@prefix vivo: . @prefix edm: . @prefix ns0: . @prefix dcterms: . @prefix dc: . @prefix skos: . vivo:departmentOrSchool "Medicine, Faculty of"@en, "Anesthesiology, Pharmacology and Therapeutics, Department of"@en ; edm:dataProvider "DSpace"@en ; ns0:degreeCampus "UBCV"@en ; dcterms:creator "Bhanot, Sanjay"@en ; dcterms:issued "2009-04-15T23:51:54Z"@en, "1995"@en ; vivo:relatedDegree "Doctor of Philosophy - PhD"@en ; ns0:degreeGrantor "University of British Columbia"@en ; dcterms:description """Although considerable data lend support to the association between insulin resistance, hyperinsulinemia and hypertension, the precise nature of this relationship remains elusive. In the present study, we examined the proposition that these metabolic defects contribute causally to the development of high blood pressure. Essentially, if these metabolic abnormalities were responsible for an increase in blood pressure, then drugs that improve these metabolic defects should also attenuate hypertension. We, therefore, examined the effects of three drugs (that are known to enhance insulin action) on insulin sensitivity, plasma insulin levels and blood pressure in two established models of experimental hypertension: (a) the spontaneously hypertensive rat and (b) the fructosehypertensive rat, where hypertension is induced in normotensive rats by feeding them a high fructose diet. The drug interventions were: (a) vanadyl sulfate, the (+lV) form of the trace element vanadium (b) bis(maltolato)oxovanadium(IV) (BMOV), an organic vanadium complex and (C) pioglitazone, a thiazolidinedione derivative that enhances peripheral insulin action. In separate studies, 6 week old spontaneously hypertensive rats (SHR) and their Wistar Kyoto (WKY) controls were started on chronic oral treatment with vanadyl sulfate (0.4-0.6 mmol/kg/day), BMOV (0.35-0.45 mmol/kg/day) or pioglitazone (0.01 -0.02 mmol/kg/day). All 3 drugs caused a sustained decrease in plasma insulin concentration in the hyperinsulinemic SHR without causing any change in the WKY. Surprisingly, all the drugs caused a marked decrease in systolic blood pressure in the SHR without causing any change in the WKY. Restoration of plasma insulin levels in the drug-treated SHR to levels that existed in the untreated SHR reversed the effect/s of the drugs on blood pressure. Low dose euglycemic insulin clamps (14 pmol/kg/min) conducted in conscious, fasted rats revealed that insulin sensitivity, expressed as steady state glucose clearance per unit of plasma insulin, was higher in the untreated SHR as compared to the untreated WKY. Although BMOV further enhanced insulin sensitivity in the SHR, pioglitazone had no effect on insulin sensitivity in the SHR or WKY. Fructose feeding induced hyperinsulinemia and increased blood pressure in normotensive Sprague Dawley rats. Vanadyl sulfate (0.4-0.6 mmol/kg/day) prevented the rise in plasma insulin and blood pressure in the fructose-fed rats. Again, restoration of plasma insulin concentration in the fructose-vanadyl treated rats to pie-treatment levels reversed the effects of vanadyl sulfate on blood pressure. Low dose insulin clamps demonstrated that insulin sensitivity was reduced in the fructose-fed rats. Vanadyl caused a marked enhancement in insulin sensitivity in the fructose-fed rats without any change in the control group. In conclusion: (i) SHR are not insulin-resistant but rather are more insulinsensitive than the WKY (ii) SHR are hyperinsulinemic and drug interventions that decrease hyperinsulinemia also attenuate hypertension in the SHR (iii) The effect of the drugs on blood pressure can be reversed by restoring plasma insulin levels in the drug-treated SHR to those observed in their untreated counterparts (iv) The antihypertensive effects of pioglitazone in the SHR are independent of its effects on insulin sensitivity, which suggests that hyperinsulinemia may be unrelated to insulin resistance in the SHR (v) Vanadyl sulfate completely prevents fructose induced insulin resistance, hyperinsulinemia and hypertension. These data indicate that either hyperinsulinemia may contribute to the development of high blood pressure in both the SHR and the fructose-hypertensive rats or that the underlying mechanism is closely related to the expression of both these disorders."""@en ; edm:aggregatedCHO "https://circle.library.ubc.ca/rest/handle/2429/7201?expand=metadata"@en ; dcterms:extent "2038240 bytes"@en ; dc:format "application/pdf"@en ; skos:note "INSULIN AND HYPERTENSION: A PHARMACOLOGICAL PERSPECTIVEbySANJAY BHANOTM.D., Government Medical College, Panjabi University Patiala, 1985A THESIS SUBMITTED IN PARTIAL FULFILLMENT OFTHE REQUIREMENTS FOR THE DEGREE OFDOCTOR OF PHILOSOPHYinTHE FACULTY OF GRADUATE STUDIESDivision of Pharmacology and ToxicologyFaculty of Pharmaceutical SciencesWe accept this thesis as conformingto the required standardTHE UNIVERSITY OF BRITISH COLUMBIASeptember 1995© Sanjay Bhanot 1995In presenting this thesis in partial fulfilment of the requirements for an advanceddegree at the University of British Columbia, I agree that the Library shall make itfreely available for reference and study. I further agree that permission for extensivecopying of this thesis for scholarly purposes may be granted by the head of mydepartment or by his or her representatives. It is understood that copying orpublication of this thesis for financial gain shall not be allowed without my writtenpermission.(Signature)________________________________Department of ‘é1 •The University of British ColumbiaVancouver, CanadaDate / 915DE-6 (2/88)11ABSTRACTAlthough considerable data lend support to the association between insulinresistance, hyperinsulinemia and hypertension, the precise nature of thisrelationship remains elusive. In the present study, we examined the propositionthat these metabolic defects contribute causally to the development of high bloodpressure. Essentially, if these metabolic abnormalities were responsible for anincrease in blood pressure, then drugs that improve these metabolic defectsshould also attenuate hypertension. We, therefore, examined the effects of threedrugs (that are known to enhance insulin action) on insulin sensitivity, plasmainsulin levels and blood pressure in two established models of experimentalhypertension: (a) the spontaneously hypertensive rat and (b) the fructose-hypertensive rat, where hypertension is induced in normotensive rats by feedingthem a high fructose diet. The drug interventions were: (a) vanadyl sulfate, the(+lV) form of the trace element vanadium (b) bis(maltolato)oxovanadium(IV)(BMOV), an organic vanadium complex and (C) pioglitazone, a thiazolidinedionederivative that enhances peripheral insulin action.In separate studies, 6 week old spontaneously hypertensive rats (SHR)and their Wistar Kyoto (WKY) controls were started on chronic oral treatment withvanadyl sulfate (0.4-0.6 mmol/kg/day), BMOV (0.35-0.45 mmol/kg/day) orpioglitazone (0.01 -0.02 mmol/kg/day). All 3 drugs caused a sustained decreasein plasma insulin concentration in the hyperinsulinemic SHR without causing anychange in the WKY. Surprisingly, all the drugs caused a marked decrease insystolic blood pressure in the SHR without causing any change in the WKY.Restoration of plasma insulin levels in the drug-treated SHR to levels that existedin the untreated SHR reversed the effect/s of the drugs on blood pressure. Lowdose euglycemic insulin clamps (14 pmol/kg/min) conducted in conscious, fasted111rats revealed that insulin sensitivity, expressed as steady state glucose clearanceper unit of plasma insulin, was higher in the untreated SHR as compared to theuntreated WKY. Although BMOV further enhanced insulin sensitivity in the SHR,pioglitazone had no effect on insulin sensitivity in the SHR or WKY.Fructose feeding induced hyperinsulinemia and increased blood pressurein normotensive Sprague Dawley rats. Vanadyl sulfate (0.4-0.6 mmol/kg/day)prevented the rise in plasma insulin and blood pressure in the fructose-fed rats.Again, restoration of plasma insulin concentration in the fructose-vanadyl treatedrats to pie-treatment levels reversed the effects of vanadyl sulfate on bloodpressure. Low dose insulin clamps demonstrated that insulin sensitivity wasreduced in the fructose-fed rats. Vanadyl caused a marked enhancement ininsulin sensitivity in the fructose-fed rats without any change in the control group.In conclusion: (I) SHR are not insulin-resistant but rather are more insulin-sensitive than the WKY (ii) SHR are hyperinsulinemic and drug interventions thatdecrease hyperinsulinemia also attenuate hypertension in the SHR (iii) The effectof the drugs on blood pressure can be reversed by restoring plasma insulin levelsin the drug-treated SHR to those observed in their untreated counterparts (iv) Theantihypertensive effects of pioglitazone in the SHR are independent of its effectson insulin sensitivity, which suggests that hyperinsulinemia may be unrelated toinsulin resistance in the SHR (v) Vanadyl sulfate completely prevents fructoseinduced insulin resistance, hyperinsulinemia and hypertension. These dataindicate that either hyperinsulinemia may contribute to the development of highblood pressure in both the SHR and the fructose-hypertensive rats or that theunderlying mechanism is closely related to the expression of both these disorders.ivTABLE OF CONTENTSAbstract iiTable of Contents ivList of Tables viiList of Figures viiiList of Abbreviations ixAcknowledgments xDedication xiINTRODUCTION I(I) Essential hypertension: a metabolic disease I(II) Insulin and hypertension: epidemiology 3(I) Studies supporting the link between insulin and hypertension 3(ii) Studies refuting the link between insulin and hypertension 4(iii) Prospective studies in normotensive offspringsof hypertensive subjects 6(Ill) Insulin and hypertension: experimental evidence 8(i) Human studies 8(ii) Animal studies 11(IV) Insulin and hypertension: the possible links 13(i) Insulin induced antinatriuresis 13(ii) Insulin and the sympathetic nervous system 14(iii)Trophic effects of insulin 15(iv)Hemodynamic effects of insulin 16(v) Insulin and the intracellular cation transport systems 18(V) Specific research problem and research strategy 20V(I) The research problem 20(ii) Rationale 21(iii) Experimental approach 21(iv) Experimental models 22(v) Drug interventions 24(VI) Working hypotheses 27MATERIALS AND METHODS 28(I) Studies in the spontaneously hypertensive rat(A) Research design and experimental protocols 28(i) Studies with vanadyl sulfate 28(a) Prevention study 28(b) Reversal study 28(c) Pair-feeding study 29(d) Insulin implant study 29(ii) Studies with bis(maltolato)oxovanadium(IV) 30(iii) Studies with pioglitazone 31(B) Methodology 32(i) Blood pressure measurement 32(a) Studies with vanadyl sulfate 32(b) Studies with MOV and pioglitazone 33(ii) Euglycemic hyperinsulinemic clamp technique 34(a) Principle underlying the method 34(b) Calculation of insulin sensitivity 37(c) Underestimation of hepatic glucose output 38(II) Studies in the fructose-hypertensive rat 42(C) Biochemical analyses 43v(D) Statistical analyses 45RESULTS(I) Studies in the spontaneously hypertensive rat(i) Studies with vanadyl sulfate 47(ii) Studies with bis(maltolato)oxovanadium(IV) 65(iii) Studies with pioglitazone 72(II) Studies in the fructose-hypertensive rat 81DISCUSSION 88(I) Effects of vanadyl sulfate in spontaneously hypertensive rats 88(II) Effects of BMOV in spontaneously hypertensive rats 94(Ill) Effects of pioglitazone in spontaneously hypertensive rats 100(IV) Effects of vanadyl sulfate in fructose-hypertensive rats 106CONCLUSIONS 113BIBLIOGRAPHY 115viiLIST OF TABLESTable Page1. Various parameters from the experimentalgroups in the exogenous insulin study 602. Plasma glucose (mmol/L) in the prevention,reversal and pair-feeding studies 613A. Food and fluid intake in the prevention study 623B. Food and fluid intake in the reversal study 633C. Food and fluid intake in the pair-feeding study 644. Characteristics of animals at 9-11 weeks ofage in the BMOV study 705. Results of glucose clamps in the BMOV study 716. Characteristics of animals at 9-1 1 weeks ofage in the pioglitazone study 787A. Results of glucose clamps at a high insulin infusionrate in the pioglitazone study 797B. Results of glucose clamps at a low insulin infusionrate in the pioglitazone study 808. Characteristics of animals at weeks 6 (baseline)and 12 in the fructose study 869. Results of glucose clamps in the fructose study 87viiiLIST OF FIGURESFiçure Pacie1. Body weights in the pair-feeding study 502. Plasma insulin levels and blood pressure in theprevention study 523. Plasma insulin levels and blood pressure in thereversal study 544. Plasma insulin levels and blood pressure in thepair-feeding study 565. Body weights in the BMOV study 666. Plasma insulin levels and blood pressure in theBMOV study 687. Body weights in the pioglitazone study 748. Plasma insulin levels and blood pressure in thepioglitazone study 769. Plasma insulin levels and blood pressure in thefructose study 84ixLIST OF ABBREVIATIONSATPase adenosine triphosphataseBMOV bis(maltolato)oxovanadium( IV)BP blood pressureCa+2 calciumcGMP guanosine 3’-5’-cyclic monophosphateDNA deoxyribonucleic acidEC50 half-maximally effective concentrationED50 half-maximally effective doseEGTA ethylene glycol bis(b-aminoethyl ether-P-N, N, N’, N’-tetraacetic acid)GINF glucose infusion rateGOT glutamic oxaloacetic transaminaseGPT glutamic pyruvic transaminaseHGO hepatic glucose outputK+ potassiumLD50 dose that is lethal to 50% of the treated animalsNa+ sodiumNIDDM non insulin dependent diabetes mellitusPE polyethyleneRa rate of glucose appearanceRd rate of glucose utilization = rate of total glucose disposalSHR spontaneously hypertensive ratsSTZ streptozotocinVSM vascular smooth muscleWKY Wistar-Kyoto ratsxACKNOWLEDGMENTSFirst and foremost, I would like to express my heartfelt gratitude to my supervisor,Dr. John H. McNeill, who was instrumental in making my present researchendeavor extremely rejuvenating and rewarding. His thoughtfulness, patience,understanding, encouragement and the trust he reposed in me made my presentacademic pursuit one of the most memorable experiences of my life.I would like to thank my Ph.D. research committee members, Dr. Jack Diamond,Dr. Kath MacLeod, Dr. Cathy Pang and Dr. Ronald Reid for their valuable inputand constructive advice towards my research project.I consider myself very fortunate to have worked with some wonderful peopleduring the course of my graduate program. The first such person whom I wouldlike to thank is Ms. Violet Yuen, who was always extremely helpful and whosetechnical expertise is commendable. I also had the opportunity to work with anexcellent summer student, Ms. Aspasia Michoulas, whose technical competenceand work ethic simply amazed me. I am indebted to Ms. Xunsheng Chen, whotaught me a lot of biochemical techniques that should prove invaluable to me inmy future research endeavors.I would like to express my sincere gratitude to Ms. Mary Battell, our laboratorymanager, for being very helpful and understanding and for her consistent supportthroughout my stay in the laboratory. The expert secretarial assistance of Ms.Sylvia Chan is gratefully acknowledged. I would also like to thank all othercolleagues and faculty members who assisted me during various aspects of myresearch project. In particular, I would like to thank Dr. Chris Orvig, Dr. RobThies, Dr. Katherine Thompson, Dr. Yong-Jiang Hei, Dr. Soter Dai, Dr. MichaelBryer Ash, Mr. Subodh Verma, Ms. Lynn Weber and Mr. Sepehr Farahbakhshian.I also appreciate the help of the various personnel in the animal care and thepurchase/ordering facilities. The assistance of the personnel in the Dean’s officeis also thankfully acknowledged.I would like to acknowledge the Medical Research Council of Canada forsupporting me with a fellowship award throughout my Ph.D. program (1991 -1 995).Finally, I consider myself blessed in having a wonderful wife, Meenakshi Bhanotand a gorgeous daughter, Puja Bhanot, who have not only made my life so joyousand blissful but have also helped me to pursue my academic goals without anyreservation.xDEDICATIONTo my grandfather Mr. L. R. Sharmawho was my pillar of strength and source of inspirationwho brought joy and happiness to all those he knewwhose love, warmth and affection will always be cherished1INTRODUCTION(I) ESSENTIAL HYPERTENSION: A METABOLIC DISEASEOne of the paradoxes in medicine is the inability of effectiveantihypertensive drugs to reduce the increased risk of coronary artery disease inhypertensive subjects (Collins et at, 1990; Koerner et al., 1982; Wikstrand et al.,1988). Several reports indicate that although lowering BP reduces the incidenceof cerebrovascular disease, it does not alter the mortality and morbidityassociated with coronary artery disease. MacMahon et al., by applying thetechnique of meta-analysis, reviewed many of these studies (MacMahon et at,1989) and found that treating hypertension reduced all cause mortality by 11%and non fatal stroke by 39%, both of which were statistically significantreductions. However, the incidence of coronary artery disease decreased by amere 8%, a difference that did not attain statistical significance. Although thisapparent paradox has received considerable attention, most of it has focused onthe notion that conventional antihypertensive therapy is associated with adverseside effects on lipid metabolism that could accentuate the risk for coronary arterydisease (MacMahon et a!., 1985; Weinberger, 1986). Although this contention isbased on good evidence, it probably addresses only one aspect of a multifacetedproblem.A finding that was overlooked until the last decade is that essentialhypertension per se is associated with multiple metabolic abnormalities incarbohydrate and lipid metabolism (Ferrannini and Natali, 1991; Reaven 1991b).The metabolic defects associated with hypertension include insulin resistance,hyperinsulinemia and dyslipidemia (DeFronzo and Ferrannini 1991; Ferrannini etal., 1987; Reaven 1991a,b), all of which have been shown to increase the risk for2coronary ischemic events (Kannell et a!., 1991; MRFIT research group 1986;Sowers et al., 1991). What arouses particular concern are findings suggestingthat these metabolic abnormalities persist and even worsen with conventionalantihypertensive agents, particularly the thiazide diuretics and the beta adrenergicblockers (Lind et al., 1994; Pollare eta!., 1989; Skarfors et a!., 1989; Weinberger1986). Of these metabolic defects, two that seem to be frequently associated withhypertension are insulin resistance (or resistance to the glucoregulatory effects ofinsulin) and hyperinsulinemia (Ferrannini and Natali, 1991; Reaven 1994;Weidmann et a!., 1993). These defects in glucose metabolism are associatedwith an atherogenic risk profile and evidence suggests that they may play a role inthe development of hypertension, dyslipidemia and atherosclerosis (DeFronzoand Ferrannini, 1991; Haffner et a!., 1992; Reaven 1988; Reaven 1994).Lending further support to this hypothesis are recent reports indicating thatinsulin resistance and hyperinsulinemia are present in the normotensive offspringof patients with hypertension as early as the second decade of life (Allemann andWeidmann 1995; Grunfeld et a!., 1994). These changes, therefore, antedate theincrease in BP and are not found in patients with secondary forms of hypertension(Shamiss et a!., 1992). The observation that insulin resistance andhyperinsulinemia occur not only in untreated human hypertensives (Ferrannini eta!., 1987; Shen et a!., 1988), but also in several rodent models of hypertension(Kotchen et al., 1991; Mondon and Reaven 1988; Reaven et a!., 1991 a)strengthens the contention that these abnormalities are intrinsically linked withhypertension and are not mere coincidental findings. Studies indicate thathyperinsulinemia is an independent risk factor for coronary artery disease(Ducimetiere et a!., 1980; Pyorala 1979) and that even a small degree of glucoseintolerance significantly increases the risk of developing coronary heart disease3(Fuller et al., 1980). Another finding that strongly supports this intriguingassociation is that patients with hypertension who have microvascular angina areinsulin-resistant and hyperinsulinemic when compared to hypertensive subjectswithout ischemic heart disease (Botker et a!., 1993; Dean et a!., 1991; Fuh et a!.,1993). Such findings are provocative and further research efforts are needed todefine these inter-relationships precisely. Given the fact that we still do not havea satisfactory answer to the coronary disease paradox, such issues deservespecial consideration. The primary objective of the current investigation was toexplore the relationship between insulin resistance, hyperinsulinemia andhypertension. Specifically, the proposition that these metabolic abnormalities maybe causally related to hypertension was examined. Before addressing thespecific research problem, it is important to review some of the epidemiologicalevidence linking insulin to high BP, which is discussed in the following section.(II) INSULIN AND HYPERTENSION: EPIDEMIOLOGY(i) Studies supporting the link between insulin and hypertensionWelborn et al. were the first to suggest an association between defectiveinsulin action and hypertension (Welborn et a!., 1966). They examined 19subjects with essential hypertension (10 of whom were untreated) and foundhigher fasting and postprandial insulin levels in both treated and untreatedhypertensive patients when compared to control, normotensive subjects. Thisinitial finding was subsequently confirmed in several large epidemiologicalstudies. Berglund et al. screened middle aged Swedish hypertensive men whohad no other overt clinical disease and who were not taking any antihypertensivetherapy (Berglund et al., 1976). They reported that both fasting insulin andglucose concentrations were higher in hypertensive patients as compared to their4normotensive controls. These results were confirmed in another extensive studyfrom Israel (Modan et al., 1985). Results from the study by Modan et al., whichinvolved about 2500 subjects, demonstrated that hypertensive patients exhibitedfasting and postprandial hyperinsulinemia independent of obesity, age ormagnitude of glucose tolerance. This observation has since been confirmed byseveral epidemiological and clinical studies (Lucas et a!., 1985; Manicardi et a!.,1986; Swislocki et a!., 1989), which have documented the presence ofirregularities in carbohydrate metabolism in young, non obese subjects withuntreated, uncomplicated essential hypertension. In a different approach, Singeret al. compared the day long insulin and glucose profiles in response to standardmeals in untreated hypertensive patients and found that the insulin response aftereach meal was markedly increased in hypertensive patients (Singer et a!., 1985).Results from their study indicated that essential hypertensive subjects not onlyhad an increased insulin response to a glucose challenge but that thisexaggerated insulin response occurred after every meal that they consumed.(ii) Studies refuting the link between insulin and hypertensionNot all hypertensive patients are hyperinsulinemic and several studiesdemonstrate a weak correlation between insulin and high BP (Asch et aL, 1991;Collins et a!., 1990). In addition, it was reported that insulinemia was not relatedto the prevalence of hypertension in Pima Indians after controlling for obesity anddrug treatment (Saad et a!., 1990). Mbanya et al. demonstrated that plasmainsulin levels were similar between hypertensive and normotensive subjects andthat insulin levels were increased only in those hypertensive patients who alsohad NIDDM (Mbanya et a!., 1988). In many studies, the correlation betweenplasma insulin levels and hypertension became weak after accounting for obesity5(Meehan et a!., 1993b). Due to a virtual flurry of reports examining therelationship between insulin and hypertension, this area has attractedconsiderable discussion and debate. In an effort to resolve this problem, Denkerand Pollock performed a meta-analysis on the various studies reported betweenthe years 1983-1991. They included only those studies that were conducted inuntreated hypertensives or in which an adequate washout period was included aspart of the study protocol and in which the subjects were neither glucoseintolerant or diabetic (Denker and Pollock, 1992). The results of their metaanalysis indicated that fasting serum insulin concentration was strongly correlatedwith both systolic and diastolic BP when data from all the studies were pooledtogether to yield a statistically meaningful result.Qespite the controversy that exists due to the disparate results of almost 50different studies reported in the last 10 years, several consistencies haveemerged regarding this association. First, as many as 50% of hypertensivepatients appear to be insulin-resistant and hyperinsulinemic when compared toage and weight matched normotensive controls (Denker and Pollock 1992;Zavaroni et a!., 1992). There is ethnic variation in this relationship such thatsignificant correlations between these metabolic defects and hypertension exist inCaucasian, Hispanic White and Japanese populations but not in Afro-Americansor Pima Indians (Miura et al., 1995; Saad et a!., 1991). However, what must beconsidered is the fact that correlations do not prove causality. A lack ofcorrelation in an epidemiological study does not rule out a role of insulin in thedevelopment of hypertension in that population. For example, although Saad etal. found that insulin concentration and hypertension were not correlated in AfroAmericans, subsequent studies demonstrated that hypertensive Afro-Americans6were both insulin-resistant and hyperinsulinemic when compared withnormotensive controls (Falkner et a!., 1990; Falkner et a!., 1993).Second, it was reported in several studies that the correlation betweeninsulin and BP became weak or even non significant after accounting for obesity.This led some investigators to propose that obesity may be the primary factor thatcauses or modulates insulin resistance in hypertensive subjects. Althoughseveral studies had indicated that insulin resistance was present even in lean,untreated hypertensive patients, the role of hypertension in modulating insulinresistance in obese subjects remained undefined. It was not clear whether theinsulin resistance observed in obese, hypertensive subjects was due to obesityper se or high BP or both. The answer to this dilemma came from a recent studyin which insulin resistance was quantified in 5 different groups of subjects:normotensive-obese, normotensive-nonobese, hypertensive-obese, hypertensivenonobese and hypertensive-obese with NIDDM (Maheux et al., 1994). Theirresults unequivocally demonstrated that the effects of obesity, hypertension andNIDDM on insulin resistance are additive and that each one of these diseasesindependently contributes towards resistance to insulin’s glucoregulatory effects.Taken together, current evidence indicates that insulin resistance andhyperinsulinemia are present in a substantial proportion of hypertensive patients.Furthermore, this relationship is stronger among some ethnic groups than inothers and is independent of age, obesity or drug treatment.(iii) Prospective studies in normotensive offsprings of hypertensive subjectsAlthough insulin resistance and hyperinsulinemia correlated positively withhigh BP in a considerable proportion of hypertensive patients, it was not clearwhether these defects were primary or whether they were a consequence of7hypertension. In the first report that dealt with this issue, insulin sensitivity andplasma insulin levels in normotensive offsprings of essential hypertensive parentswere compared with those obtained from age and weight matched normotensivesubjects with no parental history of hypertension (Ferrari et a!., 1991). It wasfound that young, lean, normotensive adults (mean age about 24 years) who hada positive family history of hypertension were insulin-resistant andhyperinsulinemic when compared to normotensive controls without a hypertensiveparent. Insulin sensitivity was decreased by about 28% and insulin levels wereincreased by about 15% in subjects with a parental history of hypertension, whichindicated that these defects antedate the increase in BP and are not secondary tohypertension. Almost identical results were reported by Facchini et al. one yearlater; in their study the majority of subjects were female as opposed to the studyby Ferrari et al. in which the subjects were predominantly male (Facchini et a!.,1992). These results indicated that the link between insulin and hypertension hada genetic basis that was independent of age, gender or body weight.Subsequent studies extended these observations and demonstrated thatthe reduced insulin sensitivity that antedated the increase in BP in normotensiveadults with a positive family history of hypertension was also accompanied by ahigher platelet intracellular calcium concentration (Ohno et a!., 1993). Thissuggested that disturbed intracellular calcium metabolism was also an inheritedtrait and raised the possibility that this intracellular defect could be a link betweeninsulin and hypertension. However, it was not clear whether the abnormalcalcium metabolism was a cause or consequence of insulin resistance, a questionthat remains unanswered to date. A recent study demonstrated thathyperinsulinemia was present even in young children (mean age about 14 years)who were normotensive, normolipidemic but had a positive family history of8hypertension, which strongly suggests that defects in insulin action antedate mostof the metabolic and hemodynamic abnormalities seen in hypertensive subjects(Grunfeld et a!., 1994). The observation that insulin resistance can be modifiedby environmental influences such as body weight or physical exercise (DeFronzoand Ferrannini, 1991) suggests that the final phenotypic expression of thesedefects is probably a combination of both genetic and acquired influences.Nevertheless, the studies discussed above clearly indicate that insulin resistanceis a genetically inherited trait and is not simply a consequence of increased BP.(Ill) INSULIN AND HYPERTENSION: EXPERIMENTAL EVIDENCE(I) Human studiesFerrannini et al. provided the first direct evidence that essentialhypertensive patients are insulin-resistant. By employing the euglycemichyperinsulinemic clamp technique, they studied insulin sensitivity in young, lean,untreated hypertensive subjects and found that insulin sensitivity was markedlydecreased (about 40%) in hypertensive patients when compared to age andweight matched controls (Ferrannini et al., 1987). Furthermore, they alsodemonstrated that the insulin resistance was tissue specific in that it occurredmainly in the skeletal muscle and that insulin-induced suppression of hepaticglucose production was normal in hypertensive subjects. Another interestingobservation that stemmed from their elegant study was that almost the entirereduction in insulin sensitivity could be accounted for by a decrease innonoxidative glucose disposal (glycogen synthesis) whereas glucose oxidationand suppression of lipolysis were unaltered in hypertensive patients. Thisobservation has subsequently been confirmed by several investigators (Cepaldoet al., 1991; Pollare et a!., 1990; Rooney et a!., 1992).9Other studies have unequivocally demonstrated that not only is insulinresistance present in hypertensive subjects but that it is not improved by loweringBP with antihypertensive drugs (Pollare et a!., 1989; Skarfors et al., 1989;Weinberger 1986). On the contrary, several studies indicate that antihypertensivetherapy worsens insulin resistance and hyperinsulinemia in essentialhypertensives (Lind et a!., 1994; Shen et a!., 1988). The view that insulinresistance is not secondary to an increase in BP is further supported by studiesindicating that insulin-mediated glucose utilization and glucose stimulated insulinsecretion is normal in patients with secondary hypertension such as reno-vascularhypertension and primary hyperaldosteronism (Shamiss eta!., 1992). More directevidence for such a link has come from studies where it was observed thatphysical training lowered BP in obese patients (without any change in bodyweight), but only in those patients who were hyperinsulinemic before the start ofthe training program (Krotkiewski et a!., 1979).However, it is important to remember that insulin resistance is not alwaysassociated with hypertension and vice versa. As was discussed earlier, there isethnic variability in this relationship such that the link appears stronger inCaucasians and Hispanic Whites. Two other arguments that have beenfrequently cited as evidence against the hypothesis that insulin plays a role in thegenesis of hypertension deserve mention. First, insulin when administeredacutely is a vasodilator and does not cause an increase in BP (Anderson andMark, 1993). The vasodilator actions of insulin will be addressed in detail in asubsequent section; however, the inability of insulin to increase BP acutely doesnot go against the view that insulin may modulate BP chronically. It rathersuggests that if insulin is a vasoactive hormone, then resistance to its vasodilatoryeffects may manifest as an increase in peripheral vascular resistance and thereby10raise BP. The finding that BP falls when the dose of insulin is decreased inobese, hypertensive patients with NIDDM (Tedde eta!., 1989) and that there is anincrease in BP when insulin treatment is started in NIDDM patients (Randeree eta!., 1992) strengthens the contention that chronic insulin therapy may exertpressor effects in humans. This view is further supported by a recent studyindicating that troglitazone, a drug that improved insulin sensitivity, also loweredBP in essential hypertensive patients with diabetes mellitus (Ogihara et al., 1995).The second argument that is often advanced against the view that insulinmodulates BP is that patients with insulinoma who are hyperinsulinemic aregenerally not hypertensive (Sawicki et a!., 1992). What is overlooked whileadvancing this argument is that patients with insulinoma do not have primaryinsulin resistance, since hyperinsulinemia in such patients is accompanied bymarked hormonal counter-regulatory responses. In addition, patients withinsulinoma lack the substrate for insulin (i.e. glucose) and it has been shown thatinsulin’s effects on metabolic and vascular smooth muscle responses aredependent on the availability of glucose (Yanagisawa-Miwa eta!., 1990). Finally,in the study that addressed this issue (Sawicki et a!., 1992), hyperinsulinemia waspresent for a relatively short period (about 18 months) as opposed to theincreased levels of insulin that are present throughout the life span ofhypertensive patients.In summary, current evidence indicates that insulin resistance inhypertension is a primary defect (independent of obesity, diabetes or drugtreatment) and is tissue specific (resides primarily in skeletal muscle) andpathway specific (involves glycogen synthesis). This defect in insulin-mediatedglucose uptake may play a role in the development of hypertension or maychronically predispose a certain proportion of subjects with a specific11neurohumoral phenotype towards an increase in BP. However, the independentcontribution of insulin resistance towards an increase in BP is probably smallerand more complex than is often emphasized and to assume that insulin is directlylinked to a rise in BP in all hypertensive subjects is oversimplistic and incorrect.(ii) Animal studiesThe association between insulin and hypertension has also beendocumented in several models of rodent hypertension. These include the Dahl rat(Kotchen et a!., 1991), the spontaneously hypertensive rat (Mondon and Reaven1988; Reaven, 1991a,b), the Milan hypertensive rat (Dall-Aglio et a!., 1991) andthe fructose-hypertensive rat (Hwang et a!., 1987). All these hypertensive ratmodels, although unrelated to each other, exhibit common defects in glucosemetabolism. In Dahl rats, insulin resistance and hyperinsulinemia occur in saltsensitive as well as salt resistant animals and are independent of the salt contentof the diet (Reaven et al., 1991). In spontaneously hypertensive rats,hyperinsulinemia precedes the development of hypertension (Reaven and Chang1991); however, the presence of insulin resistance in this rat strain remainscontroversial (Buchanan et a!., 1992a; Buchanan et a!., 1992b; Frontoni et a!.,1992; Hulman et a!., 1993). Some very relevant findings have emerged fromstudies in which insulin resistance and hyperinsulinemia were induced innormotensive Sprague Dawley rats by giving them a fructose enriched diet(Hwang at al., 1987). Induction of these metabolic defects was associated with aconcomitant increase in blood pressure in these rats. Furthermore, exercisetraining (which resulted in improved insulin sensitivity) and somatostatinadministration (which decreased hyperinsulinemia) to the fructose-fed ratsattenuated the fructose-induced increase in BP in the animals (Reaven et a!.,121988; Reaven et a!., 1 989b). Although these findings do not establish causality,they do support such a link.Results obtained from studies conducted in dogs are in contrast to thosereported in rats. Acute insulin infusion in dogs did not raise BP (Liang et a!.,1982), whereas it led to a dose-dependent increase in BP in rats (Edwards andTipton, 1989). Furthermore, chronic insulin infusion in dogs for up to 4 weeks didnot cause hypertension, although it increased plasma insulin levels almost six-fold(Hall et al., I 990a). When experiments were repeated in dogs made susceptibleto hypertension by partial nephrectomy coupled with a high salt intake, insulin stilldid not cause an increase in BP (Hall et a!., 1990b). In contrast, a chronic,physiological increase in plasma insulin concentration increased BP in rats(Brands et al., 1991). Interestingly, when dogs were fed a high fat or a highfructose diet, they became insulin-resistant, hyperinsulinemic and hypertensive(Martinez et al., 1994). Thus there are species differences with regard to theeffects of insulin on BP, which may be a result of differential effects of insulin onthe sympathetic, renal or cardiovascular systems. These disparate resultsbetween dog and rodent studies support the notion that hyperinsulinemia mayincrease BP only in conjunction with the contribution of other pressor systems andmay have a different phenotypic expression in different animal species. Takentogether, although results from animal studies appear contradictory, there issufficient evidence to suggest that the link between insulin and hypertension ismore than coincidental and that insulin may cause hypertension in certain animalspecies. The obvious issue that then needs consideration pertains to the possiblemechanisms linking insulin to an increase in BP, which is addressed in the nextsection. Hyperinsulinemia in hypertension is a reflection of resistance to theperipheral uptake and utilization of glucose, with high levels of insulin needed to13maintain and sustain euglycemia in the presence of insulin resistance. It hasbeen hypothesized that the compensatory increase in plasma insulinconcentration is not a benign phenomenon but that hyperinsulinemia maycontribute towards the development of hypertension by a variety of differentmechanisms.(IV) INSULIN AND HYPERTENSION: THE POSSIBLE LINKSThis area has been the focus of intensive investigation, which hasresulted in a lot of discussion and debate. The following section deals with theinteractions of insulin with other organ systems that could potentially result in anincrease in BP.(i) Insulin-induced antinatriuresisDeFronzo et al. were the first to demonstrate a direct sodium-retainingeffect of insulin in healthy humans. They performed euglycemic insulin clamps inyoung subjects and found that urinary sodium excretion decreased within 30-60minutes of a physiological increment in plasma insulin concentration andgradually reached a minimum, which was 50% lower than the basal rate(DeFronzo et a!., 1975). This observation has been subsequently confirmed inhumans (Gans et al., 1991), dogs (DeFronzo et a!., 1976) and rats (Kirchner,1988). The hypothesis that hyperinsulinemia leads to renal sodium and fluidretention is based on the assumption that the kidneys of hypertensive patientsmaintain normal sensitivity to the antinatriuretic effect of insulin, in contrast to theperipheral tissues which are resistant to insulin’s glucoregulatory effects. Thispremise has been directly confirmed in essential hypertensive patients, where itwas demonstrated that although insulin-mediated glucose uptake was markedlylower in hypertensives, insulin-induced sodium retention was maintained when14compared to normotensive controls (Shimamoto et a!., 1994). In addition, it wasrecently reported that in hypertensive patients, insulin directly increased sodiumreabsorption in the proximal and distal tubules (Endre ef a!., 1994; Kageyama ata!., 1994).These reports raise the possibility that insulin may cause sodium andvolume overload in hypertensive patients, which could lead to hypertension. Arecent study demonstrated for the first time that although the sodium retainingeffect of insulin was maintained in hypertensive patients, they were resistant tothe natriuretic effects of atrial natriuretic peptide (Abouchacra at a!., 1994). Thisnovel finding raises the possibility that resistance to the natriuretic effects of atrialnatriuretic peptide may be one of the mechanisms underlying the insulin-inducedincrease in BP. However, young subjects with essential hypertension do not havean increased body sodium content or an increased plasma volume or a reducedplasma renin concentration (Beretta-Piccoli et a!., 1982), indicating that theseacute effects may not be sustained or may be compensated for over a longerperiod of time.(ii) Insulin and the sympathetic nervous systemAn increase in sympathetic activity secondary to an increase in plasmainsulin concentration has been demonstrated in humans. Rowe et al. reportedthat elevations in plasma insulin levels caused a dose-dependent increase inplasma catecholamine levels and a concurrent increase in pulse and BP.However, they observed this effect only at pharmacological concentrations ofinsulin, which casts doubt with regard to the physiological relevance of theirfindings (Rowe et al., 1981). Other studies have documented that fastingdecreased catecholamine levels whereas feeding led to an increase in plasma15norepinephrine levels (DeHaven et a!., 1980; Landsberg and Drieger, 1989;O’Dea et al., 1982). This increase in sympathetic activity would be expected tocause an increase in cardiac output, peripheral vasoconstriction and aconsequent increase in BP. However, studies have demonstrated that although aphysiological increase in insulin concentration causes an increase in musclesympathetic activity and nerve firing rate, it results in a decrease in vascularresistance and either no change or a paradoxical decrease in BP (Anderson et a!.,1991; Berne et a!., 1992). In a series of elegant experiments, Baron et al.demonstrated that insulin caused a rightward shift in the norepinephrine doseresponse curve and that this effect was more pronounced in lean as compared toobese subjects (Baron et a!., 1994). They also found that insulin caused a 25%increase in the metabolic clearance of norepinephrine and that this effect wasalso blunted in obese, insulin-resistant subjects. Finally, they reported that obesesubjects were more susceptible to the pressor effects of insulin than lean insulin-sensitive humans. The reason why insulin does not increase BP, despitestimulation of the sympathetic nervous system, is that insulin causes preferentialvasodilation in the skeletal muscle vasculature and thereby leads to aredistribution of cardiac output to skeletal muscle (Baron 1993). Thus thevasodilatory effects of insulin offset the increase in cardiac output, an issue thatwill be addressed in detail in one of the subsequent sections.(iii) Trophic effects of insulinIt has been reported that insulin, via its action on insulin-like growth factorreceptors, causes an increase in vascular smooth muscle cell growth in vitro(Banskota et a!., 1989; King et al., 1985). Furthermore, it has been demonstratedthat insulin stimulates DNA synthesis in fibroblasts and vascular smooth muscle16cells (Capron et al., 1986; Rechler et al., 1974). Cruz et al. reported that chronicinsulin infusion into one femoral artery in the dog caused vascular hypertrophyonly on the ipsilateral side (Cruz et a!., 1961). Therefore, it is possible thatchronic hyperinsulinemia may cause vascular hypertrophy and lead to narrowingof the lumen of resistance vessels, consequently raising vascular resistance andBP. Although this hypothesis has not been validated, it is important to mentionthat the hypertrophic effects of insulin have been observed only atsupraphysiological concentrations and that the possibility of insulin exertingtrophic effects at physiological concentrations remains to be determined.(iv) Hemodynamic effects of insulinThe hemodynamic effects of insulin have been intensively investigatedover the last 3 years and results indicate that not only does insulin exert a varietyof metabolic effects but that it is also a powerful vasoactive hormone (Baron1993). Although some early reports had suggested that insulin exertedcardiovascular effects, most of the effects were observed at pharmacologicalconcentrations (Rowe et a!., 1981). This led to the notion that insulin did notmodulate vascular smooth muscle activity under normal physiological conditions,a notion that has been proven incorrect in recent years. The first study thatexamined this issue in detail was that of Laakso et al. who reported that insulin,when infused intravenously, caused a dose-dependent increase in leg blood flowin humans and that this effect was independent of plasma glucose concentration(Laakso et a!., 1990). Subsequently, several investigators have confirmed thisobservation and it has been demonstrated that in lean, insulin-sensitive subjects,acute insulin infusion causes a rise in peripheral blood flow with an EC50 of about40 iiU/ml (Laakso et a!., 1992). This indicates that insulin exerts potent17vasodilator effects at physiological concentrations. Furthermore, it wasdemonstrated that insulin also caused significant increases in cardiac output withan EC50 of about 70 jiU/ml (Baron and Brechtel, 1993). Studies revealed thatalthough insulin caused both systemic and peripheral vasodilation, the increase inskeletal muscle blood flow far exceeded the increment at the systemic level(Baron and Brechtel, 1993). Thus insulin, by preferentially increasing skeletalmuscle blood flow, redistributes the cardiac output to skeletal muscle (the majorsite of glucose utilization). This also explains why the insulin-induced increase incardiac output and sympathetic activity do not cause a resultant increase in BP.Even more fascinating are findings indicating that these vasodilator effectsof insulin are markedly impaired in insulin-resistant states such as obesity anddiabetes mellitus (for a review, please refer to Baron 1993). In obese subjects,the EC50 of the dose response curve to insulin’s vasodilator actions was aboutthreefold higher than that in lean controls. In NIDDM patients, the EC50 wasabout 17 fold higher, indicating a marked resistance to the vasodilator effects ofinsulin (Laakso et al. 1992). The compelling question that then comes to mind iswhether resistance to insulin’s vasodilator effects also occurs in hypertensivepatients. Indeed, it was recently demonstrated that insulin-induced vasodilation inpre-constricted dorsal hand veins was impaired in essential hypertensive patients(Feldman and Bierbrier, 1993). The mechanism/s underlying insulin-mediatedvasodilation remain to be determined and may include both systemic and localeffects (Anderson and Mark, 1993). Several recent reports suggest that insulinmay alter vascular tone via direct effects on intracellular calcium concentration invascular smooth muscle cells (Sowers eta!., 1994; Standley et al., 1993). Insulinhas also been shown to attenuate the contractile responses of vascular smoothmuscle to vasoactive amines, probably by causing changes in intracellular18calcium (Touyz and Schiffrin, 1994; Yagi et al., 1988). Other reports stronglysuggest a role for endothelium-derived nitric oxide in the insulin-inducedvasodilation (Steinberg et a!., 1994), since infusion of a nitric oxide synthaseinhibitor completely prevented the increase in insulin-mediated leg blood flow.Furthermore, methylene blue, a guanylate cyclase inhibitor, also abolishedinsulin-mediated venodilation, suggesting that insulin’s effects are cGMPdependent (Grover et a!., 1995).If the insulin-mediated decrease in vascular tone were impaired in insulin-resistant states, it could cause a resultant increase in the pressor response tovarious neurohumoral factors, thus increasing vascular resistance and BP (Baron1993). Lending support to this hypothesis are findings that obese, insulin-resistant subjects have a greater sensitivity to the pressor effects ofnorepinephrine when compared with lean, insulin-sensitive controls (Baron et a!.,1994). Insulin resistance at the vascular level could, therefore, tip the balance infavor of the pressor forces and therefore predispose an individual towardshypertension. Whether such an effect occurs in essential hypertension remainsto be elucidated, but current evidence strongly suggests such a possibility.(v) Insulin and the intracellular cation transport systemsIt was initially postulated that insulin may regulate the activity of Na-KATPase, an important cellular enzyme that extrudes sodium in exchange forpotassium and is responsible for maintaining the normal resting potential in cells.This hypothesis stemmed from observations that the activity of this pump (which isinsulin regulated) was reduced in essential hypertensive subjects as well as inexperimental models of hypertension (Boon et a!., 1985; Canessa et a!., 1984;Postnov and Orlov, 1985). Such a reduction in Na-K-ATPase activity could19lead to increased intracellular sodium levels, which could sensitize the arteriolarsmooth muscle cells to the pressor effects of catecholamines and angiotensin Il.Although this is an attractive hypothesis, evidence suggests that such anabnormality is unlikely to be the cause underlying the increase in BP inhypertensive subjects. For example, it has been demonstrated that insulinstimulated potassium uptake (a reflection of Na-K-ATPase activity) is unrelatedto insulin’s stimulatory effect on glucose metabolism (Ferrannini et al., 1988).Furthermore, insulin can promote cellular potassium uptake normally inhypertensive patients (DeFronzo and Ferrannini, 1991). Therefore, it does notseem that resistance to insulin’s glucoregulatory effects also extends to its effectson the N&K+ATPase enzyme.Another cation pump that has been examined is the Na+H+antiporter,which is responsible for maintaining intracellular sodium concentration as well asintracellular pH. Increased activity of this pump in response to increased levels ofinsulin has been observed in various cell types in hypertensive subjects (Adragnaet a!., 1982; Canessa et a!., 1987; Weder, 1985). Overactivity of this pump couldresult in increased sodium levels inside the cell, which would sensitize vascularsmooth muscle cells to the effects of various pressor amines. In addition,increased sodium levels could result in an indirect increase in intracellularcalcium concentration, which would also cause an increase in vascular tone.Finally, an increase in the activity of this proton pump would lead to intracellularalkalinization, which is a stimulus for vascular smooth muscle growth (Lever1986).Probably the most important observation that may explain the directvasodilatory effects of insulin is that insulin has marked effects on intracellularcalcium concentration (Sowers et a!., 1994). It has been reported that insulin20attenuates vascular smooth muscle calcium influx through both receptor andvoltage-operated calcium channels. In addition, insulin also modulates theactivity of Ca+2ATPase, which is responsible for the extrusion of calcium fromcells (Standley et al., 1993). Resistance to these effects of insulin would causean increase in intracellular calcium levels and a consequent enhancement ofvascular tone and BP. Thus, insulin has the ability to directly modulate severalintracellular ionic pumps and thereby alter vascular tone and BP.(V) SPECIFIC RESEARCH PROBLEM AND RESEARCH STRATEGY(i) The research problemA major limitation of studies linking insulin resistance and hyperinsulinemiato hypertension is that they do not establish causality. Although associations andcorrelations strongly favor such a link, this issue requires further experimentalevaluation. In addition, these metabolic defects are not always associated withessential hypertension, neither are all insulin-resistant subjects hypertensive.Another confounding issue is whether the insulin resistance associated withhypertension is a cause or an effect of increased BP. It has been documentedthat epinephrine, acting primarily through the beta adrenergic receptor, markedlyimpairs hepatic as well as peripheral tissue sensitivity to increments in plasmainsulin concentration (Diebert and DeFronzo, 1980). It could, therefore, beargued that a primary increase in sympathetic activity (increase in plasmacatecholamines) may antagonize insulin action and lead to secondary insulinresistance. However, decreasing blood pressure with most antihypertensivedrugs does not improve insulin sensitivity or decrease plasma insulin levels,which suggests that these metabolic defects are not secondary to hypertension.Resolution of this issue requires more direct and specific experimental21interventions, which were attempted in the present study. We examined thehypothesis that insulin resistance and hyperinsulinemia are causally related tohypertension.(ii) RationaleWe addressed this issue from a pharmacological perspective andreasoned that if insulin resistance and/or hyperinsulinemia were responsible forthe development of high BP, then a specific improvement in insulin action shouldresult in a fall in BP. However, if these metabolic defects were not causallyrelated to hypertension, or were secondary to it, then such metabolicimprovements should not cause any resultant change in BP. Recent evidencesuggests that the insulin resistance of hypertension is tissue and pathwayspecific, with the major defect residing in the glycogen synthesis pathway in themuscle (Ferrannini et a!., 1987). Therefore, drug interventions that specificallyimprove muscle glycogen synthesis and enhance insulin action should alsoprevent the increase in BP.(iii) Experimental approachTo broaden the nature of our enquiry, we used both a genetic and anacquired model of experimental hypertension, which were: (a) the spontaneouslyhypertensive rat, which is thought to closely resemble human essentialhypertension and (b) the fructose-hypertensive rat, where hypertension is inducedin normotensive rats by feeding them a high fructose diet. We attempted todirectly improve insulin sensitivity by employing three drug interventions, all ofwhich have been shown to enhance insulin action. These drugs were: (a) vanadylsulfate, the (÷lV) form of the trace element vanadium (b)bis(maltolato)oxovanadium(lV), an organic vanadium complex that is more potent22and is associated with fewer gastro-intestinal side effects than vanadyl sulfateand (c) pioglitazone, a recently synthesized thiazolidinedione derivative thatenhances insulin action, probably by sensitizing peripheral tissues to the effectsof insulin, Insulin sensitivity was quantified by using the euglycemic,hyperinsulinemic clamp in conscious rats, which is considered to be one of themost precise methods available for assessing in vivo insulin action. Multiple,direct and specific drug interventions were employed in an effort to strengthen ourexperimental approach and the anticipated results.(iv) Experimental models(a) Spontaneous hypertension: Spontaneously hypertensive rats (SHR) have agenetic predisposition towards hypertension and have been shown to be insulin-resistant and hyperinsulinemic when compared to their genetic controls, theWistar Kyoto (WKY) strain (Hulman et a!., 1991; Mondon and Reaven, 1988).Insulin stimulated glucose uptake was found to be lower in adipocytes isolatedfrom SHR as compared to the WKY rats (Reaven et al., I 989a). This decrease ininsulin stimulated glucose transport was observed despite normal insulin receptornumber, affinity and tyrosine kinase activity, suggesting that the defect in insulinaction resided distal to the insulin receptor (Reaven et al., 1 989a). Furthermore,the defect in insulin action in adipocytes preceded the development ofhypertension in the SHR (Reaven and Chang, 1991). Subsequent studiesdemonstrated that insulin clearance was decreased in the SHR, which was due todecreased removal of insulin by the kidneys and skeletal muscle rather than adefect in hepatic insulin extraction (Mondon et al., 1989). Such a reduction ininsulin clearance could also result in higher circulating insulin levels in the SHR.Reaven et al. studied the effect of a high fructose diet in the SHR and found that23although the fructose diet caused an increase in plasma insulin levels and BP inboth the SHR and WKY rats, the increase was more pronounced in the SHR ascompared with their WKY controls (Reaven et al., 1990).Although euglycemic clamp studies conducted in anesthetized SHRdemonstrate that they are insulin-resistant, recent studies done in conscious SHRhave challenged this notion (Buchanan et a!., 1992a; Buchanan et a!., 1992b;Frontoni et al., 1992). SHR have been shown to be more responsive to stress(anesthesia or restraint) as compared to the WKY (McMurty and Wexler, 1981;Shah et a!., 1977). Consequently, there could be a greater release ofendogenous catecholamines, which could, in turn, antagonize insulin action andlead to secondary insulin resistance in the anesthetized SHR. In agreement withthis concept are results from clamp studies in conscious, minimally restrained ratsthat demonstrate no difference in insulin sensitivity in the SHR as compared to theWKY (Buchanan et al., 1 992a; Frontoni et a!., 1992). However, the presence ofhyperinsulinemia in the SHR was confirmed even in the latter studies and it wasproposed that increased insulin levels may contribute to the development of highBP. This issue could be evaluated by decreasing plasma insulin levels in theSHR and studying the resultant change (if any) in blood pressure. Such anattempt was made in the present investigation.(b) Fructose-induced hypertension: The fructose-hypertensive model representsan acquired form of systolic hypertension, where the rise in BP is not geneticallydetermined but is diet-induced (Hwang et a!., 1987). The finding thathypertension can be produced in normotensive rats by an experimentalmanipulation known to result in insulin resistance and hyperinsulinemia supportsthe contention that these metabolic defects are very closely linked tohypertension. Furthermore, exercise training (which caused an improvement in24insulin sensitivity) and somatostatin administration (which decreasedhyperinsulinemia) to the fructose-fed rats attenuated the fructose-inducedincrease in BP (Reaven et al., 1988; Reaven eta!., 1989b). Addition of clonidineto the drinking water of fructose-hypertensive rats caused a decrease in BP butdid not improve insulin sensitivity or lower plasma insulin levels, suggesting thatthe metabolic abnormalities were primary and were not simply a consequence ofhigh BP (Hwang et aL, 1987). Interestingly, high fructose feeding also elicitsinsulin resistance, hyperinsulinemia and hypertension in dogs and the metabolicdefects antedate the increase in BP (Martinez et a!., 1994). Given the fact that westill do not know the relative contribution of genetic and acquired factors towardsthe insulin resistance observed in hypertension, it was important to examine thevalidity of the hypothesis under study in this model. Would the hypothesis holdtrue in a model where genetic susceptibility towards hypertensinogenicmechanisms is absent? Having discussed the experimental models that wereemployed in the study, we will next consider the drugs that were used as theexperimental interventions.(v) Drug interventions(a) Vanadyl sulfate: Vanadium (atomic weight 50.94) is a ubiquitous group Vtransition element that exists in several valence forms (-1 to +5). Its essentiality inman has not been established as yet although it is estimated that in humans, thetotal vanadium pool ranges between 100-200 pg (Byrne et a!., 1978). Over thepast decade, several laboratories have reported on the insulin-mimetic propertiesof vanadium in vitro (Clark et al., 1985; Shechter and Karlish, 1980; Tolman et a!.,1979). Heyliger et al. in our laboratory were the first to demonstrate that oralsodium orthovanadate treatment corrected both the hyperglycemia and25myocardial abnormalities in diabetic rats (Heyliger et a!., 1985). During our initialstudies, we observed that oral treatment with the vanadate (+5) form of vanadiumlead to gastrointestinal side effects. Furthermore, since the LD50 of the vanadyl(+4) form of vanadium has been reported to be 2 times greater than the vanadateform, we decided to focus on the effects of oral vanadyl sulfate in streptozotocin(STZ)-diabetic rats (Ramanadham et a!., 1989; Ramanadham et a!., 1990a,b).Work from our laboratory has established the effectiveness of oral vanadyl sulfatein correcting various abnormalities in the heart and adipose tissue in STZ-diabeticrats and in enhancing the effects of insulin in vivo (Ramanadham et a!., 1989;Ramanadham et a!., I 990a). Another observation that surfaced from our studieswas that vanadyl treatment not only decreased plasma glucose levels in STZdiabetic rats (without an increase in plasma insulin) but that it also reduced insulinlevels in control, non-diabetic rats (Heyliger et a!., 1985; Ramanadham et a!.,1990a,b). This suggested that vanadyl either potentiated or replaced the effectsof endogenous insulin, resulting in a decreased requirement of insulin in non-diabetic rats. Vanadium has also been shown to enhance glycogen synthesis inthe liver and muscle from control and diabetic rats (for a review, please refer toShechter, 1990). Vanadyl, by improving glycogen synthesis in the muscle, could,therefore improve insulin sensitivity and ameliorate hypertension (if ourhypothesis were valid).(b) Bis(maltolato)oxovanadiurn(lV) (BMOV: Inorganic vanadyl administration wasassociated with two problems: poor gastrointestinal absorption and somegastrointestinal side effects (Conklin et a!. 1982; Underwood, 1977). Wespeculated that an organic compound would be more lipophilic than its inorganiccounterpart and may, therefore, be better absorbed from the gastrointestinal tract.Subsequently, we synthesized BMOV by complexing one molecule of vanadyl with262 molecules of the common food additive maltol (McNeill et a!., 1992). Chronicoral administration of BMOV in STZ-diabetic rats normalized plasma glucoselevels without a concomitant increase in circulating insulin (Yuen et aL, 1993).Furthermore, BMOV was found to be 1.5 times more potent than vanadyl sulfatein lowering plasma glucose. The increased potency of BMOV as compared tovanadyl sulfate was also demonstrated in an acute study, where we examined theglucose-lowering effects of the two compounds after oral gavage in STZ-diabeticrats (Yuen et a!., 1995). We have also reported that the ED50 forintraperitoneally administered BMOV is 0.08 mmol/kg as compared to 0.22mmol/kg for vanadyl sulfate, indicating that BMQV is approximately 3 times morepotent than vanadyl in lowering plasma glucose (Yuen et al., 1993). Thissuggests that the increased potency of BMQV is not a result of increasedabsorption but that it either has an increased bioavailability or that the complexpermeates through the cell wall more readily. We observed 21 % mortality at thehighest administered dose of vanadyl but no mortality was noted with any dose ofBMOV (Yuen eta!., 1993). A six month assessment of the possible toxicologicaleffects of BMQV indicated that BMOV prevented the pathological changes seen inuntreated diabetic rats without producing significant toxicity in control or diabetictreated animals (Dai et a!., 1994). Of the vanadium compounds tested to date,BMOV is the most potent and the best tolerated.(C) Piocl itazone: Piogi itazone, a recently synthesized thiazolidinedionederivative, is another compound that exhibits insulin enhancing effects (Hofman etal., 1991; Ikeda et a!., 1990; Sugiyama et a!., 1990a). Pioglitazone has beenshown to improve insulin sensitivity and to attenuate hyperinsulinemia in insulinresistant animal models (Kemnitz et a!., 1994; Sugiyama et a!., I 990b). This drughas also been shown to increase the rate of glycogen synthesis in the isolated27muscle of the insulin-resistant, hyperinsulinemic Wistar Fatty rat (Sugiyama et a!.,I 990a). Furthermore, pioglitazone potentiated the insulin-mimetic effects ofvanadate on glucose metabolism in isolated adipocytes from Wistar Fatty rats(Sugiyama et al., 1990b). Implicit in the point of view outlined above is theassumption that none of these drugs possess any other antihypertensive effectsand that they selectively improve insulin action. Although these compounds havebeen studied for their insulin enhancing effects in diabetes mellitus, their effects inhypertension have not been evaluated. Since it is now known that hypertension,like obesity and type II diabetes mellitus, is an insulin-resistant state in mostinstances, it is only logical to study their effects on high BP.(VI) Working hypotheses1) Insulin resistance and hyperinsulinemia play a role in the development andregulation of high blood pressure.2) Vanadyl sulfate, BMOV and pioglitazone improve insulin sensitivity ininsulin-resistant states. These three drugs will, therefore, enhance insulinsensitivity, decrease plasma insulin levels and prevent the development ofhigh BP in the SHR if started before the SHR become hypertensive.3) By enhancing the action of insulin, vanadyl sulfate will also reversehypertension in the SHR (if given after hypertension becomes fully manifestin the SHR).4) Vanadyl sulfate will improve insulin sensitivity and prevent hypertension inthe insulin-resistant, hyperinsulinemic fructose-hypertensive rats.28MATERIALS AND METHODS(I) STUDIES IN THE SPONTANEOUSLY HYPERTENSIVE RAT(A) RESEARCH DESIGN AND EXPERIMENTAL PROTOCOLS(i) Studies with vanadyl sulfate(a) Prevention Study: 23 SHR and 18 WKY rats, all male, were procured at 4weeks of age from Charles River, Montreal, Canada and were randomlyassigned to four experimental groups: S (SHR-untreated, n=15), SV (SHRvanadyl treated, n=8), W (WKY-untreated, n=12) and WV (WKY-vanadyltreated, n=6). Systolic BP in all groups was measured before starting vanadyltreatment. Subsequently, chronic vanadyl sulfate treatment (VOSO4: nH2O,Fisher Scientific, NJ, U.S.A.) was commenced on 6 week old SV and WV rats.Vanadyl sulfate was administered at a concentration of 0.75 mg/mI ad libitum inthe drinking water. Starting at week 8, weekly measurements of systolic BP,plasma glucose and plasma insulin were done on all the groups. Food intake,fluid intake and body weight were measured once a week. In addition, fluidintake of rats on vanadyl was measured 5 times a week for calculation of thedose of vanadyl consumed.(b) Reversal Study: At the start of week 11 (weeks denote the age of the rats),the untreated SHR and WKY from the prevention study were further grouped asfollows: S (SHR-untreated, n=9), SV1 (SHR-vanadyl treated, n=6), W (WKYuntreated, n=6) and WV1 (WKY-vanadyl treated, n=6). The treated groupswere started on vanadyl sulfate (0.75 mg/mI) in the drinking water at thebeginning of week 11. Weekly measurements of BP, plasma glucose andplasma insulin were continued for the next 3 weeks. At 15 weeks of age29(termination), direct systolic BP measurements were done to validate theindirect BP readings obtained in the preceding weeks.(C) Pair-Feeding Study: Since vanadyl sulfate decreased food/fluidconsumption and body weight in the treated rats (see results below), a separatestudy was initiated in which one group of rats (SHR as well as WKY) was pair-fed with the corresponding vanadyl-treated group, but was not given vanadyl.This was done to observe if a decrease in food and fluid intake per secontributed towards the attenuation of hypertension in the vanadyl-treated rats.23 SHR and 24 WKY (all male) were procured at 5 weeks of age from CharlesRiver, Montreal, Canada and were grouped as follows: S (SHR-untreated, n=8),SV (SHR-vanadyl treated since the start of week 6, n=8), SF (SHR pair-fed withtreated rats for food and fluid consumed, n=7); W (WKY-untreated, n=8), WV(WKY-vanadyl treated since the start of week 6, n=8) and WF (WKY pair-fedwith treated rats, n=8). At weeks 5, 9, 10, 12, 15 and 16, systolic BP wasmeasured by the indirect tail-cuff method, which had already been validated bydirect arterial cannulation in the previous experiment. During the weeksmentioned above, five-hour fasted plasma samples were also collected via thetail vein and were later analyzed for glucose and insulin. At 15 weeks of age(after 10 weeks of vanadyl treatment) the rats were fasted overnight, plasmawas collected and was later analyzed for urea-nitrogen, glutamic oxaloacetictransaminase (GOT), glutamic pyruvic transaminase (GPT) and vanadiumlevels.(d) Insulin Implant Study: Vanadyl sulfate caused a decrease in plasma insulinlevels and BP in the SHR (see results below). In order to examine this issuefurther, a separate study was initiated in which 9 week old SHR and WKY (n1230in each group) were treated with vanadyl as before. Three weeks after startingvanadyl treatment, six vanadyl-treated rats in each group were administeredexogenous insulin (14000 pmol/kg/day as a subcutaneous insulin implant,Linshin Inc., Canada) while the other six were continued on vanadyl as before.This was done to restore plasma insulin levels in the vanadyl-treated rats tothose that existed before vanadyl treatment and to then observe the resultantchange in BP.During the post-implant period, plasma insulin, plasma glucose and BPwere measured one and three weeks after placement of the insulin implants. Inaddition, the rats were sampled for plasma catecholamines and glucagon beforeand three weeks after placement of the insulin implants. Plasma catecholaminesamples were collected in an identical fashion in all the rats. The rats werehandled frequently for a few days before sampling. The samples were collectedwithin 2 minutes of touching the individual rat in an undisturbed environmentand the same procedure was followed for every animal. Thus thecatecholamine concentrations do not represent the absolute stress free, basalvalues. The main objective was to look for major changes in catecholamineconcentrations that could occur secondary to hypoglycemia (in rats with insulinimplants) and therefore, the rats were not chronically cannulated forcatecholamine sampling.(ii) Studies with bis(maltolato)oxovanadium(lV)The present study was initiated with the following objectives: (a) toassess insulin sensitivity in the SHR and WKY rats by performing euglycemichyperinsulinemic clamps in conscious, minimally restrained rats and (b) to31administer BMOV to SHR and WKY rats and to examine the effects of the drugon insulin sensitivity, plasma insulin levels and systolic BP.Protocol: Five week old male SHR and WKY rats were procured from CharlesRiver, Montreal Canada and were randomly assigned to 4 experimental groups:S (SHR-untreated, n=9), SO (SHR-BMOV treated, n=1 1), W (WKY-untreated,n=11) and WO (WKY-BMOV treated, n=9). Chronic BMOV treatment (0.75mg/mi ad libitum in the drinking water) was initiated on 6 week old rats. Aconcentration of 0.75 mg/mi was chosen since previous results indicated that, atthis concentration, BMOV lowered plasma insulin levels in non-diabetic ratswhile allowing them to gain weight at rates that were comparable to untreatedcontrols (McNeill et al., 1992). Once at week 5 and again starting at week 8,systolic BP, plasma insulin (5-hour fasted) and plasma glucose were measuredweekly for the next 4 weeks. At termination (15-19 weeks of age), the rats werefasted overnight and hyperinsulinemic clamps were performed in consciousrats.(iii) Studies with pioglitazoneProtocol: Five week old male SHR and WKY were procured from Charles River,Montreal Canada and were randomly assigned to 4 experimental groups: S(SHR-untreated, n=8), SP (SHR-pioglitazone treated, n=16), W (WKYuntreated, n=8) and WP (WKY-pioglitazone treated, n=15). Pioglitazone was agenerous gift from the Upjohn Company, Kalamazoo, Michigan, U.S.A. Chronicpioglitazone treatment (0.1 mg/g in the rat chow) was initiated in 6 week old SPand WP rats, whereas the S and W groups received control rat chow of similarcomposition to that given to the treated groups, apart from the exclusion of thedrug. Once at week 5 and again starting at week 8, systolic BP, 5-hour fasted32plasma insulin and plasma glucose were measured every week. Sincepioglitazone caused a decrease in plasma insulin levels and BP in the SHR(see results below), seven pioglitazone treated rats (in each of the SP and WPgroups) were administered exogenous insulin after week 11(14000 pmol/kgldayas a subcutaneous insulin implant, Linshin Inc., Canada) while the other 8 werecontinued on pioglitazone as before. This was done in an attempt to restoreplasma insulin levels in the treated rats to those that existed before pioglitazonetreatment and to then observe the resultant change in BR During the post-implant period, plasma insulin, plasma glucose and BP were measured for 2weeks and subsequently, the animals were fasted for 20 hours and high doseinsulin clamps were performed in the S and W groups as well as the ratswithout the insulin implants in the SP and WP groups. In a separate study, wealso conducted low dose insulin clamps in SHR and WKY (11-13 weeks of age)after treating them with pioglitazone at the same dose as described above forthe first study.(B) METHODOLOGY(I) Blood pressure measurement(a) Studies with vanadyl sulfate: Indirect BP measurements (systolic) wereconducted by using the tail-cuff method. Rats were removed from the animalroom and taken to a quiet room at 0800; they were allowed free access to foodand water. The rats were prewarmed for 10 minutes in a rat holder placed on ahot plate with a surface temperature of 32 degrees Celsius. Since the rats hadbeen preconditioned to the BP measurement procedure, they became sedatewithin two minutes of being restrained in the rat holder. The cuff used was 35mm in length and was placed at the base of the tail. A pneumatic pulse sensor33was taped to the tail and was connected to a pneumatic pulse transducer(Narco Bio Systems, Houston, Texas, U.S.A.). A programmed electrosphygmomanometer PE-300 (Narco Bio Systems) was employed to keep thevarious parameters such as inflation/deflation rates and cycling intervalconstant. The reappearance of pulsations (on gradual deflation) signified thesystolic BP. In each rat, 3 consecutive readings were taken and averaged toobtain the individual BP.Direct BP was measured at the termination of the study. The rats wereanesthetized with a short acting barbiturate, sodium methohexital (Brietalsodium, Eli Lilly, Toronto, Canada), which was given at a dose of 60 mg/kg bodyweight intraperitoneally. Anesthesia was maintained with nitrous oxide and acatheter (1 meter of polyethylene tubing PE 50 joined to 7 cm of PE 10) wasintroduced via the caudal artery and placed into the abdominal aorta of each rat(Schenk et a!., 1992). Previous studies (in which this method was employed)had demonstrated that the cardiovascular status and baroreflex sensitivities ofrats were similar 5 and 72 hours after anesthesia and surgery (Bennett andGardiner 1986; Hebden et aL, 1987). Therefore, five hours after surgery, BPwas recorded continuously by connecting the catheter via a Gould p23dbpressure transducer to a polygraph (Gould TA 2000; Gould, Cleveland, OH). Atthe time of recording BP, the rats were fully conscious, freely moving and hadrecovered from the effects of anesthesia.(b) Studies with BMOV and pioclitazone: For these and all subsequent studies,indirect systolic BP was measured in conscious rats by using the indirect tailcuff method without external preheating (Bunag and Butterfield 1982). Theanimals were preconditioned to the experimental procedure before conducting34actual measurements. The equipment used includes a BP sensor/cuff, a BPamplifier and an analog/digital recorder and printer (Model 179 semi-automaticBP analyzer, IITC INC., Woodland Hills, CA, U.S.A.). The various parameterssuch as cycling interval and inflation/deflation rates are kept constant by thesemi-automatic apparatus. In this method, the reappearance of pulsations (ongradual deflation of the BP cuff) are detected by a photoelectric sensor and areamplified and recorded digitally as the systolic BP. An average of 3 suchreadings was taken as the individual systolic BP.The major advantage of this method is that the recordings are carried outat a temperature of 26-27 degrees Celsius, thus eliminating the heat stresstypical of other methods. Heat stress (occurs at temperatures greater than 30degrees Celsius) has been shown to modify the BP of rodents profoundly and isbest avoided during BP measurements. In a preliminary study, we validatedreadings obtained by this method by comparison with those recorded by directintra-arterial cannulation. Briefly, six male Wistar rats were catheterized(carotid artery) and were allowed to recover for five hours. The catheter wasexteriorized and simultaneous direct and indirect measurements were recorded.The means of twenty such readings were: indirect tail-cuff method 136±3versus direct cannulation 135±3 mmHg, P>0.05. Thus indirect tail cuffrecorded pressures were similar (within 5 mmHg) to those obtained by directcannulation; similar results have also been reported by other laboratories(Hwang eta!., 1989; Reaven eta!., 1988).(ii) Euglycemic hyperinsulinemic clamp technique(a) Principle underlying the method: This method is considered to be one of themost precise indicators of in vivo insulin action (DeFronzo et a!., 1979) and35involves raising the plasma insulin concentration acutely to a physiological leveland maintaining it at that level for 120-180 minutes. The plasma glucoseconcentration is held constant at basal level by a variable glucose infusion.Under steady state conditions of euglycemia, the glucose infusion rate equalsthe glucose uptake by all the tissues in the body (assuming that there is noendogenous hepatic glucose production). By infusing labeled glucose, hepaticglucose production can also be quantified and computed in the final results(DeFronzo eta!., 1979). This method, therefore, quantifies the ability of insulinto dispose of an infused glucose load (insulin sensitivity) during a period ofphysiological hyperinsulinemia.In all our studies, we applied this technique in conscious, minimallystressed rats. Studies document that the SHR may be more responsive tostressful stimuli (such as anesthesia or physical restraint) than the WKY(McMurty and Wexler, 1981; Shah et a!., 1977). Increased stress can cause anincrease in plasma catecholamines, which may antagonize insulin action.Therefore, it is desirable to use a method that causes the least amount of stressto the animals. Clamp studies done in conscious rats have been shown to bemuch less stressful than the application of anesthesia/physical restraint duringthe procedure (Buchanan eta!., 1992a).BMOV study: Rats were conditioned to tail restraint by a modification of theapproach of Buchanan et al. (Buchanan et a!., 1992a). In brief, the tail waspassed through a hole (approximately 1.5 cm in diameter) in the cage, afterwhich it was immobilized at a point halfway along its length by passing itthrough a soft cork and taping it distally. The rats had free access to food andwater and were conditioned for increasing periods of time (30 minutes to two36hours, thrice a day) over 3 days prior to the clamp study. The rats wereweighed daily and weight gain was comparable to that observed during the preconditioning period. The rats were fasted overnight (20 hours) before the clampstudies.Four hours before the start of the insulin infusions, each animal wasplaced in a specially designed foam rubber jacket, which allowed freemovement of all four limbs and forward vision. Subsequently, the rat wasplaced on a board with a belt positioning system which allowed it to beimmobilized in the left or right lateral and supine positions. Lidocaine 1 % wasthen infiltrated into the tissue on the ventral aspect of the tail. A 0.5 cm incisionwas made in the tail and the tail artery was cannulated with fine borepolyethylene tubing (PE 10) and flushed with heparin 50 U/mI of 0.9% saline.The tail vein was also cannulated percutaneously with a 24G Intracath (Jelco,Tampa, Florida, U.S.A.) attached to PE 50 tubing. The animal was returned tothe cage and was allowed to recover with free access to water. 50 pL of bloodwas withdrawn immediately after surgery and at 15, 30, 60, and 120 minutesafter surgery (but before the start of the insulin infusion) into tubes containing 9mM EGTA and 8 mM glutathione (in final dilution) for determination ofcatecholamine levels. During the first 15 minutes of the clamp, baseline plasmaglucose measurements were obtained. Thereafter, the following were infused:(a) Insulin, 14 pmol/kg/min from 0-120 minutes (preceded by a loading dose of3X, 2X and IX for 1 minute each).(b) Somatostatin, 920 pmol/kg/min from 0-120 minutes. Somatostatin was givento suppress endogenous insulin secretion completely. Since rat and humaninsulin share considerable immunoreactivity, it is necessary to suppress37endogenous rat insulin in order to precisely quantify circulating human insulinlevels (since human insulin is infused during the clamps).(c) D-[3-3H1-glucose at a rate of 0.10 jiCi/minute from -60 to 120 minutes, afteran initial square-wave bolus over 1 minute. The tracer infusion is started priorto the initiation of the insulin infusion in order to label the glucose pool in thebody. This allows the isotopic measurement of glucose turnover and hepaticglucose production as described by Steele (Steele, 1959). This measurementis necessary because the infused insulin may not suppress hepatic glucoseproduction, especially if there is any hepatic insulin resistance present.Furthermore, the effect of insulin on hepatic glucose production may also bedifferent in the SHR as compared to the WKY.(d) During the clamps, 20% D-glucose was infused as needed to maintainplasma glucose at the preinfusion level.30 p1 of arterial blood was sampled at 5 minute intervals fordetermination of plasma glucose. At 100, 110 and 120 minutes, 200 p1 of bloodwas withdrawn for measurement of steady state plasma insulin levels and tracerdilution and the animals were then sacrificed by an intravenous injection ofpentobarbital (250 mg/kg).(b) Calculation of insulin sensitivity: During the euglycemic clamp, two factorsdetermine insulin sensitivity : (i) the ability of insulin to enhance glucoseutilization and (ii) the ability of insulin to suppress hepatic glucose production.In an attempt to assess the individual contribution of these two factors,isotopically labeled tracer is used during the clamp procedure. Labeled glucoseis infused prior to and during the clamp as a constant infusion and plasmaspecific activity is determined. The overall rate of glucose appearance (Ra) is38then determined by using the modified equation of Steele (Steele 1959). Thismethod is based on the principle that the rates of appearance and utilization ofan unlabeled metabolite in the plasma can be determined by the rate of infusionof a labeled metabolite and its specific activity in the plasma.During the euglycemic clamp, the glucose appearance rate (Ra) in theplasma equals the sum of the hepatic glucose output (HGO) and the glucoseinfusion rate (GINF). Therefore, HGO can be calculated by subtracting GINFfrom the Ra. During conditions of steady state glucose and insulinconcentrations, the rate of glucose utilization (Rd) is equal to Ra. For eachanimal, a steady state value for plasma glucose, insulin, glucose production andglucose clearance was obtained by averaging the data recorded during the final30 minute period; glucose clearance was calculated by dividing the Rd by thesteady state glucose concentration. Insulin sensitivity was then expressed asthe ratio of the steady state glucose clearance to the steady state plasmainsulin value (Mitrakou eta!., 1992).(c) Underestimation of hepatic glucose output: As outlined in the previousparagraph, the total rate of glucose appearance during steady state equals thesum of the exogenous glucose infusion and the endogenous glucose productionby the liver, such thatRa GINF + HGOSince the rate of production and the rate of utilization are similar at steady state,Ra = Rd = GINF + HGOSince HGO can be either positive or zero, Rd must always be greater than orequal to GINF. However, when the tracer dilution method is used, calculated39total glucose uptake can be less than GINF such that artifactually negativevalues for HGO are obtained (see results of clamp studies).Thus, during steady state, HGO may actually be underestimated and thisis an inherent error in the glucose clamp technique. The reason for this error isthat the “glucose system” was modeled as a single compartment in which theinfused glucose immediately and completely mixed throughout the pool.However, it was subsequently realized that the compartment consisted of atleast two components: a rapidly mixing component (plasma) and a slowly mixingcomponent (interstitial space i.e. where insulin actually acts) (Bergman et al.,1989; Finegood et a!., 1987). During the glucose clamp, the entire glucosepool of the body is labeled by infusing tritiated glucose prior to the start of theclamp. However, when the clamp is initiated, the constant glucose infusion(20% D-glucose) increases the addition of glucose to the system by a factor of 3to 4. This leads to a sharp drop in the specific activity of the plasma glucosespace and causes dilution of the isotopically labeled tracer and anunderestimation of HGO. Thus it is apparent that the underestimation of HGOis due to the rate of change of specific activity in the space from which glucoseis infused and sampled, i.e. plasma.In an attempt to keep the specific activity of the “plasma compartment”constant, many investigators tried to infuse a labeled infusion of 20% Dglucose, however, the results have been variable (for a review, please refer toBergman et al., 1989). Although infusion of labeled glucose does lead to acertain degree of correction in the estimation of HGO, it is perhaps not a veryscientifically valid solution. This is because all it does is to reduce the dilutionof the hot tracer and therefore guarantees a positive number for HGO, which isno more meaningful than a “true” negative one. Furthermore, although such a40correction has been attempted in dogs and humans, no published data existsregarding the use of spiked cold glucose infusate for correction of negativeHGO in rats. One reason why this correction does not work well in rats is thatvery small blood samples are obtained as compared to dogs and humans andalso because rats are more insulin-sensitive and have higher clamp glucoseinfusion rates (on a weight to weight basis) as compared to the bigger species.This limitation is widely accepted and it is ethical to say that HGO was fullysuppressed during the clamp. For our studies, where cold glucose infusion rateexceeded the isotopically calculated Ra, the former figure was used to calculatetotal glucose disposal (Rd).We could not attempt to spike cold glucose during the clamps becausethe maximum amount of tritium allowed in a rat carcass at UBC is 100 jiCi/kgand our calculations indicated that we would exceed that limit. If we hadattempted to spike the glucose, we would have administered an average of 130iCi/ratIclamp while the permitted limit is only 31 iCi/ratIclamp. Therefore, in allthe clamp studies, negative HGO values were obtained and HGO has beenstated to be fully suppressed.Pioçlitazone study: Euglycemic clamps in conscious rats preconditioned tolimited tail restraint were performed by a method identical to that describedabove for the BMOV study. However, in this study, we performed clamp studiesat both high and low rates of insulin infusion, which allowed us to measureinsulin sensitivity at both physiological and pharmacological steady stateconcentrations of insulin, It has been previously reported that the stressinduced during this surgical procedure is minimal and that plasmacatecholamine concentrations return to normal within 30 minutes of line41placement in the SHR, WKY (Bhanot et a!., 1994a) as well as Sprague Dawleyrats (Bhanot eta!., 1994b).During the first 15 minutes of the clamp, baseline plasma glucose andplasma glucagon measurements were obtained. Thereafter, the following wereinfused:(a) Insulin, either 14 pmol/kg/min (for the low dose studies) or 70 pmol/kg/min(for the high dose studies) from 0-120 minutes (preceded by a loading dose of3X, 2X and IX for 1 minute each).(b) Somatostatin, 920 pmol/kg/min from 0-120 minutes.(c) D-[3-3H]-glucose at a rate of 0.10 jiCi/minute from -60 to 120 minutes, afteran initial square-wave bolus over 1 minute. This was done for isotopicdetermination of glucose turnover as described above. Where cold glucoseinfusion rate exceeded the isotopically calculated Ra, the former figure wasused to calculate total glucose disposal (Rd).(d) During the clamps, 20% D-glucose was infused as needed to maintainplasma glucose at the preinfusion level.30 iJI of arterial blood was sampled at 5 minute intervals fordetermination of plasma glucose concentration. At 100, 110 and 120 minutes,200 p1 of blood was withdrawn for measurement of steady state plasma insulinlevels and tracer dilution and the animals were then sacrificed by anintravenous injection of pentobarbital (250 mg/kg). For each animal, insulinsensitivity was expressed as the ratio of the steady state glucose clearance tothe steady state plasma insulin value, as described above for the BMOV study.42(II) STUDIES IN THE FRUCTOSE-HYPERTENSIVE RATMale Sprague Dawley rats were procured locally (body weight 180-200grams, 6 weeks of age). The animals were randomly assigned to 4experimental groups: C (control-untreated, n=8), V (control-vanadyl treated,n=12), F (fructose-untreated, n=9) and FV (fructose-vanadyl treated, n=15). Atweek 6 (weeks denote the age of the animals) BP, plasma glucose and plasmainsulin (5-hour fasted) were measured in all groups. Subsequently, chronicvanadyl sulfate treatment (VOSO4:nH2O, Fisher Scientific, NJ, U.S.A.) wasinitiated in the V and FV groups. Rats received vanadyl at a concentration of0.75 mg/mI ad libitum in the drinking water. This concentration was chosenbecause previous studies had indicated that it decreased insulin levels in non-diabetic rats without altering plasma glucose levels (Pederson eta!., 1989).One week after initiation of vanadyl treatment, the animals in the F andFV groups were started on a 66% fructose diet (Teklad Labs, Madison, WI,U.S.A.). The sodium content of the fructose diet was similar to that of thestandard rat chow (standard chow: sodium 4 g/kg; fructose diet: sodium 4.2g/kg). Systolic BP, plasma insulin (5-hour fasted) and plasma glucose weremeasured each week for the next 4 weeks. In addition, food intake, fluid intakeand body weight of the animals were recorded every week. At termination,insulin sensitivity was assessed in conscious rats by the euglycemic,hyperinsulinemic clamp technique. As observed in the previous studies,vanadyl caused a decrease in plasma insulin concentration and systolic BP inthe FV rats. To further examine this issue, V and FV rats (n=5 in each group)were administered exogenous insulin (daily subcutaneous ultralente insulin14000 pmol/kg/day, Eli Lilly, Toronto, Canada) for 3 weeks. This was done in43an attempt to restore the plasma insulin levels in the FV rats to those seen inthe untreated F group and to then observe the resultant change in BP.Indirect systolic BP was measured in conscious rats using the indirecttail-cuff method without external preheating, as described earlier. Euglycemichyperinsulinemic clamps were performed in a manner identical to that describedin the BMOV study.(C) BIOCHEMICAL ANALYSESFor all the studies, biochemical measurements were performed in an identicalfashion, as described below:(a) Plasma glucose was determined by the glucose oxidase method using kitspurchased from Boehringer Mannheim, Laval, Quebec, Canada.(b) Plasma insulin was determined by using a double antibodyradioimmunoassay against porcine insulin standards (15-1200 pmol/L) with kitsprocured from lmmunocorp, Montreal, Canada. The antiserum was generatedin guinea pigs and was titered to bind 35-50% of insulin 1251 (freshly iodinated)in the absence of nonradioactive insulin. The sensitivity of the assay, definedas the least amount of insulin that could be distinguished from zero, was 15pmol/L. This value was obtained from the 95% confidence limit of the intraassay variation of 40 zero standards. Samples from all the weeks in each studywere analyzed together at the end of the study to avoid inter-assay variations.Plasma glucagon was determined by using a double antibodyradioimmunoassay against porcine standards (25-2000 ngIL). The antiserumwas generated in rabbits by using porcine glucagon-HSA and was titered to44bind 20-40% of the glucagon-125 in the absence of nonradioactive glucagon(ICN Biomedicals Inc., Costa Mesa, U.S.A.).(c) In the pair-feeding study with vanadyl sulfate, GOT, GPT and ureanitrogen levels were assayed by colorimetric determinations with kits obtainedfrom Sigma Diagnostics, St. Louis, U.S.A..(d) Plasma vanadium analysis was done by electrochemical atomic absorptionspectrophotometery according to the method employed by Mongold et al.(Mongold eta!., 1990) with minor modifications.(e) During the clamp studies, plasma glucose was measured by the glucoseoxidase method in a YSI 23A glucose analyzer (YSI Inc., Yellow Springs, OH,U.S.A.). Plasma insulin was assayed by a double antibody radioimmunoassaytechnique using porcine insulin standards (ICN Biomedicals Inc., Costa Mesa,U.S.A.).(f) For determination of D-[3-3H]-glucose concentrations, serum was diluted1:4 with water and then added to an equal volume of perchloric acid, finalconcentration 2.5%. Proteins were precipitated by centrifugation at 2000xg for10 minutes. Aliquots of supernatant were dehydrated for 6 hours at 55 degreesCelsius and counted in a B-scintillation counter.(g) Plasma catecholamines were measured by a radioenzymatic method(Amersham Inc., U.S.A.). The assay system utilizes the enzyme catechol-Omethyltransferase to catalyze the transfer of a[3H]-methyl group from Sadenosyl-L-[methyl-3H]methionine to norepinephrine and epinephrine. The45resulting products, [3H] normetanephrine and [3H] metanephrine are convertedby periodate oxidation to [3H] vanillin and extracted.The catecholamine values reported in the various studies represent themean of values obtained 60 and 120 minutes after surgery (unless otherwisespecified), since the levels in all groups fell within the first 30 minutes andremained unchanged thereafter.(D) STATISTICAL ANALYSES(i) Samrle sizeWe initially applied a two tailed power analysis (by using the standarddeviation obtained from our preliminary study with vanadyl sulfate) for each ofthe dependent variables (insulin, glucose and BP) under study. The analysisindicated that for a probability (alpha) level of 0.05, the minimum sample sizerequired for each experimental group was 6-8 rats; therefore a minimum of 6rats/group were used in all the studies.(ii) StatisticsAll data are presented as means±SE. Data analysis for each study wasperformed by using analytical tests as described below. The independentvariable in each study was the drug intervention (treated versus control). Sincethere were several dependent variables (glucose, insulin, BP, body weight), thedifferences among various group means were studied by using a multivariateanalysis of variance (MANOVA), using the Number Cruncher StatisticalProgram (NCSS). MANOVA is the most powerful statistical procedure availablefor this type of analysis.46In MANOVA, it is assumed that individual group variance-covariancematrices are equal (a preliminary test is provided to determine the viability ofthis assumption). The test is used to determine if the mean vector (made up ofthe individual variate means) between various groups is different. The statisticused is called Wilk’s Lambda, which is the multivariate extension of R-squared(from multiple regression). Significance tests are made by using an F-approximation to Wilk’s Lambda.For all the studies, a probability of P<0.05 was taken to indicate asignificant difference between means. Once the MANOVA detected asignificant difference in the mean vector, the individual variables were analyzedby employing the Newman-Keuls test for multiple comparisons. Changes withineach group over time were analyzed by an analysis of variance (ANOVA)followed by a Newman Keuls test. Differences in the various parameters beforeand after administration of exogenous insulin (in rats given insulin implants)were compared by using the paired t-test procedure on the NCSS statisticalpackage. For the vanadyl sulfate studies in the SHR, the systolic BP andplasma insulin values reported in the text represent the mean values fromweeks 10 and 12 in the prevention study and weeks 11-13 in the reversal study.For all other studies in the SHR, the systolic BP and insulin values representthe mean values from weeks 9-11.47RESULTS(I) STUDIES IN THE SPONTANEOUSLY HYPERTENSIVE RAT(i) Studies with vanadyl sulfate(a) Body weightRepresentative results from the pair-feeding study are discussed below;similar results were obtained in the prevention and reversal studies. In agreementwith earlier studies (Mondon and Reaven 1988), the untreated W rats gainedweight at a faster rate than the untreated S group (figure 1). Although vanadylconsumption resulted in reduced weight gain in the SV and WV rats, theycontinued to gain weight throughout the study period. Body weight of the pair-fedSF and WF rats remained similar to the corresponding vanadyl-treated groupsthroughout the study.(b) Plasma insulin concentrationPrevention/Reversal Studies: The SHR were hyperinsulinemic as compared tothe WKY; vanadyl lowered plasma insulin levels in the SHR in both the prevention(SV: 252±23 versus S: 336±13 pmol/L, PcO.O1) and reversal studies (SV1:264±13 versus S: 342±7 pmol/L, PO.05 versus SV). This decrease in plasma insulinconcentration was maintained throughout the study periods. The averagepercentage decrease in the plasma insulin values (fed values) seen after vanadyltreatment was about 20% in the SHR and 4% in the WKY. The decrease inplasma insulin levels in the treated WV rats did not attain statistical significanceexcept at one time point (week 10, prevention study).48Pair-Feeding Study: Plasma insulin values in the pair-feeding study representfive-hour fasted values as opposed to fed values in the other two studies. Theuntreated SHR were hyperinsulinemic as compared to their WKY controls even asearly as 5 weeks of age (figure 4A). Vanadyl reduced the fasting plasma insulinvalues in the SHR by about 35% and this decrease persisted throughout thestudy. Plasma insulin levels in the pair-fed SHR were intermediate between thoseseen in the untreated and vanadyl-treated groups. No change in plasma insulinlevels was observed between the control, treated or pair-fed WKY. However, asthe WKY grew older and heavier, their plasma insulin levels increased, whichprobably signified the combined (negative) effect of increasing weight and age oninsulin sensitivity in these rats (DeFronzo and Ferrannini, 1991).(c) Blood pressurePrevention/Reversal Studies: In the prevention study, chronic vanadyl sulfatetreatment resulted in a marked and sustained decrease in systolic BP in the SHR,which was noticed at all time points starting at week 8 (SV: 158±2 versus S:189±1 mmHg, PO.05). Direct systolic BPmeasurements conducted at the termination of the study confirmed the previousindirect readings, which was a reflection on the validity of both the techniquesused (figure 2). There was a difference of about 8 mmHg between the direct andindirect systolic BP readings, which was consistent in all groups except theuntreated SHR. During the indirect BP measurements, the pulsations in thecaudal arteries in the untreated SHR appeared earlier than in the other groups,probably due to the increased pulse pressure and the hyperdynamic stateassociated with hypertension. Consequently, the untreated SHR had to be49restrained for a relatively shorter period of time (about 3 minutes after the initial10 minute warming period, as opposed to 5-10 minutes in the other groups),which shortened the period of restraint/external heating in that group. This maybe the reason why the modest increase in the indirect BP readings (as comparedwith the direct measurements) that occurred in the other three groups was notobserved in the untreated SHR. When vanadyl was started after the SHR werefully hypertensive (reversal study), it again caused a marked decrease in BP inthe SHR (SV1: 161±1 versus S: 188±1 mmHg, P<0.001, figure 3B) but had noeffect in the WKY.Pair-Feedincj Study: The decrease in BP in the vanadyl-treated SHR was similarin magnitude to that seen in the prevention and reversal studies (figure 4B).However, the pair-fed SHR remained as hypertensive as the control, untreatedrats and their BP remained unchanged. Also, no change in BP was noticedbetween the normotensive control, treated or pair-fed Wistar Kyoto rats (figure4B).Insulin lmrlant Study: Restoration of plasma insulin concentration in the vanadyltreated SHR (table 1) reversed the effects of vanadyl sulfate on BP and thisreversal was observed as early as 1 week after placement of the insulin implants(SV with insulin implants 190±3 versus SV without insulin implants 152±3 mmHg,P<0.O01). No change in BP was observed in the vanadyl-treated WKY treatedwith exogenous insulin. Catecholamine levels in the vanadyl-treated SHRremained unchanged as compared to those seen in the untreated SHR (SV:2470±203 versus S: 2284±377 pg/mI, P>0.05), which suggests that theantihypertensive effects of vanadyl may be independent of changes insympathetic activity.50FIGURE 1PAIR-FEEDING STUDY: Body weights in the six groups: S (SHR-untreated, n=8),W (WKY-untreated, n=8), SV (SHR-vanadyl treated, n8), WV (WKY-vanadyltreated, n=8), SF (SHR pair-fed with SV but not given vanadyl, n=7) and WE (WKYpair-fed with WE but not given vanadyl, n=7). Vanadyl treatment was initiated atthe start of week 6. Data are shown as means±SE.* P-ioo-II468101214AGE(WEEKS)54FIGURE 3REVERSAL STUDY: (A) Plasma insulin levels and (B) Systolic blood pressure inthe four groups (vanadyl treatment was started at week 11): WV1 (WKY-vanadyltreated, n=6), W (WKY-untreated, n=6), SV1 (SHR-vanadyl treated, n=6) and S(SHR-untreated, n=9). Data are shown as means±SE.* P—ioo-FIIIIIVI91011121314AGE(WEEKS)56FIGURE 4PAIR-FEEDING STUDY: (A) Plasma insulin levels and (B) Systolic blood pressurein the six groups: S (SHR-untreated, n=8), W (WKY-untreated, n=8), SV (SHRvanadyl treated, n=8), WV (WKY-vanadyl treated, n=8), SF (SHR pair-fed with SVbut not given vanadyl, n=7) and WE (WKY pair-fed with WF but not given vanadyl,n=7). Vanadyl treatment was initiated at the start of week 6. Data are shown asmeans±SE.* PC)-maP1*P17FJI*U)I-a>LS58(d) Plasma glucose concentrationThe average plasma glucose values in the various groups during thedifferent study protocols ranged from 5.8-7.7 mmol/L (table 2). All the groups inthe prevention, reversal and pair-feeding studies remained euglycemic (<8.5mmol/L) throughout the respective study periods. No changes in plasma glucoselevels were observed after vanadyl treatment in either the fed (prevention/reversalstudies) or the 5-hour fasted (pair-feeding study) states. Furthermore, plasmaglucose concentration remained unchanged in the vanadyl-treated SHR one weekafter placement of the insulin implants (6.0±0.2 versus pre-implant 6.0±0.1mmol/L, P>0.05). However, a decrease in plasma glucose was observed 3 weekspost-implant (4.0±0.9 versus pre-implant 6.0±0.1 mmol/L, P<0.05), which was notaccompanied by any change in plasma glucagon levels (table 1). Plasmacatecholamines in the vanadyl-treated SHR given insulin implants showed anincrease as compared to their pre-implant values (3 weeks post implant:4940±662 versus pre-implant 2073±245 pg/mI), which may indicate either acompensatory response to a fall in plasma glucose or a direct effect of sustainedhyperinsulinemia on the sympathetic nervous system.(e) Food and fluid intakeBoth food and fluid intake in the vanadyl-treated groups decreased ascompared to the untreated groups. This was seen in all the treated groups to asimilar extent and was noticed in the prevention, reversal and pair-feeding studies(tables 3A-C). The average decrease in food intake ranged from 10-15% and thedecrease in fluid intake ranged from 30-40%. The pair-fed SHR and WKY weregiven food and fluid in amounts that were similar to the corresponding vanadyl59treated groups, which was reflected as similar weight gain patterns between thepair-fed and the vanadyl-treated rats (figure 1).(f) Dose of vanadyl consumedDaily vanadyl dose calculated over the study periods showed an overalldowntrend with time, due to the increasing body weight of the rats rather than areduction in their fluid intake. The dose consumed ranged from 0.4-0.6mmol/kg/day and was not different between the various treated groups.(g) Hepatic functionVanadyl treatment did not cause any change in plasma GPT in either theSHR or the WKY (S: 70±4, SV: 66±7 and SF: 70±4 u/I; W: 56±4, WV: 57±4 andWF: 60±3 u/I). Furthermore, vanadyl did not affect plasma GOT levels in either ofthe two rat strains (S: 18±1, SV: 15±1 and SF: 17±2 u/I; W: 17±2, WV: 16±2 andWE: 14±3 u/I).(h) Renal functionPlasma urea-nitrogen values also remained unchanged in the treated SHRand WKY (SV: 10±1 and WV: 10±1 mmol/L) as compared with the respectiveuntreated (S: 13±1 and W: 10±1 mmol/L) or pair-fed (SF: 11±1 and WF: 9±1mmol/L) groups. Thus 10 weeks of vanadyl treatment did not cause anyimpairment of hepatic or renal function in either the SHR or WKY. In addition,none of the treated rats died or exhibited any signs of gastrointestinal disturbancethroughout the experimental period.60TABLE IVARIOUS PARAMETERS FROM THE EXPERIMENTAL GROUPS IN THEEXOGENOUS INSULIN STUDYRATS BP Insulin Glucagon(mmHg) (pmol/L) (ng/L)B Al A3 B Al A3 B A3S 205±3 191±5 197±1 330±12 402±12 396±54 398±35 310±23SV 146±4t 137±6t 159±lt 210±30t 180±12t 144±24t 405±14 296±26SIb 152±3t l90±3t l89±4 198±12t 360±484 618±42t 390±20 298±35W 130±7t 136±4t 147±lf 228±24t 264±18t 294±24t 340±30 270±22WV 135±2t 141±2t 147±2t 234±18t 276±12t 294±24t 287±16 316±25WI 135±4t 144±2t 148±3t 222±18t 324±12 48O±1O2 307±33 258±30Various parameters before (B), one week after (Al) and 3 weeks after (A3) placementof the insulin implants in the experimental groups: S (SHR-untreated, n=4); SV (SHRvanadyl treated, n=6); SI (SHR-vanadyl treated with implants, n=6); W (WKYuntreated, n=4); WV (WKY-vanadyl treated, n=6) and WI (WKY-vanadyl treated withinsulin implants, n=6). Data are shown as means±SE.*Insulin dose from the insulin implants was 14000 pmol/kg/day and implants weregiven only in the SI and WI groups. BP denotes systolic blood pressure.t P<0.05, different from S.§ P<0.05, different from the respective pre-implant values.P<0.05, SI different from SV.61TABLE 2PLASMA GLUCOSE (MMOL/L) IN THE EXPERIMENTAL GROUPS IN THEPREVENTION, REVERSAL AND PAIR-FEEDING STUDIESPREVENTION STUDY: Week 8 Week 9 Week 10 Week 12W (12) 6.7±0.1 6.6±0.2 6.7±0.2 6.5±0.1WV (6) 6.4±0.3 6.7±0.4 6.3±0.3 6.4±0.2S (15) 6.7±0.1 6.6±0.2 6.4±0.1 6.5±0.2SV (8) 6.8±0.2 6.6±0.1 6.0±0.1 6.7±0.3REVERSAL STUDY: Week 10 Week 11 Week 12 Week 13(no vanadyl)W (6) 6.4±0.3 6.3±0.2 6.2±0.1 6.2±0.1NV1 (6) 6.2±0.1 6.4±0.1 6.4±0.4 6.3±0.2S (9) 6.5±0.2 6.2±0.2 6.4±0.2 6.2±0.2SV1 (6) 6.0±0.2 5.8±0.3 6.6±0.3 6.6±0.3PAIR-FEEDING STUDY: Week 5 Week 9 Week 10 Week 12(no vanadyl)W (8) 6.9±0.2 8.2±0.2 7.1±0.2 7.0±0.1WV (8) 7.0±0.3 7.7±0.2 7.6±0.2 7.5±0.1WE (8) 7.0±0.1 7.3±0.1 7.6±0.2 7.4±0.2S (8) 6.9±0.1 8.1±0.2 7.1±0.2 6.9±0.1SV (8) 6.5±0.2 7.6±0.2 7.6±0.1 7.4±0.2SE (7) 6.2±0.1 8.0±0.3 7.6±0.2 7.0±0.3Data are shown as means±SE. The numbers in parentheses indicate the number ofanimals in each group. S=untreated SHR, W=untreated WKY, SV and SVj=vanadyltreated SHR, WV andWV1=vanadyl-treated WKY, SF=SHR pair-fed with SV but notgiven vanadyl and WF= WKY pair-fed with WV but not given vanadyl. Vanadyltreatment in the prevention and pair-feeding studies was started at the end of week5/start of week 6 (weeks denote the age of the animals) and in the reversal study atweek 11. In the pair-feeding study, plasma samples were collected after fasting theanimals for 5 hours. However, values in the prevention and reversal studies are fromfed animals.62TABLE 3AFOOD AND FLUID INTAKE IN THE VARIOUS EXPERIMENTAL GROUPS IN THEPREVENTION STUDYWeek4 Week6 Week8 Week 10 Week 12(before vanadyl)Food intake:(giday)W (12) 21±1 21±2 20±1 18±1 19±2WV (6) 21±1 16±lt 18±1 17±2 17±1S (15) 22±1 21±1 21±1 18±1 19±1SV (8) 23±2 17±2* 19±1 19±2 18±1Fluid intake:(mi/day)W (12) 31±2 34±1 39±2 37±2 40±1WV (6) 33±1 27±lt 29±lt 29±lt 28±ltS (15) 36±1 35±2 37±2 42±1 44±2SV (8) 35±2 28±2* 27±1* 28±2* 30±1*Data are shown as means±SE. The numbers in parentheses indicate the number ofanimals in each group. S=untreated SHR, W=untreated WKY, SV=vanadyl-treatedSHR, WV=vanadyl-treated WKY.Vanadyl treatment was started at the end of week 5/start of week 6 (weeks denote theage of the animals).*P<0.05 SV different from S and W; t P<0.05 WV different from W and S.63TABLE 3BFOOD AND FLUID INTAKE IN THE VARIOUS EXPERIMENTAL GROUPS IN THEREVERSAL STUDYWeek 10 Week 12 Week 14(before vanadyl)Food intake:(giday)W (6) 18±1 19±1 18±1WV1 (6) 18±1 17±2 16±2S (9) 19±1 21±1 20±2SV1 (6) 19±2 15±1*18±1Fluid intake:(mi/day)W (6) 36±2 38±2 38±2WV1 (6) 35±2 27±2t 29±2tS (9) 41±2 40±2 42±2SV1 (6) 40±2 29±2* 28±2*Data are shown as means±SE. The numbers in parentheses indicate the number ofanimals in each group. S=untreated SHR, W=untreated WKY, SV1=vanadyl-treatedSHR, WV1=vanadyl-treated WKY.Vanadyl treatment was started at the start of week 11 (weeks denote the age of theanimals).*P<0.05 SV1 different from S and W; t P<0.05 WV1 different from W and S.64TABLE 3CFOOD AND FLUID INTAKE IN THE VARIOUS EXPERIMENTAL GROUPS IN THEPAIR-FEEDING STUDYWeek5 Week6 Week 8 Week 10 Week 12(before vanadyl)Food intake:(giday)W (8) 17±1 18±2 20±2 20±2 18±2WV (8) 17±2 16±2 21±2 18±2 19±1WF (8) 17±1 (PAIR-FED WITH WV)S (8) 18±2 19±2 21±2 22±3 19±1SV (8) 17±1 14±2* 19±2 19±2 19±2SF (7) 18±2 (PAIR-FED WITH SV)Fluid intake:(mi/day)W (8) 37±2 40±3 42±1 45±2 47±3WV (8) 38±2 21±2t 31±2t 37±3t 33±ltWE (8) 38±2 (PAIR-FED WITH WV)S (8) 34±1 34±2 36±1 48±2 46±2SV (8) 35±2 20±1* 24±2* 28±2* 27±2*SF (7) 34±2 (PAIR-FED WITH SV)Data are shown as means±SE. The numbers in parentheses indicate the number ofanimals in each group. S=untreated SHRI W=untreateci WKY, SV=vanadyl-treatedSHR, WV=vanadyl-treated WKY, SF=SHR pair-fed with SV but not given vanadyl andWF=WKY pair-fed with WV but not given vanadyl.Vanadyl treatment was started at the end of week 5/start of week 6 (weeks denote theage of the animals).*P<0.05 SV different from S and W; t P<0.05 WV different from W and S.65(ii) Studies with bis(maltolato)oxovanadium(IV)(a) Blood pressure, plasma insulin and plasma glucose concentrationFive-hour fasted SHR were hyperinsulinemic as compared to the WKY(table 4). BMOV also caused a sustained decrease in plasma insulin levels in theSHR (SO: 198±6 pmol/L, P<0.0001 versus S) without having any effect in theWKY (WO: 228±5 versus W: 223±4 pmol/L, P>.0.05, table 4 and figure 6A).Interestingly, BMOV also caused a marked decrease in systolic BP in the SHR(SO: 149±3 versus S: 184±3 mmHg, P0.1 for mean of weeks 9-11 in the treatedversus the respective untreated groups for both SHR and WKY, table 4).(b) Euglycemic clamp studyDuring the 3 day conditioning period, weight gain in the rats was normaland similar to that seen during the previous weeks and none of the rats lostweight. Plasma catecholamine levels declined within the first 30 minutes ofsurgery in both the untreated SHR (plasma levels at 30 minutes: 866±50 versusthose immediately after surgery 1333±1 00 pg/mI) and WKY (plasma levels at 30minutes: 867±200 versus those immediately after surgery 1400±233 pg/mI).Steady state plasma glucose levels during clamps were similar in the fourexperimental groups (table 5) and were well matched to their corresponding basalconcentrations.66FIGURE 5BMOV STUDY: Body weights in the four groups: S (SHR-untreated, n=9), W(WKY-untreated, n=11), SO (SHR-BMOV treated, n=11), WO (WKY-BMOVtreated, n=9). BMOV treatment was initiated at the start of week 6. Data areshown as means±SE.* P <0.05; S and W different from SO and WO.67II0Os•50*vwvWO320-280-240-200 -160 -12080 -40-4I I6 8AGE (WEEKS)10 1268FIGURE 6BMOV STUDY: (A) Plasma insulin levels and (B) Systolic blood pressure in thefour groups: S (SHR-untreated, n=9), W (WKY-untreated, n=1 1), SO (SHR-BMOVtreated, n=1 1), WO (WKY-BMOV treated, n=9). BMOV treatment was initiated atthe start of week 6. Data are shown as means±SE.* P0.05 versus W,figure 8A). Pioglitazone also caused a marked decrease in systolic BP in theSHR (SP: 150±3 versus S: 195±3, P0.05). Five-hour fasted glucoselevels in all groups remained normal (<8.5 mmol/L) and no change in plasmaglucose was observed after pioglitazone treatment in either the SHR or WKY(table 6). Furthermore, the effects of pioglitazone on plasma insulin and BP wereindependent of any change in food intake, fluid intake or body weight (figure 7).(b) Insulin implant studyRestoration of plasma insulin levels in the drug-treated SHR by usingsubcutaneous insulin implants (354±30 pmol/L, P>0.05 versus S and P<0.0001versus SP without implants) reversed the effects of pioglitazone on BP and thisreversal was observed as early as 1 week after placement of the insulin implants73(SP with insulin implants 183±3 mmHg versus SP without insulin implants 150±3mmHg, P<0.001). No change in BP was observed in the treated WKY givenexogenous insulin (WP with implants 139±2 mmHg, P>O.05 versus WP withoutimplants). Catecholamine levels in the pioglitazone-treated SHR remainedunchanged as compared to those seen in the untreated SHR (SP: 1219±141versus S: 1124±101 pg/mI, P>0.05), which suggests that the antihypertensiveeffect of the drug may be independent of changes in sympathetic activity.(c) Eucilycemic clamp studiesDuring the conditioning period, weight gain in the rats was normal andsimilar to that observed during the previous weeks. Steady state plasma glucoselevels in both the low and high dose clamp studies were similar between all fourgroups (tables 7A-B) and were well matched to their basal concentrations.Similarly, steady state insulin concentrations were similar in all the groups duringboth the high and low dose insulin clamp studies. Hepatic glucose productionwas fully suppressed during insulin infusion in all groups in both studies. Duringlow dose insulin clamps, the untreated SHR were found to be more insulin-sensitive than the untreated WKY (table 7B). However, this difference was notobserved during the high dose clamps, probably because the high steady stateconcentrations of insulin resulted in maximal stimulation of glucose utilization inall the groups. Surprisingly, neither the pioglitazone-treated SHR or WKYdemonstrated any improvement in insulin sensitivity during either the low or highdose insulin clamp studies (tables 7A-B). Basal glucagon levels were significantlyhigher in the untreated SHR as compared with the untreated WKY andpioglitazone treatment had no effect on plasma glucagon concentrations in eitherthe SP or WP rats (table 6).74FIGURE 7PIOGLITAZONE STUDY: Body weights in the four groups: S (SHR-untreated,n=9), W (WKY-untreated, n=8), SP (SHR-pioglitazone treated, n=16), WP (WKYpioglitazone treated, n=15). Pioglitazone treatment was initiated at the start ofweek 6. Data are shown as means±SE.75C)E-C)II0Os.sPvwVWP320 -280-240-200 -180 -120 -80 -4 8 8 10AGE (WEEKS)1276FIGURE 8PIOGLITAZONE STUDY: (A) Plasma insulin levels and (B) Systolic blood pressurein the four groups: S (SHR-untreated, n=9), W (WKY-untreated, n=8), SP (SHRpioglitazone treated, n=16), WP (WKY-pioglitazone treated, n=15). Pioglitazonetreatment was initiated at the start of week 6. Data are shown as means±SE.* P0.05). Vanadyl also caused a modest decrease in insulin levelsin the control group (baseline: 241±15 versus mean at 9-11 weeks: 181±19pmol/L, PcO.05). As illustrated in figure 9B, BP in the fructose-fed rats increasedfrom 124±3 to 160±3 mmHg, P<0.001). This increase in systolic BP was evident 2weeks after starting the fructose diet and persisted throughout the study. Incontrast, BP did not rise in the fructose-fed rats treated with vanadyl sulfate(mean at baseline: 131±3 versus mean at weeks 9-11: 126 ±3, P>0.05). Nochanges in BP were seen in the control group (baseline: 125±5 versus mean at 9-11 weeks: 126±4 mmHg, P>0.05).(b) Insulin implant studyRestoration of plasma insulin concentration in the fructose-fed-vanadyltreated rats (FV with insulin implants 340±20 versus FV without implants 214±35pmol/L, P<0.001) reversed the effects of vanadyl sulfate and caused acorresponding increase in BP (FV with insulin implants 170±10 versus FV withoutimplants 121±3 mmHg, P<0.001). This increase in BP was independent ofchanges in body weight (FV with implants 353±8 versus FV without implants348±7 grams, P>0.05). However, no change in BP was seen in the controltreated rats (V with implants 122±5 versus V without implants 119±4 mmHg,82P>0.05). In addition, no change in 5-hour fasted plasma glucose levels wasobserved after the administration of exogenous insulin in either the controlvanadyl treated or fructose-vanadyl treated rats.Fructose feeding did not cause any change in food intake, fluid intake orbody weight when compared to the untreated controls. Vanadyl treatmentresulted in a reduction in weight gain in both the V and FV rats (table 8). Theaverage plasma glucose values in the various groups ranged from 6.3-7.8mmol/L. Five-hour fasted glucose levels in all groups remained normal (<8.5mmol/L) and no change in plasma glucose was observed after vanadyl treatmentin either the V or FV groups (table 8).(c) Euglycemic clamp studyDuring the 3 day conditioning period, weight gain in the rats was normaland similar to that seen during the previous weeks and none of the rats lostweight. Steady state plasma glucose levels were similar in the four experimentalgroups (table 9) and were well matched to their corresponding basalconcentrations. Mean plasma insulin levels during the final 30 minutes of theclamp were also similar in all four groups. Hepatic glucose production wascompletely suppressed in all groups. Negative values were obtained for HGObecause cold glucose infusate was not “spiked” with D-[3-3H]-glucose duringthese studies. As is evident from table 9, the F group was severely insulinresistant as compared to the normotensive C group. Vanadyl treatment caused amarked enhancement in insulin sensitivity in the FV rats and restored their insulinsensitivity to control levels. Plasma catecholamine levels were calculated bytaking the mean of values obtained 60 and 120 minutes after surgery, since thelevels in all groups fell within the first 30 minutes and remained unchanged83thereafter. Catecholamine levels in the FV rats did not decrease as compared tothe F rats not given vanadyl (table 8), suggesting again that the antihypertensiveeffects of vanadyl are independent of any change in sympathetic activity. Inaddition, there was no difference in catecholamine levels between the C and Vgroups (C: 915±228 versus V: 1478±493 pg/mI, P>O.05) and the values among all4 groups were similar at all time points after surgery.84FIGURE 9FRUCTOSE STUDY: (A) Plasma insulin levels and (B) Systolic blood pressure inthe four groups: C (control-untreated, n=8), V (control-vanadyl treated, n=12;vanadyl treatment started at week 6), F (fructose-untreated, n=9; fructose dietstarted at week 7) and FV (fructose-vanadyl treated, n=1 5; vanadyl started at week6 and fructose diet at week 7). Data are shown as means±SE. V and F in thefigure denote the start of the vanadyl and fructose treatments respectively.*P