neuroendocrine perturbations as a cause of insulin resistance

REVIEW PAPER

Neuroendocrine Perturbations as a Cause ofInsulin Resistance

Per BjoÈrntorp*

Department of Heart and LungDiseases, Sahlgren's Hospital,University of GoÈteborg,S-413 45 GoÈteborg, Sweden

*Correspondence to: Department ofHeart and Lung Diseases, Sahlgren'sHospital, University of GoÈteborg,S-413 45 GoÈteborg, Sweden

Received: 7 June 1999

Revised: 17 September 1999

Accepted: 5 October 1999

Published online: 19 October 1999

Summary

Insulin resistance is followed by several prevalent diseases. The most commoncondition with insulin resistance is obesity, particularly when localized toabdominal, visceral regions.

A summary of recent reviews on the pathogenesis of systemic insulinresistance indicates that major factors are decreased insulin effects onmuscular glycogen synthase or preceding steps in the insulin signallingcascade, on endogenous glucose production and on circulating free fatty acids(FFA) from adipose tissue lipolysis. Contributions of morphologic changes inmuscle and other factors are considered more uncertain.

Newly developed methodology has made it possible to determine moreprecisely the neuroendocrine abnormalities in abdominal obesity includingincreased cortisol and adrenal androgen secretions. This is probably due to ahyperactivity of the hypothalamic±pituitary±adrenal (HPA) axis, ampli®ed byinef®cient feedback inhibition by central glucocorticoid receptors, associatedwith molecular genetic defects. Secondly, secretion of gender-speci®c sexsteroid hormones becomes inhibited and the sympathetic nervous systemactivated. At this stage the HPA axis shows signs of a `burned-out' condition,and cortisol secretion is no longer elevated.

Cortisol counteracts the insulin activation of glycogen synthase in muscle,the insulin inhibition of hepatic glucose production and the insulin inhibitionof lipolysis in adipose tissue, leading to the well-established systemic insulinresistance caused by excess cortisol. This is exaggerated by increased freefatty acid mobilization, particularly with a concomitant elevation of theactivity of the sympathetic nervous system. Furthermore, capillarization and®ber composition in muscle are changed. These are the identical perturba-tions responsible for insulin resistance in recent reviews. The diminished sexsteroid secretion in abdominal obesity has the same consequences. It is thusclear that insulin resistance may be induced by neuroendocrine abnormal-ities, such as those seen in abdominal obesity. These endocrine perturbationsalso direct excess fat to visceral fat depots via mechanisms that are largelyknown, indicating why abdominal obesity is commonly associated withinsulin resistance.

This possible background to the most prevalent condition of insulinresistance has been revealed by development of methodology that allowssuf®ciently sensitive measurements of HPA axis activity. These ®ndingsdemonstrate the power of neuroendocrine regulations for somatic health.Copyright # 1999 John Wiley & Sons, Ltd.

Keywords insulin resistance; abdominal obesity; cortisol; sex steroids;glucocorticoid receptors; glycogen synthase

Introduction

Insulin resistance is the condition where a given concentration of insulin isnot showing the expected magnitude of effects on target cells. The most

DIABETES/METABOLISM RESEARCH AND REVIEWSDiabetes Metab Res Rev 1999; 15: 427±441.

CCC 1520-7552/99/060427±15$17.50Copyright # 1999 John Wiley & Sons, Ltd.

prevalent condition with insulin resistance is obesity,particularly when localized to central visceral depots [1].Insulin resistance is frequently followed by abnormalitiesin metabolic, hemostatic and hemodynamic systems,leading to prevalent diseases such as Type 2 diabetesand cardiovascular disease. Reasonably credible mechan-isms for such consequences of insulin resistance havebeen presented [2].

Insulin resistance may in fact be considered as one ofthe major background factors to some of the mostprevalent causes of morbidity and premature mortality.In spite of intensive research there is no consensus as tothe major causes of this condition, in terms of eitherenvironmental or genetic factors.

In this overview an attempt is made to summarize andanalyze the conclusions of several recent reviews in this®eld, and to try to examine the current opinions of thepathogenesis of insulin resistance. This is not a conven-tional review of primary data. Instead, conclusions byexperts in the ®eld have been extracted; conclusionsdrawn after having scrutinized the available information.

The attempt presented here to summarize theseopinions is based on a selection of the material to beanalyzed. The selection had the following reasons. First,the reviews summarized have different focuses on thediabetes/obesity problem, although all deal with insulinresistance in these conditions. Second, authors have beenchosen with a long established excellence in the ®eld. Thisdoes not mean that the conclusions in other excellentreviews of the ®eld would be of less value, but opinions inother reviews have been fairly well covered in the selectedmaterial.

Recent data on neuroendocrine causes to insulinresistance are then compared.

Current opinions on thepathogenesis of insulin resistance

The ®rst review is particularly comprehensive [3]. Itfocused mainly on insulin resistance in Type 2 diabetesand its precursor state, obesity. In brief, the localization ofinsulin resistance is considered to be muscle, and inaddition, hepatic glucose production after glucose admin-istration. Binding of insulin and the subsequent steps ofinsulin signalling are frequently less than optimal, butsecondly down-regulated by the prevalent hyperinsulin-emia, except in rare conditions where it may be primary.Glucose phosphorylation seems to be decreased.

The major abnormality is lack of adequate insulinactivation of glycogen synthase in muscle, also including®rst degree relatives. A functional genetic defect has notbeen established. An elevated mobilization of free fattyacids (FFA) with elevated lipid oxidation is of importance,and may explain most errors examined in the glycolyticpathway.

Other potential candidates are muscle ®ber andcapillarization factors, involved in circulatory variablesto determine insulin effects in muscle. Amylin and TNF-a

are discussed, but were not considered as major candi-dates in the creation of insulin resistance.

This review concludes that examination of the largenumber of reasonable candidate genes for insulin resis-tance has not yet resulted in any certain ®ndings in spiteof intensive work.

Acquired determinants were considered to be age,abdominal obesity, physical inactivity, smoking andhypertension.

The major abnormalities stressed in this review are thusan insulin resistance of the glycogen synthase activationin muscle and an elevated lipid mobilization, while otherpossibilities are considered less likely. The genetic back-ground is not known and there are several environmentalfactors involved (see also Table 1).

The next review [4] was selected because it focusesmainly on the ®rst steps of insulin action, those of theinsulin binding to the receptor and the subsequentsignalling cascade. Both were considered decreaseddepending on the severity of the insulin resistance. Thisis, however, due to a down-regulation depending on thehyperinsulinemia. FFA are important mediators whilefactors such as antibodies, abnormal insulin, counter-regulatory hormones and TNF-a are considered to beunimportant or rare causes.

This review points out FFA as important mediators(Table 1).

The next review is important because it focuses on theproblem in the most prevalent condition of pronouncedinsulin resistance, namely abdominal, visceral obesity [5].Glycogen synthase in muscle is the major locus forsystemic insulin resistance, but pyruvate dehydrogenasealso seems to be decreased. Muscle capillarization is lowand ®ber composition changed towards a decrease inType I ®bers, which is the most insulin-sensitive ®bertype. Poor suppression of hepatic glucose production byinsulin occurs. Elevated body fat mass is followed byincreased insulin secretion. Increased visceral fat massdiminishes hepatic clearance of insulin, probably due toportal FFA or, in women, a relative hyperandrogenism.

Table 1. Summary of conclusions in recent reviews on theorigin and mechanisms of insulin resistance in obesity andType 2 diabetes

Ref. [3] Ref. [4] Ref. [5] Ref. [6]

Insulin receptor andsignalling cascade

Inhib.(Sec.)

Inhib.(Sec.)

MuscleGlycogen synthase Inhib. Inhib. Inhib.Hexokinase II Inhib.Pyruvate dehydrogenase Inhib.Capillarization Inhib. Inhib.

LiverGlucose production Stim. Stim.Insulin clearance Inhib.

Adipose tissueLipid mobilization Stim. Stim. Stim.

Sec.: secondary effect.Inhib.: inhibited.Stim.: stimulated.

428 P. BjoÈrntorp

Copyright # 1999 John Wiley & Sons, Ltd. Diabetes Metab Res Rev 1999; 15: 427±441.

Insulin secretory pattern is abnormal, perhaps due todysregulation by the autonomic nervous system. Lipidoxidation due to elevated FFA mobilization is animportant factor.

This review again points out the importance of aglycogen synthase defect and of elevated lipid mobiliza-tion, but mentions also other possibilities (Table 1).

The next review was chosen because it focuses ongenetic factors [6]. These authors stress the ®ndings byseveral groups of an insuf®cient activation by insulin ofglycogen synthase in muscle, related to a polymorphismin an intron of the glycogen synthase gene. This is alsofound in ®rst degree relatives who are also insulinresistant and have an elevated waist/hip circumferenceratio (WHR) indicating abdominal obesity. It is consid-ered likely that the polymorphism indicates a linkagedysequilibrium with another defect of direct functionalimportance, yet to be identi®ed. Other molecular geneticdefects in insulin resistant subjects have not yet beende®ned. These authors consider that blood ¯ow changesin muscle are unlikely to be of importance for thepathogenesis of insulin resistance.

A defect of insulin activation of glycogen synthase inmuscle is likely to be the pathogenetic background tosystemic insulin resistance, and is based on a moleculargenetic defect yet to be identi®ed (Table 1).

These extracts from four recent reviews, havingdifferent focuses on the problem of pathogenesis ofinsulin resistance in obesity and Type 2 diabetes, arelisted in Table 1. It seems justi®ed to make the followingsummary of this table. The activation of glycogensynthase in muscle is considered to be a major locus forinsulin resistance and is likely to show a geneticpredisposition. It should be noted, however, that anapparent defect in the insulin sensitivity of the glycogensynthase system in muscle might be a consequence ofdefects in earlier steps of insulin signalling and maytherefore instead be labelled muscle insulin resistanceuntil the detailed level of the defect is revealed. Anothermajor factor is elevated lipid mobilization and oxidationwhich seems to explain de®ciencies of insulin effects onthe glycolytic pathway. Hepatic glucose production alsoseems to be involved while the insulin receptor and itssignalling are considered to be down-regulated seconda-rily to the prevailing hyperinsulinemia. The latter may beexaggerated by a diminished hepatic clearance of insulin.The potential place of diminished capillary density andother factors seems to be uncertain.

Strangely enough, an endocrine background to insulinresistance is not taken into account, or only mentionedand then immediately disregarded as unimportant [3,4].This is highly surprising for the following reasons.Obesity, particularly abdominal obesity, is the mostprevalent condition of insulin resistance and showsseveral endocrine abnormalities of importance in theaction of insulin in the body. Furthermore, some of theseendocrine abnormalities may, in fact, be major causesof insulin resistance via the mechanisms pointed outabove (Table 1). The area of insulin resistance in

obesity in relation to endocrine abnormalities is reviewedbelow.

Endocrine abnormalities in obesity

Cortisol

Cortisol secretion is regulated by stimulatory factors fromthe central nervous system (see Figure 1 for overview).The classic triggers are usually referred to collectively as`stress'. This is a vague term embracing any factor which isperceived as unpleasant, threatening or irritating, includ-ing psychological events, noise, toxins (alcohol, nicotine),pain, and mood changes. The sensitivity of this regulationis often not realized and the impact of ordinary every-day `stressors' may vary 20-fold in apparently healthyindividuals (see below). Food intake sends physiologicalsignals directly to the hypothalamic centers regulatingcortisol secretion through the hypothalamic±pituitary±adrenal (HPA) axis. These central stimulatory signals are,in turn, controlled by the adrenergic and serotonergicsystems in the central nervous system.

The output from the HPA axis is controlled by aninhibitory feedback system, which is sensitively regulatedby glucocorticoid receptors (GRs) in the central nervoussystem, particularly in the area of the hippocampus.When cortisol binds to these receptors, inhibition occursregulating output of the overriding corticotropin-releasing hormone and ACTH, resulting in diminishedcortisol secretion. This is usually tested by administrationof the synthetic glucocorticoid dexamethasone, whichwhen bound to GRs in the brain suppresses HPA axisactivity. Normally these regulatory systems are active inthe morning, occur in large peaks and are more quiescentin the afternoon and evening.

Measurements of HPA axis activity in obesity haveyielded highly varying results, although some consistencyis apparent. The studies have, in general, been performed

Figure 1. The regulation of cortisol secretion. (For details, seetext)

Neuroendocrine Perturbations 429


in obesity not taking into consideration the differencebetween central, abdominal and peripheral obesity,which, as will become apparent, is unfortunate.

Urinary output of 17-hydroxycorticosteroids, which aremainly metabolites of cortisol [7], have repeatedly shownelevated levels in obesity [8±12], even when corrected forbody surface area or creatinine output [13±16]. Urinaryfree cortisol has also been found to be elevated,particularly in abdominally obese subjects [17,18].Cortisol binding globulin does not seem to be affected,however [19,20].

Single plasma cortisol measurements are dif®cult toevaluate due to the large spontaneous variation in thesecretion pattern of cortisol. However, a reduction ofcirculating cortisol has been reported in studies wherediurnal measurements were performed, while urinarycortisol secretion was elevated [20,21]. Removal ofcortisol from the circulation has been shown to beabnormally rapid in abdominal obesity [22]. The cortisolproduction rate has repeatedly been reported to beelevated in obesity [13,23,24].

Stimulation of the HPA axis has usually been performedwith maximal or near maximal challenges, providingresults of responsiveness rather than sensitivity. ACTHstimulation has shown increased cortisol response inabdominally obese subjects [17,18,25]. After administra-tion of corticotropin-releasing hormone, ACTH or cortisolsecretion has not been abnormal when obesity in generalwas tested [26], but in abdominally obese women anelevated response has been reported [18]. This seems,however, probably not to be the case in men (Ljung et al.,unpublished).

Submaximal stimulation of the HPA axis is moreinteresting in relation to everyday situations. Studieshave shown an increase in cortisol secretion upon mentaland physical laboratory stress tests in both men andwomen who are abdominally obese [17,27,28] (RebuffeÂ-Scrive, personal communication). Furthermore, responseto food has been reported to be elevated [29,30].

The feedback control of the HPA axis has been found tobe normal when tested with a dose of 1 mg dexametha-sone [17,24] but less ef®cient in abdominally obesesubjects when tested with lower doses [31,32].

Taken together, the following conclusions can bedrawn from these studies. In obesity, in general, cortisolproduction seems to be elevated and removal rapid,resulting in normal or low circulating concentrations.Maximal and, particularly, submaximal stimulation of theHPA axis have, as a rule, produced an elevated responsein abdominal obese subjects who also tend to showblunted suppressive responses to low doses of dexa-methasone.

A critical issue with these measurements is, however,that they have not necessarily been utilizing adequatemethods, applied under optimal conditions. First, bloodsampling by itself disturbs the sensitive normal regulationof the HPA axis, particularly when performed in theunaccustomed surroundings of a laboratory or hospital.Furthermore, urinary sampling does not reveal the

kinetics of HPA axis function, and only net output ofcortisol or its metabolites are measured. Measurements ofsalivary cortisol have none of these drawbacks and closelymirror circulating, free, active cortisol [33,34].

The importance of adequate measurements of HPAaxis activity is beautifully illustrated by a recent studyfollowing salivary cortisol over a day under ordinaryliving conditions. This study shows how extremelysensitive this system is to even minor everyday events.Perceived or anticipated environmental stressors, even ofmild or moderate severity, exert profound in¯uences, andthe severity and quality of such stressors are followed by a`dose-response' of cortisol secretion. Furthermore, bothpositive and negative affects are closely mirrored by HPAaxis activities [35].

This method appears to be ideal for following everydayregulation of the HPA axis. This is important because it issuch activity that is of functional signi®cance for potentialperipheral damage of long-term elevated cortisol secre-tion.

Utilizing this method on a population basis it becomesclear that subjects with abdominal obesity have severaldisturbances in the regulation of the HPA axis. Theseabnormalities turn out to be best discovered uponchallenges in the everyday environment such as perceivedstress and food intake. Dissecting out the problem indetail, the picture becomes rather complex, probablyexplaining why previous examinations have yieldedinconsistent results and not disclosed the pathogeneticsigni®cance of dysregulation of the secretion of this well-known hormone.

First, two basic types of diurnal cortisol curves can berecognized statistically (Figure 2). One is characterizedby high morning values, a rapid decrease during the day,stimulation by a standardized lunch, and low eveningvalues. This is the well-known normal cortisol secretion.Another curve (Figure 2) is characterized by low morningvalues, with little variability, less lunch response and lowevening values.

These two basic cortisol secretion patterns were nextsubdivided into subgroups with little or frequent per-ceived stress-related cortisol secretion, reported the hoursbefore the saliva samplings. These results are summarizedin Figure 3. Men with the normal diurnal cortisol

Figure 2. Diurnal cortisol secretion in middle-aged, randomlyselected men. Upper curve shows a normal secretion withhigh morning and low evening values, and the lower a patho-logical secretion with low morning values (modi®ed fromreference [32])

430 P. BjoÈrntorp


curve, reporting little stress, show high morning values,decreasing rapidly before noon, a response to lunch andlow evening values. Men who are more sensitive to stress,but still having a basically normal cortisol secretion withthe typical diurnal pro®le with high morning, low eveningcortisols show different kinetics of their diurnal cortisolcurve (Figure 3). They fail to decrease their high morningvalues before noon and have a larger lunch peak, but thencortisol levels become low in the late afternoon. Thisresults in an elevated cortisol secretion over the dayin comparison with men with a normal cortisol curve andlow perceived stress. Both these subgroups, whether moreor less sensitive to stress, have an apparent normalsensitivity of dexamethasone suppression of cortisolsecretion.

When a similar subgrouping according to low and highreported perceived stress was performed in men with anabnormal diurnal cortisol curve, the differences were notparticularly pronounced and are therefore schematizedtogether in Figure 3. Here the dexamethasone suppres-sion appeared to be abnormal.

The proportion of men in these categories is dif®cult tocalculate precisely due to arbitrary de®nitions. Menreporting no or low stress are in a clear majority, about70%. Men with the abnormal decreased cortisol secretionare roughly estimated to be about 10% of the men studied([30,32] Rosmond and BjoÈrntorp, unpublished).

When these results of diurnal cortisol secretion areanalyzed in relation to measurements of anthropometric[body mass index (BMI), waist/hip circumference ratio(WHR) and sagittal diameter (D)], other endocrine[testosterone and insulin-like growth factor I (IGF)used as a surrogate measurement of growth hormonesecretion], metabolic (fasting glucose, insulin, lipids) andhemodynamic (systolic and diastolic blood pressures andheart rate) variables, dramatically different results wereobtained. These results were published in detail [30,32]and only summaries are shown here.

In Table 2 the results of these measurements have beenanalyzed in relation to the basic and lunch-stimulatedcortisol secretion, as well as after reported perceivedstress in the men with normal or pathological (Figure 2)HPA axis activity, and are presented as results of cor-relation analyses in each group. The number of symbolsindicate the approximate strength of the r-valuesobtained.

It is apparent that a normal HPA function with lowstress is associated with a bene®cial health pro®le,particularly visible after the physiological challenge oflunch. With stress and elevated cortisol secretion,abdominal obesity and insulin resistance appear. Withan abnormal HPA, axis metabolic values show mixedcorrelations, but insulin and the insulin/glucose ratio aswell as blood pressure and heart rate are high. As with thenormal HPA axis the associations become clearer whenexamined in relation to the lunch peak of cortisolsecretion, and are also consistent and robust in relation

Figure 3. Diurnal cortisol secretion in men with a normalfunction of the hypothalamic±pituitary±adrenal (HPA) axis(solid lines) with low (open circles) and high reported per-ceived stress (®lled circles), and in men with abnormal HPAaxis (broken line). Lunch values are those 15 min beforelunch, with summed values measured during lunch. (Fromreferences [30,32] and unpublished)

Table 2. Relations between indicated variables and cortisol secretion in men with normal and pathological function of thehypothalamic±pituitary±adrenal axis (HPA) in the basal state after lunch and in relation to reports of perceived stress in 284 ran-domly selected 51 year-old men (from references [30,32])

Normal HPA Pathological HPA

Basal Lunch Stress Basal Lunch Stress

Anthropometrica x x ++ (x) +++ +++Endocrineb + + x xInsulin x ++ ++ +++ +++Insulin/glucose ratio x ++ ++ +++ +++Lipidsc x xx +++ +++Hemodynamicd x x (+) +++ +++ +++

Signs indicate the direction and approximate strength of signi®cant correlations: low HDL indicated in lipids by positive labels.aAnthropometric variables: body mass index, waist/hip circumference ratio, abdominal, sagittal diameter.bEndocrine variables: testosterone, insulin like growth factor I.cLipids: triglycerides, total LDL and HDL cholesterol.dHemodynamic variables: Systolic and diastolic blood pressures and heart rate.



to stress-related cortisol secretion. This is summarized inTable 3 in terms of endocrine, metabolic and hemo-dynamic regulation. Cortisol secretion is elevated in menreporting elevated perceived stress with a normal HPAaxis function, but low with a pathological axis. The GRfunction, as indicated by the dexamethasone suppressiontest is blunted with a pathological HPA axis. Endocrinevalues become low and hemodynamic values consistentlyelevated with an abnormal HPA axis.

We have interpreted this as follows. An abnormal HPAaxis function is known to inhibit other central endocrineaxes [36]. In addition, in this situation the centralsympathetic nervous system is elevated, perhaps as acompensation [36], which is indicated by the elevation ofblood pressure and heart rate.

Insulin resistance is probably created by elevatedcortisol in stressed men with a normal HPA axis, but isunlikely to be responsible for this in men with apathological HPA axis, where cortisol secretion is low.In this situation other factors have to be responsible, suchas low secretion of sex steroid and growth hormones, incombination with an elevated activity of the centralsympathetic nervous system. Mechanisms for such inter-actions are known, and are brie¯y reviewed in latersections.

Insulin resistance via these mechanisms is thenpresumably generated by neuroendocrine mechanismswith HPA axis perturbations in a central position. Fromthe approximate estimations of the prevalence of the HPAaxis perturbations mentioned above, insulin resistance isrecruited in each of the normal HPA axes, high stresscortisol secretion and the pathological HPA axis groups.This is a statistical phenomenon and does not mean thatinsulin resistance is present in all men belonging tothese categories. We have abstained from trying todelineate the exact prevalence of insulin resistance inthe absence of normal borderlines. It seems likely,however, that insulin resistance via these pathogeneticmechanisms is prevalent.

To date, our observations are only cross-sectional. Wespeculate, however, based on controlled animal experi-ments, that the different HPA axis activities are the resultof a continuous derangement from normal to a low, rigidabnormal activity [36]. Such breakdown of HPA axisactivity develops after prolonged stress in animals, andresults in `burn-out' of the HPA axis. Such `burn-out' has

also been observed in humans after several conditionscharacterized by long-term psychological trauma or pain(for review, see reference [36]). The men at this stage ofHPA axis perturbations that we have studied are exposedto psychosocial and socioeconomic handicaps [37,38].In fact, both the HPA axis abnormalities and somaticsequelae seem to develop with time inversely in relationto a socioeconomic gradient (Rosmond and BjoÈrntorp,unpublished). Such observations might explain the socialinequality of disease [39].

The feedback regulation of the HPAaxis

The feedback control of the HPA axis is apparently mildlyde®cient in subgroups of men (see above), correspondingto a faulty GR function. This has been observed both indose-response experiments with dexamethasone [31] andwith a low dose of dexamethasone at a population level[30,32]. Also the function of the peripheral GR in adiposetissue appears to be abnormal. The perturbations of thecentral and peripheral GR seem, however, different,affecting sensitivity of the central GR [31] but not inadipose tissue (Ottosson et al., in preparation). Theseresults suggest the possibility of a different regulation ofcentral and peripheral GRs. Such tissue speci®city isknown from other studies [40,41].

The GR gene locus shows a polymorphism, discoveredby restriction fragment length polymorphism, utilizingthe restriction enzyme BclI (Rosmond et al., unpub-lished). Homozygotes for this polymorphism constituteno less than 14% of the Swedish male population (womenexamined currently), and are associated with poor GRfunction, abdominal obesity, insulin resistance andelevated blood pressure [41±44]. The functional impor-tance of this polymorphism in the ®rst intron of the geneis, however, not established. CAG repeats in early exonsare known to affect the transcription of steroid hormonegenes [45], but are not abnormal in the ®rst coding exon(Ottosson et al., in preparation). However, anotherpolymorphism in the 5k ¯anking region is also associatedwith cortisol secretion and might indicate functionalchanges in the regulatory promoter region (Rosmondet al., unpublished). Although these ®ndings are so far notpossible to interpret in functional terms, they suggest thepossibility of a genetic susceptibility to the developmentof the HPA axis perturbations described above, and aretherefore the focus of current developments.

Taken together, these recent studies indicate that whenadequate methodology is applied it is possible to revealthat the HPA axis is abnormal in abdominal obesity, andthat this is occurring under everyday conditions. It seemslikely that environmental stressors are involved. Thiscontention is strongly supported by experiments inanother primate, the Cynomolgus monkey. Upon stan-dardized, moderate, chronic stress these monkeysdevelop a sensitive HPA axis with a blunted, low-dosedexamethasone test, enlarged adrenals, low secretion ofsex steroid hormones, visceral accumulation of body fat,

Table 3. Interpretation of observations in Table 1

Normal HPAa Pathological HPAa

Lowstress

Highstress

Lowstress

Highstress

Cortisol secretion normal elevated low lowDexamethasone suppression normal normal low lowTestosterone and IGF-Ib normal normal low lowInsulin resistance normal elevated elevated elevatedHemodynamicc normal normal(?) elevated elevated

aHPA: hypothalamic±pituitary±adrenal axis.bIGF-I: Insulin-like growth factor I.cHemodynamic: blood pressures and heart rate.

432 P. BjoÈrntorp


insulin resistance, dyslipidemia and hypertension [46].This is an identical picture to what we see in humanssubjected to environmental stress.

The currently de®ned regulatory errors are summar-ized schematically in Figure 4.

Peripheral cortisol metabolism

There are also reports on changes in peripheral cortisolmetabolism in obesity. The reports suggesting an elevatedperipheral consumption of cortisol [21,22] might be dueto different alternative phenomena in the periphery. Oneis an increased binding to a locus with increased numberof GRs. In obesity this might be the enlarged adipose

tissue, including visceral fat, which has a particularly highdensity of GRs [47].

Another alternative might be that cortisol is rapidlymetabolized to cortisone, which has less glucocorticoidactivity [48]. Furthermore, a relative inactivity of theenzyme 21-hydroxylase, which converts 17-hydroxypro-gesterone to 11-deoxycortisol has been suggested [49].This would increase adrenal androgen production at theexpense of cortisol.

Under all these conditions it has to be assumed that acompensatory increased cortisol secretion will be aconsequence, due to a diminished feedback inhibition.This might provide an explanation of the repeatedlyobserved increased turnover of cortisol in obesity

Figure 4. Schematic overview of neuroendocrine mechanisms for insulin resistance. Normal regulation: corticotropin releasinghormone (CRH) from the hypothalamic region stimulates adrenocorticotropic hormone (ACTH) from the pituitary, which inturn drives cortisol secretion from the adrenals. This is controlled by central glucocorticoid receptors (GR). With elevatedstress this hypothalamic±pituitary±adrenal (HPA) axis is activated by central stressors. Elevated cortisol induces insulin resis-tance. Abnormal regulation: the HPA axis activity is `burned-out', presumably by prolonged stress activation, resulting in lowcortisol secretion. Gonadotropin and growth hormone releasing hormones (GnRH and GHRH) become inactivated, resulting inlow sex steroid and growth hormone (GH) secretions. The central sympathetic nervous system (SNS) becomes activated, result-ing in elevated blood pressure, heart rate and free fatty acids (FFA). Low sex steroid and growth hormones and elevated FFAbecome responsible for insulin resistance. With a dysfunction of the central GR this presumed development becomes morerapid and pronounced



[13,22±24]. This in turn would be expected to befollowed by elevated cortisol exposure of the periphery.

Cortisone is transformed to the more active gluco-corticoid cortisol by the enzyme 11-b-hydroxysteroiddehydrogenase isoform 1 which favors 11-b-reductaseactivity. This enzyme is present in a number of peripheraltissues including liver, muscle, adipose tissue andkidneys. It is thus possible that cortisol might be producedin excess locally and cause insulin resistance [40,50]. Apreponderance of this enzyme in stromal vascular cellsfrom visceral fat in culture has led to the suggestion thatthis might be a causal mechanism for speci®c accumula-tion of visceral fat, a `Cushing's disease of the omentum'[51]. The 11-b-reductase activity is stimulated by cortisoland insulin [40,50,51] and elevation of this activity maythen be secondary to hypercortisolemia from otherorigins with accompanying hyperinsulinemia, where thelocal production of cortisol ampli®es the effects of cortisolfrom the adrenals.

The BclI polymorphism referred to above has also beenfound in subjects with essential hypertension [41,43]where an increased sensitivity of the GR in the skin hasalso been found [40,41]. It seems uncertain whether theincreased GR sensitivity is tissue speci®c and thisinteresting but complex area requires further clarifyingstudies to be better understood.

Perinatal factors

Recent studies have indicated that perinatal factors maybe of importance for the development of insulinresistance and Type 2 diabetes in adult life [52]. Lowbirth weight is an indicator of such risk, and elevatedmorning cortisols have recently been found in suchsubjects at adult age [53]. The pathogenetic mechanismmight well be a programming of the HPA axis to inducehypersensitivity because exposure in utero to cytokines,which are powerful activators of the HPA axis [54], ordexamethasone, which passes the placental barrier [55]results in down-regulation of central GRs, and poorcontrol of the HPA axis. Such mechanisms are able toproduce insulin resistance and abdominal obesity inadulthood [52], resulting presumably from HPA axisperturbations similar to those described in this overview.The prevalence of insulin resistance in relation to theimpact of environmental pressure in adult life is notpossible to de®ne at present. It seems possible, however,that a perinatal programming of HPA axis function maycause a susceptibility to sensitivity to environmental stressat adult life.

In summary, there are a number of clinical featureswhich are similar in Cushing's syndrome and visceralobesity. After treatment for hypercortisolemia, patientswith Cushing's syndrome are normalized. Such clinicalobservations raise the suspicion that hypercortisolismmight also be involved in visceral obesity. There areseveral reports indicating an increased excretion ofcortisol metabolites or free cortisol particularly inabdominal obesity. Furthermore, other studies indicate

an increased production of cortisol, with normal or lowcirculating levels, in congruence with an elevated rate ofperipheral consumption or metabolism. This conditionwould presumably expose the obese organism to anabnormally elevated cortisol production irrespective ofwhether the primary cause is peripheral, with compen-sated overproduction of cortisol, or central with secon-darily changed metabolism of cortisol. There are also,however, abdominally obese, insulin resistant subjects,who display severe perturbations in the regulation of theHPA axis, which does not result in elevated cortisolsecretion, but inhibition of other hormonal secretions,and elevated activity of the sympathetic nervous system.This may well develop from the ®rst stage of elevated HPAaxis activity.

It seems likely that the primary cause of a relativehypercortisolemia in abdominal obesity is of centralorigin for the following reasons. First, elevated cortisolsecretion has been repeatedly associated with situationswhere activation of the HPA axis is known to occur,including reports of perceived stress, laboratory stresstests, as well as statistical associations to environmentalstressors. Furthermore, evidence has been found for amalfunction of the GR, controlling the HPA axis. Inaddition, this malfunction is associated with a poly-morphism of the GR gene, and although the functionalconsequences of this polymorphism are not yet known,this ®nding provides an independent credibility to thefunctional studies.

Cortisol effects on body fat distribution

With this background it seems clear that the possibilityof a perturbed HPA axis function with a stage of a relativefunctional hypercortisolism in abdominal obesity hasnow to be taken seriously. The elevated cortisol will haveat least two consequences. First, excess depot fat isdirected preferentially to abdominal visceral depots.Second, cortisol will induce insulin resistance and themechanisms in the periphery are largely identical tothose described above in the review of the consensus ofthe pathogenesis of insulin resistance, summarized inTable 1.

We will brie¯y consider the ability of cortisol to directexcess fat to visceral depots. This is seen dramatically inCushing's syndrome, disappearing with successful treat-ment [56]. The mechanisms are an elevation of lipopro-tein lipase activity, the main regulator of triglycerideaccumulation in adipocytes, by a combination of genetranscription and post-translational stabilization. Further-more, lipid mobilization is blunted in the prevailinghyperinsulinemic state resulting in a condition withpowerful lipid accumulating mechanisms. All theseevents are regulated via the local GR, which is particularlydense in visceral as compared with other depots. (Forreview and detailed references, see reference [57].)Therefore visceral accumulation of fat is occurring. Infact, visceral fat accumulation may be considered as anindex of long-term increased cortisol effects.

434 P. BjoÈrntorp


Cortisol effects on insulin actions

The interferences by cortisol on insulin effects are welldescribed. It is established that cortisol causes systemicinsulin resistance with compensatory basal and glucose-stimulated insulin secretion [58±62]. This is due tointeractions at different levels. Although the effects oninsulin receptor binding are not clear, particularly inhuman muscle and liver [63], postreceptor mechanismshave been clari®ed. Insulin stimulated glucose transportis diminished in muscle, the major site of insulinstimulated glucose uptake in vivo [64,65].

The hepatic effects of cortisol are also well described.Glucose production is enhanced by stimulation ofgluconeogenesis [66,67], by making substrate available,and by increasing the activity of gluconeogenetic enzymes[63].

Glycogen contents of the liver are increased by cortisolby increasing the activity of glycogen synthase [67,68].This may, however, not be the case in the muscle systemfor glycogen synthesis. Exposure of rats to a limited excessof glucocorticoids resulted in a decrease of insulinsensitivity after 24 h, which was parallel to a decreaseof glycogen synthesis and the insulin sensitive fraction ofglycogen synthase in, particularly, red, insulin sensitivemuscles in rats. Glucose transport and liver glycogensynthesis were not affected. After 48 h the systemicinsulin resistance was more pronounced in parallel withmuscular glycogen synthesis inhibition. Now insulin-stimulated glucose transport was also affected [65].

Another similar study has examined these eventsfurther, and these results clearly suggest that the effectsdescribed are not necessarily those of cortisol directly, butare probably mediated via an enhanced release of freefatty acids (FFA), because the cortisol effects can betotally abolished by controlling FFA mobilization [69].Whatever the direct triggering mechanism it seemsreasonably clear that cortisol excess, directly of indirectlyvia FFA, is inhibiting insulin stimulation of glycogensynthase. This seems also to be the case in the hyper-cortisolemic human [70].

Cortisol also exerts so-called permissive effects ongluconeogenesis, glycogenolysis and lipolysis [71,72]. Inthe case of lipolysis, this is of particular interest in view ofthe fact that FFA and lipid oxidation have been pointedout as central factors in the pathogenesis of insulinresistance (Table 1). Several possibilities are apparentboth on the lipolysis and reesteri®cation side of adipocytetriglyceride metabolism. The permissive effect of gluco-corticoids to facilitate catecholamine-induced lipolysis iswell-known [73]. Like most other effects of glucocorti-coids this is due to a transcription effect on the lipolyticmachinery, possibly by increasing the density of lipolyticadrenergic receptors. The b3-adrenergic receptor isknown to be expressed by glucocorticoids and has amarked lipolytic effect in humans [74]. This receptor hasa particularly high density in visceral fat, and hastherefore been considered as being responsible [75] forthe high lipolytic activity of this adipose tissue [76,77].

Lipolysis is normally markedly inhibited by insulin.Cortisol has been shown to antagonise this inhibition instudies of human adipose tissue in culture [78]. Finally,re-ester®cation of fatty acids in adipose tissue requiresglucose uptake stimulated by insulin. Thus, inhibitionof glucose uptake by cortisol [64] will diminish re-ester®cation.

Taken together, glucocorticoids exert three types ofpowerful effects on adipocyte lipid mobilization.Adrenergic lipolysis is stimulated, the antilipolytic effectof insulin blunted and reesteri®cation of fatty acidsdiminished. Each of these mechanisms would be expectedto increase fatty acid out¯ux from adipose tissue, andin combination these effects of cortisol probably explainthe elevated levels of circulating FFA in abdominal obesity,as well as contribute to the insulin resistance of thatcondition (Table 1). The enlarged adipose tissue wouldpresumably add to this effect, particularly when visceraladipose tissue is involved, which has a lively lipidmobilizing machinery. These FFAs will enter the portalvein and may have speci®c effects on hepatic metabolism,including stimulation of gluconeogenesis synthesis of verylow density lipoproteins as well as probable diminishing ofthe hepatic clearance of insulin [79±81]. An elevatedactivity of the sympathetic nervous system would beexpected to amplify the lipolytic activity further.

FFA effects on insulin sensitivity

The mechanisms whereby FFA are interfering withperipheral insulin effects are largely known. First, inmuscle critical glycolytic steps are inhibited, the so calledglucose-fatty acid cycle [82]. There is also evidence thatother metabolic pathways are inhibited by fatty acids andlipid oxidation, including glycogen synthesis [83]. Thestimulation of hepatic gluconeogenesis is another anti-insulin effect. The diminished hepatic clearance of insulin[5] would be expected to result in exaggerated hyper-insulinemia, which may down-regulate insulin receptordensity [4] to further amplify insulin resistance.

The parallel elevation of circulating insulin and FFA isan apparent paradox because insulin should be expectedto diminish FFA production by the mechanisms men-tioned above where the antilipolytic effect of insulin isparticularly powerful. A number of reports indicate thatsystemic FFA concentrations are elevated in insulinresistant conditions, including abdominal obesity, andthat insulin inhibition of these FFA concentrations is lessef®cient than normal [84±90]. This is another expressionof insulin resistance, localized to adipose tissue, mostlikely mainly localized to the antilipolytic effect of insulin.Cortisol has been shown to exert this kind of effect onadipose tissue [78], which is very closely correlated withinsulin resistance [91].

Summary of cortisol effects

Recently obtained results of sensitive measurements ofHPA axis function under everyday conditions suggest



that a considerable fraction of abdominal obesity is ahypercortisolemic condition. A smaller fraction of abdom-inal obesity is, however, associated with low cortisolsecretion. The controlling feedback is probably blunted,which may have a molecular genetic background.

Periodically elevated cortisol secretion is most likelyfollowed by accumulation of excess body fat in thevisceral region, characterizing abdominal obesity. Sys-temic insulin resistance is a well established effect ofcortisol and is localized to hepatic glucose production,muscle glycogen synthase and several steps of lipidmobilization from adipose tissue. The latter is followed byelevated circulating FFA which amplify the effects on theliver and muscles to induce insulin resistance. Theseperipheral cellular effects are identical to those describedin insulin resistance by several experts in the ®eld(Table 1), suggesting that `functionally' elevated cortisolsecretion might be a common cause to insulin resistance.In the fraction of subjects with low cortisol secretion,other mechanisms will have to be considered, as will bediscussed in the following.

Testosterone

Testosterone (T) is another hormone of major interest inthe discussion of the pathogenesis of insulin resistance. Inmales, too low and too high concentrations of T arefollowed by insulin resistance. Higher than normalconcentrations occur mainly as a result of administrationof exogenous androgens, `doping' [92], while low T levelsin men are found with aging [93] or with abdominalobesity [94]. When men with abdominal obesity and lowT levels are given T to normal for age concentrations,insulin resistance improves markedly, and insulin sensi-tivity approaches normal values. Furthermore, visceral fatmass is speci®cally diminished, as well as blood pressureand plasma lipids [95]. Other studies have shown thatT regulates glycogen synthesis and the insulin effects onglycogen synthase. When T is low, insulin effects on thestimulation of glycogen synthesis will thus be insuf®cient,resulting in systemic insulin resistance. When T isadministered, this defect is fully restituted [96].

Interestingly, muscular insulin resistance following lowT concentrations in men is also associated with enlarge-ment of visceral fat depots, diminished by T or growthhormone substitution [95,97]. T exerts powerful stimula-tion at several levels of the lipolytic pathway in adiposetissue in men. Furthermore, the lipid accumulatingpathways are inhibited. Both these effects are mediatedby an androgen receptor, which is autoregulated byandrogens, amplifying the effects. Growth hormoneexerts synergistic effects with T in these actions. (Forreview and detailed references, see reference [57].)

T is thus a hormone which diminishes fat accumula-tion by these powerful mechanisms. These effects areopposite to those of cortisol (see above), and when T andgrowth hormone secretions become too low, the cortisoleffects will dominate. Since both the GR and theandrogen receptor has higher density in visceral than

other adipose tissue regions, this imbalance will be morepronounced here, and visceral fat accumulation will bepreponderant.

In summary, in men low T secretion is followed bymuscular insulin resistance, induced by a lack of per-missive effects of insulin stimulation of glycogensynthase. In parallel, visceral accumulation of depot fatis occurring by an imbalance between cortisol on the onehand and T and growth hormone on the other, whichhave opposite effects on lipid storage, particularlypronounced in visceral fat because of high hormonalreceptor density.

In women, the prevailing pathology of androgens ishyperandrogenicity. Excess androgens of endogenous[98] or exogenous [99] origin in women or female rats[100] is followed by muscular insulin resistance, localizedto the glycogen synthase system [100]. Furthermore,glucose transport is diminished by a defect translocationof glucose transporter 4 to the cell surface [101].

At a ®rst glance it may seem paradoxical that low Tlevels in men are followed by insulin resistance while inwomen the opposite, high androgen levels, are associatedwith insulin resistance. As mentioned above, too high Tvalues in men are also followed by insulin resistance [92],apparently with the same mechanisms as in women withhyperandrogenicity [99,100]. One may therefore visua-lize gender-speci®c windows of T concentrations, whereinsulin sensitivity is optimal. For men this might beroughly in the area of 20±50 nmol/l, and in womenless than 3±4 nmol/l with no lower limit. The genderdifferences in these windows might be due to a highersensitivity of women for T exposure.

Hyperandrogenicity in women is a clinically veryimportant condition, with increased risk in developingType 2 diabetes, cardiovascular disease, prematuremortality and even certain types of cancer ([98] Lapiduset al., unpublished). The power of this risk factor may beillustrated by a 20-fold increased risk in developingdiabetes in women [98]. In fact, this might be one of themajor risk factors for prevalent disease in women. Themediating factor may well be the insulin resistancecreated by the hyperandrogenicity [99,100]. The reasonfor the increased levels of androgens is not known, andthis problem should be given a high priority for furtherresearch.

In summary, T is an important factor in the pathogen-esis of insulin resistance. The clinically importantconditions are those of low T in men, and elevated T inwomen. Insulin resistance is localized to muscles andaffects the glycogen synthase system, and in women withelevated T, also the insulin-mediated translocation ofglucose transporter 4. This condition constitutes a majorrisk for prevalent disease in the population of women.

Estrogen

Oophorectomy in female rats is followed by systemic andmuscular insulin resistance, localized mainly to glucosetransport. This is fully restituted by administration of

436 P. BjoÈrntorp


17-b-estradiol [102], indicating that estrogens areinvolved in the regulation of systemic insulin sensitivityin women. It may therefore be likely that the progressiveestrogen de®ciency with menopause is involved in therelative insulin resistance of that condition [103],although other factors are probably involved as well.

Capillarization and ®ber compositionin muscles

Endocrine effects on capillarization and muscle ®bercomposition require a speci®c comment. Capillary densityis decreased after exposure of cortisol [70], with low T inmales [96] and high T in females [100]. Insulin aloneseems to stimulate capillary outgrowth [104,105] and thesteroid hormone abnormalities discussed above thenapparently overpower this insulin effect [105]. Fibercomposition behaves differently, however, Type I ®bersdecrease in number in all the mentioned conditions inparallel with the hyperinsulinemia, and insulin itself hasa similar effect [96,100,105]. It may be that insulinregulates the synthesis of myosins.

Whether these morphological changes have anythingto do with insulin resistance or not is uncertain. (Seereviews above, references [3,5].) It seems, however, thatthese changes are also secondary to the endocrinechanges.

The sympathetic nervous system

As repeatedly mentioned above, abdominal obesity is themost prevalent condition with insulin resistance [1]. Asreviewed in a previous section there is statistical evidencethat the sympathetic nervous system activity is elevatedwhen the HPA axis function is breaking down to a `burn-out' condition, indicated by robust associations toelevated blood pressures and heart rate. In recent studiesthis has been analyzed in more detailed in about 50 menwith or without abdominal obesity. In these men,evidence of perturbed HPA axis activity is associatedwith, particularly, the regulation of central hemody-namics (Lung et al., unpublished). Elevated sympatheticnervous system activity is a hallmark of primary hyper-tension in the early, hyperkinetic state [106], and mayprovide an explanation for the commonly occurringhypertension in abdominal obesity [107]. The centersfor regulation of the HPA axis and the sympatheticnervous system are tightly connected at several levels,and activation of one of these centers is usually followedby a concomitant activation of the other. This mightexplain why hypertension is such a common comorbidityin the metabolic syndrome. This area has recently beenreviewed in detail [106].

A general increased activity of the sympathetic nervoussystem would also be expected to increase lipid mobiliza-tion. Elevated FFA mobilization has been found inabdominal obesity [5] and would be expected to amplifyinsulin resistance through established mechanisms (seeabove).

General summary of neuroendocrineeffects on the regulation of systemicinsulin actions

Cortisol clearly induces insulin resistance, localized toseveral tissues. The glycogen synthase system in muscle isinvolved as well as hepatic glucose production and theantilipolytic effect of insulin in adipose tissue. The latterresults in elevated FFA concentrations which are knownto interfere with muscular glycolysis, glycogen synthesisand probably hepatic clearance of insulin. The totalimpact of cortisol is probably a mixture of direct effectsand effects via FFA and lipid oxidation. Elevatedsympathetic nervous system activity would be expectedto amplify FFA mobilization.

Low T in men is followed by muscular insulinresistance, localized to the glycogen synthase system.This is also the case with elevated T in women where, inaddition, insulin effects on glucose transport are involved.The latter is also found with estrogen de®ciency.

Rarefaction of muscle capillaries and change in muscle®ber composition to less Type I ®bers seem to beconsequences of these endocrine abnormalities.

Interestingly, several of these endocrine changes arefollowed by accumulation of excess fat in visceral depotsthrough mechanisms which are largely known [57].This may explain why abdominal obesity is statisticallystrongly associated with insulin resistance. In factabdominal obesity is the most prevalent condition ofinsulin resistance and it seems likely that the endocrineabnormalities of that condition, a functional, moderatehypercortisolemia, low sex-speci®c steroid hormonesand hyperandrogenicity in women, are responsible forboth cardinal signs of this condition: visceral accumula-tion of body fat and insulin resistance. Insulin resistanceis likely to be ampli®ed by elevated lipid mobilizationinduced by increased activity of the sympathetic nervoussystem which will also elevate blood pressure.

The consequences of these abnormalities in terms ofcellular mechanisms are summarized in Table 4.

Comparing Table 1 with Table 4 it is apparent thatthese are essentially similar abnormalities. The conclusionmust be that the neuroendocrine abnormalities listedin Table 4 might indeed cause the insulin resistancephenomena listed in Table 1. Accumulation of visceral fatis another consequence of the endocrine perturbations.

The key issue here is the perturbed activity of the HPAaxis, because this is known to inhibit other centralendocrine axes, resulting in low secretions of gender-speci®c steroid hormones [36]. The hyperandrogenism ofwomen is most likely of adrenal origin [108], andtherefore also a likely consequence of HPA axis hyper-activity. When such a hyperactivity has occurred for asuf®ciently long period of time, a `burn-out' situation maydevelop with lower than normal cortisol secretion. In thissituation it is presumed that the sex steroid hormoneperturbations and compensatory increase of the sympa-thetic nervous system will be mainly responsible for theinsulin resistance and elevated blood pressure.



One may wonder why this neuroendocrine patho-genesis of insulin resistance has not attracted attentionpreviously. There are probably several reasons. First,methods have not been optimal for measurements ofdetails in the kinetics of HPA axis regulation (see above).Measurements of salivary cortisol to reveal kinetic detailsof diurnal cortisol secretion have probably contributedtowards a de®nite step forward in this area.

Another reason might have been a conservatism amongendocrinologists. Requirements for a diagnosis of ele-vated HPA axis activity might have been set too high, as ifCushing's syndrome were to be diagnosed. The conditionof abdominal obesity with insulin resistance and anumber of other complications may indeed be consideredas a variety of Cushing's syndrome, but the origin is not amanifest, organic perturbation, or a tumour as in classicalCushing's disease or syndrome. Instead, a functionaloversecretion is occurring, particularly as a response tostimulations. This condition requires more sensitive anddiscriminative diagnostic tools than established Cushing'ssyndrome.

Causes of HPA axis activation

A major question then becomes, why is this functionaldysregulation of the HPA axis occurring. Several factorsseem to be involved which are known to activate theHPA axis, including stressors of various types, such aspsychosocial and socioeconomic handicaps, alcohol,smoking and traits of psychiatric disease ([37,39,109]Rosmond and BjoÈrntorp, unpublished). Low socioeco-nomic status is associated with elevated stress-inducedcortisol secretion, dependent on the duration of suchhandicaps (Rosmond and BjoÈrntorp, unpublished). Theimpact of such factors probably varies depending on theindividual coping ability, as well as the genetic suscept-ibility to malfunction of the feedback control of the HPA

axis by central GRs. It seems possible that derangement ofthis control is dependent on the balance between theimpact of central stimulatory factors and the GR function,which in turn might be dependent on the status of the GRgene (Rosmond et al., unpublished).

Strong evidence for the power of psychosocial stress isobtained from controlled experiments in non-humanprimates. Monkeys in colonies form a strictly regulatedhierarchy. The monkeys at the bottom of such a hierarchybecome stressed and react with a submissive reaction.These animals show an elevation of HPA axis activity, adiminished feed-back control by central GR and enlargedadrenals. Sex steroid hormone secretion is diminished.This is followed by insulin resistance and visceral fataccumulation as well as elevated blood pressure, dimin-ished glucose tolerance and early coronary atherosclero-sis [46]. Psychosocial stress in controlled, prospectivestudies in monkeys is thus followed by exactly the sameperturbations that we see in humans subjected topsychosocial or socioeconomic handicaps.

Acquired determinants of insulinresistance

Age, smoking and physical inactivity have been consid-ered as acquired determinants of insulin effectiveness [3].Age and smoking may well be included in the generalconsideration of the endocrine basis for insulin resistance,because age is associated with decreased secretion of sexsteroids in men [93], and in women with menopause andsmoking activates the HPA axis [110].

Physical activity is, however, not possible to include inthis general scheme. Physical activity is a powerfulinducer of insulin sensitivity [111], by synergistic effectsof insulin on translocation of glucose transporter 4 inmuscle [112]. Physical inactivity is therefore an importantfactor in insulin resistance.

Concluding remarks

As brie¯y mentioned above the current synthesis of datamight at a ®rst glance be considered as an oversimpli-®cation, discussing only the effects of abnormalities inthe secretion of well-known hormones, a pathogeneticpossibility which has not been considered previously withsuf®cient seriousness. A major improvement in themethods for the diagnosis of the status of the HPA axisin the steady state and after everyday challenges has beennecessary to visualize this possibility. Another cause fornot having realized this explanation of insulin resistancemight well be that the power of the regulatory pathwaysinvolved is not suf®ciently appreciated.

The HPA axis originates in nuclei in the hypothalamicarea. In this region of the brain there are also centers forthe regulation of other vital functions such as respiration,water and electrolyte balance. The centers for regulationof the autonomic nervous system and the pituitary signals

Table 4. Effects of the hormonal abnormalities of visceralobesity on the regulation of peripheral insulin sensitivity

Testosterone

CortisolMen(low)

Women(elevated)

Insulin receptor and signallingcascade

Inhib.(Sec.)

Inhib.(Sec.)

MuscleGlycogen synthase Inhib. Inhib. Inhib.Hexokinase II Inhib.Pyruvate dehydrogenase Inhib.Capillarization Inhib. Inhib. Inhib.

LiverGlucose production Stim.Insulin clearance Inhib.

Adipose tissueLipid mobilization Stim. Inhib. Stim.

Sec.: secondary effects.Inhib.: inhibited.Stim.: stimulated.

438 P. BjoÈrntorp


to the periphery are of similarly vital importance, deter-mining circulatory and endocrine events. This regulationhas an adaptability (allostasis) [36] which aims atrestituting functions to homeostasis. When these abilitiesare exceeded abnormalities will occur. In the discussionof the pathogenesis of prevalent diseases such ascardiovascular disease and Type 2 diabetes, metabolicperturbations are usually placed in the center. Most ofthese abnormalities may well be the results of endocrinedisturbances on a neuroendocrine background, whereinsulin resistance may occupy a central role [2]. Evenmoderate abnormalities in these central regulatorysystems may, in the long run, lead to disease. Withmore severe endocrine abnormalities, disease willdevelop dramatically within a shorter time period, suchas in Cushing's syndrome. There are even acute, catas-trophic, consequences that originate in the hypothalamiccenters. Upon impacts of severe psychological trauma,such consequences may even be sudden death frommyocardial infarction (`broken heart') or cardiac arrestdue to electrical failures. These may have their origin inthrombogenetic and/or autonomic nervous systemmechanisms, aroused by hypothalamic signals. This mayoccur even in previously healthy hearts [113] and have, inessence, the same effect as execution by ®rearms or theelectric chair.

This somewhat journalistic argument might serve toteach us how endogenous mechanisms for survival can beself-destructive. It should also be realized that lessdramatic events in such dysregulations have an equallypowerful damaging effect on a longer time scale. Thereason for not having understood this might be that thedamaging signals from the central nervous system to theperiphery have been poorly de®ned. This has createddif®culties for psychologically-oriented researchers inunderstanding the statistical associations between centralabnormalities and somatic disease. Similarly, clinical andbiochemically oriented researchers have probably under-estimated the power of the central signals in theirdamaging, disease-generating actions. A closer collabora-tion between these two types of researchers will advancethis area of research.

Acknowledgements

The results of studies from the author's laboratory referred to in

this review have been obtained in collaboration with many

collaborators found in the reference list. Roland Rosmond

(MD, PhD) has been particularly instrumental in the latest

developments.

References

1. Obesity. Preventing and managing the global epidemic (1998)Report of WHO consultation on obesity. Geneva 3±5 June1997. WHO/NUT/NCD/981.

2. Reaven GM. Role of insulin resistance in human disease.Diabetes 1988; 37: 1595±1607.

3. DeFronzo R, Bonnadonna R, Ferrannini E. Pathogenesis of non-insulin dependent diabetes mellitus. In International Textbook

of Diabetes Mellitus, Alberti KGMM, Zimmet P, DeFronzo R(eds). England: Wiley, Chichester, 1997, 635±712.

4. Olefsky JM. Insulin resistance. In Ellenberg & Rifkin's DiabetesMellitus, Porte Jr D, Sherwin RS (eds). USA: Appleton & Lange,1997, 513±552.

5. Kissebah AH, Krakower GR, Sonnenberg G, Hennes MMI.Clinical manifestations of the Metabolic Syndrome. In Hand-book of obesity, Bray GA, Bouchard C, James CPT (eds). NY:Marcel Dekker, New York, 1998, 601±636.

6. Groop L, Lehto M, Orho M. Non-insulin dependent diabetesmellitus: a genetic and metabolic nightmare. Diabetes NutrMetab 1994; 7: 253±261.

7. Porter OC, Silber RH. A quantitative color reaction for cortisoneand related 17,21-dihydroxy-20-ketosteroids. J Biol Chem1950; 185: 201±207.

8. Schteingart DE, Gregerman RI, Conn JW. A comparison of thecharacteristics of increased adrenocortical function in obesityand in Cushing's syndrome. Metabolism 1963; 12: 484±497.

9. Szenas P, Pattee CJ. Studies of adrenocortical function inobesity. J Clin Endocrinol Metab 1959; 19: 344±350.

10. Dunkelman SS, Fairhurst B, Plager J, Waterhouse C. Cortisolmetabolism in obesity. J Clin Endocrinol Metab 1964; 24:832±841.

11. Posinick J, DiRaimondo V. Adrenocortical function in obesewomen. J Clin Endocrinol Metab 1956; 16: 957±958.

12. Carces LY, Kenny FM, Drash A, Taylor FA. Cortisol secretionrate during fasting of obese adolescent subjects. J ClinEndocrinol Metab 1968; 28: 1843±1847.

13. Migeon CJ, Green OC, Eckert JP. Study of adrenocorticalfunction in obesity. Metabolism 1963; 12: 718±739.

14. Murphy BPE. Clinical evaluation of urinary cortisol determina-tions by competitive protein-binding radioassay. J Clin Endo-crinol Metab 1968; 28: 343±348.

15. Streeten DHP, Stevenson CT, Dalakos TG, Nicholas JJ, DennickLG, Fellerman H. Diagnosis of hypercortisolism: biochemicalcriteria differentiating patients from lean and obese normalsubjects and from females on oral contraceptives. J ClinEndocrinol Metab 1969; 29: 1191±1211.

16. Cheek DB, Graystone JE, Seamark RF, McIntosh JE, PhillipouG, Court JM. Urinary steroid metabolites and the overgrowth oflean and fat tissues in obese girls. Am J Clin Nutr 1981; 34:1804±1810.

17. MaÊrin P, Darin N, Amemiya T, Andersson B, Jern S, BjoÈrntorpP. Cortisol secretion in relation to body fat distribution in obesepremenopausal women. Metabolism 1992; 41: 882±886.

18. Pasquali R, Cantobelli S, Casimirri F, et al. The hypothalamic-pituitary-adrenal axis in obese women with different patternsof body fat distribution. J Clin Endocrinol Metab 1993; 77:341±346.

19. Chalew SA, Zadik Z, Lozano RA, Kowarski AA. Plasma cortisollevels increase with age in obese subjects. Obes Res 1993; 1:199±202.

20. Strain GW, Zumoff B, Kream J, Strain JJ, Levin J, FukushimaDK. Sex differences in the in¯uence of obesity on the 24 hourmean plasma concentration of cortisol. Metabolism 1982; 31:209±212.

21. Strain GW, Zumoff B, Stain JJ, Levin J, Fukushima DK. Cortsolproduction in obesity. Metabolism 1980; 29: 980±985.

22. Lottenberg SA, Giannella-Neto D, Derendorf H, et al. Effect offat distribution on the pharmacokinetics of cortisol in obesity.Int J Clin Pharmacol Ther 1998; 36: 501±505.

23. Copinschi G, Cornil A, Leclerq R, Franckson JR. Cortisolsecretion rate and urinary corticoid excretion in normaland obese subjects. Acta Endocrinol (Copenh) 1966; 51:186±192.

24. Zelissen PM, Koppeschaar HP, Erkelens DW, Thijssen JH.B-Endorphin and adrenocortical function in obesity. ClinEndocrinol (Oxf) 1991; 35: 369±372.

25. Hautanen A, Adlercreutz H. Altered adrenocorticotropin andcortisol secretion in abdominal obesity: implications for theinsulin resistance syndrome. J Intern Med Suppl 1993; 234:461±469.

26. Kopelman PG, Grossman A, Lavender P. The cortisol responseto corticotrophin-releasing factor is blunted in obesity. ClinEndocrinol (Oxf) 1988; 28: 15±18.

27. Moyer A, Rodin J, Grilo C, Cummings N, Larsson L, RebuffeÂ-Scrive M. Stress-induced cortisol response and fat distributionin women. Obes Res 1994; 2: 255±262.



28. Pasquali R, Anconetani B, Chattat R, et al. Hypothalamic±pituitary±adrenal axis activity and its relationship to theautonomic nervous system in women with visceral andsubcutaneous obesity: effects of the corticotropin-releasingfactor/arginine-vasopressin test and of stress. Metabolism1996; 45: 531±536.

29. Korbonits M, Trainer PJ, Nelson ML. Differential stimulation ofcortisol and dehydroepiandrosterone levels by food in obeseand normal subjects: relation to body fat distribution. ClinEndocrinol (Oxf) 1996; 45: 699±706.

30. Rosmond R, Holm G, BjoÈrntorp P. Food-induced cortisolsecretion in relation to anthropometric, metabolic andhemodynamic variables in men. Int J Obes Relat MetabDisord, (in press).

31. Ljung T, Andersson B, Bengtsson B-AÊ, BjoÈrntorp P, MaÊrin P.Inhibition of cortisol secretion by dexamethasone in relation tobody fat distribution: a dose-response study. Obes Res 1996; 4:277±282.

32. Rosmond R, Dallman MF, BjoÈrntorp P. Stress-related cortisolsecretion in men: relationships with abdominal obesity andendocrine, metabolic and hemodynamic abnormalities. J ClinEndocrinol Metab 1998; 83: 1853±1859.

33. Vining RF, McGinley, Maksvytis JJ, Ho KY. Salivary cortisol: abetter measure of adrenal cortisol function than serum cortisol.Ann Clin Biochem 1983; 20: 329±335.

34. Laudat MH, Cerdas S, Fournier C, Guiban D, Guilhaume B,Luton JP. Salivary cortisol measurement: a practical approachto assess pituitary±adrenal function. J Clin Endocrinol Metab1988; 66: 343±348.

35. Smyth J, Ockenfels MC, Porter L, Kirschbaum C, HellhammerDH, Stone AA. Stressors and mood measured on a momentarybasis are associated with salivary cortisol secretion. Psycho-neuroendocrinology 1998; 23: 353±370.

36. McEwen BS. Protective and damaging effects of stressmediators. N Engl J Med 1998; 338: 171±179.

37. Rosmond R, Lapidus L, MaÊrin P, BjoÈrntorp P. Mental distress,obesity and body fat distribution in middle-aged men. Obes Res1996; 4: 245±252.

38. Rosmond R, Lapidus L, BjoÈrntorp P. The in¯uence ofoccupational and social factors on obesity and body fatdistribution in middle-aged men. Int J Obes Rel Metab Disord1996; 20: 599±607.

39. Brunner E, Marmot MG, Nanchahal K, Shipley MJ, Stans®eldSA, Jujena M, Alberti KGMM. Social inequality in coronary risk:central obesity and the Metabolic Syndrome. Evidence from theWhitehall II study. Diabetologia 1997; 40: 1341±1349.

40. Walker BR, Phillips DIW, Noon JJP, et al. Increased gluco-corticoid activity in men with cardiovascular risk factors.Hypertension 1998; 31: 891±895.

41. Panarelli M, Holloway CD, Fraser R, et al. Glucocorticoidreceptor polymorphism, skin vasoconstriction, and othermetabolic intermediate phenotypes in normal subjects. J ClinEndocrinol Metab 1998; 83: 1846±1852.

42. Weaver JU, Hitman GA, Kopelman PG. An association betweena BclI restriction fragment length polymorphism of theglucocorticoid receptor gene locus and hyperinsulinaemia inobese women. J Mol Endocrinol 1992; 9: 295±300.

43. Watt GCH, Harrap SB, Foy CJW, et al. Abnormalities ofglucocorticoid metabolism and a renin-angiotensin system: afour-corners approach to the identi®cation of genetic determi-nants of blood pressure. J Hypertension 1992; 10: 473±482.

44. Buemann B, Vohl M-C, Chagnon M, et al. Abdominal visceralfat is associated with a Bcl I restriction fragment lengthpolymorphism at the glucocorticoid receptor gene locus. ObesRes 1997; 5: 186±192.

45. Riggins GJ, Lokey LK, Chastain JL, et al. Human genescontaining polymorphic trinucleotide repeats. Nat Genet1992; 2: 186±191.

46. Shively CA, Laber-Laird K, Anton RF. Behavior and physiologyof social stress and depression in female Cynomolgus monkeys.Biol Psychiatry 1997; 41: 871±882.

47. RebuffeÂ-Scrive M, BroÈnnegaÊrd M, Nilsson A, Eldh J, GustafssonJAÊ, BjoÈrntorp P. Steroid hormone receptors in human adiposetissues. J Clin Endocrinol Metab 1990; 71: 1215±1219.

48. Walker RB, Connacher AA, Webb DJ, Edwards CRW. Gluco-corticoids and blood pressure: a role for the cortisol/cortisoneshuttle in control of vascular tone in man. Clin Sci 1992; 83:171±178.

49. Hautanen A, RaÈikkoÈnen K, Adlercreutz H. Associationsbetween pituitary±adrenocortal function and abdominal obe-sity, hyperinsulinemia and dyslipidemia in normotensivemales. J Int Med Res 1997; 241: 451±461.

50. Seckl JR, Chapman KE. The 11-b hydroxysteroid dehydrogen-ase system, a determinant of glucocorticoid and mineralocor-ticoid action. Eur J Biochem 1997; 249: 361±364.

51. Bajalska JJ, Kumars, Stewart PM. Does central obesity re¯ext`Cushing's disease of the omentum'. Lancet 1997; 349:1210±1213.

52. Barker DJP. Fetal origins of coronary heart disease. Br Med J1995; 311: 171±174.

53. Phillips DIW, Barker DJP, Fall CHD, et al. Elevated plasmacortisol concentrations: a link between low birth weight andthe insulin resistance syndrome? J Clin Endocrinol Metab 1998;83: 757±760.

54. Reul JMHM, Stec J, Wiegers GJ, et al. Prenatal immunechallenge alters the hypothalamic±pituitary axis in adult rats.J Clin Invest 1994; 93: 2600±2607.

55. Levitt NS, Lindsay RS, Holmes GE, Seckl JR. Dexamethasone inthe last week of pregnancy attenuates hippocampal glucocor-ticoid receptor gene expression and elevates blood pressure inthe adult offspring of rats. Neuroendocrinolology 1996; 64:412±418.

56. LoÈnn L, Kvist H, Ernest I, SjoÈstroÈm L. Changes in bodycomposition and adipose tissue distribution after treatment ofwomen with Cushing's syndrome. Metabolism 1994; 43:1517±1522.

57. BjoÈrntorp P. Review: the regulation of adipose tissue distribu-tion in humans. Int J Obes Rel Metab Disord 1996; 20: 291±302.

58. Perley M, Kipnis D. Effect of glucocorticoids on plasma insulin.N Engl J Med 1966; 274: 1237±1241.

59. Kitabchi A, Jones G, Duckworth W. Effect of hydrocortisoneand corticotropin on glucose induced insulin and proinsulinsecretion in man. J Clin Endocrinol Metab 1973; 37: 79±84.

60. Fantus I, Ryan J, Hizuka N, Gordon P. The effect ofglucocorticoids on the insulin receptor: an in vivo and in vitrostudy. J Clin Endocrinol Metab 1981; 52: 953±960.

61. Kalhan S, Adam P. Inhibitory effect of prednisone on insulinsecretion in man: model for duplication of blood glucoseconcentration. J Clin Endocrinol Metab 1975; 41: 600±610.

62. Wise J, Hendler R, Felig P. In¯uence of glucocortcoids onglucagon secretion and plasma amino acid concentrations inman. J Clin Invest 1973; 52: 2774±2782.

63. McMahon M, Gerich J, Rizza R. Effects of glucocorticoids oncarbohydrate metabolism. Diabetes Metab Rev 1988; 4: 17±30.

64. Riddick F, Reisler D, Kipnis D. The sugar transport system instriated muscle. Diabetes 1962; 11: 171±178.

65. HolmaÈng A, BjoÈrntorp P. The effects of cortisol on insulinsensitivity in muscle. Acta Physiol Scand 1992; 144: 425±431.

66. Exton J, Mallettie L, Jefferson L, et al. The hormonal control ofhepatic gluconeogenesis. Recent Prog Horm Res 1970; 26:411±457.

67. Kraus-Friedmann N. Hormonal regulation of hepatic gluco-neogenesis. Physiol Rev 1984; 64: 170±259.

68. DeWulff H, Hers H. The stimulation of glycogen synthase andof glycogen synthase in the liver by glucocorticoids. EurJ Biochem 1967; 2: 57±60.

69. Guillaume-Gentil C, Assimacopoulos-Jeanneau IF, JeanrenaudB. Involvement of non-esteri®ed fatty acid oxidation inglucocorticoid-induced peripheral insulin resistance in vivo inrats. Diabetologia 1993; 36: 899±906.

70. RebuffeÂ-Scrive M, Krotkiewski M, Elfverson J, BjoÈrntorp P.Muscle and adipose tissue morphology and metabolism inCushing's syndrome. J Clin Endocrinol Metab 1988; 67:1122±1128.

71. Exton J, Corbin J, Park C. Control of gluconeogenesis in liver.J Biol Chem 1969; 244: 4095±4102.

72. Schaeffer L, Chenoweth M, Dunn A. Adrenal corticosteroidinvolvement in the control of liver glycogen phosphorylaseactivity. Biochim Biophys Acta 1969; 192: 292±303.

73. Reynisdottir S, Wahrenberg H, Bylin G, Arner P. Effect ofglucocorticoid treatment on b-adrenoceptor subtype functionin adipocytes from patients with asthma. Clin Sci 1993; 85:237±244.

74. LoÈnnqvist F, Krief S, Strosberg A, Nyberg B, Emorine L, Arner P.

440 P. BjoÈrntorp


Evidence for a functional b3-adrenoceptor in man. BrJ Pharmacol 1998; 110: 929±936.

75. LoÈnnqvist F, ThoÈrne A, Nilsell K, Hoffstedt J, Arner P. Apathogenic role of visceral b3-adrenoceptor in obesity. J ClinInvest 1995; 95: 1109±1116.

76. RebuffeÂ-Scrive M, Andersson B, Olbe L, BjoÈrntorp P. Metabo-lism of adipose tissue in intraabdominal depots of non-obesemen and women. Metabolism 1989; 38: 453±458.

77. RebuffeÂ-Scrive M, Andersson B, Olbe L, BjoÈrntorp P. Metabo-lism of adipose tissue in intraabdominal depots in severelyobese men and women. Metabolism 1990; 39: 1021±1025.

78. Cigolini M, Smith U. Human adipose tissue in culture. VIII.Studies on the insulin-antagonistic effect of glucocorticoids.Metabolism 1979; 20: 502±510.

79. BjoÈrntorp P. `Portal' adipose tissue as a generator of risk factorsfor cardiovascular disease and diabetes. Arterioscler ThrombVasc Biol 1990; 10: 493±496.

80. Banerji M, Buckley C, Chaiken R, Gordon D, Lebovitz H, Kral J.Liver fat, serum triglycerides and visceral adipose tissue ininsulin-sensitive and insulin-resistant black men with NIDDM.Int J Obes Relat Metab Disord 1995; 19: 846±850.

81. BjoÈrntorp P. Guest commentary: liver triglycerides andmetabolism. Int J Obes Relat Metab Disord 1995; 19: 839±840.

82. Randle P, Garland P, Hales N, Newsholme E. The glucose-fattyacid cycle: its role in insulin sensitivity and the metabolicdisturbances of diabetes mellitus. Lancet 1963; 785±789.

83. Wititsuwannakul D, Ki-Han. Mechanism of palmityl coenzymeA inhibition of liver glycogen synthase. J Biol Chem 1977; 252:7812±7817.

84. Groop LC, Saloranta C, Shank M, Bonadonna RC, Ferrannini E,DeFronzo RA. The role of the free fatty acid metabolism in thepathogenesis of insulin resistance in obesity and non insulin-dependent diabetes mellitus. J Clin Endocrinol Metab 1991; 72:96±107.

85. Bonadonna RC, Groop L, Kraemer N, Ferrannini E, Del Prato S,DeFronzo RA. Obesity and insulin resistance in humans: adose-response study. Metabolism 1990; 39: 452±459.

86. Howard B, Klimies J, Vasques B, Brady D, Nagulesparan M,Unger RH. The antilipolytic action of insulin in obese subjectswith resistance to its glucoregulatory action. J Clin EndocrinolMetab 1984; 58: 544±551.

87. Chen Y-DI, Golay A, Swislocki ALM, Reaven GM. Resistance toinsulin suppression of plasma free fatty acid concentrations andinsulin stimulation of glucose uptake in noninsulin-dependentdiabetes mellitus. J Clin Endocrinol Metab 1987; 64: 17±21.

88. Zancanaro C, Cigolini M, Bonora E, Moghetti P, Querena M,Muggeo M. Glucose and insulin suppression of plasma freefatty acids in obese subjects with normal glucose tolerance ormild, newly diagnosed Type 2 (non-insulin-dependent) dia-betes. J Endocrinol Invest 1990; 13: 55±59.

89. Felber JP, Golay A, Felley C, Jecquier E. Regulation of glucosestorage in obesity and diabetes: metabolic aspects. DiabetesMetab Rev 1988; 4: 691±700.

90. Coppack SW, Evans RD, Fischer RM, et al. Adipose tissuemetabolism in obesity: lipase action in vivo before and after amixed meal. Metabolism 1992; 41: 264±272.

91. MaÊrin P, Andersson B, Ottosson M, et al. The morphology andmetabolism of intraabdominal adipose tissue in men. Metabo-lism 1992; 41: 1242±248.

92. Cohen JC, Hickman R. Insulin resistance and diminishedglucose tolerance in power lifters ingesting anabolic steroids.J Clin Endocrinol Metab 1987; 64: 960±971.

93. Vermeulen A. Androgens in the aging male. J Clin EndocrinolMetab 1991; 73: 221±224.

94. Seidell JC, BjoÈrntorp P, SjoÈstroÈm L, Kvist H, Sannerstedt R.

Visceral fat accumulation in men is positively associated withinsulin glucose and C-peptide levels, but negatively with insulinlevels. Metabolism 1990; 39: 897.

95. MaÊrin P, HolmaÈng S, Gustafsson C, et al. Androgen treatmentof abdominally obese men. Obes Res 1993; 1: 245±251.

96. HolmaÈng A, BjoÈrntorp P. The effects of testosterone on insulinsensitivity in male rats. Acta Physiol Scand 1992; 146: 505±510.

97. Johannsson G, MaÊrin P, LoÈnn L, et al. Growth hormonetreatment of abdominally obese men reduces abdominal fatmass, improves glucose and lipoprotein metabolism, andreduces diastolic blood pressure. J Clin Endocrinol Metab1997; 82: 727±734.

98. Lindstedt G, Lundberg PA, Lapidus L, Lundgren H, BengtssonC, BjoÈrntorp P. Low sex-hormone-binding globulin concentra-tion as independent risk factor for development of NIDDM12-yr follow-up of population study of women in Gothenburg,Sweden. Diabetes 1991; 40: 123±128.

99. Polderman K, Gooren L, Asscheman H, Bakker A, Heine R.Induction of insulin resistance by androgens and estrogens.J Clin Endocrinol Metab 1994; 79: 265±271.

100. HolmaÈng A, Svedberg J, Jennische E, BjoÈrntorp P. Effects oftestosterone on muscle insulin sensitivity and morphology infemale rats. Am J Physiol 1990; 259: E555±E560.

101. Rincon J, HolmaÈng A, Odegaard WahlstroÈm E, et al. Mechan-isms behind insulin resistance in rat skeletal muscle afteroophorectomy and additional testosterone treatment. Diabetes1996; 45: 615±621.

102. Kumagai S, HolmaÈng A, BjoÈrntorp P. The effects of oestrogenand progesterone on insulin sensitivity in female rats. ActaPhysiol Scand 1993; 149: 91±97.

103. Barrett-Connor E, Laakso M. Ischemic heart disease risk inpost-menopausal women. Effects of estrogen use on glucoseand insulin levels. Arterioscler Thromb Vasc Biol 1990; 10:531±534.

104. HolmaÈng A, Brzezinska Z, BjoÈrntorp P. Effects of hyperinsu-linemia on muscle ®ber composition and capillarization in rats.Diabetes 1993; 42: 1073±1081.

105. HolmaÈng A, Jennische E, BjoÈrntorp P. Rapid formation ofcapillary endothelial cells in rat skeletal muscle after exposureto insulin. Diabetologia 1996; 39: 206±211.

106. BjoÈrntorp P, Rosmond R, Holm G, Folkow B. Hypertension andthe Metabolic Syndrome. J Hypertension (in press).

107. Rosmond R, BjoÈrntorp P. Blood pressure in relation to obesity,insulin and the hypothalamic-pituitary-adrenal axis in Swedishmen. J Hypertension 1998; 16: 1721±1726.

108. Vague J, Vague P, Boyer J, Cloix M. Anthropometry of obesity,diabetes, adrenal, and beta-cell functions. Exc Med Int Congr1970; 23: 517±525.

109. BjoÈrntorp P. Visceral fat accumulation: the missing linkbetween psycho-social factors and cardiovascular disease?J Int Med 1991; 230: 195±201.

110. Gossain VV, Sherma NK, Srivastava L, Michelakis AM, RovnerDR. Hormonal effects of smoking. II: effects on plasma cortisol,growth hormone, and prolactin. Am J Med Sci 1986; 291:325±327.

111. BjoÈrntorp P, de Jounge K, SjoÈstroÈm L, Sullivan L. The effect ofphysical training on insulin production in obesity. Metabolism1970; 19: 631±638.

112. Hayashi T, Wojtaszewski JF, Goodyear LJ. Exercise regulationof glucose transport in skeletal muscle. Am J Physiol 1997; 273:E1039±E1051.

113. Uno H, Tarara R, Else JG, Suleman MA, Sapolsky RM.Hippocampal damage associated with prolonged and fatalstress in primates. J Neurosci 1989; 9: 1705±1711.



neuroendocrine perturbations as a cause of insulin resistance

Documents