Glucose-induced microvascular functional changes in nondiabetic rats are stereospecific and are prevented by an aldose reductase inhibitor.

Exposure of skin chamber granulation tissue vessels in nondiabetic rats to 11 or 15 mM D-glucose (but not L-glucose or 3-O-methylglucose) twice daily for 10 d induces vascular functional changes (increased albumin permeation and blood flow) identical to those in animals with mild or severe streptozotocin diabetes, respectively. These vascular changes are strongly linked to increased metabolism of glucose via the sorbitol pathway and are independent of nonenzymatic glycosylation as well as systemic metabolic and hormonal imbalances associated with the diabetic milieu. (J. Clin. Invest. 1990. 85:1167-1172.)

Galactose ingestion increases vascular permeability and collagen solubility in normal male rats.

In view of the similarity of cataracts and neuropathy in galactose-fed and diabetic rats, the present experiments were undertaken to determine whether consumption of galactose-enriched diets (10, 25, or 50% by weight) also increases collagen crosslinking and permeation of vessels by 125I-albumin analogous to that observed in diabetic rats. The observations in these experiments: demonstrate that consumption of galactose-enriched diets for 3 wk selectively increases 125I-albumin permeation of the same vascular beds affected in diabetic rats and by diabetic vascular disease in humans (i.e., the aorta and vessels in the eye, kidney, sciatic nerve, and new tissue formed in the diabetic milieu); demonstrate that the susceptibility of the vasculature to aldose reductase-linked injury (increased permeability) varies greatly in different tissues; indicate that collagen solubility (crosslinking) changes in galactose-fed rats differ sharply from those in diabetic rats; and provide new evidence that consumption of galactose-enriched diets induces a hypogonadal state in male rats.

Capillary basement membranes in diabetes.

In discussions of vascular complications of diabetes the fact that capillary basement membranes are, in general, thickened (CBMT) in poorly controlled diabetics is no longer at issue. However, three important questions concerning the pathophysiologic significance of CBMT remain unanswered: (1) How and why do capillary basement membranes thicken in diabetes? (2) What is the functional significance of capillary basement membrane changes in diabetes? (3) What is the nature of the relationship of CBMT to other forms of diabetic vascular disease; in particular, is CBMT observed in tissues amenable to needle biopsy, i.e., skeletal muscle, useful in identifying individuals at high risk for developing clinically significant retinopathy, nephropathy, or atherosclerotic vascular disease? In this survey, we will consider the nature of capillary basement membrane changes in diabetes and subsequently address the above questions.

Basement-membrane thickening and diabetic microangiopathy.

This report considers the pathophysiologic significance of capillary basement-membrane thickening in diabetic nephropathy and retinopathy and the relationship of capillary basement-membrane thickening to increased susceptibility to infections and to increased vascular permeability in diabetes. The evidence available (1) indicates that basement-membrane thickening affects most if not all capillaries of the diabetic and may contribute to increased susceptibility to infection and (2) suggests that increased capillary permeability in diabetes need not be attributed to basement-membrane changes per se, but rather may be due to changes in the cellular elements of the capillary wall.

Histocompatibility antigens and diabetic retinopathy.

Of 160 patients with onset of diabetes at or after 30 years of age, the 84 with no evidence prevalences of HLA-A1 and B8 when compared with the 76 with retinal complications or with the 282 healthy blood donors. In addition, in 90 patients with onset of diabetes before age 30 years, we could confirm the reported significant increase of HLA-B8 and decrease of B7, but no differences were noted between those juvenile-onset diabetics with and those without retinopathy.

Influence of fixation and morphometric technics on capillary basement-membrane thickening prevalence data in diabetes.

Muscle tissue obtained by needle biopsy from 20 diabetic subjects and from 20 age-matched control subjects was divided into two portions, one of which was fixed and processed by our routine procedure (primary glutaraldehyde fixation followed by osmium fixation and embedment in araldite) and the other was fixed initially in osmium tetroxide and embedded in maraglas, the procedure employed by Siperstein et al. Basement-membrane width of capillaries was measured by the 20-point method of Siperstein et al. and by the two-minimum-point technic developed in our own laboratory. Contrary to the experience of Siperstein et al., the prevalence of basement-membrane thickening in diabetic subjects based on mean width values and/or standard deviations in excess of 95 per cent tolerance intervals was highest (65 per cent) in minimum measurements of glutaraldehyde-fixed tissues and lowest (30 per cent) in osmium-fixed tissues (X2 = 4.9123, p less than 0.05). Internal discrepancies in the data of Siperstein et al. indicate that (1) their basement-membrane-width values derived from multiple measurements from control subjects are anomalous (low) and (2) the very high prevalence of basement-membrane thickening they reported in diabetic and in prediabetic subjects and considered as strong support for their conclusion that basement-membrane disease is independent of and precedes glucose intolerance is suspect.

Quantitative vitreous fluorophotometry. A sensitive technique for measuring early breakdown of the blood-retinal barrier in young diabetic patients.

Vitreous and aqueous humor fluorescein concentrations were measured one hour after graded intravenous fluorescein was given to 20 juvenile diabetics, ages 20 to 40, with and without retinopathy, and to 12 controls of similar age. Vitreous fluorescein concentrations were significantly higher in diabetics, indicating breakdown of the blood-retinal barrier. Mean vitreous fluorescein values were 10.66 +/- 0.65 for the diabetics and 4.28 +/- 0.37 ng./ml. for the controls. Breakdown of the blood-retinal barrier was also confirmed in diabetics under the age of 20 without retinopathy. The blood-aqueous barrier was similarly altered in diabetics. Vitreous fluorophotometry quantitatively measures breakdown of the blood-retinal barrier, possibly the earliest detectable ocular vascular abnormality in juvenile diabetic patients.

Albumin permeation of new vessels is increased in diabetic rats.

125I-bovine serum albumin (BSA) permeation of the vasculature of 3-wk-old granulation tissue (induced by subcutaneous implantation of polyester fabric) formed in the diabetic milieu was assessed in female BB/W, spontaneously diabetic rats and in male, Sprague-Dawley rats with streptozocin-induced diabetes as well as in corresponding nondiabetic controls. Albumin permeation of new granulation tissue vessels was markedly increased in both groups of diabetic animals relative to that of nondiabetic controls, while albumin permeation of vessels in most other tissues did not differ for controls and diabetics. These observations indicate that the functional integrity of new vessels formed in the diabetic milieu is impaired: (1) to a greater extent than that of older vessels formed before induction of diabetes and (2) relative to new vessels in nondiabetics. The implication of these observations is that molecular constituents of vessels synthesized in the diabetic milieu are quantitatively and/or qualitatively abnormal and/or their incorporation into vessels is defective.

Hemoglobin AIc as an indicator of the degree of glucose intolerance in diabetes.

Hemoglobin AIc concentration, fasting blood sugar, response to an oral glucose tolerance test, and skeletal muscle capillary basement membrane thickness were measured in diabetic patients. Hemoglobin AIc concentration correlates with both response to a glucose tolerance test (r = 0.82, p less than 0.001) and fasting blood sugar (r = 0.62, p less than 0.001). The correlation of hemoglobin AIc concentration with glucose tolerance is independent of fasting blood sugar concentration (partial r = 0.61, p less than 0.005), whereas that of hemoglobin AIc with fasting blood sugar probably reflects the relationship between fasting blood sugar levels and glucose tolerance (partial r = 0.22, p less than 0.05). Hemoglobin AIc levels do not correlate with basement membrane thickness ( r = 0.15, p less than 0.05).

Monozygotic triplets with discordance for diabetes mellitus and diabetic microangiopathy.

A set of monozygotic triplets (PE.K., P.K., S.K.) has been studied. There is no diabetes in first-degree relatives. PE.K. developed insulin-requiring (60 U. NPH) diabetes at the age of 13 years. Over a period of 11 years since that time, numerous studies of insulin and growth-hormone secretion were performed on P.K. and S.K., including multiple oral glucose tolerance tests (OGTTs), cortisone-primed oral glucose tolerance tests (C-OGTTs), intravenous glucose tolerance tests (IVGTTs), and intravenous tolbutamide tests (IVTTs). The results of each test were compared with age- and sex-matched control subjects. P. K. developed insulin-requiring (56 U. NPH) diabetes after remaining discordant for eight years. Glucose, insilin, and growth-hormone responses during all tests were normal except during the IVGTT performed four months prior to the onset of diabetes. This last IVGTT revealed a glucose disappearance rate of 0.98 per cent per minute, and the slope of the regression line of serum-insulin response (IRI) on blood glucose (BG) was markedly decreased to 0.005 micronU./ml. IRI/mg./dl. BG (controls 0.340 +/- 0.04; mean +/- S.E.M.). The insulin responses in P.K. and S.K. were similar during all OGTTs, C-OGTTs, and IVTTs. S.K. has continued to maintain normal glucose tolerance and normal insulin and growth-hormone responses during all tests. The histocompability antigen studies have revealed HLA-A2, AW24, BW15, and BW40 phenotype in these monozygotic triplets. Muscle capillary basement membranes of the nondiabetic triplet were normal, whereas both diabetic triplets manifested evidence of capillary basement membrane thickening. The clinical and biochemical profiles in these triplets and the capillary basement membrane data lend strong credence to the role of "nongenetic" determinants in the development of "genetic" diabetes as well as diabetic microangiopathy in juvenile-onset-type diabetes.

Thin muscle capillary basement membranes in myotonic dystrophy.

Muscle capillary basement membrane width (MCBMW) was measured in 18 myotonic dystrophy patients and compared with that in age- and sex-matched normal and diabetic subjects. The MCBMW in myotonic dystrophy patients (773 +/- 258 A) was significantly thinner than in normal subjects (925 +/- 181 A, P less than 0.05) or in diabetics (1224 +/- 614 A, P less than 0.01). An increase in MCBMW with advancing age was present in all groups but was greatest in the myotonic dystrophy groups (r = +0.59, P less than 0.01). There was no relation between MCBMW and either the degree of glucose intolerance or insulin hypersecretion in the myotonic dystrophy group, though none had fasting hyperglycemia. This is the first report of a condition associated with thinner-than-normal capillary basement membrane.

Increased collagen cross-linkages in experimental diabetes: reversal by beta-aminopropionitrile and D-penicillamine.

The effects of diabetes on collagen cross-link formation and solubility were investigated in granulation tissue collagen induced by polyester fabric implanted subcutaneously in rats at the same time diabetes was produced by injection of streptozotocin. Thus, all the collagen analyzed was formed in a diabetic milieu. Ten days later the implants were removed and the total collagen content as well as the fraction soluble in 0.5 M acetic acid was determined. Predominantly type I collagen accumulated in the implants. Total collagen content was the same in diabetics and controls; however, the acid-soluble fraction in diabetic animals was only half that of controls (8.5% and 17.7%, respectively), and the ratio of beta chains to alpha chains in the acid-soluble fraction was higher in diabetics (0.89) than in controls (0.69). In animals treated with beta-aminopropionitrile or D-penicillamine the acid-soluble fraction of collagen from diabetics equaled that from controls. These observations indicate that both intramolecular and intermolecular cross-links are increased in type I collagen from diabetic animals. Since these cross-links interfere with degradation of collagen by collagenase, they may contribute to accelerated intimal sclerosis of arteries and to capillary basement membrane thickening in diabetes.

Effects of islet isografts on hemodynamic and vascular filtration changes in diabetic rats.

To assess the reversibility of diabetes-induced increases in regional vascular albumin permeation and blood flow and changes in kidney filtration function, islet isografts were given via the portal vein after 2 mo of streptozocin-induced diabetes in male Lewis rats. One month later, vascular function was assessed in control rats, islet-transplanted diabetic rats, and untreated diabetic rats (6-9 rats/group). Untreated diabetic rats were markedly hyperglycemic, hyperphagic, and polyuric. Transplanted rats were euglycemic within 6 days; 24-h urine volumes were virtually normalized by 2 wk and food consumption was normalized 4 wk after transplantation. Vascular albumin permeation in diabetic rats was significantly increased 1.4- to 1.7-fold in anterior uvea, choroid, retina, sciatic nerve, new granulation tissue, and kidney and was increased 1.1- to 1.3-fold in diaphragm, cecum, and optic nerve. Albumin permeation was not increased in aorta, brain, heart, or forelimb skeletal muscle. Islet transplants significantly reduced but did not completely normalize vascular albumin permeation in most tissues in which it was increased by diabetes but had no effect on albumin permeation in optic nerve, sciatic nerve, and diaphragm. Urinary excretion of endogenous albumin and IgG in diabetic rats was significantly increased 19- and 14-fold, respectively, and was virtually normalized 4 days after islet transplantation. Marked (1.8-fold) increases in glomerular filtration rate (GFR) in diabetic rats were also substantially reduced by islet transplants but remained elevated 1.4-fold control values. Likewise, diabetes-induced increases in regional blood flow were reduced in general but not normalized by islet transplants. These observations indicate that 1) diabetes-induced hemodynamic changes and alterations in vascular filtration function are not rapidly reversed by euglycemia after islet transplantation, 2) diabetes-induced increases in urinary albumin and IgG excretion are more readily normalized by euglycemia than increases in GFR and renal 125I-labeled bovine serum albumin (125I-BSA) filtration, and 3) significant increases in GFR and renal 125I-BSA filtration may not be manifested by albuminuria.

Modulation of hemodynamic and vascular filtration changes in diabetic rats by dietary myo-inositol.

To assess the potential of myo-inositol-supplemented diets to prevent diabetes-induced vascular functional changes, we examined the effects of diets supplemented with 0.5, 1, or 2% myo-inositol on blood flow and vascular filtration function in nondiabetic control rats and rats with streptozocin-induced diabetes (STZ-D). After 1 mo of diabetes and dietary myo-inositol supplementation, 1) 131I-labeled bovine serum albumin (BSA) permeation of vessels was assessed in multiple tissues, 2) glomerular filtration rate (GFR) was estimated as renal plasma clearance of 57Co-labeled EDTA, 3) regional blood flows were measured with 15-microns 85Sr-labeled microspheres, and 4) endogenous albumin and IgG urinary excretion rates were quantified by radial immunodiffusion assay. In STZ-D rats, 131I-BSA tissue clearance increased significantly (2- to 4-fold) in the anterior uvea, choroid-sclera, retina, sciatic nerve, aorta, new granulation tissue, diaphragm, and kidney but was unchanged in skin, forelimb muscle, and heart. myo-Inositol-supplemented diets reduced diabetes-induced increases in 131I-BSA clearance (in a dose-dependent manner) in all tissues; however, only in new granulation tissue and diaphragm did the 2% myo-inositol diet completely normalize vascular albumin permeation. Diabetes-induced increases in GFR and in urinary albumin and IgG excretion were also substantially reduced or normalized by dietary myo-inositol supplements. Increased blood flow in anterior uvea, choroid-sclera, kidney, new granulation tissue, and skeletal muscle in STZ-D rats also was substantially reduced or normalized by the 2% myo-inositol diet. myo-Inositol had minimal if any effects on the above parameters in control rats. These observations indicate that diabetes-induced increases in regional blood flow, 131I-BSA permeation, GFR, and urinary protein excretion can be markedly reduced or normalized by consumption of myo-inositol-supplemented diets that raise plasma myo-inositol levels approximately fivefold. The failure of the 2% myo-inositol diet to normalize GFR and blood flow and albumin permeation in several tissues despite markedly elevated plasma myo-inositol levels and normal or elevated tissue myo-inositol levels indicates that if vascular functional changes in these tissues are linked to altered myo-inositol levels, they are resistant to normalization by elevation of plasma myo-inositol levels. These results suggest that other factors independent of changes in relative or absolute tissue myo-inositol levels may play an important role in the pathogenesis of diabetes-induced vascular functional changes in these tissues.(ABSTRACT TRUNCATED AT 400 WORDS)

Neural dysfunction and metabolic imbalances in diabetic rats. Prevention by acetyl-L-carnitine.

The rationale for these experiments is that administration of L-carnitine and/or short-chain acylcarnitines attenuates myocardial dysfunction 1) in hearts from diabetic animals (in which L-carnitine levels are decreased); 2) induced by ischemia-reperfusion in hearts from nondiabetic animals; and 3) in nondiabetic humans with ischemic heart disease. The objective of these studies was to investigate whether imbalances in carnitine metabolism play a role in the pathogenesis of diabetic peripheral neuropathy. The major findings in rats with streptozotocin-induced diabetes of 4-6 weeks duration were that 24-h urinary carnitine excretion was increased approximately twofold and L-carnitine levels were decreased in plasma (46%) and sciatic nerve endoneurium (31%). These changes in carnitine levels/excretion were associated with decreased caudal nerve conduction velocity (10-15%) and sciatic nerve changes in Na(+)-K(+)-ATPase activity (decreased 50%), Mg(2+)-ATPase (decreased 65%), 1,2-diacyl-sn-glycerol (DAG) (decreased 40%), vascular albumin permeation (increased 60%), and blood flow (increased 65%). Treatment with acetyl-L-carnitine normalized plasma and endoneurial L-carnitine levels and prevented all of these metabolic and functional changes except the increased blood flow, which was unaffected, and the reduction in DAG, which decreased another 40%. In conclusion, these observations 1) demonstrate a link between imbalances in carnitine metabolism and several metabolic and functional abnormalities associated with diabetic polyneuropathy and 2) indicate that decreased sciatic nerve endoneurial ATPase activity (ouabain-sensitive and insensitive) in this model of diabetes is associated with decreased DAG.

Muscle capillary basement membrane width and its relationship to diabetes mellitus in monozygotic twins.

Quadriceps (Q) and gastrocnemius (G) muscle capillary basement membrane width (CBMW) were measured in 18 pairs of monozygotic (MZ) twins. Thirteen of these twin pairs were discordant for insulin-dependent diabetes (IDD) and five pairs were concordant for either IDD (two pairs) or for non-insulin-dependent diabetes (NIDD). In 12 of the 13 nondiabetic (ND) twin mates of IDD, 50 oral glucose tolerance tests performed in the years before or after determination of CBMW revealed mean blood glucose levels in the 36-52 percentile range, compared with normal controls. The mean (+/-SD) age at the onset of IDD in discordant twins was 18.7 +/- 10.1 (range 8-37) yr and the mean duration of discordance at the time of biopsy was 13.6 +/- 8.3 (range 3-32) yr. CBMW data were compared within each twin (Q versus G) and between twin mates and age- and sex-matched controls. Overall, CBMW of IDD twins was greater than that of their ND twin mates. Differences between IDD and ND twins, however, were much more marked in gastrocnemius (1859 +/- 643 versus 1222 +/- 307 A, P less than 0.0003) than in quadriceps (1291 +/- 319 versus 1112 +/- 302 A; P less than 0.04). CBMW in gastrocnemius was significantly thicker than that in the quadriceps of IDD twins (t = 4.55, P less than 0.0008) but not in their ND twin mates (t = 1.15, P less than 0.27). CMBW was significantly thicker in IDD than in their ND twin mates (in quadriceps and/or gastrocnemius) in 10 of the 12 twin pairs.(ABSTRACT TRUNCATED AT 250 WORDS)

Muscle capillary basement membrane in identical twins discordant for insulin-dependent diabetes.

Although hereditary factors clearly modulate susceptibility to develop diabetes, their role as determinants of vascular complications associated with diabetes remains unclear. These studies were undertaken to further assess the extent to which capillary basement membrane thickening (CBMT) is governed by metabolic derangements associated with relative or absolute insulin deficiency versus genetic determinants of vascular disease closely linked to but independent of those modulating susceptibility to develop relative or absolute insulin deficiency. Quadriceps muscle capillary basement membranes obtained by needle biopsy were examined in eight pairs of identical twins discordant for insulin-dependent diabetes (IDD) for 11-29 yr. Biopsy material from one of the diabetic twins was technically unsuitable for study. The average CBM width of the IDD twins was found to be significantly thicker than that of their nondiabetic (ND) twin mates (t = 2.50, P less than 0.025). Three IDD, but none of the ND twins, had basement membrane width values in excess of 95% upper tolerance intervals for age- and sex-matched controls with no family history of diabetes. The absence of CBMT in all of the ND twins and in four of the IDD twins with diabetes of 15-24 yr duration argues against the existence, in this group of subjects, of hereditary determinants of diabetic vascular disease linked to those governing susceptibility to develop diabetes. In addition, the absence of CBMT in four subjects with IDD of 15-24 yr duration is consistent with evidence from other studies indicating that diabetic microangiopathy is not an inevitable consequence of the diabetic milieu.(ABSTRACT TRUNCATED AT 250 WORDS)

Hyperglycemic pseudohypoxia and diabetic complications.

Vasodilation and increased blood flow are characteristic early vascular responses to acute hyperglycemia and tissue hypoxia. In hypoxic tissues these vascular changes are linked to metabolic imbalances associated with impaired oxidation of NADH to NAD+ and the resulting increased ratio of NADH/NAD+. In hyperglycemic tissues these vascular changes also are linked to an increased ratio of NADH/NAD+, in this case because of an increased rate of reduction of NAD+ to NADH. Several lines of evidence support the likelihood that the increased cytosolic ratio of free NADH/NAD+ caused by hyperglycemia, referred to as pseudohypoxia because tissue partial pressure oxygen is normal, is a characteristic feature of poorly controlled diabetes that mimics the effects of true hypoxia on vascular and neural function and plays an important role in the pathogenesis of diabetic complications. These effects of hypoxia and hyperglycemia-induced pseudohypoxia on vascular and neural function are mediated by a branching cascade of imbalances in lipid metabolism, increased production of superoxide anion, and possibly increased nitric oxide formation.

Sorbinil prevents diabetes-induced increases in vascular permeability but does not alter collagen cross-linking.

In recent studies we have demonstrated a marked increase in albumin permeation of new vessels formed by angiogenesis (in subcutaneous tissue) in the diabetic milieu. Likewise, lysyl oxidase-mediated collagen cross-linking is markedly increased in the scar tissue associated with angiogenesis. The present studies were undertaken to determine whether sorbinil, a chemical inhibitor of aldose reductase that has been shown to prevent and reverse diabetic cataracts and neuropathy, also could prevent the vascular permeability and collagen cross-linking changes in this model. Vascular permeation by 125I-BSA, collagen cross-linking, and tissue levels of sorbitol, myo-inositol, and scyllo-inositol were assessed in male Sprague-Dawley rats 3 wk after injection of streptozocin and induction of angiogenesis and collagen synthesis in polyester fabric implanted subcutaneously. Sorbinil (approximately 25 mg/kg/day) added to the diet of diabetic rats reduced the diabetes-induced increases in albumin permeation by 80%, completely prevented diabetes-induced changes in tissue levels of sorbitol and myo-inositol, and markedly reduced diabetes-induced changes in tissue levels of scyllo-inositol. In contrast, sorbinil had no effect on plasma glucose levels or collagen solubility (an index of collagen cross-linking). These observations indicate that increased vascular permeability associated with diabetes is linked to imbalances in sorbitol/inositol metabolism. These findings also indicate that diabetes-induced increases in vascular permeability and in collagen cross-linking are independent phenomena and diabetes-induced increases in vascular permeability are largely preventable by treatment with an aldose reductase inhibitor in the face of high plasma glucose levels.

Increased vascular permeability in spontaneously diabetic BB/W rats and in rats with mild versus severe streptozocin-induced diabetes. Prevention by aldose reductase inhibitors and castration.

125I-labeled albumin permeation (IAP) has been assessed in various tissues in spontaneously diabetic insulin-dependent female BB/W rats and in male Sprague-Dawley rats with severe or mild forms of streptozocin-induced diabetes (SS-D and MS-D, respectively). In BB/W diabetic rats and in rats with SS-D, indices of IAP were significantly increased in tissues and vessels predisposed to diabetic vascular disease in humans, including the eyes (anterior uvea, posterior uvea, and retina), sciatic nerve, aorta, kidney, and new vessels formed after induction of diabetes. No evidence of increased IAP was observed in heart, brain, testes, or skeletal muscle in BB/W or SS-D rats. In MS-D rats, indices of IAP were increased only in the kidney and in new vessels formed after the onset of diabetes. Marked tissue differences were observed in the effects of two structurally different aldose reductase inhibitors (sorbinil and tolrestat) and of castration on diabetes-induced increases in IAP and in tissue levels of polyols in SS-D rats. Both aldose reductase inhibitors and castration completely prevented diabetes-induced increases in IAP in new vessels and in sciatic nerve in BB/W and SS-D rats. Both aldose reductase inhibitors also markedly decreased IAP in the anterior uvea (approximately 85%), posterior uvea (approximately 65-75%), retina (approximately 65-70%), and kidney (approximately 70-100%); castration reduced IAP in the anterior uvea (approximately 55%), kidney (approximately 50%), and retina (approximately 30%) but had no effect on the posterior uvea. The diabetes-induced increases in IAP in the aorta were reduced only slightly (approximately 20%) by aldose reductase inhibitors and castration.(ABSTRACT TRUNCATED AT 250 WORDS)