Contraction and relaxation of aortas from galactosaemic rats and the effects of aldose reductase inhibition

European Journal of Pharmacology, 243 (1993) 47-53

47

© 1993 Elsevier Science Pubhshers B.V. All rights reserved 0014-2999/93/$06.00

EJP 53326

Contraction and relaxation of aortas from galactosaemic rats and the effects of aldose reductase inhibition N o r m a n E. C a m e r o n * and M a r y A. C o t t e r Department of Biomedical Sctences, Umcerstty of Aberdeen, Manschal College, Aberdeen AB9 1AS, Scotland, UK Recewed 21 July 1993, accepted 27 July 1993

Rats were fed for 10 days with a 40% galactose diet, in order to chronically stimulate the polyoi pathway. Thoracic aorta contraction and relaxation were studied. Compared to controls, galactosaemla did not influence contractions to phenylephrine or serotonm. Acetylcholine produced concentration-dependent relaxation of aortic rectangles precontracted with phenylephrlne, galactosaemla caused a 25% deficit in maximum relaxation to acetylchohne (P < 0.01) and a 168% increase in ECs0. There was a similar 25% reduction in relaxation to 3 /zM calcium ionophore A23187 (P < 0.05). By contrast, there were no significant differences m endothehum-mdependent relaxation to nitroglycerine or cromakahm. The abnormalities m endothehum-dependent relaxation were completely prevented by treating galactosaemlc rats with the aldose reductase inhibitor, ponalrestat. Thus, the data demonstrate that elevated polyol pathway activity contributes to reduced endothelium production, release or the action of mtric oxide in experimental galactosaemla, and suggest that this mechanism could also contribute to the vascular defects seen in diabetes melhtus. Aldose reductase, Aorta; EDRF (endothelium-derived relaxing factor); Galactosaemia; Nitric oxide (NO); Polyol pathway

1. Introduction

One of the metabohc factors suggested to be of primary importance for the development of complications of diabetes mellitus, which includes microvascular disease, is increased flux through the polyol pathway (Dvornik, 1987). Thus, hyperglycaemia stimulates the production of sorbitol from glucose by aldose reductase (L-alditol:NADP + 1-oxidoreductase; [EC 1.1.1.21]); sorbitol is further metabolized to fructose by the second enzyme in the pathway, sorbitol dehydrogenase (L-iditol:NAD + 5-oxidoreductase; [EC 1.1.1.14]). A1dose reductase inhibitors ameliorate some of the deficits related to nerve (Cameron et al., 1986; 1989a; Boulton et al., 1990), renal (Larkins and Dunlop, 1992), retinal (Robison et al., 1986) and cardiac and skeletal muscle (Cameron et al., 1989b; 1990; Cotter et al., 1993) function associated with diabetes in experimental models and patients. Dysfunction has been demonstrated in vessels and vascular beds of diabetic animals and patients, in vitro and in vivo (Tomlinson et al., 1992). Both endothelium and smooth muscle are involved. Thus, experiments

* Corresponding author Tel. 44 224 273013, fax 44 224 273019.

have ~dentified altered responses to vasoconstrictors such as a-adrenoceptor agonists and serotonin (Scarborough and Carrier, 1983; Hagen et al., 1985), increased Ca 2÷ channel activity (White and Carrier, 1990), reduced responses to ATP-sensitive K ÷ channel agonists (Kamata et al., 1989b; Cameron and Cotter, 1992a), and impaired production of endothelium-dependent relaxing factor ( E D R F ) and prostacyclin (Funakawa et al., 1983; Oyama et al., 1986; Durante et al., 1988; Peiper and Gross, 1988; Kamata et al., 1989a; Ward et al., 1989; Hattori et al., 1991; McVeigh et al., 1992). Recently, we have observed that aldose reductase inhibition prevents the development of several of these abnormalities (Cameron and Cotter, 1992a). Diabetes mellitus involves complex metabolic and hormonal changes not restricted to enhanced polyol pathway activity. Several investigators have reported that many diabetic complications, such as neuropathy, nephropathy, myopathy and retinopathy may at least in part be duplicated in animals fed a high galactose diet (Low and Schmelzer, 1983; Robison et al., 1986; Larkins and Dunlop, 1992; Cameron et al., 1992a; Cotter et al., 1992). Galactose is a better substrate for aldose reductase than glucose, and is converted to galactitol. However, galactitol is only poorly metabolized by sorbitol dehydrogenase (Dvormk, 1987). Thus, galactosaemia

48 provides a model of enhanced flux through the first half of the polyol pathway. Furthermore, it does not involve the hyperglycaemia, hypoinsulinaemia and tissue wasting found in some experimental models of diabetes and is, therefore, suitable for studying the relatively 'pure' effects of polyol pathway activity (Dvornik, 1987; Cameron et al., 1992b). The polyol pathway is present in both vascular endothelium and smooth muscle; aortas from galactosaemic animals contain high ( > 5.0 n m o l / m g ) concentrations of galactitol (Dvornik, 1987). Similar concentrations in sciatic nerve give rise to diabetes-like reductions in conduction velocity that are prevented by aldose reductase inhibitor treatment (Cameron et al., 1992). Thus, the aim of the investigation was to examine whether galactosaemia produced functional deficits in aorta similar to those seen in experimental diabetes, the specificity of effects for polyol pathway activity being further tested using the aldose reductase inhibitor, ponalrestat (Stribling et al., 1985).

2. Materials and methods

Mature male Sprague-Dawley rats, aged 19 weeks at the start of the experiments, were obtained from the Aberdeen University breeding colony. One group of rats acted as untreated controls. Two other groups were fed normal rat chow (Labsure CRM, Special Diet Services, Manea, Cambridgeshire, UK), to which 40% galactose was added, for 10 days. This duration of galactosaemia was chosen as some deficits in skeletal muscle function, which depend largely on vascular abnormalities, are present after 7 days galactosaemia (Cameron et al., 1992a). One of the galactosaemic groups was also treated with the aldose reductase inhibitor, ponalrestat, ICI 128,436, 3-(4-bromo-2-fluorobenzyl)-4-oxo-3H-phthalazin-l-ylacetic acid (ICI Pharmaceuticals, Macclesfield, Cheshire, UK) at a dose of 100 mg kg -1 daily by gavage from the start of galactosaemia. Previous experimentation has shown that this results in a high ( > 90%) level of polyol pathway blockade in several tissues (Stribling et al., 1985; Dvornik, 1987; Cameron and Cotter, 1992b; Cameron et al., 1992). Methods have been previously described in detail (Cameron and Cotter, 1992a). In final experiments, thoracic aortas were removed under 2 - 5 % halothane anaesthesia, cleaned of fat and connective tissue, cut into 4 mm rings, then cut longitudinally to form rectangles which were mounted using small tissue clips in 20 ml organ baths. They were bathed in modified Krebs solution (144.0 Na ÷, 5.0 K ÷, 2.5 Ca 2+, 1.1 Mg 2+, 25.0 HCO~-, 1.1 PO 3 - , 1.1. SO42-, 11.0 glucose, in mM) gassed with 95% 0 2 5% CO z (pH 7.35) at 37°C. Tension was monitored using isometric transducers,

the outputs being displayed on chart recorders. Resting tension was maintained at 1 g. Aortas were allowed to equilibrate for 45 min and were then tested for contraction with 1 IzM phenylephrine. They were then washed and left to equilibrate for a further 30 min. Bathing fluid was changed every 20-30 min except during cumulative dose-response experiments. Relaxation was assessed against a background of precontraction by phenylephrine (100 nM). At the end of the experiment, whilst still under tension, the length of tissue was measured with vernier callipers. Aortic rectangles were gently blotted to remove excess bathing medium, and weighed. Cross sectional areas were calculated from length and weight, assuming a tissue density of 1.05 (Peiper and Gross, 1988), and developed tensions were expressed per cross sectional area. Drugs came from the following sources. Acetylcholine, Ca 2+ ionophore A23187, serotonin, phenylephrine and flurbiprofen were obtained from Sigma, Poole, Dorset, UK. Nitroglycerine, in 5 mg tablets, was obtained from Aberdeen Royal Infirmary Pharmacy. Tablets were crushed in Krebs solution to obtain stock solutions for serial dilution. Ponalrestat was obtained from ICI Pharmaceuticals, Macclesfield, Cheshire, UK.

2.1. Stattsttcal analysts Results are expressed as group means + S.E.M. Data were subjected to one way analysis of variance (Moore and Edwards, 1965), and any statistically significant differences were assigned to individual between group comparisons using Student t-tests and the Bonferroni correction for multiple comparisons (Graphpad, San Diego, Ca, USA).

3. Results

Body weights were 492 + 13, 519 + 14 and 459 + 11 g for control, galactosaemic and ponalrestat-treated galactosaemic groups respectively, which were not significantly different. Galactosaemic rats approximately doubled their food intake over the first 3 days. There were no visible signs of lens sugar cataract development over the 10 day period. Concentration-response curves for phenylephrine (fig. 1A) and serotonin (fig. 1B) did not show significant differences between the different groups. ECs0 values for phenylephrine were 42.1 + 5.9, 47.4 + 4.1 and 48.1 + 7.8 nM for control, galactosaemic and ponalrestat-treated galactosaemic groups respectively. Corresponding values for serotonin were 2.77 + 0.45, 2.43 __+0.33, and 3.05 + 0.58 ~M. Concentration-response curves for acetylcholine (fig. 2) revealed that maximum relaxation was 77.5 + 1.9% for controls. This was reduced to 57.7 + 5.1% with 10

49 300

A

250

200

"~ 150

,o,

100 0

-9

-8

I

I

I

I

-7

-6

-5

-4

50

i

i /

C

GAL

FBN

ARI

F~g 3. ECs0 values for aorta relaxation to acetylchohne. C, controls (open bar), n = 14, galactosaemlc group (filled bars) without (GAL) and with (FBN) 30 mln pretreatment with 3 / z M flurblprofen, n = 11; ARI, galactosaemlc group treated with the aldose reductase inhibitor ponalrestat (cross-hatched bar), n = 6 Error bars are S E M.

--

08

I

-7

-6

-5

-4

Concentrabon (Log10 M) Fig. 1. Concentration-response curves for aorta contractions to A, phenylephrine, and B, serotonm. Controls ( • ) , n = 12; galactosaemlc group (o), n = 11; ponalrestat-treated galactosaemlc group (zx), n = 6. Error bars are S.E M

100

80

60

n"

40

20

0~

-9

-8

-7

-6

-5

-4

Concentrabon (Log10 M) Fig. 2. Concentration-response curves for aorta relaxation to acetylcbohne following precontractlon with t00 n M phenylephrme. Controls ( • ) , n = ]4; galactosaemic group (©), n = ] ] , galactosaemic group after pretreatment for 30 rain with 3 /J,M flurblprofen (e), ponalrestat-treated galactosaemic group (A), n = 6. E r r o r bars are S.E.M.

days galactosaemia (P < 0.01). Aortas from ponalrestat-treated galactosaemic rats had a maximum relaxation of 77.1 + 2.7%, not significantly different from controls, but greater than for untreated galactosaemia (P < 0.05). For both control and ponalrestat-treated galactosaemic aortas, there were significant differences (P < 0.05) compared to untreated galactosaemic aortas for all concentrations of acetylcholine used (1 nM to 100/zM). In separate tissue samples from galatosaemic rats, acetylcholine relaxation was studied after preincubation for 30 min with 3 p,M flurbiprofen. There were no significant effects of flurbiprofen on the relaxation deficit. ECs0 values for acetylcholine (fig. 3) were elevated approximately 168-220% by galactosaemia compared to controls (P < 0.01). By contrast, ECs0 for the ponalrestat-treated diabetic group was not significantly different from controls. Relaxation to 3 ~M calcium ionophore A23187 after precontraction with phenylephrine (fig. 4) showed greater variability than for acetylcholine, however, there was a similar 25% deficit in relaxation with galactosaemia (P < 0.05). With ponalrestat treatment, relaxation was not significantly different from controls. For endothelium-independent relaxation to nitroglycerine, curves for individual groups (not shown) markedly overlapped and did not reveal any significant differences in nitrovasodilator responses. Thus, maximum relaxations were 100.8 + 0.8, 100.5 + 0.6 and 100.1 ___ 0.7% for control, galactosaemic and ponalrestat-treated galactosaemic groups respectively. Corresponding ECs0 values were 6.54 + 1.11, 7.66 + 1.14 and 5.49 + 0.79 nM.

5O 1oo

4. Discussion

90

80

n"~ -

7O

60

50 GAL ARI c Fig 4 Aorta relaxation to 3 tzM calcium lonophore A23187 after precontrachon with 100 nM phenylephrlne C, controls (open bar), n = 10; GAL, galactosaemlc group (filled bar), n = 11, ARI, galactosaemlc group treated with the aldose reductase inhibitor ponalrestat (cross-hatched bar), n = 6 Error bars are S E M

Maximum relaxation to cromakalim (fig. 5) was around 92% and was not significantly different for the three groups. ECs0 values were 212 + 58, 343 + 57 and 112 + 32 nM for control, galactosaemic and ponalrestat-treated galactosaemic aortas respectively. Whilst not showing any significant differences from controls, ECs0 values for galactosaemia were elevated compared to ponalrestat-treated galactosaemia (P < 0.05), which was also apparent for the mid-range of the concentration-response curve (0.1 and 0.3/xM). 100 T

80

/ o~

nr

60

/ /

40

20

0

-9

-8

-7

I

I

I

-6

-5

-4

Concentration (Loglo M)

Fig 5 Concentration-response curves for aorta relaxatton to cromakahm following precontraction with 100 nM phenylephrlne. Controls ( • ), n = 10; galactosaemlc group (©), n = 11; ponalrestat-treated galactosaemlc group ( zx ), n = 6 Error bars are S.E.M

The data demonstrate that short-term galactosaemia causes abnormal in vitro responses of aortas. All the effects were prevented by aldose reductase inhibitor treatment, which strongly supports the notion of polyol pathway involvement. Galactosaemia did not alter responses to phenylephrine, and had no significant effect on serotonin contractions. This contrasts with the reduced serotonin responses found for experimental diabetes (Hagen et al., 1985; Head et al., 1987; Cameron and Cotter, 1992a). The literature for adrenoceptor agonists lS complex, with sensitivity or contractile force increased (Scarborough and Carrier, 1982; Peiper and Gross, 1988; White and Carrier, 1990), unchanged (Mulhern and Docherty, 1989; Cameron and Cotter, 1992a), or reduced (Head et al., 1987). It is possible that these changes in diabetes do not depend on polyol pathway activity and, therefore, would not be expected in a galactosaemic model. However, against this argument, similar abnormahties in diabetic aorta were partially prevented by aldose reductase inhibition (Cameron and Cotter, 1992a). Thus, it is likely that polyol pathway activity is involved in contractile dysfunction, which was not evident in the present experiments, perhaps because of the short duration of galactosaemia. The data demonstrate an early effect of polyol pathway activtty on relaxation to acetylcholine and calcium lonophore A23187. A similar deficit m diabetes is characterised by a shift in the concentration-response curve for acetylcholine to the right, which may or may not be accompanied by a reduction in maximal relaxation (Oyama et al., 1986; Durante et al., 1988; Peiper and Gross, 1988; Kamata et al., 1989a; Saenz de Tejada et al., 1989; Hattori et al., 1991; Cameron and Cotter, 1992a; Taylor et al., 1992). The findings for galactosaemia suggest polyol pathway involvement in this relaxation deficit, which is emphasised by prevention with ponalrestat. A structurally unrelated sulphonylnitromethane-based aldose reductase inhibitor also prevented the development of an endothelium-dependent relaxation deficit consequent on 3 months streptozotocin-diabetes in rats (Cameron and Cotter, 1992a). For diabetic rabbit aorta, a similar relaxation deficit was prevented by zopalrestat treatment (Tesfamariam et al., 1993). The lack of effect of galactosaemia or diabetes on responses to nitrovasodilators (Durante et al., 1988; Kamata et al., 1989a) in vitro rules out changes m the smooth muscle soluble guanylate cyclase system. Similarly, the deficit in relaxation to calcium ionophore A23187 argues against effects dependent on abnormalities of acetylcholine receptors or transduction mechanisms. Investigations into the mechanism(s) underlying impaired endothelium-dependent relaxation in aorta from

51 diabetic rabbits and cerebral vasculature in diabetic rats have identified enhanced synthesis of vasoconstrictor prostanoids, stimulated by acetylcholine, which cause contraction to oppose relaxation. This effect may be inhibited by indomethacin or a prostaglandin H z/ thromboxane A z blocker (Tesfamariam et al., 1989; Mahan et al., 1991). In contrast, cyclo-oxygenase inhibition by flurbiprofen did not block the relaxation deficit in galactosaemic aorta, therefore, the mechanism of impaired relaxation is likely to be different. The argument is further supported by the deficit in response to A23187 with galactosaemia as this compound does not stimulate abnormal prostanoid synthesis in diabetic rabbit aorta (Tesfamarium et al., 1989). Thus, the likely cause of the deficit in endothelium-dependent relaxation in galactosaemic rats is a reduction in the formation or release of the EDRF, nitric oxide (NO) (Moncada et al., 1991). Several mechanisms could cause polyol-pathway-dependent deficits in endothelium-dependent relaxation. They relate to alterations in cell redox state (Dvornik, 1987), ATP production (Davidson and Murphy, 1985), protection against superoxide damage (Baynes, 1991) and neutralization of NO (Moncada et al., 1991), and glycoxylation of cell proteins and lipids (Bucala et al., 1991). Thus, the polyol pathway is present in endothelial cells (Ohtaka et al., 1992) and enhanced flux depletes NADPH (Dvornik, 1987). As this co-factor is necessary for the conversion of L-arginine to NO (Moncada et al., 1991), therefore, enhanced polyol pathway flux could reduce NO production. In addition, NADPH is required for glutathione synthesis, which aids the removal of oxygen free radicals that would otherwise neutralize NO (Moncada et al., 1991; Hattori et al., 1991). Endothelium from diabetic rats shows increased susceptibility to free radical damage (Peiper and Gross, 1988). Exposure of cultured endothelial cells to elevated glucose concentrations caused impaired glutathione-peroxidase-dependent degradation of hydrogen peroxide which was prevented by aldose reductase inhibition (Kashiwagi and Kikkawa, 1991). Rabbit aortas subjected to elevated glucose concentrations in vitro showed an aldose reductase inhibitor preventable deficit in endothelium-dependent relaxation (Tesfamariam et al., 1992) which was also reduced when rabbits were pretreated with the anti-oxidant, probucol (Tesfamariam and Cohen, 1992). The NADPH requirement of the polyol pathway also diverts glucose metabolism to the pentose phosphate shunt (Dvornik, 1987) which reduces ATP synthesis (Davidson and Murphy, 1985). Agents that inhibit endothelial energy metabolism interfere with NO production (Richards et al., 1991; Weir et al., 1991). In addition, advanced glycation products resulting from non-enzymatic reactions of reducing sugars with proteins can quench NO. Aminoguanidine treatment,

which inhibits production of advanced glycation products, prevented impaired depressor responses to acetylcholine in diabetic rats (Bucala et al., 1991). Polyol pathway activity in galactosaemia would be expected to enhance glycation because products associated with metabolic changes (e.g. pentoses) are good substrates and the process is also accelerated by oxidative stress (Baynes, 1991). However, it is not clear whether the duration of galactosaemia in this investigation would allow sufficient advanced glycation end product formation to alter endothelium-dependent relaxation (Bucala et al., 1991), although the potential importance of short-term intracellular protein glycation by reactive glucose and polyol pathway-related metabolites has been stressed (Brownlee, 1992). Short-term galactosaemia did not significantly change endothelium-independent relaxation to cromakalim. However, there was a modest effect compared to ponalrestat-treated aortas. Long-term diabetes results in reduced relaxation mediated by ATPsensitwe K ÷ channels (Kamata et al., 1989b) that is prevented by aldose reductase inhibition (Cameron and Cotter, 1992a). This effect may depend on elevated voltage-dependent calcium channel number or activity (White and Carrier, 1990). Thus, it is possible that smooth muscle relaxation abnormalities develop more slowly than endothelium-dependent dysfunction in galactosaemia and diabetes. In conclusion, the data provide support for a polyol-pathway-related reduction in NO production that occurs early in experimental galactosaemia and parallels the development of vascular-mediated abnormalities in some other tissues (Cameron et al., 1992a). Similar findings have been described for experimental diabetes (Cameron and Cotter, 1992a) and could be important for complications such as neuropathy (Cameron at al., 1989a; 1991) and myopathy (Cameron et al., 1989b; 1990; Cotter et al., 1992, 1993). Thus, some beneficial effects of aldose reductase inhibition in diabetes and galactosaemia could be explained by a vascular action.

Acknowledgement We are grateful for Dr. Don Mlrrlees of ICI Pharmaceuticals for the generous gift of ponalrestat

References

Baynes, J.W, 1991, Role of oxidative stress m the developmentof comphcaUons in diabetes, Diabetes 40, 405. Boulton,A J.M., S Levmand J. Comstock,1990,A multtcentretrial of the aldose-reductaseinhibitor, tolrestat, in pahents w~thsymptomatic diabetic neuropathy,Dlabetologla33, 431

52 Brownlee, M , 1992, Glycatlon products and the pathogenesis of diabetic comphcations. Diabetes Care 15, 1835. Bucala, R., K.J. Tracey and A. Cerami, 1991, Advanced glycosylation products quench nitric oxide and mediate defective endotheliumdependent vasodilatlon in experimental diabetes, J Chn. Invest 87, 432. Cameron, N E. and M.A Cotter, 1992a, Impaired contraction and relaxation in aorta from streptozotocm-dlabetic rats role of the polyol pathway, Dmhetologia 35, 1011 Cameron, N.E. and M A Cotter, 1992b, Dissociation between biochemical and functional effects of the aldose reductase inhibitor ponalrestat on peripheral nerve in experimental diabetes, Br J Pharmacol. 107, 939 Cameron, N E , M.B. Leonard, I Ross and P Whiting, 1986, The effects of sorbind on peripheral nerve conduction velocity, polyol concentrations and morphology m the streptozotocln-diabetlc rat, Dlabetologla 29, 168 Cameron, N E., M A. Cotter and S Robertson, 1989a, The effect of aldose reductase inhibition on the pattern of nerve conduction deficits in diabetic rats, Q.J Exp. Physiol. 74, 917 Cameron, N.E., M.A Cotter and S Robertson, 1989b, Contractile properties of cardiac papillary muscle m streptozotocln-diabetic rats and the effects of aldose reductase inhibition, Diabetologia 32, 365 Cameron, N E , M A. Cotter and S Robertson, 1990, Changes in skeletal muscle contractile properties in streptozocln-induced diabetic rats and role of polyol pathway and hypolnsuhnemla, Diabetes 39, 460 Cameron, N.E, M.A Cotter and P A Low, 1991, Nerve blood flow m early experimental diabetes in rats relation to conduction deficits, Am. J. Physiol. 261, E1 Cameron, N E , M A. Cotter, S Robertson and D Cox, 1992a, Muscle and nerve dysfunction in rats with experimental galactosaemia, Exp. Physiol. 77, 89. Cameron, N E , M A Cotter, K C Dines and A. Love, 1992b, Effects of ammoguamdine on peripheral nerve function and polyol pathway metabolites in streptozotocm-dmbetlc rats, Dlabetologia 35, 946 Cotter, M A , N E. Cameron and S Robertson, 1992, Polyol pathway mediated changes in cardmc muscle contractile properties studles in streptozotocln-dmbetlc and galactose-fed rats, Exp. Physiol. 77, 829 Cotter, M A., N E Cameron, S. Robertson and I Ewing, 1993, Polyol pathway related skeletal muscle contractile and morphological abnormalities in diabetic rats, Exp Physiol. 78, 139 Davldson, W.S. and D G Murphy, 1985, Aldehyde reductases and their involvement m muscular dystrophy, in: Enzymology of Carbonyl Metabolism 2 Aldehyde Dehydrogenase, Aldo-keto Reductase, and Alcohol Dehydrogenase, eds. T.G Flynn and H Wemer (Llss, New York) p 251 Durante, W , A.K. Sen and F A Sunahara, 1988, Impairment of endothehum-dependent relaxation in aortae from spontaneously diabetic rats, Br J Pharmacol 94, 463 Dvornik, D., 1987, Hyperglycemm in the pathogenesls of dmbetlc complications, in. Aldose Reductase Inhibition An Approach to the Prevention of Diabetic Comphcatlons, ed. D Porte (Biomedical Information Corporation, New York) p 69 Funakawa S., T. Okahara, M. Imanlshl, T. Komorl, K Yamamoto and Y. Tochmo, 1983, Renln-angiotensin system and prostacychn biosynthesis in streptozotocln-diabetlc rats, Eur J Pharmacol. 94, 27 Hagen, A.A, Y Shlrasawa and R P White, 1985, Experimental diabetes" reduction of serotonin-induced vasoconstriction by meclofenamlc acid in vitro, Pharmacology 30, 197 Hattorx, Y., H. Kawasakl, A Kazuhlro and M. Kanno, 1991, Superoxide dlsmutase recovers altered endothehum-dependent relaxation in diabetic rat aorta, Am. J Physiol 261, H1086

Head, R J , P A Longhurst, R L Panek and R E. Stitzel, 1987, A contrasting effect of the diabetic state upon the contractile responses of aortic preparations from the rat and rabbit, Br J Pharmacol. 91, 275. Kamata, K., N Mlyata, Y. Kasuya, 1989a, Impairment of endothehum-dependent relaxation and changes in levels of cyclic GMP in aorta from streptozotocm-mduced diabetic rats, Br J Pharmacol 97, 614. Kamata, K , N. Mlyata, Y Kasuya, 1989b, Functional changes in potassium channels in aortas from rats with streptozotocm-mduced diabetes, Eur. J Pharmacol 166, 319. Kashlwagl, A and R Kakkawa, 1991, Abnormal glutathione redox cycle in cultured human endothelial cells for the pathogenesis of diabetic vascular complication, in' Diabetes 1991, eds. H Rlfkm, J A Colwell and S.I Taylor (Excerpta Medica, Elsevier, Amsterdam) p 679. Larklns, R.G. and M.E Dunlop, 1992, The link between hyperglycaemla and diabetic nephropathy, Diabetologla 35, 499. Low, P A and J D Schmelzer, 1983, Peripheral nerve conduction studies in galactose-poisoned rats, J Neurol. Scl 59, 415 Mahan, W.G., L.K Simmons and G.M Sharpe, 1991, Mechanism of impaired responses of cerebral arterioles during diabetes melhtus, Am. J. Physiol. 260, H319. McVelgh, G E , G M Brennan, G.D Johnston, B J McDermott, L.T McGrath, W.R Henry, J W Andrews and J.R Hayes, 1992, Impaired endothehum-dependent and independent vasoddatlon in patients with Type 2 (non-msuhn-dependent) diabetes melhtus, Diabetologla 35, 771 Moncada, S., Palmer, R and E. Higgs, 1991, Nitric oxide physiology, pathophysiology and pharmacology, Pharmacol Rev. 43, 109. Moore, P.G. and D E. Edwards, 1965, Standard Statistical Calculations (Pitman, London) p 13 Mulhern, M and J R Docherty, 1989, Effects of experimental diabetes on the responsiveness of rat aorta, Br J Pharmacol. 97, 1007 Ohtaka, M., M. Tawata, Y. Hosaka and T Onaya, 1992, Glucose modulation of aldose reductase mRNA expression and its activity in cultured calf pulmonary artery endothelial cells, Dlabetologla 35, 730 Oyama, Y., H. Kawasaki, Y. Hattori and M Kanno, 1986, Attenuation of endothellum-dependent relaxation in aorta from dmbetic rats, Eur J. Pharmacol 131, 75 Pelper, G M and G J. Gross, 1988, Oxygen free radicals abolish endothehum-dependent relaxation m dmbetlc rat aorta, Am J Physiol. 255, H825 Richards, J M , I F Gibson and W Martin, 1991, Effects of hypoxla and metabolic lnhibitors on production of prostacychn and endothehum-derlved relaxing factor by pig aortic endothelial cells, Br. J Pharmacol 102, 203 Roblson, W.G., P.F. Kador and Y Akagi, 1986, Prevention of basement membrane thickening in retinal capillaries by a novel inhibitor of aldose reductase, tolrestat, Diabetes 35, 295. Saenz de Tejada, I, I. Goldstem, K. Azadzol, R J. Krane and R.A. Cohen, 1989, Impaired neurogenlc and endothehum-mediated relaxation of penile smooth muscle from diabetic men with impotence, N Eng. J Med. 320, 1025. Scarborough, N L. and G O Carrier, 1983, Increased alpha-2 adrenoceptor mediated vascular contraction in diabetic rats, J Autonom. Pharmacol. 3, 177 Strlbhng, D , D.J Mlrrlees, H.E Harrison and D.C.N Earl, 1985, Properties of ICI 128,436, a novel aldose reductase inhibitor and its effects on diabetic complications in the rat, Metabolism 34, 336. Taylor, P.D, A.L. McCarthy, C.R Thomas and L Poston, 1992, Endothehum-dependent relaxation and noradrenahne sensitivity m mesenteric resistance arteries of streptozotocm-mduced diabetic rats, Br J Pharmacol 107, 393

53 Tesfamarlam B. and R.A Cohen, 1992, Free radicals me&ate endothelial cell dysfunction caused by elevated glucose, Am. J Physiol. 263, H321. Tesfamanam, B, J.A. Jakubowskl and R.A. Cohen, 1989, Contraction of diabetic rabbit aorta caused by endothelium-derwed PGH2-TxA2, Am J Physiol. 257, H1327. Tesfamanam B., M L Brown and R.A. Cohen, 1992, Aldose reductase and myo-lnosltol in endothelial cell dysfunction caused by elevated glucose, J Pharmacol Exp Therapeut 263, 153. Tesfamanam B, J.J Palaclno, R M Welsbrod and R A. Cohen, 1993, Aldose reductase inhibition restores endothelial cell function m diabetic rabbit aorta, J. Cardlovasc Pharmacol. 21, 205 Tomhnson, K C , S M. Gardmer, R A. Hebden, T. Bennett, 1992,

Functional consequences of streptozotocm-mduced &abetes melhtus, with particular reference to the cardiovascular system, Pharmacol. Rev 44, 103 Ward, K K., P.A. Low, J.D Schmelzer and D.W. Zochodne, 1989, Prostacyclin and noradrenahne m peripheral nerve of chronic experimental diabetes in rats, Brain 112, 197 Weir, C J , I.F. Gibson and W Martin, 1991, Effects of metabolic Inhlbltors on endothehum-dependent and endothehum-lndependent vasodtlatlon of rat and rabbit aorta, Br J Pharmacol. 102, 162. White, R.E. and G.O. Carrier, 1990, Vascular contraction induced by activation of membrane calcium ion channels is enhanced In streptozotocln-dlabetes, J Pharmacol Exp Ther 253, 1057