type mice (Table). Calculated peripheral vascular resistance (obtained from mean arteriolar pressure and cardiac output data) was significantly elevated in db/db mice compared with wild-type animals (Table).
Hemodynamic Parameters, Basal Tone, and Agonist-Induced Changes in Diameter of Isolated Skeletal muscle Arterioles of Mice
Pressure-Induced Arteriolar Response
After a 1-hour incubation period, spontaneous myogenic tone developed in isolated skeletal muscle arterioles without the use of any vasoactive agent. At 80 mm Hg, the diameter of arterioles of db/db mice was significantly reduced compared with that of arterioles from wild-type mice (Table). There were no significant differences between passive arteriolar diameters in the 2 groups of animals obtained in Ca2+-free PSS at 80 mm Hg (Table). Stepwise increases in intraluminal pressure from 20 to 120 mm Hg elicited significantly greater reductions in the diameter of arterioles from db/db mice compared with control vessels at each pressure step (Figure 1A and 1B). The passive pressure-diameter curves of arterioles (obtained in Ca2+-free solution) were not different in the 2 groups of animals (Figure 1A). The calculated myogenic index was also not significantly different in arterioles isolated from control and db/db mice (Figure 1C).
Figure 1. Diameters of skeletal muscle arterioles isolated from control (n=15) and db/db (n=15) mice in response to stepwise increases (20 to 120 mm Hg) in intraluminal pressure in the presence or absence of extracellular Ca2+ (A). Normalized arteriolar diameters of control (n=15) and db/db (n=15) mice are expressed as percentages of the passive diameter (B). Calculated myogenic index values of arterioles from control and db/db mice developed in response to stepwise increases in intraluminal pressure (C). Data are mean±SEM. *Significant difference (P<0.05). Abbreviations are as defined in text.
Role of COX Isoforms in Pressure-Induced Arteriolar Tone
To elucidate the role of PGs in pressure-induced arteriolar tone development in control and db/db mice, selective inhibitors of PG receptors and COX enzymes were used. The presence of the PGH2/TxA2 receptor antagonist SQ29548 did not affect the pressure-induced responses of arterioles of control mice, but it reduced the tone of arterioles of db/db mice back to control levels (Figure 2A and 2B). The presence of the selective inhibitor of cyclooxygenase-1 (COX-1), SC-560, did not affect the basal tone of arterioles in either control or db/db mice (Figure 2C and 2D). On the other hand, the presence of NS-398, a selective inhibitor of COX-2, caused a significant upward shift in the arteriolar pressure-diameter curve of vessels from db/db mice, but did not significantly affect that of arterioles isolated from control animals (Figure 2C and 2D).
Figure 2. Normalized arteriolar diameters of control (n=9) and db/db (n=9) mice in the absence or presence of SQ 29,548, a PGH2/TXA2 receptor antagonist (A and B); NS-398, a selective inhibitor of COX-2; and SC-560, a selective inhibitor of COX 1 (C and D). Data are mean±SEM. *Significant difference (P<0.05). Abbreviations are as defined in text.
Role of COX Isoforms in Arachidonic Acid–Induced Arteriolar Response
In the next series of experiments, we obtained arteriolar responses to exogenously administered arachidonic acid, the precursor of PGs, in the absence and presence of the specific inhibitors. In arterioles of control mice, arachidonic acid in a concentration-dependent manner elicited dilation; however, it caused significant constriction in arterioles isolated from db/db mice (Figure 3A). The presence of the PGH2/TxA2 receptor antagonist SQ29548 did not affect the arachidonic acid–induced dilation of arterioles of control mice but reduced the arachidonic acid–induced constriction in arterioles of db/db mice (Figure 3A). Also, we found that removal of the endothelium reduced arachidonic acid–induced arteriolar responses in control vessels but did not affect responses significantly in arterioles of db/db mice (Figure 3B). In control mice, arachidonic acid–induced dilations were significantly reduced by the selective inhibitor of COX-1, SC-560, but not by NS-398, a selective inhibitor of COX-2 (Figure 4A). In contrast, in arterioles of db/db mice, the arachidonic acid–induced constriction was significantly reduced by the COX-2 inhibitor NS-398, whereas the COX-1 inhibitor SC-560 had no effect (Figure 4B).
Figure 3. Arachidonic acid-induced changes in arteriolar diameters from control (n=7) and db/db (n=7) mice in the absence or presence of SQ 29,548, a PGH2/TXA2 receptor antagonist (A), and after endothelium removal (B, n=5). Data are mean±SEM. *Significant difference from control; #Significant difference from db/db arterioles (P<0.05). Abbreviations are as defined in text.
Figure 4. Arachidonic acid–induced changes in arteriolar diameters from control (n=7) and db/db (n=7) mice in the absence or presence of NS-398, a selective inhibitor of COX-2, and SC-560, a selective inhibitor of COX 1 (A and B). Data are mean±SEM. *Significant difference from control (P<0.05). Abbreviations are as defined in text.
Western blot analysis was performed on aortas from both control and db/db mice. There were no significant differences in total COX-1 protein levels in the 2 groups, whereas COX-2 expression was significantly greater in aortas from db/db mice (Figure 5).
Figure 5. Western blot analysis of the expression of COX-1 (A and B) and COX-2 (C and D) in aortas from control and db/db mice. Anti–?-actin was used to normalize for loading variations. Bar graphs represent the summary of normalized densitometric ratios (n=5 for each group). Data are mean±SEM. *Significant difference (P<0.05). Abbreviations are as defined in text.
The main findings of the present study are that mice with T2-DM have elevated systolic blood pressures and increased peripheral vascular resistance. In T2-DM mice, these alterations are associated with enhanced skeletal muscle arteriolar tone, which is likely to be attributable to increased release of COX-2–derived constrictor PGs within the skeletal muscle arteriolar wall.
T2-DM is frequently associated with elevated systemic blood pressure; however, the nature of the mechanisms have not yet been fully elucidated. A key role for altered regulation of microvascular resistance has been suggested by several earlier investigations that found specific impairment of microvascular vasoregulatory mechanisms in subjects with T2-DM.4 Accordingly, a reduced endothelium-dependent arteriolar vasodilation6,9–11 and/or an enhanced smooth muscle–dependent vasoconstriction5–7 in microvessels has been demonstrated in T2-DM, alterations that could influence total peripheral resistance.
In the present study, hemodynamic parameters were obtained first in awake mice. Using a tail-cuff method, we found a significant rise in systolic and mean arterial blood pressures in db/db mice compared with control animals (Table). Furthermore, by using echocardiography, a decreased cardiac output was also observed in db/db mice (Table). On the basis of these parameters, the total peripheral resistance was calculated and found to be significantly elevated in db/db mice compared with that of control animals (Table). This study is the first to demonstrate that at 12 weeks of age, in association with obesity, hyperglycemia, and hyperinsulinemia, db/db mice exhibit an enhanced peripheral vascular resistance and elevated systemic blood pressure. These results are in accordance with previous clinical observations of the prevalence of high blood pressure in patients with obesity and T2-DM.2,27
In the next series of experiments, we aimed to elucidate the possible underlying mechanisms responsible for the enhanced peripheral vascular resistance. It is well known that arterioles respond to an increase in transmural pressure by constriction, a response termed myogenic constriction, which plays a key role in the local regulation of tissue blood flow.28,29 Earlier it had been found that an elevation of systemic blood pressure is associated with a rise in resistance of the skin microcirculation in T2-DM patients.30 It has been proposed that enhanced arteriolar tone could protect the distal part of the microcirculation from the increased intraluminal pressure.31 Indeed, previous studies have demonstrated an enhanced tone in skeletal muscle5 and mesenteric6 arterioles of T2-DM animals. An enhanced arteriolar tone, especially in the skeletal muscle circulation, may further increase vascular resistance, which, if uncompensated, could ultimately result in the elevation of systemic blood pressure.
In the present study, we demonstrated that in response to stepwise increases in intraluminal pressure (from 20 to 120 mm Hg), the diameter of isolated skeletal muscle arterioles was significantly reduced in db/db compared with control mice, whereas the passive diameter (obtained in Ca2+-free solution) of arterioles was not significantly different between the 2 groups (Figure 1). These findings indicate that active tone is greater in the arterioles of db/db mice, which, however, is unlikely to be the result of an altered arteriolar compliance. Next, myogenic indexes28,29 were calculated, allowing us to assess the pressure-sensitive behavior of skeletal muscle arterioles. We found that myogenic indexes were not significantly different in the 2 groups, suggesting that mechanisms other than pressure-sensitive myogenic regulation are responsible for the reduction in the basal diameter of arterioles from db/db mice.
Arteriolar diameter is continuously modulated by dilator and constrictor factors, many of them intrinsic to the vascular wall. Early investigations reported enhanced release of a constrictor prostanoid from diabetic vessels.18 On the basis of previous observations, we hypothesized that vascular production of constrictor PGs is increased in T2-DM, thereby contributing to the reduced diameter of skeletal muscle arterioles of db/db mice. Indeed, we found that a PGH2/TxA2 receptor antagonist increased the diameter of arterioles of T2-DM mice back to control levels, whereas it did not affect the diameter of vessels from control animals (Figure 2A and 2B). These findings indicate that endogenous release of constrictor PGs, PGH2/TxA2, may be responsible for the reduced diameter of arterioles from T2-DM mice.
Recently, a role for prostanoid-mediated vascular inflammation has been shown to be associated with the development of vascular complications in T2-DM.13 Prostanoids generated by COXs from arachidonic acid20 are important mediators of several inflammatory mechanisms. Two isoforms of the COX enzyme, encoded by distinct genes, have been isolated in mammalian cells.20 COX-1 is constitutively expressed in most tissues, such as vascular endothelial cells, and is involved in the maintenance of cellular homeostasis.23 In contrast, under normal conditions, COX-2 is expressed only at low or undetectable levels but is readily upregulated by inflammatory, mitogenic, and physical stimuli.32 Only a limited number of biochemical studies have investigated alterations in COX-2–dependent mechanisms related to DM. In this context, it has been found that high-glucose treatment caused increases in expression of COX-2 protein in mesangial cells.33 Also, in cultured human endothelial cells, high-glucose treatment elicited an enhanced production of TxA2 in association with an upregulation of COX-2.21 Because in the db/db mouse model of T2-DM the level of vascular COX-2 expression is not known, we aimed to measure COX-2 expression in intact vessels from control and db/db mice. COX-2 protein levels in the aortas of T2-DM mice were markedly increased compared with those of control animals (Figure 5). Although COX-2 expression has been found to be associated with enhanced production of constrictor prostanoids under normal34 and certain pathologic22,35 conditions, the specific role of COX-1– and COX-2–dependent prostanoid synthesis in the mediation of arteriolar diameter changes in T2-DM has not yet been elucidated.
In control arterioles, selective inhibition of COX-1 did not affect the diameter of arterioles (Figure 2C), whereas it reduced arachidonic acid–induced dilation (Figure 4A). Because removal of the endothelium also reduced arachidonic acid–induced arteriolar responses in control vessels (Figure 3B), we concluded that dilator PGs, most likely endothelium-derived prostacyclin, were produced by the metabolism of arachidonic acid. In contrast, in arterioles of T2-DM mice, both basal tone (Figure 2D) and arachidonic acid–induced constrictions were reduced by the selective inhibitor of COX-2, but not that of COX-1, or endothelium removal (Figures 3B and 4B). We interpret these findings to mean that in arterioles of db/db mice, COX-2–dependent release of constrictor PGs, most likely PGH2/TxA2, derived primarily from vascular smooth muscle cells, mediate both the enhanced pressure- and arachidonic acid–induced reductions of arteriolar diameter.
Recently, an important role for reactive oxygen species in the regulation of arteriolar tone has received a great deal of attention.36 In T2-DM rats, reactive oxygen species have been proposed to play a role in myogenic activation of skeletal muscle arterioles.5 In this context, we previously found that in T2-DM (db/db) mice, owing to the reduced activity of vascular superoxide dismutase and catalase together with enhanced activation of vascular NAD(P)H oxidase, vascular production of superoxide was increased.24 Thus, one can speculate that vascular oxidative stress in T2-DM or other disease conditions37 may also be associated with alteration in COX-2–dependent synthesis of PGs. Indeed, recently it has been found that in high glucose–treated mesangial cells, mitochondrial superoxide production was associated with enhanced COX-2 expression.33 Also, in cultured human endothelial cells, high glucose elicited enhanced production of reactive oxygen species, resulting in increased production of TxA2, which was also associated with an upregulation of COX-2.21 However, in T2-DM, the interrelation between vascular oxidative stress and altered prostanoid metabolism needs to be addressed in future investigations.
Taken together, we propose an important role for COX-2–derived constrictor PGs in the altered regulation of skeletal muscle arteriolar resistance in T2-DM, obese mice. It still remains a question, however, whether and to what extent changes in arteriolar PG synthesis could contribute to the alterations of total peripheral resistance and blood pressure in T2-DM. Nevertheless, based on the present studies, alterations in COX-2–dependent and PG-mediated modulation of vasomotor function should be taken into consideration in future investigations of T2-DM.
This study was supported by a grant from the American Heart Association, Northeast Affiliate (0555897T), and National Institutes of Health grants HL-43023, HL-46813, OTKA T-034779, 048376, F-048837, and 711-K-84289.
Turner RC, Millns H, Neil HA, Stratton IM, Manley SE, Matthews DR, Holman RR. Risk factors for coronary artery disease in non-insulin dependent diabetes mellitus: United Kingdom Prospective Diabetes Study (UKPDS: 23). BMJ. 1998; 316: 823–828.
Sowers JR, Epstein M, Frohlich ED. Diabetes, hypertension, and cardiovascular disease: an update. Hypertension. 2001; 37: 1053–1059.
Stehouwer CD, Lambert J, Donker AJ, van Hinsbergh VW. Endothelial dysfunction and pathogenesis of diabetic angiopathy. Cardiovasc Res. 1997; 34: 55–68.
De Vriese AS, Verbeuren TJ, Van de Voorde J, Lameire NH, Vanhoutte PM. Endothelial dysfunction in diabetes. Br J Pharmacol. 2000; 130: 963–974.
Frisbee JC, Maier KG, Stepp DW. Oxidant stress-induced increase in myogenic activation of skeletal muscle resistance arteries in obese Zucker rats. Am J Physiol Heart Circ Physiol. 2002; 283: 2160–2168.
Lagaud GJ, Masih-Khan E, Kai S, van Breemen C, Dube GP. Influence of type II diabetes on arterial tone and endothelial function in murine mesenteric resistance arteries. J Vasc Res. 2001; 38: 578–589.
Schofield I, Malik R, Izzard A, Austin C, Heagerty A. Vascular structural and functional changes in type 2 diabetes mellitus: evidence for the roles of abnormal myogenic responsiveness and dyslipidemia. Circulation. 2002; 106: 3037–3043.
Caballero AE, Arora S, Saouaf R, Lim SC, Smakowski P, Park JY, King GL, LoGerfo FW, Horton ES, Veves A. Microvascular and macrovascular reactivity is reduced in subjects at risk for type 2 diabetes. Diabetes. 1999; 48: 1856–1862.
Erdos B, Miller AW, Busija DW. Impaired endothelium-mediated relaxation in isolated cerebral arteries from insulin-resistant rats. Am J Physiol Heart Circ Physiol. 2002; 282: 2060–2065.
Pannirselvam M, Verma S, Anderson TJ, Triggle CR. Cellular basis of endothelial dysfunction in small mesenteric arteries from spontaneously diabetic (db/db –/–) mice: role of decreased tetrahydrobiopterin bioavailability. Br J Pharmacol. 2002; 136: 255–263.
Bagi Z, Koller A, Kaley G. Superoxide-NO interaction decreases flow- and agonist-induced dilation of coronary arterioles in type 2 diabetes mellitus. Am J Physiol Heart Circ Physiol. 2003; 285: H1404–H1410.
Caballero AE. Endothelial dysfunction, inflammation, and insulin resistance: a focus on subjects at risk for type 2 diabetes. Curr Diab Rep. 2004; 4: 237–246.
Helmersson J, Vessby B, Larsson A, Basu S. Association of type 2 diabetes with cyclooxygenase-mediated inflammation and oxidative stress in an elderly population. Circulation. 2004; 109: 1729–1734.
Higgs GA, Moncada S, Vane JR. Eicosanoids in inflammation. Ann Clin Res. 1984; 16: 287–299.
Koller A, Kaley G. Prostaglandins mediate arteriolar dilation to increased blood flow velocity in skeletal muscle microcirculation. Circ Res. 1990; 67: 529–534.
Koller A, Kaley G. A new role for prostaglandins in the regulation of peripheral resistance. Adv Prostaglandin Thromboxane Leukot Res. 1991; 595–598.
Myers TO, Messina EJ, Rodrigues AM, Gerritsen ME. Altered aortic and cremaster muscle prostaglandin synthesis in diabetic rats. Am J Physiol. 1985; 249: 374–379.
Sterin-Borda L, Franchi AM, Borda ES, Del Castillo E, Gimeno MF, Gimeno AL, Sterin Borda I. Augmented thromboxane generation by mesenteric arteries from pancreatectomized diabetic dogs is coincident with the vascular tone enhancement evoked by Na-arachidonate and prostacyclin. Eur J Pharmacol. 1984; 103: 211–221.
Okon EB, Szado T, Laher I, McManus B, van Breemen C. Augmented contractile response of vascular smooth muscle in a diabetic mouse model. J Vasc Res. 2003; 40: 520–530.
Mitchell JA, Warner TD. Cyclo-oxygenase-2: pharmacology, physiology, biochemistry and relevance to NSAID therapy. Br J Pharmacol. 1999; 128: 1121–1132.
Cosentino F, Eto M, De Paolis P, van der Loo B, Bachschmid M, Ullrich V, Kouroedov A, Delli Gatti C, Joch H, Volpe M, Luscher TF. High glucose causes upregulation of cyclooxygenase-2 and alters prostanoid profile in human endothelial cells: role of protein kinase C and reactive oxygen species. Circulation. 2003; 107: 1017–1023.
Quilley J, Chen Y-J. Role of COX-2 in the enhanced vasoconstrictor effect of arachidonic acid in the diabetic rat kidney. Hypertension. 2003; 42: 837–843.
Davidge ST. Prostaglandin H synthase and vascular function. Circ Res. 2001; 89: 650–660.
Bagi Z, Koller A, Kaley G. PPAR- activation, by reducing oxidative stress, increases NO bioavailability in coronary arterioles of mice with type 2 diabetes. Am J Physiol Heart Circ Physiol. 2004; 286: H742–H748.
Li W, Jue T, Edwards J, Wang X, Hintze TH. Changes in NO bioavailabilty regulate cardiac O2 consumption: control by intramitochondrial SOD2 and intracellular myoglobin. Am J Physiol Heart Circ Physiol. 2004; 286: H47–H54.
Jackson PA, Duling BR. Myogenic response and wall mechanics of arterioles. Am J Physiol. 1989; 257: H1147–H1155.
Sorof J, Daniels S. Obesity hypertension in children: a problem of epidemic proportions. Hypertension. 2002; 40: 441–447.
Davis MJ, Hill MA. Signaling mechanisms underlying the vascular myogenic response. Physiol Rev. 1999; 79: 387–423.
Hill MA, Zou H, Potocnik SJ, Meininger GA, Davis MJ. Invited review: arteriolar smooth muscle mechanotransduction: Ca2+ signaling pathways underlying myogenic reactivity. J Appl Physiol. 2001; 91: 973–983.
Jaap AJ, Shore AC, Tooke JE. The influence of hypertension on microvascular blood flow and resistance to flow in the skin of patients with type 2 (non-insulin-dependent) diabetes. Diabet Med. 1994; 11: 883–887.
Koller A. Signaling pathways of mechanotransduction in arteriolar endothelium and smooth muscle cells in hypertension. Microcirculation. 2002; 9: 277–294.
Parente L, Perretti M. Advances in the pathophysiology of constitutive and inducible cyclooxygenases: two enzymes in the spotlight. Biochem Pharmacol. 2003; 65: 153–159.
Kiritoshi S, Nishikawa T, Sonoda K, Kukidome D, Senokuchi T, Matsuo T, Matsumura T, Tokunaga H, Brownlee M, Araki E. Reactive oxygen species from mitochondria induce cyclooxygenase-2 gene expression in human mesangial cells: potential role in diabetic nephropathy. Diabetes. 2003; 52: 2570–2577.
Baber SR, Champion HC, Bivalacqua TJ, Hyman AL, Kadowitz PJ. Role of cyclooxygenase-2 in the generation of vasoactive prostanoids in the rat pulmonary and systemic vascular beds. Circulation. 2003; 108: 896–901.
Belton OA, Duffy A, Toomey S, Fitzgerald DJ. Cyclooxygenase isoforms and platelet vessel wall interactions in the apolipoprotein E knockout mouse model of atherosclerosis. Circulation. 2003; 108: 3017–3023.
Cseko C, Bagi Z, Koller A. Biphasic effect of hydrogen peroxide on skeletal muscle arteriolar tone via activation of endothelial and smooth muscle signaling pathways. J Appl Physiol. 2004; 97: 1130–1137.
Wang D, Chabrashvili T, Wilcox CS. Enhanced contractility of renal afferent arterioles from angiotensin-infused rabbits: roles of oxidative stress, thromboxane prostanoid receptors, and endothelium. Circ Res. 2004; 94: 1436–1442.
From the Division of Cardiovascular Research (K.F.K., R.P., M.B.-M., R.K., Y.Y., C.C., M.S., M.K., T.A., D.W.L.), St. Elizabeth Medical Center, Boston, Mass; and Curis Inc (K.L.A., W.M.), Cambridge, Mass.
Objective— The embryonic morphogen sonic hedgehog (SHh) has been shown to induce neovascularization of ischemic tissue but has not been shown to play a role in regulating vascular nerve supply. Accordingly, we investigated the hypothesis that systemic injection of SHh protein could improve nerve blood flow and function in diabetic neuropathy (DN).
Methods and Results— Twelve weeks after induction of diabetes with streptozotocin, motor and sensory nerve conduction velocities (MCV and SCV) of the sciatic nerves were significantly reduced in diabetic rats. SHh-treated diabetic rats demonstrated marked improvement of both MCV and SCV (P<0.05). Laser Doppler perfusion imaging showed that nerve blood flow was significantly reduced in the diabetic rats but was restored in SHh-treated diabetic rats (P<0.05 versus diabetic saline-treated rats) to levels similar to those achieved with vascular endothelial growth factor-2 (VEGF-2) gene therapy. In vivo perfusion of Bandeuraea simplicifolia (BS)-1 lectin showed marked reduction in the vasa nervora in diabetic sciatic nerves but restoration of nerve vasculature to nondiabetic levels in the SHh-treated and plasmid DNA encoding human VEGF-2 (phVEGF-2)–treated diabetic nerves. Interestingly, the SHh-induced vasculature was characterized by larger diameter and more smooth muscle cell-containing vessels, compared with VEGF-2 gene-treated diabetic rats.
Conclusions— These data indicate that Shh induces arteriogenesis and restores nerve function in DN.
We administered Sonic Hedgehog (SHh) in a rat model of diabetic neuropathy (DN) and found that it replenishes the depleted vasa nervora that are depleted by diabetes and restores nerve function. Notably, the neurovasculature induced by SHh is composed of significantly greater numbers of arterioles than in VEGF-treated rats.
Key Words: angiogenesis ? diabetes mellitus ? cytokine ? microcirculation ? peripheral vasculature
In the United States alone, >18 million people have diabetes.1 Diabetic neuropathy (DN) is a frequent complication of diabetes, affecting 1 to 7 million people, including 7% within 1 year of diagnosis and 50% of patients after 25 years. It has also been reported that up to 90% of patients have subclinical levels of neuropathy.2 Although several factors have been reported to contribute to diabetic polyneuropathy,3–9 the pathogenic basis has remained uncertain.10 An association between changes in the vasa nervorum and DN has been noted in multiple previous reports;11–17 however, the pathophysiologic importance of these observations remains uncertain. The possibility that attenuation of the vasa nervorum might be a major factor in the development of DN is suggested by several recent studies. Impaired ischemia-induced angiogenesis was noted in animal models of diabetes,18 and more recently we have reported that both ischemic19 and DN20 are associated with attenuation of the vasa nervorum and that local delivery of naked DNA encoding for vascular endothelial growth factor (VEGF-1 and VEGF-2) restores the vascular supply and has a favorable effect on the nerve conduction velocities. These observations, documenting the loss of vasa nervorum in diabetic animals, and restoration of neural vascularity by VEGF, associated with a return of nerve function, suggested that the microangiopathic abnormality is one of the critical factors that cause DN.
Sonic hedgehog (SHh) is a prototypical morphogen known to regulate epithelial/mesenchymal interactions during embryonic development of limb, lung, gut, hair follicles, and bone.21–23 The hedgehog (Hh) pathway also plays an essential inductive and morphogenetic role in the developing central24–26 and peripheral nervous system.27 Recently, we have also reported that SHh protein has an indirect but powerful angiogenic effect in a mouse hind-limb ischemia model.28
Together, these previous studies suggested to us the possibility that diabetic polyneuropathy results, at least in part, from attenuation of vasa nervorum, that restoration of nerve blood flow supply can mitigate neuropathy despite persistent diabetes, and that SHh can exert angiogenic effects that could mitigate DN. Accordingly, we performed a series of investigations to test the hypothesis that SHh could replenish vasa nervorum in diabetes, thereby restoring nerve blood flow and nerve function in DN.
All protocols were approved by St. Elizabeth’s Institutional Animal Care and Use Committee. In all experiments, investigators performing the follow-up examinations were blinded to identify of the treatment administered.
Male Sprague-Dawley rats (Charles River Laboratories, Wilmington, Mass) weighing 200 to 225 grams were used. Rats were fed standard laboratory rodent chow and water ad libitum and housed individually.
Induction of Diabetes
Rats were made diabetic by a single intraperitoneal injection of streptozotocin (75 mg/kg in 0.9% sterile saline) into anesthetized rats (5 mg/100 g pentobarbital).
Systemic Treatment With SHh Protein
Human SHh proteins were used to construct SHh rat IgG fusion proteins to increase the half-life, as described.29 Systemic injection of SHh–rat IgG fusion protein was started 12 weeks after the induction of diabetes. After completion of baseline nerve conduction measurements, animals received subcutaneous injection of SHh proteins (1.0 mg/kg) or saline using 27-gauge needle 3 times per week for 4 weeks.
phVEGF-2 Plasmid and Gene Transfer
As a positive control, we used naked plasmid DNA encoding human VEGF-2 (phVEGF-2), as described previously.20
Nerve motor and sensory conduction velocity was measured as described previously20,30 in all rats at baseline (before treatment) and then at 2 and 4 weeks after treatment. All procedures and analyses were performed by an experienced researcher who was blinded to treatment.
In Vivo Assessment of Perfusion and Vascularity: Laser Doppler Imaging of Vasa Nervorum Blood Flow
Blood perfusion of the sciatic vasa nervorum was measured unilaterally in the hind limb of the rats with a laser Doppler perfusion imager (LDPI) system (Moor Instruments, Wilmington, Del) as described previously,20 programmed to measure perfusion of surrounding tissue as zero, or background. All perfusion measurements, as well as neurophysiological examinations, were performed with the animal placed on a heating blanket underneath a warming lamp controlled by a thermistor probe applied to the proximal nerve to maintain temperature at 37°C. All procedures and analyses were performed by an experienced researcher who was blinded to treatment assignment.
To insure that blood pressure was not affected by treatment, subgroups of animals from all treatment groups underwent analysis of blood pressure and heart rate. At the time of euthanization, a 2.0-French high-fidelity Millar pressure catheter (Millar Instruments) was inserted from the left ventricular apex to the ascending aorta, and systolic aortic pressure and heart rate were recorded. Calibration of the Millar catheter was verified before and after each measurement.
Sciatic Nerve Histology: Fluorescent Imaging of Vasa Nervorum
Vascularity of sciatic nerves from both normal and diabetic rats were assessed by in situ fluorescent staining using the endothelial cell-specific marker Bandeuraea simplicifolia (BS)-1 lectin conjugated to fluorescein isothiocyanate (Vector Laboratories, Burlingame, Calif) as described previously.20
Sciatic nerves were fixed in 100% methanol and paraffin-embedded sections of 5-μm thickness were stained for murine-specific endothelial marker isolectin B4 (Vector Laboratories), factor VIII (Signet Laboratories, Dedham, Mass), or alpha-smooth muscle actin (Sigma Chemical Co, St. Louis, Mo) and counterstained with eosin to detect capillary endothelial cells or smooth muscle cells in the vasa nervora.
Reverse-Transcription Polymerase Chain Reaction
Total RNA was extracted from sciatic nerves or L4,5 dorsal root ganglia 1 week after treatment using the Ambion Isolation kit (RNAqueousTM) according to the manufacturer’s instructions. DNAase digestion was performed after RNA extraction. Reverse-transcription polymerase chain reaction was performed according to the manufacturer’s instructions (Clontech, Palo Alto, Calif). All procedures and analyses were performed by an experienced researcher who was blinded to treatment assignment.
Cultured Nerve Fibroblasts
Primary cultured nerve fibroblasts were obtained from 250- to 350-gram male Sprague Dawley rats according to the method of Bolin.31 Cells were harvested after 48 hours and reverse-transcription polymerase chain reaction was performed.
All results were expressed mean±SD. Statistical comparisons between groups were performed by ANOVA with Bonferroni correction. P<0.05 was considered statistically significant.
DN Model: Treatment With SHh Versus VEGF Versus Saline
As shown in Table I (available at http://atvb.ahajournals.org), the serum glucose and blood urea nitrogen were elevated in the diabetic versus nondiabetic rats, as expected. Weight was also reduced in all diabetic animals. There were no significant differences between any of the diabetic treatment groups in these parameters or in blood pressure or heart rate (Table I).
Depletion of Vasa Nervorum Accompanies DN: SHh Replenishes Nerve Vascular Supply: In Vivo Staining of Vasa Nervorum by BS-1 Lectin Perfusion
Whole-mount staining reveals restoration of vasa nervorum by SHh (Figure 1A). Four weeks after treatment, an endothelial-specific marker, fluorescein isothiocyanate-conjugated BS-1 lectin was injected to permit documentation of vasa nervora. The nondiabetic rat in both saline and SHh showed a regular pattern of vascularity including a superficial longitudinal network and penetrating branches responsible for providing blood flow to the endoneurial vascular network. However, in nerves of diabetic rats treated with saline, the total number of vasa nervora was markedly decreased and the vascular network was substantially destroyed, resulting in an irregular distribution pattern and areas of nonvascularized nerve tissue. In SHh-treated diabetic rats, the vascular network was restored, with both superficial and penetrating branches. Similar findings were disclosed with VEGF-2 gene transfer, which was included as a positive control.20
Figure 1. In vivo perfusion imaging reveals attrition of vasa nervorum by diabetes and recovery induced by SHh. A, Representative fluorescent BS-1 lectin-perfused rat sciatic nerves (longitudinal views). In the saline-treated diabetic rat, the total network of vasa nervorum is markedly disrupted. SHh and phVEGF-2 treatment resulted in significant restoration of vasa nervorum. SHh administration to nondiabetic rats had no effect. B, Representative fluorescent BS-1 lectin-perfused rat sciatic nerves (cross-section). A reduced number of epineurial/perineurial and endoneurial vessels (C) are observed in diabetic rats. SHh-treated (and phVEGF-2 gene therapy as a positive control20) rats showed replenished vascularity. The total number of epineurial/perineurial vessels was decreased in saline-treated diabetic rats; however, in SHh-treated diabetic rats, the number of vessels was similar to nondiabetic controls. Similar recovery is noted in VEGF-2 gene therapy-treated diabetic rats. *P<0.01 vs nondiabetic plus saline, #P<0.05 vs diabetic plus saline. C, Representative factor VIII immunostaining of rat sciatic nerve (cross-section). Endoneurial vessels are reduced in diabetic rats. SHh (and phVEGF-2 gene therapy, as a positive control) induced recovery of endoneurial vascularity in diabetic rats. SHh administration to nondiabetic animals had no effect. *P<0.01 vs nondiabetic plus saline, #P<0.05 vs diabetic plus saline.
Quantification of epineurial/perineurial and endoneurial capillaries in sciatic nerves documents recovery of vasa nervora in SHh-treated rats (Figure 1B and 1C). To analyze the sciatic nerve capillaries, we counted the number of vessels using cross-section slides. Figure 1B clearly showed much more epineural/perineural capillaries in the nondiabetic nerves compared with saline-treated diabetic nerves (epineural/perineural vasa/cross-section: 138.0±8.0 in nondiabetic plus saline, n=7; 142.0±12.0 in nondiabetic plus SHh, n=6; and 62.2±11.0 per section in diabetic plus saline, n=5; P<0.01). There was no significant difference between saline-treated and SHh-treated nondiabetic nerves. Endoneural capillaries were also significantly reduced in saline-treated diabetic rats (endoneural vasa/cross-section: 37.8±3.3 in nondiabetic plus saline, n=7; 38.2±4.5 in nondiabetic plus SHh, n=6; and 21.0±2.4 per section in diabetic plus saline, n=5; P<0.01). SHh treatment resulted in recovery of both epineural/perineural and endoneural capillaries (epineural/perineural: 105.7±14.0; endoneural: 36.3±2.4 in SHh; n=6 per section). Similar findings were noted in VEGF-2 gene therapy-treated animals as shown previously (epineural/perineural: 108.3±22.3; endoneural: 35.8±7.1 per section in phVEGF-2; n=5) (Figure 1B). Endoneural capillaries were also counted using factor VIII staining. As shown in Figure 1C, factor VIII-positive vessels were also reduced in saline-treated diabetic rats (34.0±4.5 in nondiabetic plus saline, n=5, and 37.4±7.9 in nondiabetic plus SHh, n=6, and 14.2±3.5 per section in diabetic, n=5; P<0.01). SHh treatment resulted in recovery of endoneurial capillaries (25.8±4.8, n=5) similar to the results of VEGF-2 gene therapy (25.6±6.4 per section, n=5).
LDPI of Sciatic Nerve Blood Flow
LDPI was performed to evaluate blood flow in the sciatic nerves of rats in all treatment groups (Figure 2). This blinded analysis revealed markedly reduced nerve blood flow in saline-treated diabetic rats (401.0±106.3 LDPI units versus 1185.2±370.1 LDPI units in nondiabetic controls; P<0.01) as described previously.20 SHh treatment in diabetic rats resulted in substantial restoration of sciatic nerve perfusion (791.0±351.4 LDPI units, P<0.05, versus saline-treated diabetic rats; Figure 2B). VEGF-2 gene transfer also restored perfusion of sciatic nerves to a level similar to that seen in SHh-treated diabetic rats (816.8±310.1 LDPI units, P<0.05, versus saline-treated diabetic rats). To further validate the usefulness of LDPI measurements as an indicator of vascular recovery, the capillary counts and LDPI measurements were correlated in randomly selected subgroups from all treatment groups. As shown in Figure 2C, there was a significant (P<0.01) correlation between total (epineural/perineural and endoneural) capillary density in the nerve and LDPI measurements in each animal.
Figure 2. Laser Doppler perfusion imaging (LDPI) documents recovery of sciatic nerve blood flow after SHh treatment. A, Representative images of in vivo LDPI depicting blood flow in rat sciatic nerve 4 weeks after treatment. The lowest blood flow is indicated by blue color and maximum blood flow in red. B, Summarized results of LDPI measurements taken from both rat sciatic nerves. Compared with nondiabetic control (n=20), saline-treated diabetic rats (n=10) showed markedly reduced perfusion. Significant improvement in nerve perfusion was observed in SHh-treated (n=12) and phVEGF-2–treated (n=8) diabetic rats. C, Cross-sectional total capillary density and LDPI were assessed in individual animals from all treatment groups (n=23) and were found to exhibit significant correlation (P<0.01), indicating that LDPI assessment of flow was corroborated by anatomic evidence of recovery of the vasa nervorum and vice versa.
SHh-induced neovasculature is morphologically distinct (Figure 3). During our initial analysis of capillary density, we noted that the vasculature of the epineurium/perineurium appeared larger in size than the vessels in the other treatment groups (Figure 3A top and middle) We measured vessel diameter and found that the epineurial/perineurial vessels in the SHh group were significantly larger than those in the phVEGF-2–treated rats and were similar in size to those in the nondiabetic control rats (mean vessel diameter 15.3 μm in phVEGF-2 group versus 26.4 μm in SHh-treated group, P<0.05) (Figure 3B). Moreover, staining for -smooth muscle actin revealed that the SHh-treated nerves contained a greater number of -actin-positive cells colocalized in the epineurial/perineurial vessels than in nerves from phVEGF-2–treated rats.(Figure 3A bottom). We then measured the total area of -actin–positive vasculature in all treatment groups and found that the -actin–positive vasculature in SHh-treated nerves was significantly closer to the nondiabetic nerves than after VEGF gene therapy. These data indicated that treatment with SHh resulted in a vessel morphology that was distinct from that induced by gene transfer of a single angiogenic cytokine.
Figure 3. SHh treatment results in larger neovasculature with greater smooth muscle content. A, Representative photomicrographs of longitudinal views of BS-1 lectin staining (upper), cross-sectional views of isolectin B4 staining (middle), and cross-sectional views of alpha-smooth muscle actin staining (lower). Compared with phVEGF-2–treated nerve (left), the SHh-induced vasculature appears larger and contains a greater number of vessels with a smooth muscle cell layer (right). B, Epineurial/perineurial vessel diameter (upper). We measured 20 randomly selected vessels from each sample (5 sections per nerve) and calculated the mean vessel diameter (x60 objectives). The vasculature of the SHh-treated nerves (n=6) was significantly larger in diameter than in the diabetic phVEGF-2–treated nerves (n=5) and was similar to the nondiabetic nerves. Total area of -actin–positive vasculature in the epineurium/perineurium in all treatment groups (lower). As shown, VEGF treatment did not increase the -actin–positive vasculature, whereas SHh treatment resulting in a significant restitution of larger, -actin–positive vessels. Together these findings indicate that SHh induces the formation of neovessels that are similar in multilayered appearance to the native vasculature before diabetes.
SHh restores nerve function in DN (Figure 4). Within 12 weeks of the onset of diabetes induced by streptozotocin, a severe peripheral neuropathy developed in rats, as described previously.20 Electrophysiological recordings revealed that significant slowing of motor nerve conduction velocity (MCV) and sensory nerve conduction velocity (SCV) was observed in diabetic rats (MCV=35.0±2.9 m/s versus 46.2±3.1 m/s , SCV=34.2±2.5 m/s , and 48.1±3.7 m/s ; P<0.01 for both). Saline-treated diabetic rats showed no change in nerve conduction velocities during the 4 weeks of treatment (MCV=35.2±2.5 m/s and SCV=35.6±3.0 m/s). In contrast, 4 weeks after treatment with systemic injection of SHh protein, all nerve conduction velocities demonstrated a marked improvement (Figure 4). Specifically, MCV in diabetic rats treated with SHh protein increased to 44.9±4.2 m/s and SCV increased to 47.5±7.0 m/s (both P<0.01 versus saline-treated diabetic rats, and P=NS versus nondiabetic rats). phVEGF-2–treated diabetic rats also showed significant improvement in both MCV and SCV 4 weeks after injection. (MCV=42.5±4.6 m/s and SCV=44.5±7.5 m/s).
Figure 4. Motor and sensory nerve conduction is restored by SHh treatment. Before treatment (week 0), both MCV and SCV in diabetic rats (n=11) were significantly decreased compared with age-matched nondiabetic rats (n=22). However, 4 weeks after treatment with systemic injection of SHh protein (n=13), both MCV and SCV were improved significantly. Local gene transfer of VEGF-2 (n=12) also improved function as previously shown and was included as a control.
SHh upregulates expression of multiple angiogenic and neurotrophic cytokines (Figures I and II, available online at http://atvb.ahajournals.org). To identify potential mechanisms responsible for the therapeutic effect of SHh, we evaluated the expression of the Hh-related transcriptional factor Gli-1 and certain neurotrophic factors (BDNF and IGF-1) and angiogenic cytokines (VEGF-1, angiopoietin-1, and angiopoietin-2) in treated and control rats. As shown in Figure I, endogenous Gli expression was downregulated in saline-treated diabetic rats, suggesting that the Hh pathway was inactivated in the nerves of diabetic rats. The expression of angiogenic factors and neurotrophic cytokines were also downregulated (Figure I). However, SHh treatment resulted in a significant increase in the expression of mRNA of both endogenous angiogenic cytokines (VEGF-1, angiopoietin-1, and angiopoietin-2) and neurotrophic factors (BDNF, IGF-1), as well as upregulation of Gli-1 mRNA expression to nondiabetic levels. In contrast, phVEGF-2 did not upregulate the expression angiogenic cytokines or neurotrophic factors, except BDNF.
To verify these findings and to establish a direct effect of SHh on gene expression, we repeated reverse-transcription polymerase chain reaction on primary cultured rat nerve fibroblasts. Expression of Gli-1 was not detected in the cultured fibroblasts (Figure II). However the expression of mRNA for angiogenic cytokines (VEGF-1, angiopoietin-1, and angiopoietin-2) and neurotrophic factors (BDNF, IGF-1), as well as Gli-1, were upregulated by SHh protein (Figure II) in a dose-dependent manner (1, 5, 10 μg/mL), suggesting that SHh stimulation of neural fibroblasts can modulate expression of multiple factors with the potential to promote nerve recovery.
The peripheral neuropathy that complicates diabetes results in major morbidity, contributing to the leading cause of hospitalization among diabetic subjects and loss of tissue integrity in the lower extremities. The magnitude of this public health problem has led to aggressive efforts to define the cause and develop preventative measures or treatment strategies for this disabling condition. Despite the identification of multiple potential mechanisms, no therapy attempting to address individual causative factors has proven successful.
Our results demonstrate that SHh induces functional recovery in DN by simultaneously normalizing a repertoire of vascular and neural growth and survival factors and cytokines and replenishing a more mature-appearing vasa nervorum in both endoneurial and epineurial/perineurial capillaries. Notably, and in contrast to a recent report,29 our data reveal that DN is associated with vascular pathology. Specifically, disruption and loss of vasa nervorum accompany the onset of neuropathy in multiple animal models of DN (and ischemic neuropathy),20,32 and restitution of vascular architecture and nerve perfusion have now been repeatedly shown to be a consistent component of neurological recovery. These findings are consistent with developmental models that have verified the requirement for coordination between vascular and neurological elements.33 The role of vascular recovery in the restoration of neurophysiologic function induced by SHh in diabetes is underscored by the observed decrease in angiogenic factor expression in the effected nerves in diabetic animals and the recovery of expression after SHh treatment. Consistent with the central role of vascular recovery, direct replenishment of an angiogenic cytokine by VEGF-2 gene therapy also results in significant neurophysiologic recovery. Although the neovasculature induced by SHh and VEGF was different in appearance, the recovery of perfusion is similar, as are the degree and rate of physiological recovery. These data support a vasculogenic cause of DN.
The ability of the morphogen SHh to normalize expression of numerous factors downregulated in diabetic subjects resulted in the restoration of vasa nervora that appeared morphologically distinct and more similar in appearance to normal vessels than did the VEGF induced vessels. This is consistent with previous studies in which multiple cytokines were shown to induce formation of multilayered vessels.34 This observation regarding the vasculature induced by SHh may provide clues to the cause of diabetes-induced attrition of the vasa nervorum and to a better understanding of the mechanisms of neovascularization in vivo.
Downregulation of Angiogenic Cytokines, Neurotrophic Factors, and Hh Pathway in the Diabetic Sciatic Nerve
Multiple mechanisms have been implicated in the pathogenesis of DN, including modification and inactivation of proteins critical to neural function by nonenzymatic glycosylation,8 altered neural polyol metabolism,6,7 reductions in neurotrophin or the availability of neurotrophic factors, and microvascular disease including reduced vasa nervora in the diabetic nerve.20,35 However, debate still oscillates between propositions based on neurochemical versus vascular events. Our data demonstrate that not only neurotrophic factors but also various angiogenic cytokines were significantly reduced in the diabetic sciatic nerves. These data reveal that downregulation of both neurochemical and of vascular factors is related to the development of DN. After injection of SHh, expression of the Gli-1 transcription factor was upregulated and the expression of multiple endogenous angiogenic cytokines (angiopoietin-1, angiopoietin-2, and VEGF-1) and neurotrophic factors (BDNF and IGF-1) was restored to nondiabetic levels. These observations were also confirmed in vitro. However, phVEGF-2 treatment did not induce upregulation of endogenous cytokines or neurotrophic factors but did restore the vasa nervora with an equal impact on nerve physiology. These data suggest that the vascular pathology plays a key role in the advent of DN.
Anatomically, in situ fluorescent imaging of whole-mounted nerves (Figures 1 and 3) revealed that diabetes resulted in attrition of the vasa vasorum (both epineurium/perineurium and endoneurium) and disruption of the nerve architecture that is also characteristic of ischemic neuropathy, as has been documented previously in this model,20 resulting in decreased nerve perfusion. All of these phenomena were reversed by SHh. Interestingly, the morphological features of the vasa in SHh-treated rats seemed to more closely resemble the native vasculature, with a range of vessels sizes, in comparison to the restored vasculature in phVEGF-2–treated rats. Because we show here that SHh upregulates multiple endogenous angiogenic cytokines, including VEGF and angiopoietin-1, the observed differences in morphology appear consistent with the effect of SHh on multiple downstream targets. Similar observations were reported in a model of acute hind-limb ischemia in mice.28
In conclusion, these data suggest that SHh targets multiple signaling pathways that can influence the recovery of nerve perfusion in DN. These findings also highlight the potential for SHh to promote the development of a neovasculature that exhibits morphological features of the mature native vasculature and may therefore provide clues to the signaling mechanisms that distinguish arteriogenesis from angiogenesis.
This work supported by National Institutes of Health grants HL 53354, 63695, 66957, and 57516. The authors gratefully acknowledge the assistance of Mickey Neely and Deirdre Costello in the preparation of this manuscript.
Reiber GE. The epidemiology of diabetic foot problems. Diabet Med. 1996; 13 (suppl 1): S6–S11.
Partanen J, Niskanen L, Lehtinen J, Mervaala E, Siitonen O, Uusitupa M. Natural history of peripheral neuropathy in patients with non-insulin-dependent diabetes mellitus. N Engl J Med. 1995; 333: 89–94.
Dyck PJ. Hypoxic neuropathy: does hypoxia play a role in diabetic neuropathy. Neurology. 1989; 39: 111–118.
Rodriguez-Pena A, Botana M, Gonzalez M, Requejo F. Expression of neurotrophins and their receptors in sciatic nerve of experimentally diabetic rats. Neurosci Lett. 1995; 200: 37–40.
Apfel SC, Arezzo JC, Brownlee M, Federoff H, Kessler JA. Nerve growth factor administration protects against experimental diabetic sensory neuropathy. Brain Res. 1994; 634: 7–12.
Cameron NE, Cotter MA. The relationship of vascular changes to metabolic factors in diabetic mellitus and their role in the development of peripheral nerve complications. Diabetes Metab Rev. 1994; 10: 189–224.
Greene DA, Lattimer SA, Sima AA. Sorbitol, phosphoinositides, and sodium-potassium-ATPase in the pathogenesis of diabetic complications. N Engl J Med. 1987; 316: 599–606.
Cullum NA, Mahon J, Stringer K, McLean WG. Glycation of rat sciatic nerve tubulin in experimental diabetes mellitus. Diabetologia. 1991; 34: 387–389.
Schmidt RE, Dorsey DA, Beaudet LN, Plurad SB, Parvin CA, Miller MS. Insulin-like growth factor I reverses experimental diabetic autonomic neuropathy. Am J Pathol. 1999; 155: 1651–1660.
Simmons Z, Feldman EL. Update on diabetic neuropathy. Curr Opin Neurol. 2002; 15: 595–603.
Reske-Nielsen E, Lundbaek K. Pathological changes in the central and peripheral nervous system of young long-term diabetics. II. The spinal cord and peripheral nerves. Diabetologia. 1968; 4: 34–43.
Powell H, Knox D, Lee S, Charters AC, Orloff M, Garrett R, Lampert P. Alloxan diabetic neuropathy: electron microscopic studies. Neurology. 1977; 27: 60–66.
Tuck RR, Schmelzer JD, Low PA. Endoneurial blood flow and oxygen tension in the sciatic nerves of rats with experimental diabetic neuropathy. Brain. 1984; 107: 935–950.
Powell HC, Rosoff J, Myers RR. Microangiopathy in human diabetic neuropathy. Acta Neuropathol (Berl). 1985; 68: 295–305.
Cameron NE, Cotter MA, Low PA. Nerve blood flow in early experimantal diabetes in rats: relation to conduction deficits. Am Journal of Physiology (Heart Circ Physiol 34). 1991; 261: E1–E8.
Tesfaye S, Harris N, Jakubowski JJ, Mody C, Wilson RM, Rennie IG, Ward JD. Impaired blood flow and arterio-venous shunting in human diabetic neuropathy: a novel technique of nerve photography and fluorescein angiography. Diabetologia. 1993; 36: 1266–1274.
Stevens EJ, Carrington AL, Tomlinson DR. Nerve ischemia in diabetic rats: time-course of development, effect of insulin treatment plus comparison of streptozotocin and BB models. Diabetologia. 1994; 37: 43–48.
Rivard A, Silver M, Chen D, Kearney M, Magner M, Annex B, Peters K, Isner JM. Rescue of diabetes-related impairment of angiogenesis by intramuscular gene therapy with adeno-VEGF. Am J Pathol. 1999; 154: 355–363.
Schratzberger P, Schratzberger G, Silver M, Curry C, Kearney M, Magner M, Alroy J, Adelman LS, Weinberg DH, Ropper AH, Isner JM. Favorable effect of VEGF gene transfer on ischemic peripheral neuropathy. Nat Med. 2000; 6: 405–413.
Schratzberger P, Walter DH, Rittig K, Bahlmann FH, Pola R, Curry C, Silver M, Krainin JG, Weinberg DH, Ropper AH, Isner JM. Reversal of experimental diabetic neuropathy by VEGF gene transfer. J Clin Invest. 2001; 107: 1083–1092.
Johnson RL, Tabin CJ. Molecular models for vertebrate limb development. Cell. 1997; 90: 979–990.
St-Jacques B, Dassule HR, Karavanova I, Botchkarev VA, Li J, Danielian PS, McMahon JA, Lewis PM, Paus R, McMahon AP. Sonic hedgehog signaling is essential for hair development. Curr Biol. 1998; 8: 1058–1068.
St-Jacques B, Hammerschmidt M, McMahon AP. Indian hedgehog signaling regulates proliferation and differentiation of chondrocytes and is essential for bone formation. Genes Dev. 1999; 13: 2072–2086.
Chiang C, Litingtung Y, Lee E, Young KE, Corden JL, Westphal H, Beachy PA. Cyclopia and defective axial patterning in mice lacking sonic hedgehog gene function. Nature. 1996; 383: 407–413.
Wang M, Jin P, Bumcrot DA, Marigo V, McMahon AP, Wang EA, Woolf T, Pang K. Induction of the dopaminergic neuron phenotype in the midbrain by sonic hedgehog protein. Nat Mede. 1995; 1: 1184–1188.
Roelink H, Augsburger A, Heemskerk J, Korzh V, Norlin S, Ruiz i Altaba A, Tanabe Y, Placzek M, Edlund T, Jessell TM. Floor plate and motor neuron induction by Vhh-1, a vertebrate homologue of hedgehog expressed by the notochord. Cell. 1994; 76: 761–755.
Parmantier E, Lynn B, Lawson D, Turmaine M, Namini SS, Chakrabarti L, McMahon AP, Jessen KR, Mirsky R. Schwann cell-derived Desert hedgehog controls the development of peripheral nerve sheaths. Neuron. 1999; 23: 713–724.
Pola R, Ling LE, Silver M, Corbley MJ, Taylor FR, Baker DP, Asahara T, Isner JM. The morphogen sonic hedgehog is an indirect angiogenic agent upregulating two families of angiogenic growth factors. Nat Med. 2001; 7: 706–711.
Calcutt NA, Allendoerfer KL, Mizisin AP, Middlemas A, Freshwater JD, Burgers M, Ranciato R, Delcroix JD, Taylor FR, Shapiro R, Strauch K, Dudek H, Engber TM, Galdes A, Rubin LL, Tomlinson DR. Therapeutic efficacy of sonic hedgehog protein in experimental diabetic neuropathy. J Clin Invest. 2003; 111: 507–514.
Stanley EF. Sensory and motor nerve conduction velocities and the latency of the H reflex during growth of the rat. Exp Neurol. 1981; 71: 497–506.
Bolin LM, Iismaa TP, Shooter EM. Isolation of activated adult Schwann cells and a spontaneously immortal Schwann cell clone. J Neurosci Res. 1992; 33: 231–238.
Mukouyama YS, Shin D, Britsch S, Taniguchi M, Anderson DJ. Sensory nerves determine the pattern of arterial differentiation and blood vessel branching in the skin. Cell. 2002; 109: 693–705.
Nusslein-Volhard C WE. Mutations affecting segment number and polarity in Drosophila. Nature. 1980; 287: 795–801.
Visconti RP, Richardson CD, Sato TN. Orchestration of angiogenesis and arteriovenous contribution by angiopoietins and vascular endothelial growth factor (VEGF). Proc Natl Acad Sci U S A. 2002; 99: 8219–8224.
Isner JM, Ropper A, Hirst K. VEGF gene transfer for diabetic neuropathy. Hum Gene Ther. 2001; 12: 1593–1594.
From the Gonda Diabetes Research Center (R.N.), Beckman Research Institute of City of Hope, Duarte Calif; and the Diabetes and Hormone Center of Excellence (J.L.N.), University of Virginia School of Medicine, Charlottesville, Va.
Series Editor: Richard A. Cohen
ATVB in Focus
Diabetic Vascular Disease: Pathophysiological Mechanisms in the Diabetic
Milieu and Therapeutic Implications
Previous Brief Review in this Series:
?Naka Y, Bucciarelli LG, Wendt T, Lee LK, Rong LL, Ramasamy R, Yan SF, Schmidt AM. RAGE axis: animal models and novel insights into the vascular complications of diabetes. 2004;24:1342–1349.
Type 2 diabetes is associated with significantly accelerated rates of macrovascular complications such as atherosclerosis. Emerging evidence now indicates that atherosclerosis is an inflammatory disease and that certain inflammatory markers may be key predictors of diabetic atherosclerosis. Proinflammatory cytokines and cellular adhesion molecules expressed by vascular and blood cells during stimulation by growth factors and cytokines seem to play major roles in the pathophysiology of atherosclerosis and diabetic vascular complications. However, more recently, data suggest that inflammatory responses can also be elicited by smaller oxidized lipids that are components of atherogenic oxidized low-density lipoprotein or products of phospholipase activation and arachidonic acid metabolism. These include oxidized lipids of the lipoxygenase and cyclooxygenase pathways of arachidonic acid and linoleic acid metabolism. These lipids have potent growth, vasoactive, chemotactic, oxidative, and proinflammatory properties in vascular smooth muscle cells, endothelial cells, and monocytes. Cellular and animal models indicate that these enzymes are induced under diabetic conditions, have proatherogenic effects, and also mediate the actions of growth factors and cytokines. This review highlights the roles of the inflammatory cyclooxygenase and 12/15-lipoxygenase pathways in the pathogenesis of diabetic vascular disease.
Evidence suggests that inflammatory responses in the vasculature can be elicited by small oxidized lipids that are components of oxidized low-density lipoprotein or products of the lipoxygenase and cyclooxygenase pathways of arachidonic and linoleic acid metabolism. This review evaluates these inflammatory and proatherogenic pathways in the pathogenesis of diabetic vascular disease.
Key Words: lipoxygenase ? diabetes ? diabetes complications ? inflammation ? lipids
Diabetes is associated with significantly accelerated rates of cardiovascular complications such as atherosclerosis and hypertension. In particular, type 2 diabetes is associated with 2- to 4-fold increase in coronary artery disease.1 This has been attributed to the clustering of several risk factors, including insulin resistance, hypertension, obesity, and dyslipidemia.2,3 Multiple mechanisms contribute to vascular and arterial disease in the diabetic population.2 Basic biochemical mechanisms have been described by which hyperglycemia-induced oxidant stress activates several downstream signals that mediate diabetic complications.4–8 Furthermore, advanced glycation end products formed by glucose-induced modification of proteins can act via their receptors such as RAGE and induce cellular oxidant stress, inflammation, and vascular dysfunction in diabetes.9–12 Recent evidence from laboratory and clinical studies demonstrates that diabetic atherosclerosis is not simply a disease of hyperlipidemia but also an inflammatory disorder involving multiple mediators such as C-reactive protein, cytokines such as tumor necrosis factor alpha, and interluekin-6 (IL-6).2,13,14 A recent gene profiling study showed that high glucose treatment of monocytes leads to increased expression of multiple inflammatory cytokines, chemokines, and related factors, many of which are regulated by the proinflammatory transcription factor, nuclear factor-kappa B (NF-B).15 The recognition now that highly effective antidiabetic agents, such as thiazolidinediones, and lipid-lowering agents, such as statins, possess antiinflammatory properties underscores the major role played by inflammatory mediators in the cardiovascular complications of diabetes. However, although the actions of inflammatory peptide growth factors, cytokines, and acute phase reactants have been fairly well studied, much less is known about the actions of the oxidized lipids and eicosanoids generated by these factors during cellular activation. These oxidized lipids, generated by the action of the lipoxygenase and cyclooxygenase enzymes, are produced under diabetic conditions and have potent proatherogenic effects in vessel wall cells. This review highlights the role of these pathways and their lipid products in the pathogenesis of diabetic vascular disease.
The Cyclooxygenase and Lipoxygenase Pathways
When growth factors and cytokines bind to their cell surface receptors, they can activate several phospholipases, which act on membrane phospholipids to release arachidonic acid, a precursor for several eicosanoids with potent biological effects.16,17 Arachidonic acid can be metabolized by 3 major oxidative pathways: the cyclooxygenase (COX) pathway that forms prostaglandins; the lipoxygenase (LO) pathway, which forms hydroxyeicosatetraenoic acids (HETEs) and leukotrienes; and thirdly, the cytochrome P-450 monooxygenase pathway that forms epoxides and HETEs18 (Figure 1). Products of the cytochrome P450 metabolic pathway have potent vasoactive properties, particularly in the kidney,19 but there are no reports of their involvement in diabetic vascular disease. COX-1 and COX-2 enzymes catalyze the first step in the biosynthesis of prostaglandins (PGs) by converting arachidonic acid to PGH2.20–24 PGH2 is further converted into other PGs and eicosanoids such as PGE2, PGD2, PGF2, PGI2 (prostacyclin), and thromboxane18,20 (Figure 1). COX-1 is constitutively expressed in most cells and plays a role in basal physiological functions in several cells and tissues. COX-2, however, is usually expressed at low or undetectable levels in most tissues and cells but is significantly induced by stimuli such as lipopolysaccharide, cytokines such as interleukin (IL)-1, IL-1?, and tumor necrosis factor-, and by growth factors.20–24 An exception is seen in some tissues,24 including the pancreatic islet that constitutively and dominantly expresses COX-2,25,26 and where its products such as PGE2 are believed to play a role in inflammation, islet destruction, and inhibition of insulin secretion associated with type 1 diabetes.25–28 COX-2 and its products also have renal functions and vascular effects.29,30 They are implicated in the pathogenesis of several inflammatory diseases, and selective inhibition of COX-2 is effective in reversing inflammation without gastric side effects.20,21,24,30 Although COX-2 can form the vasodilatory and protective prostacyclin, it also produces the potent inflammatory prostaglandin, PGE2.21–24
Figure 1. Metabolism of arachidonic acid. The cellular actions of growth factors, cytokines, and other agonists can lead to the activation phospholipases and thereby release arachidonic acid. Arachidonic acid can then be metabolized by the cyclooxygenase, lipoxygenase, and cytochrome P-450 enzymes to various bioactive molecules. Note that certain LO enzymes, including 12- and 15-LO, can also react with other fatty acid substrates, such as linoleic acid, to yield additional products.
The lipoxygenase (LOs) are mainly classified as 5-, 8-, 12- or 15-LO, based on their ability to insert molecular oxygen at the corresponding carbon position of arachidonic acid (Figure 1). 31–33 The 5-LO pathway leads to the formation of 5(S)-HETE and leukotrienes. Proinflammatory leukotrienes have been implicated in the pathogenesis of atherosclerosis, but very little is known regarding their role in diabetes. The 12- and 15-LOs can form 12(S)- and 15(S)-HETEs from arachidonic acid. The production of 12(S)- and 15(S)-HETE has been shown in several vascular tissues and cells, including cultured vascular smooth muscle cells (VSMC), endothelial cells, and monocytes. LO products may play important roles in the pathogenesis of hypertension, atherosclerosis, and diabetes, as discussed more in detail later in the review. Functionally distinct isoforms of 12-LO have been cloned, including platelet, leukocyte, and epidermal 12-LOs.31–37 Human and rabbit 15-LOs and the leukocyte 12-LO have high homology and are classified as 12/15-LOs because they can form both 12(S)-HETE and 15(S)-HETE from arachidonic acid via their hydroperoxy precursors and mainly hydroperoxyocatadecadienoic acids and hydroxyocatadecadienoic acids from linoleic acid.31,38 The 12/15-LO has been detected in porcine leukocytes,34 VSMC,39 endothelial cells,40–42 and in several rat and mouse tissues.43–45
The Cyclooxygenase Pathway in Diabetic Vascular Disease
COX-2 and its proinflammatory products have been implicated in the pathogenesis of several inflammatory diseases including atherosclerosis because COX-2 products such as PGE2 and thromboxane have potent proinflammatory and vasoconstrictor properties.20–24,46 Furthermore, augmented expression of COX-2 was noted in atherosclerotic lesions,47 and COX-2 could promote lesion formation in low-density lipoprotein (LDL) receptor-deficient mice.48 Because COX-2 inhibitors also block formation of the protective prostacyclin (PGI2), studies have been performed to determine whether these inhibitors could worsen atherosclerosis.30 Earlier studies showed that elevated glucose can stimulate the generation of endothelium-derived vasoconstrictor prostanoids such as thromboxane-A2.49 However, the potential involvement of COX-2 in diabetic vascular complications, diabetic atherosclerosis, or the regulation of COX-2 in relevant cells under diabetic conditions is only now becoming evident. In endothelial cells, high-glucose (HG) treatment increased COX-2 expression and decreased nitric oxide availability.50 Very recently, COX-2 activity and expression were shown to be upregulated by high glucose as well as ligands of the receptor for advanced glycation end products (RAGE) in monocytes, and this appeared to be primarily mediated by NF-B activation.51,52 Increased oxidant stress and protein kinase C activation under diabetic conditions could be contributory factors. Interestingly, COX-2 expression was also markedly increased in monocytes from diabetic patients.51,52 Furthermore, new data show that diabetic conditions can lead to chromatin remodeling and histone acetylation at the COX-2 gene promoter at NF-B binding sites.53 COX-2 also seemed to mediate monocyte adhesive interactions.51 A recent report demonstrated that in humans, RAGE overexpression is associated with enhanced inflammatory reactions and COX-2 expression in diabetic plaque macrophages, and that this effect could also contribute to plaque destabilization by inducing metalloproteinase expression.54 There was a significant correlation between plasma levels of hemoglobin A1c and RAGE and COX-2 expression. These results suggest that apart from its documented role in pancreatic islet dysfunction, COX-2 may be a key inflammatory mediator in the pathogenesis of diabetic atherosclerosis. Thus, the diabetic state can increase COX-2 expression and activity in vascular cells and monocytes and thereby aggravate downstream inflammatory and vascular events. It is also possible that 12/15-LO activation can increase COX-2 transcription based on studies in pancreatic islet beta cells.55
The Lipoxygenase Pathway in Atherosclerosis, Restenosis, Diabetes, and Insulin Resistance
LO enzymes and their products, namely HETEs and hydroxyocatadecadienoic acids, have been implicated in the pathogenesis of atherosclerosis. The 12/15-LO enzyme can mediate the oxidative modification of low-density lipoprotein (LDL) to the atherogenic oxidized LDL.56,57 Furthermore, angiotensin II (AII) could increase macrophage-mediated modification of LDL via the 12/15-LO pathway.58 Animal models have demonstrated the key role of the LO pathway in the pathogenesis of atherosclerosis and restenosis. Overexpression of 15-LO in the vascular endothelium could accelerate atherosclerosis in LDL receptor-deficient mice.59 Leukocyte-type 12/15-LO mRNA and protein were observed in porcine atherosclerotic lesions, which were greatly augmented in diabetic and hyperlipemic pigs displaying accelerated atherosclerosis.60 LO activation may also play a role in neointimal thickening associated with restenosis because there was a marked increase in 12/15-LO expression in balloon-injured rat carotid arteries relative to uninjured. Furthermore, pretreatment with a ribozyme targeted to rat 12/15-LO could significantly reduce neointimal thickening in this rat model.61 Convincing evidence supporting a pathological role for leukocyte 12/15-LO in atherosclerosis comes from recent reports showing marked decrease in atherosclerosis in apo E–/– mice and LDLR–/– that were cross-bred with leukocyte 12/15-LO–/– mice.62,63 Furthermore, a novel inflammatory link was suggested because the macrophages from 12/15-LO–/– mice had a selective defect in lipopolysaccharide-induced IL-12 synthesis.64 An interesting genetic study suggests that 5-LO may be an important proatherogenic gene locus.65 Also, 5-LO was abundantly expressed in atherosclerotic lesions,66 and it has been suggested that 5-LO may mediate specific stages of atherosclerosis.67 However, the role of 5-LO in diabetic vascular disease is not yet known. Overall, available evidence suggests that the LOs can contribute to the pathology of atherosclerosis and diabetic vascular disease by virtue of their capacity to oxidize LDL, to induce growth and inflammatory events, and by being in an atherogenic gene locus. The relative importance of the different LOs in this regard is not yet clear. Because the 12/15-LO pathway can be upregulated by hyperglycemia, growth factors, and cytokines, it is likely that it can augment diabetic atherosclerosis and vice versa, thereby setting off a vicious loop of events.
HG culture enhanced 12/15-LO pathway activation and expression in VSMC39 and endothelial cells.42 Furthermore, AII-induced 12/15-LO activity in VSMC was greater in HG relative to normal glucose.39 Apart from AII, 12/15-LO activity and expression in VSMC could also be potently upregulated by platelet-derived growth factor (PDGF) BB and by cytokines such as IL-1, IL-4, and IL-8 in VSMC.68,69 The 15-LO expression was induced in monocytes and endothelial cells by IL-4 or IL-13.41,70–72 Thus, 12/15-LO in vascular and mononuclear cells can be induced by diabetic conditions, growth factors, and cytokines, and may contribute to their biological and atherogenic effects.
The LO pathway may therefore play a role in the cardiovascular complications associated with diabetes. Endothelial cells and VSMC cultured under hyperglycemic conditions produced increased amounts of HETEs.39,42,73 Furthermore, HG-induced adhesion of monocytes to endothelial cells could be mediated by the LO pathway.42,74 The 12/15-LO products also appear to mediate minimally modified LDL-induced monocyte binding to endothelial cells.75 LO products have potent chemotactic and hypertrophic effects in VSMC.76–78 The hypertrophic effects of 12(S)-HETE in VSMC were markedly enhanced under HG culture conditions in a manner similar to those of angiotensin II.77 There is now considerable evidence to support a role for 12/15-LO in promoting the development of diabetes and atherosclerosis. Bleich et al noted that 12/15-LO–deficient mice were resistant to the development of diabetes.79 In vivo relevance of 12/15-LO in human diabetes was suggested by a study demonstrating increased urinary excretion of 12(S)-HETE in diabetic subjects compared with matched nondiabetic controls.80 Interestingly, these diabetic subjects also had decreased urinary prostacyclin levels, suggesting a potential shunting into the 12/15-LO pathway. Recently, increased 12/15-LO expression was noted in a swine model of hyperlipidemia and diabetes-induced accelerated atherosclerosis.60 Diabetes and hyperlipidemia alone increased both monocyte oxidant stress and 12/15-LO expression in arteries, but the combination of these 2 risk factors in this swine model led to not only a marked acceleration of atherosclerosis but also a synergistic increase in oxidant stress and 12/15-LO activation.60 A recent report demonstrated increased expression of 12/15-LO in urine and endothelial cells from diabetic db/db mice.81 Interestingly, it was noted that the increased production of 12/15-LO products by the endothelial cells of the db/db mice was responsible for the observed increased in binding of monocytes to the endothelial cells from db/db versus those from control mice, and it was concluded that the 12/15-LO pathway is important for mediating early vascular changes and inflammatory reactions in diabetes.81 Taken together, these results suggest an in vivo role for leukocyte type 12/15-LO in diabetic atherosclerosis.
Emerging evidence supports a clear role of insulin resistance and diabetes in leading to accelerated cardiovascular disease. As discussed, elevated glucose and diabetes increase the expression and activity of 12/15-LO. However, fewer reports have evaluated the role of 12/15-LO in metabolic disturbances seen in the insulin resistance syndrome. Of interest are studies showing that masoprocol, a LO inhibitor, can reduce triglycerides, free fatty acids, and improve insulin action in both fructose-fed and fat-fed rat models of insulin resistance and type 2 diabetes.82,83 In addition, 12-LO products can downregulate glucose transport in VSMC.84 To further evaluate the effect of insulin resistance on vascular injury responses and 12/15-LO expression, we studied the effect of carotid balloon injury in lean and obese insulin-resistant Zucker rats. After injury, the intima-to-media ratio of obese Zucker rats was significantly greater than leans starting at 14 days after injury and persisting up to at least day 30. The expression of inflammatory mediators including 12/15-LO and IL-6 were markedly increased in obese compared with lean animals suggesting that vascular injury in obese Zucker rats is associated with inflammation. Increased macrophage and p-selectin staining was also seen. These studies (unpublished) indicate an exaggerated injury response in the insulin resistant obese Zucker rat model and that inflammation may play a major role in mediating neointimal growth under these conditions. In addition, 12/15-LO was one of the few genes upregulated in the pancreatic beta cell of the insulin resistant prediabetic Zucker diabetic fatty rat,85 thereby suggesting that 12/15-LO expression is enhanced in the prediabetic metabolic syndrome condition before frank hyperglycemia. Thus 12/15-LO may have a role in the excess cardiovascular disease seen even before diabetes is diagnosed. Because hyperglycemia alone also increases 12/15-LO expression in vascular cells, it is likely that 12/15-LO can participate in the development of type 2 diabetes and atherosclerosis, whereas the associated hyperglycemia, dyslipidemia, and insulin resistance can further augment 12/15-LO pathway activation to set off a vicious loop of inflammatory events. Furthermore, factors such as growth factors, cytokines, and advanced glycation end products, all of which are relevant to the pathogenesis of diabetes, can also upregulate the activity and expression of 12/15-LO (Figure 2).
Figure 2. Actions of 12/15-LO in the vessel wall. Induction of 12/15-LO and its products in endothelial cells by factors such as HG and AGEs can lead to oxidant stress, release of chemokines, activation of monocyte integrins, and key endothelial adhesive molecules and thereby lead to endothelial dysfunction, monocyte activation, and adhesion. In VSMC, similarly, 12/15-LO and its products can induce oxidant stress, adhesion molecules, extracellular matrix proteins, release of inflammatory cytokines, and chemokines, thereby leading to VSMC hypertrophy, migration, and inflammatory responses. 12/15-LO in monocyte/macrophages, endothelial cells, and VSMC can mediate LDL oxidation to oxidized LDL.
Lipoxygenase Products Have Growth, Chemotactic, Adhesive, and Inflammatory Effects in VSMC and Endothelial Cells
Treatment of human aortic endothelial cells with 12(S)-HETE, but not 12(R)-HETE, increased monocyte binding to the endothelial cells, a key early step in the development of atherosclerosis.74 Furthermore, the 12/15-LO ribozyme blocked high-glucose–induced binding of monocytes to endothelial cells, suggesting that glucose-induced LO activation in endothelial cells may lead to endothelial activation and dysfunction.42 The 12(S)-HETE increased the expression of CS-1 fibronectin on endothelial cells, which could be a key mechanism for inducing monocyte adhesion. Certain LO products also increased the surface expression of key inflammatory adhesion molecules such as VCAM-1 via activation of the transcription factor, NF-B.86,87 LO products also directly increased migration, cellular hypertrophy, and fibronectin synthesis in VSMC.76–78 Angiotensin II (AII)-induced increases in total cellular protein content of porcine VSMC were significantly attenuated by a specific LO inhibitor.77 Furthermore, direct addition of the 12-LO product, 12(S)-HETE, increased total cell protein content and fibronectin levels to nearly the same extent as AII.77,78 A rat 12/15-LO ribozyme could significantly inhibit HG-induced fibronectin production.61 Because AII and HG culture can increase the formation of LO products, it is attractive to speculate that the enhanced growth-promoting and matrix effects of the LO products formed by AII and HG are potential mechanisms for the accelerated growth of VSMC and enhanced hypertrophic effects of AII under HG conditions. In support of this, it was noted that rat VSMC and cardiac fibroblasts stably expressing mouse 12/15-LO showed increased growth properties.78,88 In addition, pharmacological LO inhibitors, as well as the 12/15-LO ribozyme, could also significantly inhibit PDGF-induced migration of VSMC.42,68 Because PDGF can upregulate 12/15-LO,68 LO activation may mediate, at least in part, the chemotactic effects of PDGF.
LO products also have proinflammatory effects in VSMC. Thus the 12/15-LO product of linoleic acid, 13-HPODE, led to a significant increase in the activation of the redox-sensitive and inflammatory transcription factor, NF-B in VSMC.87 This was associated with increased transcription of the inflammatory adhesion molecule VCAM-1 and the potent chemokines monocyte chemoattractant protein-1 via an NF-B–dependent mechanism.87,89
Signal Transduction and Gene Regulation Mechanisms by Which LO Products Mediate Their Cellular Actions
HETEs can activate certain isoforms of protein kinase C directly or indirectly by incorporating into membrane phospholipids, which then generate HETE-containing diacylglycerol species to activate protein kinase C.90 They can also activate several mitogen activated protein kinases (MAPKs) and thereby activate key transcription factors that mediate the expression of growth and inflammatory genes.78,90 In VSMC, 12(S)-HETE could lead to the transcription of the fibronectin gene, and this was regulated by the transcription factor CREB in a p38MAPK-dependent manner.78 However, the hydroperoxy LO product, 13(S)-HPODE, could increase the expression of VCAM-1 in an NF-B–dependent manner and partly via p38MAPK.87 Thus these oxidized lipids can serve as novel signal transducers, regulators, and amplifiers of gene induction by high glucose, growth factor, and cytokine actions. Interestingly, a novel role for HPODE and hydroperoxy precursor as seeding molecules responsible for LDL oxidation by artery wall cells and associated oxidative events related to the pathogenesis of atherosclerosis has been demonstrated.91 In monocytes, 9-hydroxyocatadecadienoic acid and 13-hydroxyocatadecadienoic acid (12/15-LO products of linoleic acid metabolism) induced the expression of the scavenger receptor, CD36, apparently via activation of the nuclear receptor, peroxisome proliferator activator-gamma.92 IL-4–induced 12/15-LO activation was also implicated in this process.93
Coffey et al demonstrated that 12/15-LO can lead to the catalytic consumption of the vasodilator, nitric oxide, and prevent nitric oxide-mediated soluble guanylate cyclase activation.94 This suggests that 12/15-LO may mediate the pathology of vascular diseases such as atherosclerosis, hypertension, and diabetes not only by the bioactivity of their lipid products but also by limiting the availability of nitric oxide in the vessel wall. Reactive oxygen species generated during LO pathway activation95 may mediate growth and inflammatory effects in VSMC and endothelial cells. Conversely, high-glucose induced oxidant stress, and reactive oxygen species in diabetes can lead to the induction of 12/15-LO in VSMC and endothelial cells and promote cellular dysfunction. Recent reports show that VSMC derived from 12/15-LO–/– mice grow slower than those derived from genetic control mice, produce much lesser amounts of superoxide, and have reduced activation of MAPKs,96 whereas endothelial cells derived from these 12/15-LO KO mice display decreased binding to monocytes compared with those from control mice.97 However, new data show that endothelial cells from 12/15-LO transgenic mice reciprocally display enhanced monocyte binding relative to those from control mice.97 Interestingly, these 12/15-LO transgenic mice also developed more atherosclerotic lesions.97 Overall, it appears that 12/15-LO can participate in an inflammatory loop with cytokines and other inflammatory genes to amplify or modulate their cellular responses and thereby accelerate the development of cardiovascular complications in diabetes.
In summary, LO and COX-2 enzymes in vascular, inflammatory, and other cells can form products with pleiotropic physiological and pathological effects. These include vasoactive, growth, adhesive, chemotactic, oxidative, and inflammatory responses, which therefore implicate them in the pathogenesis of diabetic vascular disease such as atherosclerosis and hypertension. Although COX-2–specific inhibitors are now available for clinical use, there are currently no clinically safe, pharmacologically selective, and optimally bioavailable inhibitors of 12/15-LO. Hence, the development of 12/15-LO inhibitors, including novel ribozymes, may lead to new antiinflammatory therapies for diabetic vascular complications.
We acknowledge grant support from the National Institutes of Health (PO1 HL55798, RO1 DK55240, RO1 DK065073, and RO1 DK58191) and the Juvenile Diabetes Research Foundation International. We thank Dr Q. Cai for his help with the manuscript.
Beckman JA, Craeger MA, Libby P. Diabetes and atherosclerosis. epidemiology, pathophysiology, and management. JAMA. 2002; 287: 2570–2581.
Libby P, Plutzky J. Diabetic macrovascular disease. The glucose paradox? Circulation. 2002; 106: 2760–2763.
Mokdad AH, Bowman BA, Ford ES, Vinicor F, Marks JS, Koplan JP. The continuing epidemics of obesity and diabetes in the United States. JAMA. 2001; 286: 1195–1200.
Brownlee M. Biochemistry and molecular cell biology of diabetic complications. Nature. 2001; 414: 813–820.
Ido Y, Kilo C, Williamson JR. Cytosolic NADH/NAD+, free radicals, and vascular dysfunction in early diabetes mellitus. Diabetologia. 1997; 40 (suppl 2): S115–S117.
Du X, Matsumura T, Edelstein D, Rossetti L, Zsengeller Z, Szabo C, Brownlee M. Inhibition of GAPDH activity by poly (ADP-ribose) polymerase activates three major pathways of hyperglycemic damage in endothelial cells. J Clin Invest. 2003; 112: 1049–1057.
Way KJ, Katai N, King GL. Protein kinase C and the development of diabetic vascular complications. Diabet Med. 2001; 18: 945–959.
Tesfamariam B, Brown ML, Cohen RA. Elevated glucose impairs endothelium-dependent relaxation by activating protein kinase C. J Clin Invest. 1991; 87: 1643–1648.
Brownlee M, Cerami A, Vlassara H. Advanced glycosylation end products in tissue and the biochemical basis of diabetic complications. N Engl J Med. 1988; 318: 1315–1321.
Schmidt A, Hori O, Brett J, Yan SD, Wautier JL, Stern D. Cellular receptors for advanced glycation end products. Implications for induction of oxidant stress and cellular dysfunction in the pathogenesis of vascular lesions. Arteriosclerosis Thromb. 1994; 14: 1521–1528.
Schmidt AM, Hori O, Chen J, Li JF, Crandall J, Zhang J, Cao R, Yan SD, Brett J, and Stern D. Advanced glycation endproducts interacting with their endothelial receptor induce expression of vascular cell adhesion molecule-1 (VCAM-1): a potential mechanism for the accelerated vasculopathy of diabetes. J Clin Invest. 1995; 96: 1395–1403.
Wautier JL, Zoukourian C, Chappey O, Wautier MP, Guillausseau PJ, Cao R, Hori O, Stern D, Schmidt AM. Receptor-mediated endothelial cell dysfunction in diabetic vasculopathy. Soluble receptor for advanced glycation end products blocks hyperpermeability in diabetic rats. J Clin Invest. 1996; 97: 238–243.
Ridker PM, Rifai N, Rose L, Buring JE, Cook NR. Comparison of C-reactive protein and low-density lipoprotein cholesterol levels in the prediction of first cardiovascular events. N Engl J Med. 2002; 347: 1557–1565.
Jialal I, Devaraj S, Venugopal SK. Oxidative stress, inflammation and diabetic vasculopathies: the role of alpha tocopherol therapy. Free Radic Res. 2002; 36: 1331–1336.
Shanmugam N, Reddy MA, Guha M, Natarajan R. High Glucose induced expression of pro-inflammatory cytokine and chemokine genes in monocytic cells. Diabetes. 2003; 52: 1256–1264.
Liscovitch M. Crosstalk among multiple signal-activated phospholipases. Trends Biochem Sci. 1992; 17: 393–399.
Di Marzo V. Arachidonic acid and eicosanoids as targets and effectors in second messenger interactions. Prostaglandins Leukot Essent Fatty Acids. 1995; 53: 239–254.
Smith WL. The eicosanoids and their biochemical mechanisms of action. Biochem J. 1989; 259: 315–324.
Roman RJ. P-450 metabolites of arachidonic acid in the control of cardiovascular function. Physiol Rev. 2002; 82: 131–185.
Smith WL, Dewitt DL. Prostaglandin endoperoxide H synthases (cyclooxygenases)-1 and -2. Advances Immunol. 1996; 62: 167–215.
Fu JY, Masferrer JL, Seibert K, Raz A, Needleman P. The induction and suppression of prostaglandin H2 synthase (cyclooxygenase) in human Monocytes. J Biol Chem. 1990; 265: 16737–16740.
Herschman HR. Prostaglandin Synthase 2. Biochim Biophys Acta. 1996; 1299: 125–140.
Hla T, Ristimaki A, Appleby SB, Barriocanal JG. Cyclooxygenase gene expression in inflammation and angiogenesis. Ann N Y Acad Sci. 1993; 696: 197–204.
Vane JR, Bakhle YS, Botting RM. Cyclooxygenases 1 and 2. Ann Rev Pharmacol Toxicol. 1998; 38: 97–120.
Robertson RP. Dominance of cyclooxygenase-2 in the regulation of pancreatic islet prostaglandin synthesis. Diabetes. 1998; 47: 1379–1383.
McDaniel ML, Kwon G, Hill JR, Marshall CA, Corbett JA. Cytokines and nitric oxide in islet inflammation and diabetes. Proc Soc Exp Biol Med. 1996; 211: 24–32.
Tran PO, Gleason CE, Poitout V, Robertson RP. Prostaglandin E (2) mediates inhibition of insulin secretion by interleukin-1beta. J Biol Chem. 1999; 274: 31245–31248.
Kwon G, Corbett JA, Hauser S, Hill JR, Turk J, McDaniel ML. Evidence for involvement of the proteasome complex (26S) and NF-kappaB in IL-1?-induced nitric oxide and prostaglandin production by rat islets and RINm5F cells. Diabetes. 1998; 47: 583–591.
Harris RC, McKanna JA, Akai Y, Jacoban HR, Dubois RN, Breyer MD. Cyclooxygenase-2 is associated with the macula densa of rat kidney and increases with salt restriction. J Clin Invest. 1994; 94: 2504–2510.
McAdam BF, Catella-Lawson F, Mardini IA, Kapoor S, Lawson JA, FitzGerald GA. Systemic biosynthesis of prostacyclin by cyclooxygenase (COX)-2: The human pharmacology of a selective inhibitor of COX-2. Proc Natl Acad Sci U S A. 1999; 96: 272–277.
Yamamoto S. Mammalian lipoxygenases: molecular structures and functions. Biochim Biophys Acta. 1992; 1128: 117–131.
Kuhn H, Thiele BJ. The diversity of the lipoxygenase family. Many sequence data but little information on biological significance. FEBS Lett. 1999; 449: 7–11.
Funk CD. The molecular biology of mammalian lipoxygenases and the quest for eicosanoid functions using lipoxygenase-deficient mice. Biochim Biophys Acta. 1996; 1304: 65–84.
Yoshimoto T, Suzuki H, Yamamoto S, Takai T, Yokoyama C, Tanabe T. Cloning and sequence analysis of the cDNA for arachidonate 12-lipoxygenase of porcine leukocytes. Proc Natl Acad Sci U S A. 1990; 87: 2142–2146.
Sigal E. The molecular biology of mammalian arachidonic acid metabolism. Am J Physiol. 1991; 260: L13–L28.
Funk CD, Furci L, FitzGerald. Molecular cloning, primary structure, and expression of the human platelet/erythroleukemia cell 12-lipoxygenase. Proc Natl Acad Sci U S A. 1990; 87: 5638–5642.
Izumi T, Hoshiko S, Radmark O, Samuelsson B. Cloning of the cDNA for human 12-lipoxygenase. Proc Natl Acad Sci U S A. 1990; 87: 7477–7481.
Sigal E, Craik CS, Highland E, Grunberger D, Costello LL, Dixon RA, Nadel JA. Molecular cloning and primary structure of human 15-lipoxygenase. Biochem Biophys Res Commun. 1988; 157: 457–464.
Natarajan R, Gu JL, Rossi J, Gonzales N, Lanting L, Xu L, Nadler J. Elevated glucose and angiotensin II increase 12-lipoxygenase activity and expression in porcine aortic smooth muscle cells. Proc Natl Acad Sci U S A. 1993; 90: 4947–4951.
Kim JA, Gu JL, Natarajan R, Berliner JA, Nadler JL. A leukocyte type of 12-lipoxygenase is expressed in human vascular and mononuclear cells. Evidence for upregulation by angiotensin II. Arterioscler Thromb Vasc Biol. 1995; 15: 942–948.
Lee YW, Kuhn H, Kaiser S, Hennig B, Daugherty A, Toborek M. Interleukin 4 induces transcription of the 15-lipoxygenase I gene in human endothelial cells. J Lipid Res. 2001; 42: 783–791.
Patricia MK, Natarajan R, Dooley AN, Hernandez F, Gu JL, Berliner JA, Rossi JJ, Nadler JL, Meidell RS, Hedrick CC. Adenoviral delivery of a leukocyte-type 12-lipoxygenase ribozyme inhibits effects of glucose and platelet-derived growth factor in vascular endothelial and smooth muscle cells. Circ Res. 2001; 88: 659–665.
Watanabe T, Medina JF, Haeggstrom JZ, Radmark O, Samuelsson B. Molecular cloning of a 12-lipoxygenase cDNA from rat brain. Eur J Biochem. 1993; 212: 605–612.
Katoh T, Lakkis FG, Makita N, Badr KF. Co-regulated expression of glomerular 12/15-lipoxygenase and interleukin-4 mRNAs in rat nephrotoxic nephritis. Kidney Int. 1994; 46: 341–349.
Chen XS, Kurre U, Jenkins NA, Copeland NG, Funk CD. cDNA cloning, expression, mutagenesis of C-terminal isoleucine, genomic structure, and chromosomal localizations of murine 12-lipoxygenases. J Biol Chem. 1994; 269: 13979–13987.
Linton MF, Fazio S. Cyclooxygenase-2 and inflammation in atherosclerosis. Curr Opin Pharmacol. 2004; 4: 116–123.
Schonbeck U, Sukhova G, Graber P, Coulter S, Libby P. Augmented expression of cyclooxygenase-2 in human atherosclerotic lesions. Am J Pathol. 1999; 155: 1281–1291.
Burleigh ME, Babaev VR, Oates JA, Harris RC, Gautam S, Riendeau D, Marnett LJ, Morrow JD, Fazio S, Linton MF. Cyclooxygenase-2 promotes early atherosclerotic lesion formation in LDL receptor-deficient mice. Circulation. 2002; 105: 1816–1823.
Tesfamariam B, Brown ML, Deykin D, Cohen RA. Elevated glucose promotes generation of endothelium-derived vasoconstrictor prostanoids in rabbit aorta. J Clin Invest. 1990; 85: 929–932.
Cosentino F, Eto M, De Paolis P, van der Loo B, Bachschmid M, Ullrich V, Kouroedov A, Gatt CD, Joch H, Volpe M, Luscher TF. High glucose causes upregulation of cyclooxygenase-2 and alters prostanoid profile in human endothelial cells: role of protein kinase C and reactive oxygen species. Circulation. 2003; 107: 1017–1023.
Shanmugam N, Kim YS, Lanting L, Natarajan R. Regulation of cyclooxygenase-2 expression in monocytes by ligation of the receptor for advanced glycation end products. J Biol Chem. 2003; 278: 34834–34844.
Shanmugam N, Gaw-Gonzalo I, Natarajan R. Molecular mechanisms of high glucose-induced COX-2 regulation in monocytes. Diabetes. 2004; 53: 795–802.
Miao F, Gonzalo IG, Lanting L, Natarajan R. In vivo chromatin remodeling events leading to inflammatory gene transcription under diabetic conditions. J Biol Chem. 2004; 279: 18091–18097.
Cipollone F, Iezzi A, Fazia M, Zucchelli M, Pini B, Cuccurullo C, De Cesare D, De Blasis G, Muraro R, Bei R, Chiarelli F, Schmidt AM, Cuccurullo F, Mezzetti A. The receptor RAGE as a progression factor amplifying arachidonate-dependent inflammatory and proteolytic response in human atherosclerotic plaques: role of glycemic control. Circulation. 2003; 108: 1070–1077.
Han X, Chen S, Sun Y, Nadler JL, Bleich D. Induction of cyclooxygenase-2 gene in pancreatic beta cells by 12-lipoxygenase pathway product 12-hydroxyeicosatetraenoic acid. Mol Endocrinol. 2002; 16: 2145–2154.
Parthasarathy S, Wieland E, Steinberg D. A role for endothelial cell lipoxygenase in the oxidative modification of low density lipoprotein. Proc Natl Acad Sci U S A. 1989; 86: 1046–1050.
Folcik VA, Nivar-Aristy RA, Krajewski LP, Cathcart MK. Lipoxygenase contributes to the oxidation of lipids in human atherosclerotic plaques. J Clin Invest. 1995; 96: 504–510.
Scheidegger KJ, Butler S, Witztum JL. Angiotensin II increases macrophage-mediated modification of low-density lipoprotein via a lipoxygenase-dependent pathway. J Biol Chem. 1997; 272: 21609–21615.
Harats D, Shaish A, George J, Mulkins M, Kurihara H, Levkovitz H, Sigal E. Overexpression of 15-lipoxygenase in vascular endothelium accelerates early atherosclerosis in LDL receptor-deficient mice. Arterioscler Thromb Vasc Biol. 2000; 20: 2100–2105.
Natarajan R, Gerrity RG, Gu JL, Lanting L, Thomas L, Nadler JL. Role of 12-lipoxygenase and oxidant stress in hyperglycaemia-induced acceleration of atherosclerosis in a diabetic pig model. Diabetologia. 2002; 45: 125–133.
Gu JL, Pei H, Thomas L, Nadler JL, Rossi JJ, Lanting L, Natarajan R. Ribozyme-mediated inhibition of rat leukocyte-type 12-lipoxygenase prevents intimal hyperplasia in balloon-injured rat carotid arteries. Circulation. 2001; 103: 1446–1452.
Cyrus T, Witztum JL, Rader DJ, Tangirala R, Fazio S, Linton MF, Funk CD. Disruption of the 12/15-lipoxygenase gene diminishes atherosclerosis in apo E-deficient mice. J Clin Invest. 1999; 103: 1597–1604.
George J, Afek A, Shaish A, Levkovitz H, Bloom N, Cyrus T, Zhao L, Funk CD, Sigal E, Harats D. 12/15-Lipoxygenase gene disruption attenuates atherogenesis in LDL receptor-deficient mice. Circulation. 2001; 104: 1646–1650.
Zhao L, Cuff CA, Moss E, Wille U, Cyrus T, Klein EA, Pratico D, Rader DJ, Hunter CA, Pure E, Funk CD. Selective interleukin-12 synthesis defect in 12/15-lipoxygenase-deficient macrophages associated with reduced atherosclerosis in a mouse model of familial hypercholesterolemia. J Biol Chem. 2002; 277: 35350–35356.
Mehrabian M, Allayee H, Wong J, Shi W, Wang XP, Shaposhnik Z, Funk CD, Lusis AJ, Shih W. Identification of 5-lipoxygenase as a major gene contributing to atherosclerosis susceptibility in mice. Circ Res. 2002; 91: 120–126.
Spanbroek R, Grabner R, Lotzer K, Hildner M, Urbach A, Ruhling K, Moos MP, Kaiser B, Cohnert TU, Wahlers T, Zieske A, Plenz G, Robenek H, Salbach P, Kuhn H, Radmark O, Samuelsson B, Hadenicht AJ. Expanding expression of the 5-lipoxygenase pathway within the arterial wall during human atherogenesis. Proc Natl Acad Sci U S A. 2003; 100: 1238–1243.
Mehrabian M, Allayee H. 5-Lipoxygenase and atherosclerosis. Curr Opin Lipidol. 2003; 14: 447–457.
Natarajan R, Bai W, Rangarajan V, Gonzales N, Gu JL, Lanting L, Nadler JL. Platelet-derived growth factor BB mediated regulation of 12-lipoxygenase in porcine aortic smooth muscle cells. J Cell Physiol. 1996; 169: 391–400.
Natarajan R, Rosdahl J, Gonzales N, Bai W. Regulation of 12-lipoxygenase by cytokines in vascular smooth muscle cells. Hypertension. 1997; 30: 873–879.
Conrad DJ, Kuhn H, Mulkins M, Highland E, Sigal E. Specific inflammatory cytokines regulate the expression of human monocyte 15-lipoxygenase. Proc Natl Acad Sci U S A. 1992; 89: 217–221.
Nassar GM, Morrow JD, Roberts LJ, 2nd, Lakkis FG, Badr KF. Induction of 15-lipoxygenase by interleukin-13 in human blood monocytes. J Biol Chem. 1994; 269: 27631–27634.
Roy B, Cathcart MK. Induction of 15-lipoxygenase expression by IL-13 requires tyrosine phosphorylation of Jak2 and Tyk2 in human monocytes. J Biol Chem. 1998; 273: 32023–32029.
Brown ML, Jakubowski JA, Leventis LL, Deykin D. Elevated glucose alters eicosanoid release from porcine aortic endothelial cells. J Clin Invest. 1988; 82: 2136–2141.
Patricia MK, Kim JA, Harper CM, Shih PT, Berliner JA, Natarajan R, Nadler JL, Hedrick CC. Lipoxygenase products increase monocyte adhesion to human aortic endothelial cells. Arterioscler Thromb Vasc Biol. 1999; 19: 2615–2622.
Honda HM, Leitinger N, Frankel M, Goldhaber JI, Natarajan R, Nadler JL, Weiss JN, Berliner JA. Induction of monocyte binding to endothelial cells by MM-LDL: role of lipoxygenase metabolites. Arterioscler Thromb Vasc Biol. 1999; 19: 680–686.
Nakao J, Ooyama T, Ito H, Chang WC, Murota S. Comparative effect of lipoxygenase products of arachidonic acid on rat aortic smooth muscle cell migration. Atherosclerosis. 1982; 44: 339–342.
Natarajan R, Gonzales N, Lanting L, Nadler J. Role of the lipoxygenase pathway in angiotensin II-induced vascular smooth muscle cell hypertrophy. Hypertension. 1994; 23: I142–I147.
Reddy MA, Thimmalapura PR, Lanting L, Nadler JL, Fatima S, Natarajan R. The oxidized lipid and lipoxygenase product 12(S)-hydroxyeicosatetraenoic acid induces hypertrophy and fibronectin transcription in vascular smooth muscle cells via p38 MAPK and cAMP response element-binding protein activation. Mediation of angiotensin II effects. J Biol Chem. 2002; 277: 9920–9928.
Bleich D, Chen S, Zipser B, Sun D, Funk CD, Nadler JL. Resistance to type 1 diabetes induction in 12-lipoxygenase knockout mice. J Clin Invest. 1999; 103: 1431–1436.
Antonipillai I, Nadler J, Vu EJ, Bughi S, Natarajan R, Horton R. A 12-lipoxygenase product, 12-hydroxyeicosatetraenoic acid, is increased in diabetics with incipient and early renal disease. J Clin Endocrinol Metab. 1996; 81: 1940–1945.
Hatley ME, Srinivasan S, Reilly KB, Bolick DT, Hedrick CC. Increased production of 12/15 lipoxygenase eicosanoids accelerates monocyte/endothelial interactions in diabetic db/db mice. J Biol Chem. 2003; 278: 25369–25375.
Reed MJ, Meszaros K, Entes LJ, Claypool MD, Pinkett JG, Brignetti D, Luo J, Khandwala A, Reaven GM. Effect of masoprocol on carbohydrate and lipid metabolism in a rat model of Type 2 diabetes. Diabetologia. 1999; 42: 102–106.
Scribner K, Gadbois T, Gowri M, Azhar Sl, Reaven G. Masoprocol decreases serum triglyceride concentration in rats with fructose-induced hypertriglyceridemia. Metabolism. 2000; 9: 1106–1110.
Alpert E, Gruzman A, Totary H, Kaiser N, Reich R, Sasson S. A natural protective mechanism against hyperglycemia in vascular endothelial and smooth muscle cells: role of glucose and 12-hydroxyeicosatetraenoic acid. Biochem J. 2002; 362: 413–422.
Tokuyama Y, Sturis J, DePaoli AM, Takeda J, Stoffel M, Tang J, Sun X, Polonsky KS, Bell GI. Evolution of beta cell dysfunction in the male Zucker diabetic fatty rat. Diabetes. 1995; 44: 1447–1457.
Sultana CY, Shen V, Rattan V, Kalra K. Lipoxygenase metabolites induced expression of adhesion molecules and transendothelial migration of monocyte-like HL-60 cells is linked to protein kinase C activation. J Cell Physiol. 1996; 167: 477–487.
Natarajan R, Reddy MA. Malik KU, Fatima S, Khan BV. Signaling mechanisms of nuclear factor-kappab-mediated activation of inflammatory genes by 13-hydroperoxyoctadecadienoic acid in cultured vascular smooth muscle cells. Arterioscler Thromb Vasc Biol. 2001; 21: 1408–1413.
Wen Y, Gu J, Peng X, Zhang G, Nadler JL. Overexpression of 12-lipoxygenase and cardiac fibroblast hypertrophy. Trends in Cardiovasc Med. 2003; 134: 129–136.
Dwarakanath RS, Sahar S, Reddy MA, Castanotto D, Rossi JJ, Natarajan R. Regulation of monocyte chemoattractant protein-1 by the oxidized lipid,13-hydroperoxyoctadecadienoic acid, in vascular smooth muscle cells via nuclear factor-kappa B (NF-kB). Mol Cell Cardiol. 2004; 36: 585–595.
Natarajan R, Nadler JL. Lipoxygenases and Lipid Signaling in Diabetes. Frontiers in Biosciences. 2003; 8: S783–S795.
Navab M, Hama SY, Cooke CJ, Anantharamaiah GM, Chaddha M, Jin L, Subbanagounder G, Faull KF, Reddy ST, Miller NE, Fogelman AM. Normal high-density lipoprotein inhibits three steps in the formation of mildly oxidized low-density lipoprotein: step 1. J Lipid Res. 2000; 41: 1481–1494.
Tontonoz P, Nagy L, Alvarez JG, Thomazy VA, Evans RM. PPARgamma promotes monocyte/macrophage differentiation and uptake of oxidized LDL. Cell. 1998; 93: 241–252.
Huang JT, Welch JS, Ricote M, Binder CJ, Willson TM, Kelly C, Witztum JL, Funk CD, Conrad D, Glass CK. Interleukin-4-dependent production of PPAR- ligands in macrophages by 12/15-lipoxygenase. Nature. 1999; 400: 378–382.
Coffey MJ, Natarajan R, Chumley PH, Coles B, Thimmalapura PR, Nowell M, Kuhn H, Lewis MJ, Freeman BA, O’Donnell VB. Catalytic consumption of nitric oxide by 12/15- lipoxygenase: inhibition of monocyte soluble guanylate cyclase activation. Proc Natl Acad Sci U S A. 2001; 98: 8006–8011.
Roy P, Roy SK, Mitra A, Kulkarni AP. Superoxide generation by lipoxygenase in the presence of NADH and NADPH. Biochim Biophys Acta. 1994; 1214: 171–179.
Reddy MA, Kim YS, Lanting L, Natarajan R. Reduced growth factor responses in vascular smooth muscle cells derived from 12/15-lipoxygenase deficient mice. Hypertension. 2003; 41: 1294–1300.
Reilly KB, Srinivasan S, Hatley ME, Kim Patricia M, Lannigan J, Bolick DT, Vandenhoff G, Pei H, Natarajan R, Nadler JL, Hedrick CC. 12/15 lipoxygenase activity mediates inflammatory monocyte: endothelial interactions and atherosclerosis in vivo. J Biol Chem. 2003; 279: 9440–9450.
S A Vernon
Department of Ophthalmology, Queens Medical Centre, University Hospital, Nottingham NG7 2UH, UK; firstname.lastname@example.org
Which dosage should we use?
Keywords: diabetic retinopathy; corticosteroids; macular oedema; intravitreal drug delivery
It is less than 4 years ago since publication of the first report on the use of crystalline cortisone in the form of triamcinolone acetonide to treat recalcitrant macular oedema in patients with diabetes.1 This report of a single eye was rapidly followed by a case series from the same author (Jonas),2 who has remained faithful to a dosage regimen of 20 mg in a number of publications documenting the efficacy and side effects of this novel form of therapy.3–8 In parallel with Jonas, Martidis and co-workers in the United States9 reported on the use of a 4 mg dosage in a similar clinical scenario. Since then there have been many reports, including early results from a randomised controlled trial (RCT), utilising this somewhat lower dosage.10–15 All studies have thus far indicated a significant improvement in macular function and/or structure following injection, at least in the short term.
Why have different dosage regimens been employed? Triamcinolone acetonide is conveniently and affordably available in concentrations of 40 mg/ml in a sterile preparation (Kenolog, or Volon A or Kenacort depending on country, Bristol-Myers-Squibb) used commonly in other specialties such as orthopaedics. As 0.1 ml is the maximum volume most eyes can tolerate when injected into the vitreous cavity without causing inevitable central retinal artery occlusion, the maximum dosage of unadulterated triamcinolone one can give at any one time without resorting to paracentesis is clearly 4 mg. Jonas concentrates the triamcinolone crystals using a filter and then injects (after a routine paracentesis) 0.2 ml of a suspension of triamcinolone and Ringer’s solution (0.4 ml in the comparative dosage study in this issue of the BJO (p 999)).
In the aforementioned study, a small RCT comparing the efficacy of three dosages of triamcinolone (assayed to be 2 mg, 5 mg, and 13 mg of injected triamcinolone), Spandau and colleagues from Jonas’s team make a case for the use of the higher dosage in diabetic macular oedema (13 mg was found to be the equivalent of a 20 mg stated dose in Jonas’s previous studies). This was based upon a significantly improved outcome in terms of maximum distance visual acuity in the 13 mg group compared with the 2 mg group (but notably not the 5 mg group—a dosage closest to that probably injected in all other series), and a finding of a significant dose/duration effect correlation when all three doses were considered. There are a number of unanswered questions posed by the study such as the duration of oedema before treatment and the numbers in each group who had received laser treatment. It would also be tempting to discount this small study after viewing the data points on the figures where it can be seen that one or two eyes in the 13 mg group appear to be having a disproportionate influence on the results. The results, however, appeal somewhat intuitively, and merit further consideration.
Whether vitrectomy or intravitreal triamcinolone, perhaps combined with cataract surgery, proves to be optimal for an individual patient/eye will require much more research
Examination of the literature, comparing Jonas’s results with those of others, fails to create a clearcut difference in outcome between dosage regimens for many reasons. These include differing entry criteria (with or without previous macular laser, duration of oedema, levels of acuity at baseline), methods of acuity measurement (EDTRS chart versus Snellen chart), the presence or lack of supporting evidence of efficacy such as optical coherence tomography (OCT), and methods of comparison (randomised trial versus non-randomised case or case-control series). However, a common feature of all series extending to 6 months or longer is a rapid improvement phase, followed by a plateau phase and then regression, at least in some eyes.
Martidis et al,9 using 4 mg, reported that three of the eight eyes at 6 months required re-injection for recurrence of oedema with loss of vision. In the study by Massin et al11 (4 mg), the difference in macular thickness (which correlates well with visual improvement) measured by OCT at 6 months compared with baseline had become non-significant owing to recurrence of oedema in five of the 12 injected eyes. In addition, at no stage was there a significant difference between the acuity in treated and the control (untreated) eyes. However, Sutter et al (RCT with 4 mg versus subconjuctival placebo),10 at 3 months, showed a significant visual improvement in treated eyes with 24% improving by 10 or more EDTRS letters. In the study by Ciardella et al,12 8/30 (27%) of eyes had received at least one re-injection (again 4 mg) because of recurrence of oedema at a median time of 6 months between the first and second injection. Audren et al,16 used pharmacokinetic-pharmacodynamic modelling of OCT readings, estimated the mean maximum duration of effect of a 4 mg injection to be 140 days. In Jonas’s largest series to date, in which triamcinolone was the only treatment for diabetic macular oedema, a dosage of 20 mg (probably 13 mg active triamcinlone) resulted in an improvement of at least two Snellen lines in 68% of the 97 treated eyes with a mean increase of 2.6 lines at best.4 There was no tendency to regress over the first 4 months following injection but by 6 months the effect of injection on acuity had become insignificant in a group analysis, a finding similar to our own series13 utilising a 4 mg injection.
The lack of any comparative data on lens morphology and density and macular thickness in Spandau’s study is unfortunate as acuity some months after injection may be compromised by cataract formation.17 Comparative data on all aspects of visual function, including the much neglected reading ability13,18 and complication rates, in particular cataract and glaucoma, will only be answered by further prospective RCTs, which should begin to utilise subjective patient generated outcome measures relating visual improvement to time of perceived benefit.
The use of intravitreal triamcinolone has given new hope to many patients with chronic diabetic macular oedema often permitting them to read again, even if only for a few months.13,18 Whether vitrectomy19 or intravitreal triamcinolone, perhaps combined with cataract surgery,20 proves to be optimal for an individual patient/eye will require much more research. The National Eye Institute in the United States has funded studies examining the efficacy of intraocular steroids in diabetic retinopathy and retinal vein occlusion.21 Those practising elsewhere should take into account the results from Professor Jonas’s team when formulating their protocols for similar research.
Jonas JB, Sofker A. Intraocular injection of crystalline cortisone as adjunctive treatment of diabetic macular edema. Am J Ophthalmol 2001;132:425–7.
Jonas JB, Kreissig I, Sofker A, et al. Intravitreal injection of triamcinolone for diffuse diabetic macular edema. Arch Ophthalmol 2003;121:57–61.
Jonas JB, Harder B, Kamppeter BA. Inter-eye difference in diabetic macular edema after unilateral intravitreal injection of triamcinolone acetonide. Am J Ophthalmol 2004;138:970–7.
Jonas JB, Akkoyun I, Kreissig I, et al. Diffuse diabetic macular oedema treated by intravitreal triamcinolone acetonide: a comparative, non-randomised study. Br J Ophthalmol 2005;89:321–6.
Jonas JB, Kreissig I, Degenring RF, et al. Repeated intravitreal injection of triamcinolone acetonide for diffuse diabetic macular oedema. Br J Ophthalmol 2005;89:122.
Jonas JB, Degenring R, Kreissig I, et al. Safety of intravitreal high-dose reinjections of triamcinolone acetonide. Am J Ophthalmol 2004;138:1054–5.
Jonas JB, Degenring RF, Kamppeter BA, et al. Duration of the effect of intraviteal triamcinolone acetonide as treatment for diffuse diabetic macular edema. Am J Ophthalmol 2004;138:158–60.
Jonas JB, Kreissig I, Degenring R. Intraocular pressure after intravitreal injection of triamcinolone acetonide. Br J Ophthalmol 2003;87:24–7.
Martidis A, Duker JS, Greenberg PB, et al. Intravitreal triamcinolone for refractory diabetic macular edema. Ophthalmology 2002;109:920–7.
Sutter FKP, Simpson JM, Gillies MC. Intravitreal traimcinolone for diabetic macular edema that persists after laser treatment. Ophthalmology 2004;111:2044–9.
Massin P, Audren F, Haouchine B, et al. Intravitreal triamcinolone acetonide for diffuse diabetic macular edema—preliminary results of a prospective controlled trial. Ophthalmology 2004;111:218–25.
Ciardella AP, Klancnik J, Schiff W, et al. Intravitreal triamcinolone for the treatment of refractory diabetic macular oedema with hard exudates: an optical coherence tomography study. Br J Ophthalmol 2004;88:1131–6.
Negi AK, Vernon SA, Lim CS, et al. Intravitreal triamcinolone improves vision in eyes with chronic diabetic macular oedema refractory to laser photocoagulation. Eye. 2004 .
Lam DSC, Chan CKM, Tang EWH, et al. Intravitreal triamcinolone for diabetic macular oedema in Chinese patients: six-month prospective longitudinal pilot study. Clin Exp Ophthalmol 2004;32:569–72.
Karacorlu M, Ozdemir H, Karacorlu S, et al. Intravitreal triamcinolone as a primary therapy in diabetic macular oedema. Eye. 2004 .
Audren F, Tod M, Massin P, et al. Pharmacokinetic-Pharmacodynamic modelling of the effect of triamcinolone acetonide on central macular thickness in patients with diabetic macular edema. Invest Ophthalmol Vis Sci 2004;45:3435–41.
Gillies MC, Simpson JM, Billson FA, et al. Saftey of an intravitreal injection of triamcinolone—results from a randomised clinical trial. Arch Ophthalmol 2004;122:336–40.
Islam MS, Negi A, Vernon SA. Improved visual acuity and macular thickness 1 week after intravitreal triamcinolone for diabetic macular oedema. Eye. 2004 .
Yamamoto T, Akabane N, Takeuchi S. Vitrectomy for diabetic macular edema: the role of posterior vitreous detachment and epimacular membrane. Am J Ophthalmol 2001;132:369–77.
Lam DSC, Chan CKM, Mohamed S, et al. Phacoemulsification with intravitreal triamcinolone in patients with cataract and coexisting diabetic macular oedema: a six-month prospective pilot study. Eye. 2004 .
Flynn HW, Scott IU. Intravitreal triamcinolone acetonide for macular edema associated with diabetic retinopathy and venous occlusive disease. Arch Ophthalmol 2005;123:258–9.
1 Gloucestershire Eye Unit, Cheltenham General Hospital, Cheltenham GL53 7AN, UK
2 Gloucestershire R & D Support Unit, Cheltenham General Hospital, Cheltenham GL53 7AN, UK
Dr P H Scanlon
Gloucestershire Eye Unit, Cheltenham General Hospital, Sandford Road, Cheltenham GL53 7AN, UK; email@example.com
Accepted for publication 3 February 2005
Aims: To determine how the workload of an ophthalmology department changed following the introduction of an organised retinal screening programme.
Methods: Information was collected from the hospital medical record of people with diabetes attending eye clinics over 4 years. The first year was before screening, the next 2 years the first round, and the fourth year the second round.
Results: The total number of people with diabetes referred each year over the 4 year period was 853, 954, 974, 1051 consecutively. The number of people with diabetes in the county rose by 1400 per annum. The total number of referrals for an opinion about diabetic retinopathy was 227, 333, 363, 368, for cataract was 64, 57, 77, 93, and for glaucoma was 57, 62, 61, 68. The total number of patients referred for laser treatment over the 4 years was 77, 124, 111, and 63
Conclusion: This study suggests that the workload in the eye clinic increases in the first round of screening but in subsequent rounds it does not fall below the pre-screening level, except for laser treatment. This may be partly because of increasing numbers of people with diabetes. With the introduction of a national screening programme, this has significant workload implications for the National Health Service.
Abbreviations: DR, diabetic retinopathy; GDESS, Gloucestershire Diabetic Eye Screening Service
Keywords: screening; diabetic retinopathy; digital photography
In 1995-6 Gloucestershire Primary Care Clinical Audit Group coordinated a countywide audit,1 which identified 9556 people with diabetes over the age of 16 (2.1% of the county’s population aged 16 and over). In October 1998, a mobile digital photographic screening programme was introduced, funded by Gloucestershire Health Authority, with contributions from charitable sources.
In 2000, the National Screening Committee produced their recommendations2–4 for a national screening programme and estimated the effect on treatment workload and associated costs.
The aim of this study was to determine how the workload of an ophthalmology department changes following the introduction of an organised retinal screening programme.
MATERIALS AND METHODS
A retrospective collection of data from the hospital medical records after electronic identification of diabetic patients attending the eye clinic was carried out by matching retinal screening numbers with eye clinic codes.
The sample group consisted of all people known to have diabetes in Gloucestershire aged 16 years and over attending as new patients to eye clinics in Gloucestershire over 4 years. This case review included the year before the introduction of screening, the 2 years of the first round of screening, and first year of the second round of screening.
Those aged less than 16 years. We do not collect information on people with diabetes under 16 years for our screening service at present. There are approximately 170 children and teenagers in this age group under the care of Gloucestershire paediatricians.
Justification of sample size
The number of people with diabetes attending as new patients to eye clinics in Gloucestershire over a 3 year period was 2700 and it was expected to be 3600 for the 4 years when the data was collected in October 2001.
In all, 8566 people with diabetes were screened in the first round of screening in Gloucestershire with 434 (5.1%) referred with referable diabetic retinopathy (33 proliferative, 102 pre-proliferative, and 299 maculopathy). Our previous study5 showed a sensitivity of 88% in detection of referable diabetic retinopathy (DR). It follows that a missed case (false negative) might occur in five proliferative patients out of 8566 (0.06%) people with diabetes and in 41 maculopathy patients (0.48%). Patients missed in screening, by being false negatives, will be picked up subsequently by natural presentation via other routes of referral and to identify these small numbers required looking at all 3600 records. In this study complete coverage was intended and there was therefore no sampling. When the study was designed we anticipated an annual screening service and, hence, data collection for a 3 year period and the design was altered to 4 years when it became clear that the first round of screening would take 2 years to complete (because of the higher than anticipated workload for the screeners).
All people with diabetes aged 16 years and over in Gloucestershire have been given a retinal screening number. This includes those responding and those not responding to the screening invitation. The diabetic registers of all of the 85 general practices in Gloucestershire were used to compile a list of all people known to have diabetes. A total of 12 300 people with diabetes (2.6% of the population) had been identified when this grant application was considered in February 2001 reaching 13 239 in October 2001. Patients with retinal screening numbers on the east Gloucestershire PAS system were matched against all new patient attendance at eye clinic codes on the same system in February 2001 and again in October 2001. For west Gloucestershire, all patients with retinal screening numbers were matched against all new patient attendance to west Gloucestershire eye clinics using NHS number, surname, forename, and date of birth. The retrospective identification of these people with diabetes attending eye clinics was undertaken once the first round of screening had been completed and an accurate list of people with diabetes aged >16 years in the county had been collected.
The medical records from all new diabetic referrals to eye outpatient departments in Gloucestershire over 4 years were examined and the following information was recorded:
Type of diabetes care
Date of referral letter
Date of clinic appointment
Route of referral
Reason given for referral
Date of previous retinal examination (if known)
Date of previous retinal photography (if applicable)
Best corrected visual acuity recorded in each eye (Snellen)
Grade of retinopathy in worse eye
Other eye disease
The definition of referable DR used in this study was described in two previous studies.5,6 The reason for using referable DR in this study was because the screening service has a level at which referral to an eye department is felt to be necessary, which may not be at such an advanced level as may require laser treatment. In the context of the current study, sight threatening DR was used to describe a retinopathy level thought to be at a stage requiring laser treatment for maculopathy or pre-proliferative/proliferative DR. In Gloucestershire, laser is usually applied to treat maculopathy as described in the Early Treatment Diabetic Retinopathy Study7,8 and pre-proliferative DR is usually followed up carefully and laser applied when proliferative DR is detected.
Annual new diabetic referral rate to eye clinics in Gloucestershire over a 4 year period
Over the 4 year period 3877 people with diabetes were identified electronically as having attended Eye clinics in Gloucestershire as new patients (table 1). This included patients who were already registered at the eye clinic who attended as an emergency at a date between their booked appointments, as routinely recorded as a "new patient referral" on our PAS system.
Table 1 New referrals in people with diabetes
There has been a progressive rise in new patient attendances of people with diabetes in the eye clinic during the 4 years starting in 1997–8 with annual rises from 853 (pre-screening) to 954, 974, and 1051.
Of the 3877 patients attending eye clinics it was possible to examine the notes of 3832. Notes were not available for 45 patients over the 4 year period, and for eight patients there was no evidence of a clinic appointment in the notes.
During the first 3 years of screening by the Gloucestershire Diabetic Eye Screening Service (GDESS) there was an increase in the numbers screened per annum from 4157 to 4520 (table 2).
Table 2 Numbers screened by GDESS (bi-annual service)
The number of people with diabetes in the county: annual increases
In the 1996 Gloucestershire audit there were 9566 people with diabetes identified in the county. After the first round of screening in October 2000 there were 11 909 identified. The following October 2001 there were 13 239. This number has continued to rise at a rate of approximately 1400 per annum and was 15 433 in May 2003 (table 3).
Table 3 Number of adult diabetic people in Gloucestershire (aged >16 years)
The reason given for attendance
The commonest reasons for referral are "reduced vision" and "retinopathy seen."
Other reasons given for referral are shown in table 4 and the reason itemised where it gave rise to more than 100 referrals.
Table 4 Reason given for referral
The workload and source of the referral for "reduced vision"
The numbers referred with "reduced vision" decreased over the 4 years from 288 in year 1 (pre-screening) to 272, 232, and 221 and as a percentage of those referred, this was 33.8%, 28.5%, 23.8%, and 21% (2 for trend 44.1, p<0.0001).
The common outcomes for these patients are shown in table 5 (NB, more than one outcome is possible).
Table 5 Common outcomes in those referred with reduced vision
The proportion of patients referred with reduced vision who were discharged at first clinic attendance decreased from 20.1% (pre-screening) to 17.3%, 13.8%, and 8.6% in subsequent screening years (2 for trend 13.7, p<0.001).
The number of patients referred for reduced vision who were listed for laser for DR at first clinic attendance decreased over the 4 years from 26 (pre-screening) to 18, eight, and seven and as a percentage of those referred with reduced vision this was 8%, 5.8%, 3.1%, and 2.9% (2 for trend 9.1, p = 0.003).
The number of patients referred for reduced vision who were listed for cataract surgery at first clinic attendance decreased over the 4 years from 112 (pre-screening) to 100, 89, and 99 and as a percentage of those referred with reduced vision this was 34.3%, 32.1%, 34.9%, and 41.6% (2 for trend 3.27, p = 0.07).
The workload and source of the referral for background and referable diabetic retinopathy (tables 4 and 5)
The number of patients referred who were found to have no DR during the 4 years did not vary greatly (484 pre-screening, 473, 466, and 512) but since the number of people referred increased the percentage found to have no DR decreased from 56.7% to 49.6%, 47.8%, 48.7% (2 for trend 11.5, p = 0.001).
The number of patients referred who were found to have mild to moderate background DR increased during the first round of screening and remained at a higher level (162 pre-screening, 196, 194, and 212) and as a percentage of those referred, this was 19%, 20.5%, 19.9%, and 20.2% (2 for trend 0.22, p = 0.64).
The number of patients referred who were found to have referable retinopathy showed a rise during the first two rounds of screening. There was then a slight reduction but not to the level of the year before the screening programme commenced (172 pre-screening, 247, 268, 236) and as a percentage of those referred, this was 20.2%, 25.9%, 27.5%, and 22.5% (2 for trend 1.22, p = 0.27).
The source of referral where referable DR was found in the eye clinic was GP (38, 19, 16, 21), optometrist (via GP) (60, 49, 30, 34), physician (35, 49, 28, 24) and eye screening service (GDESS) (0 pre-screening, 71, 155, 115).
The sources of referral where "an opinion about DR seen" was requested or where any diabetic retinopathy was found in the eye clinic are summarised in table 6.
Table 6 The change in workload during the first round of screening (middle 2 years)
The source of referral for "an opinion about DR seen" or where any diabetic retinopathy was found in the eye clinic decreased from general practitioners but not from optometrists or physicians.
The eye screening service (GDESS) showed a rise in referrals for "an opinion about DR seen" from 103 in the first year of screening to 237 in the second and decreased to 190 in the first year of the second round of screening. As a percentage of those referred for "an opinion about DR seen" the rise was from 30.9% to 65.3% and decreased in the third year to 51.6% (2 for trend 27.6, p<0.001). Similarly, The eye screening service (GDESS) showed a rise in numbers where diabetic retinopathy was found in the eye clinic from 104 in the first year to 233 in the second and decreased to 184 in the first year of the second round of screening. As a percentage of those referred where diabetic retinopathy was found in the eye clinic this was 25.7% rising to 54.2% and falling in the third year to 45.2% (2 for trend 31.7, p<0.001). For patients referred from screening with maculopathy or pre-proliferative DR (92% of referable DR), there was a lag time of approximately 13 weeks from screening to being seen in the eye clinic.
The workload and source of referral of patients with diabetes referred for cataract opinion and reduced vision and the numbers listed for cataract surgery
The total number of referrals specifically for cataract opinion was 66, 59, 77, 98. The source of these was GP (19, 14, 13, 19), optometrist (via GP) (40, 35, 39, 62), physician (5, 1, 2, 2), eye screening service (GDESS) (0 pre-screening, 4, 17, 12), and other sources (2, 5, 6, 3).
The total number of people with diabetes listed for cataract surgery at first visit remained fairly stable (158 pre-screening, 143, 145, and 168) and as a percentage of those referred, this was 18.5%, 15%, 14.9%, and 16% (2 for trend 1.73, p = 0.19).
The workload and source of referral of patients with diabetes referred specifically for a glaucoma opinion and the numbers of patients with diabetes actually diagnosed with glaucoma
The total number of referrals specifically for glaucoma opinion was 57, 62, 61, and 68. The source of these was GP (2, 8, 3, 2), optometrist (via GP) (51, 42, 25, 52), physician (3, 3, 2, 3), eye screening service (GDESS) (0, 8, 28, 6), and other sources (1, 1, 3, 5).
The number of patients with diabetes actually diagnosed with glaucoma was 55, 54, 55, and 62. The source of referral of those actually diagnosed with glaucoma (even if the referral reason did not suggest this) was GP (11, 12, 8, 12), optometrist (via GP) (42, 32, 19, 34), physician (2, 6, 5, 5), eye screening service (GDESS) (0 pre-screening, 3, 17, 7), and other sources (0, 1, 6, 4).
Over the 4 years the false positive referrals for glaucoma were five out of 15 (33%) for GPs, 88 out of 170 (52%) for optometrists, six out of 11 (55%) for physicians, and 26 out of 42 (62%) for the 3 years of the diabetic eye screening service.
Non-attenders for screening presenting with retinopathy
During the year 2000–1, 55 people who declined their invitation to attend the first round of screening presented in clinic with referable DR.
Number of new patients referred for laser treatment (and type of laser treatment) and numbers of laser treatment sessions required over the following year
The total number of new patients referred for laser treatment over the 4 years was 77 pre-screening, 124, 111, and 63 and as a percentage of those referred the percentage was 9%, 13%, 11.4%, and 6%.
Numbers of new patients referred for laser treatment for maculopathy over the 4 years were 61, 94, 81, and 36. The total number of laser treatment sessions required for these patients was 89, 136, 119, and 62.
Numbers of new patients referred for laser treatment for pre-proliferative/proliferative over the 4 years were 16, 30, 30, and 27. The total number of laser treatment sessions required for these patients was 82, 146, 146, and 137.
The total number of laser treatment sessions for sight threatening diabetic retinopathy for new patients over the 4 years was 171, 282, 265, and 199.
The number of people with diabetes referred to the eye department continued to rise over the 4 years of the study, as did the number of referrals for an opinion about diabetic retinopathy. The figures have been measured against a background of a rise of approximately 1400 per annum in the number of people with diabetes in the county. The number of patients referred for laser treatment rose from a pre-screening figure of 77 to 124 and 111 during the first round of screening and returned to 63 in the fourth year. A 2 for trend test was not performed as there was no expectation of an increasing trend over the 4 years of the study.
The percentage of referrals found to have no DR decreased from 56.7% to 48.7% over the 4 years and those referred with reduced vision declined from 33.8% to 21% during this period. Of those who were referred with reduced vision, the percentage of these who were discharged at first clinic visit decreased from 20.1% to 8.6% and the percentage of these patients listed for laser for sight threatening DR decreased from 9% to 3.2%. These trends suggest to the authors that the quality of referral criteria improved during this time.
The figures consistently show a doubling of the numbers referred by the eye screening service in each category from the first year of the first round of screening to the second year of the first round of screening, despite the same numbers of patients being screened. Screening in the first round commenced with one camera in October 1998 and a second camera was introduced in January 1999 with the rate of screening increasing during the first year for both screeners as they became more experienced. Hence, a proportionately greater number were screened in the latter 3 months of the first year of screening. For this group there was a lag time of 13 weeks from screening to being seen in the eye clinic, which explains the increased numbers appearing in the eye clinic in the second year of screening.
Following a survey in 2000 of screening programmes in the South-West Region, Freudenstein and Verne published an editorial in the BMJ.9 They found a significant difference in rates of referral from screening programmes for cataract and glaucoma depending on criteria used for referral. In this study, the eye screening service did not have a significant impact on referrals for cataract or glaucoma.
This workload study has shown continuing referrals from other sources after the introduction of the screening programme. Referral for eye conditions other than retinopathy accounts for a proportion of these. Although referrals for an opinion about diabetic retinopathy decreased from GPs there was no decrease from optometrists who have always been a greater source of referral or from physicians who refer a smaller number. Our current screening programme is only offered every 2 years and this may have been an additional factor in the continued numbers of referrals from optometrists and in the ophthalmology workload. It is hoped that, with an annual programme, a significant number of annual review patients might be discharged to an annual screening photographic review.
The current study supports the reports from the literature that show a steady rise in incidence of diabetes10–13 and related complications14 in all age groups, both in the United Kingdom and worldwide. This is mostly because of a rising incidence of type 2 diabetes associated with an epidemic of obesity.15 A change in the WHO definition of diabetes16 and improved identification and reporting of diabetes have made a much smaller contribution.
This study has shown a progressive rise in workload related to the epidemic of diabetes, which will increase the pressure on ophthalmology services.
We thank Mark Histed for his assistance with examining case records and Solon Asteriadis and Denize Atan for their work on a laser treatment audit in Gloucestershire, which improved the quality of the data presented on laser treatment.
Clifford R, PCCAG. An audit on the care of adult diabetic patients in Gloucestershire. Cheltenham: Gloucestershire Primary Care Clinical Audit Group (PCCAG), 1996, 1–54 (all).
National Screening Committee. Preservation of sight in diabetes: a risk reduction programme: www.diabetic-retinopathy.screeening.nhs.uk 2000.
Gillow JT, Gray JA. The National Screening Committee review of diabetic retinopathy screening. Eye 2001;15 (Pt 1) :1–2.
Garvican L, Clowes J, Gillow T. Preservation of sight in diabetes: developing a national risk reduction programme. Diabet Med 2000;17:627–34.
Scanlon PH, Malhotra R, Thomas G, et al. The effectiveness of screening for diabetic retinopathy by digital imaging photography and technician ophthalmoscopy. Diabet Med 2003;20:467–74.
Scanlon PH, Malhotra R, Greenwood RH, et al. Comparison of two reference standards in validating two field mydriatic digital photography as a method of screening for diabetic retinopathy. Br J Ophthalmol 2003;87:1258–63.
Early Treatment Diabetic Retinopathy Study Research Group. Photocoagulation for diabetic macular edema. Early Treatment Diabetic Retinopathy Study report number 1. Arch Ophthalmol 1985;103:1796–806.
Early Treatment Diabetic Retinopathy Study Research Group. Treatment techniques and clinical guidelines for photocoagulation of diabetic macular edema. Early Treatment Diabetic Retinopathy Study Report Number 2. Ophthalmology 1987;94:761–74.
Freudenstein U, Verne J. A national screening programme for diabetic retinopathy. Needs to learn the lessons of existing screening programmes. BMJ 2001;323:4–5.
Amos AF, McCarty DJ, Zimmet P. The rising global burden of diabetes and its complications: estimates and projections to the year 2010. Diabet Med 1997;14 (Suppl 5) :S1–85.
Mokdad AH, Ford ES, Bowman BA, et al. Diabetes trends in the US: 1990–1998. Diabetes Care 2000;23:1278–83.
Mokdad AH, Ford ES, Bowman BA, et al. The continuing increase of diabetes in the US. Diabetes Care 2001;24:412.
Ehtisham S, Barrett TG, Shaw NJ. Type 2 diabetes mellitus in UK children—an emerging problem. Diabet Med 2000;17:867–71.
Fitzsimons B, Wilton L, Lamont T, et al. The Audit Commission review of diabetes services in England and Wales, 1998–2001. Diabet Med 2002;19 (Suppl 4) :73–8.
Sorensen TI. The changing lifestyle in the world. Body weight and what else? Diabetes Care 2000;23 (Suppl 2) :B1–4.
World Health Organization. Definition, diagnosis and classification of diabetes mellitus and its complications. Geneva: WHO, 1999:1–66.
1 Ophthalmic Research Center, Shaheed Beheshti University of Medical Sciences, Tehran, Iran
2 Department of Ophthalmology, Kerman University of Medical Sciences, Kerman, Iran
3 Department of Biostatics and Epidemiology, School of Health, Tehran University of Medical Sciences, Tehran, Iran
MD, Labbafinejad Medical Center, Pasdaran Ave, Boostan 9 Street, Tehran 16666, Iran; firstname.lastname@example.org
Accepted for publication 1 February 2005
Aims: To evaluate the effect of tranexamic acid on early postvitrectomy haemorrhage in diabetic patients.
Methods: In a clinical trial, 62 diabetic patients scheduled for vitrectomy were randomly assigned to two groups. The treatment group (32 eyes) received two doses of tranexamic acid (10 mg/kg) shortly before and after the operation intravenously, continued orally for 4 days (20 mg/kg/8 hours). The control group (30 eyes) received no medication. Both media clarity and visual acuity were compared during 4 weeks.
Results: Four weeks after surgery visual acuity was low (1 metre counting fingers) in 21.4%, moderate (>1 metre counting fingers but<20/200) in 14.3%, and good (20/200) in 64.3% of the treated group. Corresponding figures in the control group were 26.1%, 26.1%, and 47.8%, respectively. These differences were of no statistical significance. The ratio of mild to severe vitreous haemorrhage during the first 4 days and after 4 weeks was 79% to 21% and 82% to 18% in the treatment group and 76.7% to 23.3% and 78.3% to 21.7% in the control group respectively, which showed no statistically significant difference.
Conclusion: Tranexamic acid, with the method of administration in this study, had no effect on reducing early postvitrectomy haemorrhage in diabetic patients.
Abbreviations: EACA, -aminocaproic acid; IOP, intraocular pressure; PT, prothrombin time; PTT, partial thromboplastin time
Keywords: tranexamic acid; antifibrinolytic agent; diabetic retinopathy; vitreous haemorrhage; vitrectomy
Early vitreous haemorrhage, within a week after vitrectomy, is a common complication in diabetic patients, with an incidence of 29–75%.1 It may cause severe visual impairment (especially important in monocular patients), interfere with examination and laser therapy, induce ghost cell glaucoma,2 increase need for vitrectomy,3 and stimulate the growth of epiretinal membranes and fibrous tissue.
Antifibrinolytic drugs, like tranexamic acid and EACA (-aminocaproic acid), inhibit clot lysis through interference with plasmin action.1 The haemostatic effect of EACA has been proved in different types of operations such as prostatectomy, dental, cardiac, and orthopaedic operations.4,5 In addition, the role of these drugs in decreasing rebleeding in hyphaema is clear.2
These two medications have been used after vitrectomy for diabetic patients in two separate studies and good results were observed in one of them.6,7 The present study evaluated the effect of tranexamic acid on early vitreous haemorrhage after vitrectomy in diabetic patients with proliferative retinopathy.
MATERIALS AND METHODS
This randomised clinical trial was conducted on diabetic patients scheduled for vitrectomy for advanced retinopathy including non-clearing vitreous haemorrhage, tractional retinal detachment, and progressive fibrovascular proliferation. All registered patients were fully informed of the side effects of tranexamic acid.
After complete history taking and ophthalmic examination, laboratory tests were performed for each enrolled patient that included blood cell and platelet counts, serum creatinine and fasting blood sugar level, prothrombin and partial thromboplastin time. Exclusion criteria consisted of history of cataract surgery or vitrectomy, haemodialysis, pregnancy, history of disorders such as deep vein thrombosis, myocardial infarction, or cerebrovascular accidents.
Before surgery, patients were randomly assigned to receive tranexamic acid or no medication. In the case group, the first drug administration (10 mg/kg intravenously) was performed just before transferring the patient to the operating room. The second tranexamic acid intravenous injection was performed after surgery with the same dose. From the day after vitrectomy, the medication was continued orally in the form of 250 mg capsules (20 mg/kg every 8 hours) for 4 days during hospitalisation. The drug dosage was adjusted according to serum creatinine level (table 1).
Table 1 Tranexamic acid dosage adjustment according to serum creatinine level
Standard three port vitrectomy was performed for all patients under local or general anaesthesia. Only Ringer’s solution was used to avoid the anticoagulative effect of citric acid present in balanced salt solution. Additional procedures like endolaser, membrane dissection, etc, were performed if needed. Intraoperative bleeding was controlled by either raising intraocular pressure or endodiathermy. If use of an internal tamponade such as air, gas, or silicone oil was mandatory, that case would be excluded from the study. Blood pressure was monitored during hospitalisation including operation time. Eye pressure was checked at the end of the surgery with the Schiotz device. The surgeons were masked to the randomisation.
All patients were hospitalised after surgery and for 4 days, all had bed rest in a semi-sitting position. Complete examination was performed during the admission period, and 1 week and 4 weeks after surgery. The degree of intravitreal haemorrhage was scaled according to the diabetic retinopathy vitrectomy study grading system (table 2).8 For statistical analysis, eyes with grade 0 or 1 haemorrhage were classified as mild, and cases with grade 2 or 3 as severe haemorrhage.
Table 2 Intravitreal haemorrhage scaling according to the diabetic retinopathy vitrectomy study
Visual acuity was also graded into low (1 metre counting fingers), moderate (more than 1 metre counting fingers but less than 20/200), and good (20/200). The amount of vitreous haemorrhage and visual acuity were compared between the two groups at three different times: hospitalisation period (according to the mean of 4 days), after 1 week and 4 weeks. All the examiners were masked to patients’ allocation into each group.
This clinical trial was approved by the review board/ethics committee of the ophthalmic research centre.
Taking into account the reported 70% incidence of postvitrectomy diabetic haemorrhage and reducing this to 35% as our main goal, a minimal sample size of 32 eyes in each group was required to detect a significant difference at the two sided 5% level with a study power of 80%.
Independent samples t test was employed for evaluating quantitative variables. The 2 test was used to compare the severity of preoperative and postoperative vitreous haemorrhage and visual acuity levels. Statistical level of significance was preset at 0.05. Data were analysed using SPSS 9.
Sixty two eyes of 62 patients (32 cases and 30 controls) were included in the study. Some relevant individual and clinical characteristics of each group are presented in table 3. It may be noted that both groups were similar in all factors except for age. Mean age of the treatment group was 6 years older than the controls (p = 0.04).
Table 3 Individual and clinical characteristics of 62 patients (%)
Means of probable confounding factors are shown in table 4. All of the mentioned parameters displayed no significant difference between the groups except mean intraocular pressure at the first week.
Table 4 Confounding factors in the study groups (SD)
The presence and severity of vasoproliferative or fibroproliferative changes at the time of surgery were also similar in both groups (data not shown).
Visual acuity levels, evaluated at three different times, are presented for each group separately in table 5. There was no significant statistical difference at any of the examination intervals.
Table 5 Visual acuity levels at three different times in each group (%)
Overall, visual acuity during the first 4 days was low in 22 eyes (35.5%), moderate in 28 eyes (42.5%), and good in 12 eyes (19.4%). Corresponding figures with correction after 4 weeks were 12 eyes (23.5%), 10 eyes (19.6%), and 29 eyes (56.9%) respectively.
Severity of vitreous haemorrhage is presented in table 6. The prevalence of severe vitreous haemorrhage in the treated and observation groups was 6/32 versus 7/30, 3/27 versus 4/27, and 4/27 versus 5/23 after 4 days, 1 week, and 4 weeks, respectively. No statistically significant difference was observed between the two groups.
Table 6 Vitreous haemorrhage severity compared between groups at three different intervals (%)
Overall, the presence and severity of retinal neovascularisation were similar in eyes with either mild or severe postoperative haemorrhage during the first 4 days. Although eyes with mild postvitrectomy haemorrhage had more severe fibrovascular changes, this difference was not significant (p = 0.052). The amount of haemorrhage during the first 4 days in both groups had no correlation with intraoperative bleeding or performing membrane dissection.
Early postoperative vitreous haemorrhage was correlated with the presence of fresh preoperative vitreous haemorrhage. Of eyes with mild postoperative haemorrhage 61.2% (30 of 49) had fresh preoperative haemorrhage; however, 92.3% (12 of 13 eyes) of eyes with severe postoperative haemorrhage had such a finding (p = 0.03).
No evidence of thrombotic or thromboembolic complications was seen in the treated group. Among the cases, three complained of nausea, two of vomiting, and one of diarrhoea. One of the controls also reported nausea.
This clinical trial showed that tranexamic acid did not reduce early vitreous haemorrhage or improve visual acuity after vitrectomy in patients with proliferative diabetic retinopathy.
The source of early postvitrectomy diabetic haemorrhage is retained blood in the vitreous cavity or rebleeding from cut edges of fibrovascular tissue during surgery. Some methods used to decrease this complication are: (1) adding thrombin to the irrigation fluid during surgery, which decreased bleeding time during the operation9,11; (2) use of sodium hyaluronate (Healon) intravitreally at the end of the surgery to mechanically prevent dispersion of haemorrhage12,13; (3) injection of silicone oil to prohibit the spread of coagulative elements14; (4) fluid exchange with air or gas to produce tamponade, however one study showed an opposite result.15
The main cause of early postoperative haemorrhage is lysis of the clot, which is usually formed at the edges of cut vessels or dissected fibrovascular tissue during surgery. Therefore, antifibrinolytic drugs that inhibit clot lysis might decrease rebleeding. In the present study, we preferred to use tranexamic acid rather than EACA, because of its lower dosage and less side effects.
Both groups in our study were matched according to basic and confounding factors except in two aspects, age, and IOP at 1 week. On average, treated patients were 6 years older than controls. Although this difference was statistically significant, it could not have been an important confounding factor clinically. Contrary to the present study, average age in the treated group was 10 years lower in the study performed by Laatikainen et al.6 Since in both studies tranexamic acid had no beneficial effect, these two papers could complement each other regarding the confounding factor of age.
We believe that the 3 mm Hg difference in the average first week IOP has no clinical importance either. In addition, IOP before and after the first week, showed no meaningful difference. There is also a low probability that increased IOP in the treated group was caused by tranexamic acid; as far as we know, such a side effect has not previously been reported.
To the best of our knowledge, antifibrinolytic agents had been evaluated for postoperative vitreous haemorrhage in only two studies.6,7 In 1987, Laatikainen et al performed a similar study with tranexamic acid on fewer patients (31 cases). They administered the first dose at the end of the operation and the second one 12 hours later. However, in our study, cases received the first dose before going to the operating room and the second on returning. Although rates of vitreous haemorrhage in their study were 44% in treated patients versus 60% in controls, the difference was not significant.6 No important side effect was seen in either of these studies with tranexamic acid, and they both concluded that this drug had no effect on early postvitrectomy vitreous haemorrhage and final visual acuity.
A beneficial effect of EACA on early vitreous haemorrhage was reported in 1985 in study by de Bustros et al on 96 patients.7 The potency of this drug is 10 times less than tranexamic acid, so it must be administered in much higher doses. In spite of some side effects (nausea, vomiting, and diarrhoea in 21% and postural hypotension in 6%), there was no need for discontinuation of the drug in any of their cases. The study concluded that EACA reduced postvitrectomy haemorrhage in the first 4 days (p = 0.002); however, no significant difference was noted in the second and sixth weeks. The authors believed that disappearance of intergroup difference was in part caused by rebleeding in treated patients after EACA was discontinued, and spontaneous clearing of haemorrhage in untreated eyes. They concluded that although EACA could not decrease the recurrence of bleeding, its ability to reduce vitreous haemorrhage during the early postoperative period permits better evaluation of the fundus during this crucial time when media clarity is required for detection of retinal detachment or performing laser therapy.
Since the mechanism of these drugs is inhibition of clot lysis, they would not have any effect unless bleeding has already happened. As a result, it is logical to expect their effect only in cases with bleeding during vitrectomy, but the statistical analysis was not able to show any benefit, even in such cases. In Laatikainen’s study also, even though bleeding and need for diathermy were more common in the control group, the incidence of vitreous haemorrhage was greater in the first week in the treated patients.6
Among all probable confounding factors, only fresh vitreous haemorrhage before surgery affected the first 4 day bleeding. Such an effect is an expected finding, because fresh vitreous haemorrhage implies activity of retinopathy with more predisposition to bleeding.
When evaluating the effect of tranexamic acid on early postvitrectomy haemorrhage in diabetic patients some factors must be kept in mind:
The role of improper clot formation as a result of coagulopathy in diabetic patients who might already have a renal problem.
Inadequate drug dose or intraocular drug concentration. Although drug concentration is low in the vitreous cavity at the end of surgery, it seems that the fibrinolysis is more active at the vessel side than the vitreous side of a clot. Therefore, blood drug concentration must be more important.
Short drug administration period. Most bleedings occur after the hospitalisation period (the first 4 or 5 days), at which time the drug is discontinued and has no more effect. In de Bustros’s study, the benefit of drug was shown only during its administration and repeat bleeding occurred after its discontinuation. Therefore, they recommended a longer period of drug administration.7
Some instances of early vitreous haemorrhage are merely the result of dispersion of pre-existing blood in the vitreous cavity and not because of rebleeding after clot lysis. We should remind ourselves that all of our patients were phakic and complete clearing of the peripheral vitreous cavity from blood was impossible.
Because of dysfunction of coagulative and fibrinolytic systems in diabetics (especially in cases with renal failure), the effect of tranexamic acid may be different in these patients. Moreover, drug effect may differ specifically in the eye compared to other organs.
Although attempts were made to control for confounding for both groups, their effects could not be ignored completely. Two of these factors were blood pressure rise after surgery both with general or local anaesthesia (noting that most of our patients were hypertensive) and coagulative disorders. As a result, administration of a drug with only relative inhibitory effect on clot lysis may not have had the ability to compete with such factors, with greater potential to induce bleeding.
In conclusion, we would not recommend tranexamic acid for decreasing postvitrectomy diabetic vitreous haemorrhage. Further studies with other drugs may be suggested.
We gratefully acknowledge the cooperation of Dr Arash Anisian in the final preparation of the paper.
Mieler W, Wolf M. Management of postvitrectomy diabetic vitreous hemorrhage. In: Lewis H, Rayan ST, eds. Medical and surgical retina . St Louis: Mosby, 1994:29.
Rahmani B, Jahadi HR. Comparison of tranexamic acid and prednisolone in the treatment of traumatic hyphema: a randomized clinical trial. Ophthalmology 1999;106:375–9.
Sanislo SR, Blumenkranz MS. Diabetic vitrectomy. In: Duane’s clinical ophthalmology . Philadelphia: Lippincott Williams & Wilkins Publishers, 2004:57.
Ekback G, Axelsson K, Ryttberg L. et al. Tranexamic acid reduces blood loss in total hip replacement surgery. Anesth Analg 2000;91:1124–30.
Hiippala ST, Strid LJ, Wennerstrand MI, et al. Tranexamic acid radically decreases blood loss and transfusion associated with total knee arthroplasty. Anesth Analg 1997;84:839–44.
Laatikainen L, Summanen P, Immonen I. Effect of tranexamic acid on postvitrectomy haemorrhage in diabetic patients. Int Ophthalmol 1987;10:153–5.
De Bustros S, Glaser BM, Michels RG, et al. Effect of epsilon-aminocaproic acid on postvitrectomy hemorrhage. Arch Ophthalmol 1985;103:219–21.
Diabetic Retinopathy Vitrectomy Study (DRVS). Two-year course of visual acuity in severe proliferative diabetic retinopathy with conventional management. Report No 1. Ophthalmology 1985;92:492–502.
Verdoorn C, Hendrikse F. Intraocular human thrombin infusion in diabetic vitrectomies. Ophthalmic Surg 1989;20:278–9.
Kim SH, Cho YS, Choi YJ. Intraocular hemocoagulase in human vitrectomy. Jpn J Ophthalmol 1994;38:49–55.
Thompson JT, Glaser BM, Michels RG, et al. The use of intravitreal thrombin to control hemorrhage during vitrectomy. Ophthalmology 1986;93:279–82.
Packer AJ, McCuen BW 2nd, Hatton WL. et al. Procoagulant effects of intraocular sodium hyaluronate (Healon) after phakic diabetic vitrectomy: a prospective randomized study, Ophthalmology 1989;96:1491–4.
Folk JC, Packer AJ, Weingeist TA, et al. Sodium hyaluronate (Healon) in closed vitrectomy. Ophthalmic Surg 1986;17:299–306.
Chairs S. Vitreous microsurgery. 2nd ed. Baltimore: Williams and Wilkins, 1987.
Brine C, Joondeph BC, Blankenship GW. Haemostatic effect of air versus fluid in diabetic vitrectomy. Ophthalmology 1989;96:1710–16.
1 Department of Ophthalmology, University of Udine, Udine, Italy
2 Department of Medical and Morphological Research, University of Udine, Udine, Italy
MD, Department of Ophthalmology, University of Udine, P le S Maria della Misericordia, 33100 Udine, Italy; email@example.com
Accepted for publication 1 October 2004
Aim: To compare the effectiveness of "light" versus "classic" laser photocoagulation in diabetic patients with clinically significant macular oedema (CSMO).
Methods: A prospective randomised pilot clinical trial in which 29 eyes of 24 diabetic patients with mild to moderate non-proliferative diabetic retinopathy (NPDR) and CSMO were randomised to either "classic" or "light" Nd:YAG 532 nm (frequency doubled) green laser. "Light" laser treatment differed from conventional ("classic") photocoagulation in that the energy employed was the lowest capable to produce barely visible burns at the level of the retinal pigment epithelium. Primary outcome measure was the change in foveal retinal thickness as measured by optical coherence tomography (OCT); secondary outcomes were the reduction/elimination of macular oedema on contact lens biomicroscopy and fluorescein angiography, change in visual acuity, contrast sensitivity, and mean deviation in the central 10° visual field. Examiners were masked to patients’ treatment.
Results: 14 eyes were assigned to "classic" and 15 were assigned to "light" laser treatment. At 12 months, seven (50%) of 14 eyes treated with "classic" and six (43%) of 14 eyes treated with "light" laser had a decrease of foveal retinal thickness on OCT (p = 0.79). A comparison of reduction/elimination of oedema, visual improvement, visual loss, change in contrast sensitivity, and mean deviation in the central 10° showed no statistical difference between the groups at 12 months (p>0.05 for all groups).
Conclusions: This study suggests that "light" photocoagulation for CSMO may be as effective as "classic" laser treatment, thus supporting the rationale for a larger equivalence trial.
Abbreviations: CSMO, clinically significant macular oedema; ETDRS, Early Treatment Diabetic Retinopathy Study; FA, fluorescein angiography; FTH, foveal thickness; MD, mean deviation; NPDR, non-proliferative diabetic retinopathy; OCT, optical coherence tomography; PEDF, pigment epithelium derived factor; RPE, retinal pigment epithelium; VA, visual acuity
Keywords: clinically significant macular oedema; laser treatment
Macular oedema is the most common cause of visual loss in patients with diabetic retinopathy and its prevalence in the diabetic population has increased approximately fivefold over the past 10 years, from 0.4% to 2.1%.1
The Early Treatment Diabetic Retinopathy Study (ETDRS) demonstrated the benefit of a specific strategy of laser treatment on reducing the risk of moderate visual loss in eyes with clinically significant macular oedema (CSMO).2 Treatment with a grid pattern in the parafoveal region up to and including the edge of the foveal avascular zone in case of diffuse thickening has also been advocated in non-randomised series.3,4 However, the beneficial effect of laser photocoagulation is associated with severe destruction of retinal photoreceptors and considerable side effects, such as post-treatment atrophic scarring causing paracentral dense scotomas, generalised loss of the central 10° threshold sensitivity, choroidal neovascularisation, and subfoveal fibrosis.5,6,7,8,9,10,11 Moreover, a significant lateral spread of retinal pigment epithelium (RPE) atrophy may occur over time thus impairing visual function even if visual acuity (VA) is relatively preserved.8,12 These potential complications may also negatively affect the decision to apply additional laser treatments whenever required.
Recently, some authors have shown that barely perceivable very light threshold treatment and non-visible end point subthreshold treatment promoted resolution of CSMO in small uncontrolled series.13–15 This therapeutic effect has been correlated with the selective treatment of the RPE and subsequent restoration of its barrier function and production of growth factors.16–20 Among these, the pigment epithelium derived factor (PEDF) has been shown to have a strong antiangiogenic activity and to be upregulated after photocoagulation.21,22 Based on these data, conventional ETDRS-like laser burns may not be necessary in the treatment of CSMO.
When treating or re-treating eyes with CSMO and leaking abnormalities close to the foveal centre or diffuse macular thickening, very light laser burns may have the theoretic advantage of producing less invalidating paracentral scotomata or significant reduction in the central sensitivity. We present here results of a prospective, randomised pilot trial comparing the effectiveness of barely visible or "light" laser treatment versus conventional or "classic" laser photocoagulation in CSMO.
Patients were enrolled from 1 September 2001 to 30 November 2002. Eligibility criteria included a diagnosis of either type 1 or type 2 diabetes mellitus and non-proliferative diabetic retinopathy with CSMO documented by slit lamp contact lens biomicroscopy, as defined by the ETDRS,3 and confirmed by optical coherence tomography (OCT). Increased macular thickness was established if the foveal thickness (FTH), defined as the mean thickness of the central 1 mm diameter disc of the retinal map, exceeded two SD the mean normal value (that is, >210 μm).23,24 Other inclusion criteria were haemoglobin A1c equal or less than 10%, diastolic blood pressure less than 90 mm Hg, and visual acuity of at least 20/200 on the ETDRS chart. Patients with previous laser treatment, proliferative diabetic retinopathy, history of retinal detachment, glaucoma or any other clinically relevant ocular disease, cataract extraction or lens implantation within the past 12 months, or significant media opacities were excluded. The study was conducted according to the tenets of the Declaration of Helsinki and all subjects gave informed consent after the intent of the study had been explained.
At the baseline examination, an independent examiner refracted both eyes, measured distance visual acuity and contrast threshold using ETDRS visual acuity chart at 4 metres and Pelli-Robson charts at 1 metre, respectively. Best corrected visual acuity was scored based on the total number of correct letters identified at 4 metres plus 30.25 The first examination also included automated static threshold perimetry, stereo fundus photography, fluorescein angiography (FA), and OCT. If FA disclosed a macular area of non-perfusion at least twice as large as the foveal avascular zone, patients were excluded from the study. Automated static threshold perimetry was performed using the Humphrey field analyser (Humphrey-Allergan Medical Instruments, Irvine, CA, USA) 10-2 program. The mean deviation (MD) over the central 10° was recorded. OCT scans were obtained using the OCT 2000 scanner (Zeiss Humphrey Instruments, Dublin, CA, USA) with the A5 version software. OCT scanning was performed by selecting the "radial lines" scan pattern, which acquires six linear scans 6 mm long centred on the fovea at equally spaced angular orientation. Acceptable scans were automatically analysed by the OCT computer software and retinal thickness maps were generated. Images were judged to be "acceptable" on the basis of the following acceptance criteria: good demarcation of the vitreoretinal and chorioretinal interface allowing for a correct identification of the two interfaces by the software and absence of artefacts caused by eye motion or unstable fixation. The FTH was recorded for analysis from the retinal thickness maps.
Patients who were believed to satisfy all eligibility criteria were assigned randomly at enrolment to receive "classic" or "light" laser treatment. If a patient had both eyes simultaneously eligible for the study, the right eye was assigned randomly to "classic" or "light" laser treatment and the left eye received the opposite assignment. All the visits and tests were performed by the same masked examiners (AP, MDB).
Photocoagulation was performed by the same surgeon (FB) using a Nd:YAG 532 nm (frequency doubled) green laser. "Light" laser treatment differed from "classic" in that the energy employed was the lowest capable to produce barely visible burns at the level of the retinal pigment epithelium. For primary treatment with "classic" photocoagulation, 47 (SD 26) spots were applied with power ranging between 100 mW and 250 mW (median 140 mW), and for primary treatment with "light" photocoagulation, 92 (SD 36) spots were applied with power ranging between 50 mW and 100 mW (median 50 mW). Follow up visits were performed every 3 months after treatment and included a protocol refraction, best corrected visual acuity and contrast threshold measurement, fundus biomicroscopy, FA, and OCT. Automatic static threshold perimetry of the central 10° was performed at the 3 month and 12 month follow up visit. Supplemental treatment with the same protocol of the primary treatment was considered at each visit in those eyes with treatable lesions on FA and either increased foveal thickening on fundus examination or persistent foveal thickening and decreased vision. For supplemental treatment, 45 (SD 10) spots were applied in "classic" group, and 106 (SD 44) spots were applied in the "light" group. The study included all patients who completed the 12 month follow up examination.
The primary efficacy outcome was the proportion of patients with significant decrease in FTH on OCT retina thickness maps. A significant change in FTH was defined as a change greater than 10%, which is slightly higher than the reproducibility of the instrument (that is, reproducibility of plus or minus 6%)26 in an attempt to exclude changes caused by spontaneous inter-visit variability. Secondary efficacy outcomes included the proportion of patients with reduction elimination of CSMO on biomicroscopy and fluorescein leakage on FA compared with baseline examination at 3, 6, and 12 months after study entry, the proportion of eyes that experienced a visual gain or loss of five or more letters (approximately one line) on the ETDRS chart, mean changes in visual acuity, contrast threshold, MD of the central 10° sensitivity, and number of local losses greater than 5 dB at each test point of the central 10°, suggestive of post-treatment scotomata.
Comparison between groups for categorical variables was performed with 2 test; when assumptions for 2 test were not verified, Fisher’s exact test was used. The normality of continuous variables distributions was checked by the Shapiro-Wilk test. Comparisons between groups for continuous variables were performed with t test or the Mann-Whitney U test depending on Shapiro-Wilk test results. Comparisons within groups for continuous variables on differences from baseline were performed with paired t test or Wilcoxon test also depending on Shapiro-Wilk test results. All analyses were performed using the statistical software SPSS version 11.1.
A total of 29 eyes of 24 patients were assigned randomly to "classic" or "light" laser photocoagulation; 14 eyes were assigned to "classic" and 15 to "light" laser treatment. Five patients were treated bilaterally and 19 patients were treated unilaterally. OCT data were not available for one eye in the "light" group; therefore it was excluded from OCT data analysis. Baseline variables for each group are given in table 1. Systemic, functional, and morphological characteristics were similar and not statistically different between the two groups. Table 2 summarises the results at 12 months for individual patients. Table 3 compares the outcome measurements at 3, 6, and 12 months between the two study groups. No statistically significant differences could be found for any of the outcome measurements except for the proportions of eyes that experienced a visual gain of five or more letters at 6 months, which was larger in the "light" group. The average FTH in both groups did not change significantly from baseline at any follow up visit. The initial median visual acuity of 20/32 in both groups remained unchanged at 12 months. At the 3, 6, and 12 month follow up, five (36%), seven (50%), and seven (50%) of 14 eyes treated with "classic" and four (29%), six (43%), and six (43%) of 14 eyes treated with "light" laser had a decrease of FTH on OCT (p = 1, 0.50 and 0.79, respectively, Fisher’s exact test). At 12 month follow up, five (36%) of 14 eyes treated with "classic" and five (33%) of 15 eyes treated with "light" laser showed one line or more improvement in the visual acuity and two (14%) in the "classic" group and two (13%) eyes in the "light" group showed one line or more decrease in the visual acuity (p = 1, Fisher’s exact test). The average MD decrease over the central 10° at 12 months after "classic" and "light" laser treatment was 0.04 (SD 1.39) dB and 0.03 (SD 1.84) dB, respectively (p = 0.99, independent samples test). The number of local losses greater than 5 dB at each test point of the central 10°, suggestive of post-treatment paracentral scotomata, was 5.00 and 4.78 in the "classic" and "light" laser treatment group, respectively, at 12 months (p = 0.40, Mann-Whitney U test). One and three eyes were retreated at 3 months and four and three eyes at 6 months in the "classic" and "light" treatment group, respectively (p = 0.60 and 0.68, respectively, Fisher’s exact test).
Table 1 Baseline characteristics*
Table 2 Details of the patients who underwent "classic" and "light" treatment
Table 3 Outcome measurements by time and treatment group*
A 55 year old man presented with a 8 year history of type 2 diabetes mellitus and non-proliferative diabetic retinopathy in both eyes. On examination, his visual acuity was 20/50 in the right eye and 20/64 in the left eye. Fundus biomicroscopy revealed diffuse retinal thickening in both eyes (figs 1A, 2A). FA showed extensive perifoveal leaking microaneurysms bilaterally (figs 1B, 2B). OCT retina thickness maps were similar in both eyes and displayed severe diffuse macular thickening (figs 1C, 2C). Two sessions of "classic" and "light" laser treatment were applied to the right and left eye, respectively, at study entry and after 6 months. The degree and extent of retinal and RPE whitening produced by laser burns, which were greater in the right eye, were documented immediately after the first laser session by red-free photographs (figs 1D, 2D). One year later his visual acuity had decreased to 20/80 in the right eye and improved to 20/50 in the left eye. The fundus examination disclosed marked decrease in foveal thickness in both eyes and multiple laser scars, which where more prominent and partly hyperpigmented in the right eye (fig 1E). FA disclosed a significant reduction of late leakage in both eyes (figs 1F, 2F). The OCT retinal thickness map documented a marked decrease of foveal thickening bilaterally (figs 1G, 2G). The central 10° MD decreased from –5.4 dB to –3.45 dB in the right eye and from –5.9 dB to –3.60 dB in the left eye (figs 1H, 2H). The number of paracentral scotomata was six and two in the right and left eye, respectively.
Figure 1 Case report, right eye. Clinically significant macular oedema before and after "classic" laser treatment. (A) Fundus photograph. (B) Fluorescein angiography shows late leakage surrounding the fovea. (C) OCT retina thickness map shows severe diffuse macular thickening. (D) Red-free photograph immediately after "classic" laser treatment demonstrates multiple, ETDRS level grey-white burns applied to the areas of fluorescein leakage. (E) Fundus photograph 1 year after treatment shows multiple laser scars associated with several areas of hyperpigmentation. (F) Late phase fluorescein angiogram obtained 1 year after treatment demonstrates only minimal residual leakage temporal to the fovea. (G) OCT retina thickness map at 1 year shows a significant decrease in foveal thickening with minimal residual thickening temporal to the fovea. (H) "Total deviation" display of the central 10° visual field before (left) and 1 year after laser treatment (right). A mean deviation (MD) decrease from –5.4 dB to –3.45 dB and six scotomata (focal losses greater than 5 dB, indicated by circles) mostly located on the nasal hemifield can be observed at 1 year.
Figure 2 Case report, left eye. Clinically significant macular oedema before and after "light" laser treatment. (A) Fundus photograph. (B) Fluorescein angiography reveals areas of focal and diffuse leakage. (C) OCT retina map demonstrates severe diffuse macular thickening. The topographic pattern of retinal thickening was very similar to the other eye. (D) Immediate post-"light" laser treatment photograph demonstrates barely visible laser lesions. (E) and (F) Fundus photograph and late phase fluorescein angiogram obtained 1 year after treatment show very faint laser scar and significant reduction in fluorescein leakage. (G) 1 year after "light" laser treatment the oedema is almost completely resolved in the central fovea and residual thickening temporal to the fovea is present. (H) "Total deviation" display of the central 10° visual field before (left) and 1 year after laser treatment (right). A mean deviation (MD) decrease from –5.9 dB to –3.6 dB and two scotomata (focal losses greater than 5 dB, indicated by circles) both located on the nasal hemifield can be seen at 1 year.
Recent reports have suggested that the energy employed by the ETDRS in patients with CSMO may not be necessary to obtain a therapeutic effect and that these patients may be overtreated.13–15 We designed this prospective, randomised pilot trial specifically to determine whether a significant difference existed between "classic" and "light" laser photocoagulation for CSMO and to assess the feasibility of a definite trial. Since visual acuity remains stable for a rather long time in the majority of patients with CSMO, regardless of treatment, the detection of small differences in the proportions of eyes losing vision requires large sample sizes and a long follow up time. Therefore, we used as our primary outcome a morphological parameter, which has been shown by the ETDRS to be useful in demonstrating a beneficial treatment effect. Biomicroscopic/photographic evaluation of macular oedema is semi-quantitative at best, while OCT retinal thickness measurements are objective, quantitative, and reproducible.26–29 In our study, OCT examinations were performed at every visit by an experienced independent examiner who accepted only scans fulfilling well defined acceptance criteria. Since it has been shown that in eyes undergoing grid laser photocoagulation the average threshold sensitivity across the central 10° significantly decreases,5 we also measured the perimetric central sensitivity in order to identify theoretical differences of the effect of the two treatments on the central visual field.
The results of this randomised trial suggest that there is no significant difference between eyes with CSMO treated with either "classic" or "light" laser treatment within the first year. However, it is possible that significant differences might have been identified if a larger sample was used or follow up was continued beyond 1 year. In particular, the greater gain in mean number of letters, larger decrease in mean FTH, and greater proportion of eyes with reduction in CSMO on biomicroscopy in the "classic" group at 1 year, though not reaching statistical significance, may suggest a trend towards a larger beneficial effect from "classic" treatment. On the other hand, a significantly larger number of patients experiencing a visual gain in the "light" group at 6 months may suggest an earlier benefit from this treatment, which might decrease after 6 months. Also, the failure to detect a significant difference in change in MD from baseline and in number of scotomata between the two treatment groups at 1 year could depend upon the fact that the central 10° sensitivity showed no significant decrease following "both" laser treatments, in contrast with the findings reported by Striph et al.5 This may be because of the application of "lighter" intensity and fewer burns even in the "classic" group either because of a different technique or because fewer eyes with diffuse CSMO requiring extensive treatment were included in our study. Moreover, a more sensitive test, such as the microperimetry, might have been better in identifying differences between treatments in terms of focal losses on the visual field, since the point density of the central 10-2 test may not be sufficient to detect all the localised defects.
We also found a discrepancy between the number of eyes with reduction of CSMO based on biomicroscopic examination and OCT in the "classic" group. This finding was the result of the resolution of paracentral oedema and persistence of central foveal thickening in three patients (patients 6, 9, 10, in table 2), as confirmed by reviewing their retinal thickness maps, which revealed a marked decrease in the paracentral areas of the map but not in the central region.
In conclusion, this study, although underpowered and unlikely to yield definite conclusions as most pilot trials, seems to rule out large differences between the two treatments. Given the potential complications from the "classic" ETDRS level treatment, in particular the relatively high risk of invalidating paracentral scotomata, lighter intensity laser modalities should be investigated. Our preliminary data, suggesting a similar therapeutic effect between "classic" and "light" laser treatment, supports the further investigation of "light" laser photocoagulation with a larger equivalence trial.
Lee PP, Feldman ZW, Ostermann J, et al. Longitudinal prevalence of major eye diseases. Arch Ophthalmol 2003;121:1303–10.
Early Treatment Diabetic Retinopathy Study Research Group. Photocoagulation for diabetic macular edema. Early Treatment Diabetic Retinopathy Study report number 1. Arch Ophthalmol 1985;103:1796–806.
Lee CM, Olk RJ. Modified grid laser photocoagulation for diffuse diabetic macular edema. Long-term visual results. Ophthalmology 1991;98:1594–602.
Akduman L, Olk RJ. Diode laser (810 nm) versus argon green (514 nm) modified grid photocoagulation for diffuse diabetic macular edema. Ophthalmology 1997;104:1433–41.
Striph GG, Hart WM Jr, Olk RJ. Modified grid laser photocoagulation for diabetic macular edema. The effect on the central visual field. Ophthalmology 1988;95:1673–9.
Sinclair SH, Alaniz R, Presti P. Laser treatment of diabetic macular edema: comparison of ETDRS-level treatment with threshold-level treatment by using high-contrast discriminant central visual field testing. Sem Ophthalmol 1999;4:214–22.
Hudson C, Flanagan JG, Turner GS, et al. Influence of laser photocoagulation for clinically significant diabetic macular edema (DMO) on short-wavelenght and conventional automated perimetry. Diabetologia 1998;41:1283–92.
Schatz H, Madeira D, McDonald HR, et al. Progressive enlargement of laser scars following grid laser photocoagulation for diffuse diabetic macular edema. Arch Ophthalmol 1991;109:1549–51.
Lewis H, Schachat AP, Haimann MH, et al. Choroidal neovascularization after laser photocoagulation for diabetic macular edema. Ophthalmology 1990;97:503–10.
Guyer DR, D’Amico DJ, Smith CW. Subretinal fibrosis after laser photocoagulation for diabetic macular edema. Am J Ophthalmol 1992;113:652–6.
Han DP, Mieler WF, Burton TC. Submacular fibrosis after photocoagulation for diabetic macular edema. Am J Ophthalmol 1992;113:513–21.
Mainster MA, White TJ, Tips JH, et al. Retinal-temperature increases produced by intense light sources. J Opt Soc Am 1970;60:264–70.
Friberg TR, Karatza EC. The treatment of macular disease using a micropulsed and continuous wave 810-nm diode laser. Ophthalmology 1997;104:20–308.
Roider J, Brinkmann R, Wirbelauer C, et al. Subthreshold (retinal pigment epithelium) photocoagulation in macular diseases: a pilot study. Br J Ophthalmol 2000;84:40–7.
Moorman CM, Hamilton AM. Clinical applications of the MicroPulse diode laser. Eye 1999;13:145–150.
Bresnick GH. Diabetic maculopathy. A critical review highlighting diffuse macular edema. Ophthalmology 1983;90:1301–17.
Wallow IH, Bindley CD. Focal photocoagulation of diabetic macular edema. A clinicopathologic case report. Retina 1988;8:261–9.
Del Priore LV, Glaser BM, Quigley HA, et al. Response of retinal pigment epithelium to laser photocoagulation in organ culture. Arch Ophthalmol 1989;107:119–22.
Marshall J, Clover G, Rothtery S. Some new findings of retinal irradiation by krypton and argon lasers. In: Birngruber R, Gabel V-P, eds. Laser treatment and photocoagulation of the eye. Doc Ophthalmol Proc Series 36. The Hague: Junk, 1984:21–37.
Roider J, Michaud N, Flotte T, et al. Response of the RPE to selective photocoagulation of the RPE by repetitive short laser pulses. Arch Ophthalmol 1992;110:1786–92.
Dowson DW, Volpert OV, Gillis P, et al. Pigment epithelium-derived factor: a potent inhibitor of angiogenesis. Science 1999;285:245–8.
Ogata N, Tombran-Tink J, Jo N, et al. Upregulation of pigment epithelium-derived factor after laser photocoagulation.
Hee MR, Puliafito C, Duker JS, et al. Topography of diabetic macular edema with optical coherence tomography. Ophthalmology 1998;105:360–70.
Massin P, Erginay A, Haouchine B, et al. Retinal thickness in healthy and diabetic subjects measured using optical coherence tomogroaphy mapping software. Eur J Ophthalmol 2002;12:102–8.
Ferris FL 3rd, Kassoff A, Bresnick GH, et al. New visual acuity charts for clinical research. Am J Ophthalmol 1982;94:91–6.
Massin P, Vicaut E, Haouchine B, et al. Reproducibility of retinal mapping using optical coherence tomography. Arch Ophthalmol 2001;119:1135–42.
Chauhan DS, Marshall J. The interpretation of optical coherence tomography images of the retina. Invest Ophthalmol Vis Sci 1999;40:2332–42.
Polito A, Shah SM, Haller JA, et al. Comparison between retinal thickness analyzer and optical coherence tomography for assessment of foveal thickness in eyes with macular disease. Am J Ophthalmol 2002;134:240–51.
Browning DJ, McOwen MD, Bowen R, et al. Comparison of the clinical diagnosis of diabetic macular edema with diagnosis by optical coherence tomography. Ophthalmology 2004;111:712–15.
1 Department of Ophthalmology, Asahikawa Medical College, Asahikawa, Japan
2 Faculty of Applied Biological Science, Gifu University, Gifu, Japan
3 Preveqol Co, Tokyo, Japan
MD, Department of Ophthalmology, Asahikawa Medical College, 2-1-1-1 Midorigaoka Higashi, Asahikawa, 078-8510 Japan; firstname.lastname@example.org
Accepted for publication 1 December 2004
Aim: To investigate the serum levels of prorenin and its correlation with the severity of diabetic retinopathy (DR).
Methods: 248 patients with diabetes and 108 control subjects were divided into four groups: no-DR (n = 146), no proliferative diabetic retinopathy (no-PDR) (n = 78), PDR (n = 24), and controls (n = 108). Serum levels of prorenin from all subjects were measured using the new antibody activating direct kinetic (AAD-PR) assay. The serum prorenin levels were compared among the groups.
Results: The serum levels of prorenin in the control, no-DR, no-PDR, and PDR groups, respectively, were 109.1 (66.1), 194.6 (160.4), 271.5 (220.3), and 428.4 (358.4) pg/ml (mean (SD)). Prorenin in the PDR group was remarkably high compared with the control and no-DR groups (p<0.0001) and with the no-PDR group (p = 0.002). Serum levels of prorenin increased with increasingly severe retinopathy. No correlation was found between the prorenin level and the duration of disease or HbA1c.
Conclusions: The serum levels of prorenin in patients with PDR were found to be markedly high using the AAD-PR assay. Increased levels of prorenin in diabetes may have an important role in the pathogenesis of DR.
Abbreviations: AAD-PR, antibody activating direct prorenin; DR, diabetic retinopathy; PDR, proliferative diabetic retinopathy; RAS, renin-angiotensin system
Keywords: antibody activating direct prorenin assay; diabetes; diabetic retinopathy; prorenin; renin-angiotensin system
Diabetic retinopathy (DR) is a major cause of blindness worldwide. Although strict glycaemic control is thought to be essential to prevent the occurrence of DR,1 some cases unfortunately develop advanced proliferative diabetic retinopathy (PDR).2 In fact, because it is difficult to confirm if diabetic microangiopathy including retinopathy is progressive or not, a useful predictor that is well correlated with the occurrence of diabetic microangiopathy is needed to prevent the development of diabetic microvascular complications.
Prorenin is an inactive precursor of renin. The circulating prorenin level is five to 10 times higher than the active form of renin. Although little is known about the biological function of prorenin, it reportedly increases in diabetes and is associated with the occurrence of DR and nephropathy.3–5 Furthermore, in adolescents with diabetes, higher serum levels of prorenin occur several years before diabetic nephropathy6–8 and retinopathy.9 This modulation of prorenin in diabetes indicates that prorenin is involved in the occurrence and the progression of diabetic microangiopathy. Although measuring prorenin seems to be a good method to determine if diabetic microangiopathy is present or not, the method of measuring prorenin in previous reports has been complicated. Until recently, the level of prorenin was determined by measuring the total renin level and subtracting the active rennin level.10–12 Total renin was measured after activating inactive prorenin by trypsin or non-proteolytically. At the same time, active renin was measured independently, and the difference in the levels between total renin and active renin was defined as the prorenin level.
A new method called the antibody activating direct prorenin (AAD-PR) assay, developed by Suzuki et al,13 enables direct measurement of the concentration of prorenin using an antibody to the prorenin profragment, which detects prorenin in serum and confirms the complex to the prorenin. This complex has renin-like activity—that is, the ability to convert angiotensinogen to angiotensin I. The generated angiotensin I is measured with the enzyme linked immunosorbent assay. The prorenin level can be calculated by the amount of generated angiotensin I. The AAD-PR assay was reported to have higher sensitivity than previous methods.14
In this study, we focused on the relation between the serum levels of prorenin and the severity of DR. We measured serum levels of prorenin in patients with type 2 diabetes and estimated the clinical implication of prorenin in DR using the AAD-PR assay.
PATIENTS AND METHODS
In all, 248 patients with diabetes and 108 control subjects from Asahikawa Medical College Hospital were included. The control subjects had a normal examination that included urinalysis, blood chemistry, and blood pressure measurement and had never had type 2 diabetes. Patients with diabetes who were followed by physicians at Asahikawa Medical College Hospital all satisfied the criteria for diagnosis of diabetes by the World Health Organization. The subjects received a detailed explanation of the aims of the study and provided informed consent. This study protocol was reviewed by the ethics committee of our institution. All procedures adhered to the tenets of the Declaration of Helsinki.
The subjects were divided into four groups: patients without DR (no-DR group), those with retinopathy but no proliferative DR (no-PDR group), patients with proliferative DR (PDR group), and controls. The characteristics of these groups are shown in table 1. Sera were obtained from all subjects and then treated as described by Kawazu et al to measure the serum levels of prorenin.14 The distribution of serum prorenin levels in the four groups was compared using one way of analysis variance and Scheffe’s test. A p value of 0.05 or lower was considered significant. The Pearson correlation coefficient (r) was calculated to determine whether there were close associations among the variables.
Table 1 Characteristics of the study groups (SD)
The characteristics of the subjects are shown in table 1. There is no statistical difference between males and females, which has been reported to affect the serum levels of prorenin.14 Systolic blood pressure and diastolic blood pressure were not significantly different among the four groups. The distribution of the serum levels of prorenin in the four groups is shown in figure 1. The serum levels of prorenin in the control, no-DR, no-PDR, and PDR groups were 109.1 (66.1), 194.6 (160.4), 271.5 (220.3), and 428.4 (358.4) pg/ml (mean (SD)), respectively. The serum levels of prorenin were markedly higher in the PDR group than in the control and no-DR groups (p<0.0001, Scheffe’s test) and the no-PDR (p = 0.002) group. The serum levels of prorenin were higher with increasingly severe retinopathy. No significant correlation was found between the serum prorenin level and disease duration (r = 0.17, p = 0.04) or the HbA1c level (r = 0.05, p = 0.56).
Figure 1 The distribution of serum levels of prorenin; *p<0.05, **p<0.01.
In this study, we evaluated the serum levels of prorenin in patients with diabetes using the newly developed AAD-PR assay. This study showed that the serum levels of prorenin in patients with diabetes with PDR were remarkably higher than in the normal healthy subjects, patients without DR, and those with no PDR. The serum concentration of prorenin in patients with diabetes was higher than in control subjects, and a high serum concentration of prorenin in patients with diabetes increased with increasingly severe retinopathy. These results supported previous reports that had shown the clinical implication of prorenin in the occurrence and the development of diabetic microangiopathy.3–9,15
Recent studies investigated the relation between the concentration of prorenin and the occurrence or the development of DR.3–9,15 Franken et al reported that a high plasma prorenin level is associated with DR, particularly PDR.4 Makimattila et al reported that the serum total renin level increased and was a useful marker of activity and the severity of DR.15 Total renin is composed of renin and prorenin, and 90% of total renin is prorenin.16 The active renin level in diabetes does not increase.17,18 An increase in the total renin level was thought to be the result of the increased level of prorenin in diabetes. These reports showed the close relation between the concentration of prorenin and the severity of DR4,15 and supported our results. Although those previous reports showed higher levels of prorenin in diabetes with retinopathy, the conventional measurement method was more complicated and less sensitive for determining the concentration of prorenin than the AAD-PR assay.14
In the present study, we showed that there was no close relation between the serum levels of prorenin and HbA1c or duration of diabetes. Franken et al reported that the plasma concentration of prorenin was not correlated with HbA1c and the duration of diabetes.5 On the other hand, Makimattila et al reported that the serum concentration of total renin was correlated with HbA1c.15 Luetscher et al also demonstrated a positive correlation between HbA1c and the plasma concentration of prorenin.3 HbA1c and the duration of diabetes are key risk factors for diabetic microangiopathy and are thought to be associated with the occurrence of DR.1,19 Although HbA1c is an important indicator for determining the degree of glycaemic control in diabetes, this is not sufficient to be associated with the occurrence and the severity of DR.20 Higher serum levels of prorenin in diabetes might be more appropriate for estimating the occurrence and the severity of DR than HbA1c. In this study, the duration of diabetes was longer in patients with PDR than other patients who had no retinopathy or in whom retinopathy was not proliferative; however, there was no close relation between the serum levels of prorenin and the duration of diabetes. Duration, as mentioned previously, is also an important key factor for the occurrence of DR,19 but it does not seem to affect the serum concentration of prorenin.
In this study, we did not measure renin at the same time to determine if the serum level of renin in diabetes increased or not. Renin is well known to be a key enzyme in the cleavage of angiotensinogen to angiotensin I, and this reaction is a rate limiting step to generate angiotensin II in the renin-angiotensin system (RAS). Previous reports showed that the concentration of renin in diabetes does not increase,21 although RAS has been implicated in the pathogenesis of DR.3–5,15,22–25 The fact that renin does not increase in diabetes seems to be a discrepancy, but RAS is activated in diabetes. Our study, as other previous reports showed,3–5,13,26 might indicate the involvement of increased prorenin in the development of DR. In addition, as mentioned previously, the plasma concentration of prorenin precedes the occurrence of diabetic nephropathy by several years.7,8 Increasing prorenin in diabetes may trigger microangiopathy and promote the development of diabetic microangiopathy through the activation of RAS.
Recently, prorenin was reported to have enzymatic activity that generated angiotensin I and activated the RAS thorough the generation of angiotensin II.26–28 Prorenin is composed of two components, the profragment of prorenin and mature renin. Suzuki et al demonstrated that prorenin has a key region in its profragment for non-proteolytic activation with protein interaction.26 Furthermore, Nguyen et al, who investigated the renin/prorenin receptor, showed that the prorenin binding this receptor activated the conversion of angiotensinogen to angiotensin I. High levels of this receptor mRNA were detected in the heart, brain, and placenta and lower levels in the kidney and liver.27 Recently, Ichihara et al proved the non-proteolytic activation of prorenin and the presence of renin/prorenin receptor in the kidney using streptozotocin induced diabetic rats. They reported that the interference of prorenin with peptide which inhibits an interaction of prorenin with renin/prorenin receptor, inhibited the local generation of angiotensin II and improved diabetic nephropathy in streptozotocin induced diabetic rats.18 Angiotensin I generated with non-proteolytic activation of prorenin is transformed to angiotensin II by soluble or endothelium specific angiotensin converting enzyme. Angiotensin II exhibits pathological effects in the retina in diabetes through binding angiotensin II type 1 receptor, which is thought to be the most important receptor of all subtypes to exhibit the physiological and pathological effects. Angiotensin II is associated with overexpression of some angiogenic factors—that is, vascular endothelial growth factor (VEGF),29–33 and angiopoietin 2.34 VEGF and angiopoietin 2 have a crucial role in the development of retinal neovascularisation,35–38 a main feature of PDR. Taken together, it is possible that a high concentration of prorenin in patients with diabetes activates the local RAS in the eyes through its binding renin/prorenin receptor and promotes the pathogenesis of DR through the generation of angiotensin II.
In this study, we evaluated the serum levels of prorenin in patients with type 2 diabetes with a newly developed method, the AAD-PR assay. The serum levels of prorenin in patients with PDR were markedly high. High levels of prorenin in diabetes increase with increasingly severe retinopathy. We showed that a high concentration of circulating prorenin may be involved in the pathogenesis of DR. Further prospective study is needed to investigate the relation between modulation of the serum levels of prorenin and the severity of DR, and in turn, whether patients without DR with a higher level of prorenin will develop retinopathy.
The Diabetes Control and Complications Trial Research Group. The effect of intensive treatment of diabetes on the development and progression of long-term complications in insulin-dependent diabetes mellitus. N Engl J Med 1993;329:977–86.
Mattock MB, Cronin N, Cavallo-Perin P, et al. Plasma lipids and urinary albumin excretion rate in Type 1 diabetes mellitus: the EURODIAB IDDM Complications Study. Diabet Med 2001;18:59–67.
Luetscher JA, Kraemer FB, Wilson DM, et al. Increased plasma inactive renin in diabetes mellitus. A marker of microvascular complications. N Engl J Med 1985;312:1412–17.
Franken AA, Derkx FH, Schalekamp MA, et al. Association of high plasma prorenin with diabetic retinopathy. J Hypertens Suppl 1988;6:S461–3.
Franken AA, Derkx FH, Blankestijn PJ, et al. Plasma prorenin as an early marker of microvascular disease in patients with diabetes mellitus. Diabete Metab 1992;18:137–43.
Allen TJ, Cooper ME, Gilbert RE, et al. Serum total renin is increased before microalbuminuria in diabetes. Kidney Int 1996;50:902–7.
Deinum J, Ronn B, Mathiesen E, et al. Increase in serum prorenin precedes onset of microalbuminuria in patients with insulin-dependent diabetes mellitus. Diabetologia 1999;42:1006–10.
Chiarelli F, Pomilio M, De Luca FA, et al. Plasma prorenin levels may predict persistent microalbuminuria in children with diabetes. Pediatr Nephrol 2001;16:116–20.
Kordonouri O, Wladimirowa A, Danne T. High total serum renin concentrations are associated with the development of background retinopathy in adolescents with type 1 diabetes. Diabetes Care 2000;23:1025–6.
Sealey JE. Plasma renin activity and plasma prorenin assays. Clin Chem 1991;37:1811–19.
Cooper RM, Murray GE, Osmond DH. Trypsin-induced activation of renin precursor in plasma of normal and anephric man. Circ Res 1977;40:I171–9.
Derkx FH, de Bruin RJ, van Gool JM, et al. Clinical validation of renin monoclonal antibody-based sandwich assays of renin and prorenin, and use of renin inhibitor to enhance prorenin immunoreactivity. Clin Chem 1996;42:1051–63.
Suzuki F, Hatano Y, Nakagawa T, et al. Non-proteolytic activation of human prorenin by anti-prorenin prosegment (pf#1: 1P-15P) antiserum. Biosci Biotechnol Biochem 1999;63:550–4.
Kawazu S, Minagawa S, Yazawa M, et al. Sex difference and possible relationship to microvascular complications of serum prorenin levels in type 2 diabetic patients, measured by a novel antibody-activating direct enzyme kinetic assay. Journal of Diabetes and Its Complications 2004;18:275–81.
Makimattila S, Summanen P, Matinlauri I, et al. Serum total renin, an independent marker of the activity and severity of retinopathy in patients with IDDM. Br J Ophthalmol 1998;82:939–44.
Sealey JE, Glorioso N, Itskovitz J, et al. Prorenin as a reproductive hormone. New form of the renin system. Am J Med 1986;81:1041–6.
Ubeda M, Hernandez I, Fenoy F, et al. Vascular and adrenal reninlike activity in chronically diabetic rats. Hypertension 1988;11:339–43.
Ichihara A, Hayashi M, Kaneshiro Y, et al. Inhibition of diabetic nephropathy by a decoy peptide corresponding to the "handle" region for nonproteolytic activation of prorenin. J Clin Invest 2004;114:1128–35.
Chaturvedi N, Sjoelie AK, Porta M, et al. Markers of insulin resistance are strong risk factors for retinopathy incidence in type 1 diabetes. Diabetes Care 2001;24:284–9.
Kohner EM, Aldington SJ, Stratton IM, et al. United Kingdom Prospective Diabetes Study, 30: diabetic retinopathy at diagnosis of non-insulin-dependent diabetes mellitus and associated risk factors. Arch Ophthalmol 1998;116:297–303.
Derkx FH, Schalekamp MA. Human prorenin: pathophysiology and clinical implications. Clin Exp Hypertens A 1988;10:1213–25.
Drury PL, Bodansky HJ, Oddie CJ, et al. Increased plasma renin activity in type 1 diabetes with microvascular disease. Clin Endocrinol 1982;16:453–61.
Moravski CJ, Skinner SL, Stubbs AJ, et al. The renin-angiotensin system influences ocular endothelial cell proliferation in diabetes: transgenic and interventional studies. Am J Pathol 2003;162:151–60.
Funatsu H, Yamashita H. Pathogenesis of diabetic retinopathy and the renin-angiotensin system. Ophthalmic Physiol Opt 2003;23:495–501.
Sjolie AK, Chaturvedi N. The retinal renin-angiotensin system: implications for therapy in diabetic retinopathy. J Hum Hypertens 2002;16 (Suppl 3) :S42–6.
Suzuki F, Hayakawa M, Nakagawa T, et al. Human prorenin has "gate and handle" regions for its non-proteolytic activation. J Biol Chem 2003;278:22217–22.
Nguyen G, Delarue F, Burckle C, et al. Pivotal role of the renin/prorenin receptor in angiotensin II production and cellular responses to renin. J Clin Invest 2002;109:1417–27.
Prescott G, Silversides DW, Reudelhuber TL. Tissue activity of circulating prorenin. Am J Hypertens 2002;15:280–5.
Wilkinson-Berka JL, Kelly DJ, Gilbert RE. The interaction between the renin-angiotensin system and vascular endothelial growth factor in the pathogenesis of retinal neovascularization in diabetes. J Vasc Res 2001;38:527–35.
Sarlos S, Rizkalla B, Moravski CJ, et al. Retinal angiogenesis is mediated by an interaction between the angiotensin type 2 receptor, VEGF, and angiopoietin. Am J Pathol 2003;163:879–87.
Otani A, Takagi H, Suzuma K, et al. Angiotensin II potentiates vascular endothelial growth factor-induced angiogenic activity in retinal microcapillary endothelial cells. Circ Res 1998;82:619–28.
Moravski CJ, Kelly DJ, Cooper ME, et al. Retinal neovascularization is prevented by blockade of the renin-angiotensin system. Hypertension 2000;36:1099–104.
Williams B. A potential role for angiotensin II-induced vascular endothelial growth factor expression in the pathogenesis of diabetic nephropathy? Miner Electrolyte Metab 1998;24:400–5.
Otani A, Takagi H, Oh H, et al. Angiotensin II induces expression of the Tie2 receptor ligand, angiopoietin-2, in bovine retinal endothelial cells. Diabetes 2001;50:867–75.
Oh H, Takagi H, Suzuma K, et al. Hypoxia and vascular endothelial growth factor selectively up-regulate angiopoietin-2 in bovine microvascular endothelial cells. J Biol Chem 1999;274:15732–9.
Takagi H, Koyama S, Seike H, et al. Potential role of the angiopoietin/tie2 system in ischemia-induced retinal neovascularization. Invest Ophthalmol Vis Sci 2003;44:393–402.
Umeda N, Ozaki H, Hayashi H, et al. Colocalization of Tie2, angiopoietin 2 and vascular endothelial growth factor in fibrovascular membrane from patients with retinopathy of prematurity. Ophthalmic Res 2003;35:217–23.
Oshima Y, Deering T, Oshima S, et al. Angiopoietin-2 enhances retinal vessel sensitivity to vascular endothelial growth factor. J Cell Physiol 2004;199:412–17.