An implicit hypothesis guiding the continued development of novel insulin delivery strategies is that precise minute-to-minute control of plasma glucose will be necessary for the amelioration of secondary consequences of insulin-dependent diabetes mellitus (IDDM).⁵ The Diabetes Control and Complications Trials (DCCT) incontrovertibly established the importance of improved glycemic control in the reduction of renal, retinal and neural pathologies of IDDM (DCCT 1993, Santiago 1993). Yet, intensive insulin therapy (IIT) was not without consequence, because an increase in body weight was characterized by a disproportionate gain in fat mass with no improvement in lean body mass (Carlson and Campbell 1993, DCCT 1988). These metabolic and nutritional consequences of IIT are favored because of a reduction in glycosuria and an increase in carbohydrate utilization with a concomitant fall in proteolysis and lipolysis. With an enhanced energetic efficiency, fat accretion ensues at the expense of a gain in lean body mass.

As an experimental paradigm of improved insulin delivery in IDDM, the islet-transplanted mouse, rat, dog and subhuman primate provided further data in support of the glycemic control hypothesis. In animal models of diabetes, reductions in renal, retinal, neural and cardiovascular complications have been achieved after gaining improved glycemic control with pancreatic islet transplantation (Krupin et al. 1979, Lopaschuk et al 1993, Mauer et al. 1974, Schmidt et al. 1983). In addition, we recently demonstrated that male rats transplanted with 3000 syngeneic islets of Langerhans will regain body weight, demonstrating a proportional gain in both carcass protein and fat mass (Tobin and Marchello 1995). Thus, certain metabolic and nutritional complications ascribed to IIT per se are not associated with a more physiological delivery of insulin achieved through pancreatic islet transplantation in experimental IDDM.

Although studies in male animals indicate improved body composition with islet transplantation, similar studies have not been performed in female animals, and energetic efficiency has not been determined in any model. This knowledge gap is clinically relevant for two reasons. First, withholding insulin for the purpose of avoiding weight gain in young adult and adolescent diabetic women has been estimated to exist for 11-15% of patient populations and is associated with poor glycemic control and an increase in secondary diabetic pathologies (Biggs et al. 1994, Polonsky et al. 1994). Thus, the belief that IIT promulgates body fat gain may serve as a clinical barrier to effective IDDM therapy. Second, retrospective analyses of several animal studies (Finegood et al. 1992, Ryan et al. 1993, Tobin et al. 1993) and subsequent gender-specific investigations (Bell et al. 1994) illustrate physiologically different responses to pancreatic islet transplantation in female animals. The improved glycemic normalization demonstrated in female rats suggests that gender-specific hormonal advantages may promulgate enhanced graft function (Brelje and Sorenson 1991, Brelje et al. 1994) and subsequently may improve overall post-transplant metabolic sequelae. Thus, although IIT may be associated with gender-specific clinical obstacles, islet transplantation may conversely be more efficacious in females than males.

The purpose of the present studies, therefore, was to examine the effects of short-term diabetes and improved glycemic control achieved though pancreatic islet transplantation on the weight gain of female rats. Specifically, we wished to determine if previously diabetic, islet-transplanted female rats would regain body weight as a proportional accretion in both fat and protein mass. To elucidate the mechanisms of weight gain, we additionally investigated food intake and energetic efficiency to determine if hyperphagia is normalized concomitant with an improved energetic efficiency after pancreatic islet transplantation. We hypothesized that improved glucose control achieved through islet transplantation in female rats would be associated with a proportional gain in body fat and protein mass, a reduction in hyperphagia and a concomitant normalization of feed efficiency.

MATERIALS AND METHODS

At the time of analysis, three aliquots of finely ground and homogenized samples were used for proximate chemical determination (AOAC 1990, Hartsook and Hershberger 1963). The energy value of the feed and carcass were determined by adiabatic bomb calorimetry using benzoic acid standards. Dry matter was evaluated following lyophilization and oven drying at 105^oC. Protein was determined by the micro-Kjeldahl method, and total fat by the Foss-let procedure. Mineral content was determined by ashing the samples in a muffle furnace (470^oC) for 18 h.

During the energy efficiency trials, rat body weights were calculated by data smoothing using the Optimal Segments routine (Finegood and Bergman 1983, Finegood et al. 1988) as previously described for energy efficiency trials (Tobin et al. 1993) to minimize the effect of animal weighing error. A complete growth curve analysis was performed on individual animals to yield a smoothed data set and to determine the amount of weight gained during the energy efficiency trial period. Energy efficiency was calculated as the percentage of kilojoules retained in the carcass per gross energy intake [(kJ gained/kJ ingested) × 100] during the feed intake trial.

RESULTS

At the end of the acclimation period (wk 0), there were no significant differences in body weights among experimental groups (P > 0.05, Fig 2). However, 1 wk after streptozotocin administration, all diabetic rats (including the pre-Transplant group) had an average weight loss of 6.5 g (P < 0.05), yet there were no differences among hyperglycemic groups (P > 0.05, Diab-3 vs. Diab-9 vs. pre-Transplant). During the final 5 wk of the study, the control and Transplant rats demonstrated similar growth; the Diab-9 group gained only 2.9 g overall, whereas the control rats gained 19.6 g by 9 wk.

Glucose tolerance tests were performed at 3 or 8 weeks of study and analyzed by one-way ANOVA (Table 1). Following five additional weeks of diabetes, the plasma glucose concentration was 31% greater (P = 0.069) in Diab-9 vs. Diab-3 rats; Sham-9 and transplant groups did not differ (P > 0.05, Table 1). Basal plasma insulin did not differ among groups (P > 0.05). The AIR_g was calculated as the incremental increase in plasma insulin above base line during the first hour after oral glucose injection. We observed peak insulin responses in virtually all of the rats during the first hour after glucose injection, with the exception of two rats. The AIR_g in Transplant rats was 100 times what their value would have been (Diab-9) if they had not acquired an islet transplant (P < 0.05). The AIR_g of the Sham-9 group was a similar magnitude greater than that of Diab-9 rats and was not significantly different than that of Transplant rats (P > 0.05). The K_g of islet transplant rats was not significantly different than that of Sham-9 or Diab-9 rats (P > 0.05).

Table 1. Effect of diabetes and islet transplantation of indices of metabolic control in rats studied for 3 or 9 wk1,2 [View Table]

One week following OGTT, rats were killed and body composition was subsequently determined by proximate analysis in eviscerated animal carcasses (Table 2). Sham-3 rats weighed 134% of Diab-3 rats (P < 0.05); however, the final body weights of Sham-9 and Transplant rats did not differ. Carcass moisture was greater in Diab-3 and Diab-9 rats than in Sham-3, Transplant and Sham-9 rats (P < 0.05). The proportion of carcass protein demonstrated a similar trend, in that the diabetic rats had a larger fractional protein content than the Sham-3, Sham-9 and Transplant groups (P < 0.05). In all diabetic rats, the proportion of carcass fat was less than that of Sham or Transplant groups (P < 0.05). The proportion of body mineral content (ash) did not differ among groups (P > 0.05).

Table 2. Effect of diabetes and islet transplantation on body weight and percentage body composition in rats studied for 3 or 9 weeks1,2 [View Table]

Further analysis of animal body composition on a total mass basis (Fig. 3) demonstrated greater carcass protein and fat content for Transplant rats than for either diabetic group (P < 0.05). The carcass protein in the Sham-9 group averaged 33.3 g and did not differ from Sham-3 or Transplant groups (P > 0.05). At 6 wk post-transplantation, previously diabetic rats that received 3519 ± 838 islets of Langerhans demonstrated a carcass protein content that was 22% greater than Diab-3 rats (P < 0.05). After a 6-wk period, carcass fat in Sham-9 rats was 81% greater in mass than that of Diab-9 rats; Transplant rats were also 81% greater in fat mass than Diab-9 rats and did not differ significantly from the Sham-9 group (P > 0.05).

Energy efficiency (Table 3), was determined over a 5- to 7-d period the week before OGTT. As expected, diabetic rats demonstrated pronounced hyperphagia, with energy intake in all diabetic rats greater than Sham or Transplant rats (P < 0.05). Five weeks after islet transplants, energy intake was 54% less than Diab-3 rats, but was not different from Sham-9 rats (P > 0.05), demonstrating a complete amelioration of diabetic hyperphagia. Carcass energy did not differ between the diabetic groups, despite six additional weeks of diabetes in the Diab-9 rats, nor was it different among the Sham-3, Sham-9 and Transplant groups (P > 0.05). However, 6 wk after diabetic rats had received islet transplants, carcass energy of Transplant rats was 134% of Diab-3 and Diab-9 rats (P < 0.05) and was not different from Sham-9 rats. Sham-3 rats demonstrated the highest efficiency of energy storage during growth (Table 3). At 6 wk post-transplantation, energy efficiency of Transplant rats was 408% of Diab-9 rats (P < 0.05). However, during this same time that diabetic rats continued to exhibit an inefficiency in energy utilization, Transplant rats were not different than Sham-3 or Sham-9 rats (P > 0.05).

Table 3. Effect of diabetes and islet transplantation on indices of food intake, carcass energy, and energy efficiency in rats studied for 3 or 9 wk1,2 [View Table]

DISCUSSION

In the present study, we demonstrated that islet transplantation was associated with a proportionate gain in both body protein and body fat. These proportionate growth results are similar to previously published studies using male animals (Tobin and Marchello 1995). Some differences are noteworthy, however. In those studies, previously diabetic islet-transplanted male rats had a carcass protein content 15% greater than that observed at 2 wk of diabetes and 61% greater than that of age-matched diabetic rats at 7 wk. In the present investigation, islet transplantation improved protein content 22% above that observed at 3 wk of diabetes yet resulted in a final protein mass only 23% greater than that of age-matched diabetic animals. Although it may appear that the female rats are at a functional disadvantage compared with the male rats, a closer examination of the data reveals that this is not due to a difference in protein accretion, but occurs because female rats fail to engender a further loss of body protein during successive weeks of diabetes. In addition, our previous investigations illustrated no differences in the final percentage of body fat of islet-transplanted rats vs. age-matched controls, a result identical to the present studies. Thus, although gender may influence glycemic normalization (Bell et al. 1994), the present studies do not provide data supporting the hypothesis that female animals are at a distinct advantage over males when considering the post-transplant gain in body protein or the gain in percentage of body fat. The present investigations, however, were conducted using an islet mass approximately equivalent to 60% of the normal pancreatic islet content. Thus, the gender-specific effects that have been previously illustrated (Bell et al. 1994) at more dramatically reduced islet masses (10% of controls) may be ameliorated (or obscured) by transplanting a more substantial islet mass.

Experimental diabetes is associated with pronounced hyperphagia, yet, insulinopenia precludes weight gain in the face of a dramatically increased energy intake. In the present studies, the energy consumption of diabetic rats was roughly twice that of age-matched sham control rats. Islet transplantation, however, was associated with a reduction in energy intake to a level not significantly different than that of controls. What is particularly noteworthy in the present investigations is that this amelioration of diabetic hyperphagia is associated with an energetic compensation that results in an energy efficiency that is improved to, but not in excess of control levels 5 wk after islet transplantation. Thus, it is apparent that in this paradigm, concomitant alterations in both energy intake and energy efficiency are responsible for the return to normal body weight following pancreatic islet transplantation. These are the first studies to demonstrate this phenomenon.

The occurrence of weight gain and disproportionate body fat accumulation with IIT has been the subject of several investigations (Carlson and Campbell 1993, DCCT 1988, Leiter 1995), and the mechanisms involved in these effects have been previously described. A recent study illustrates the metabolic basis of some of these detrimental effects of IIT on IDDM patients. These studies by Carlson and Campbell (1993) have delineated that the disproportionate fat accretion associated with this therapy can be ascribed to a shift in energy balance. Seventy percent of this positive energy balance is explained by a reduction in glycosuria, and the remainder is attributed to a reduction in the resting metabolic rate. In addition, IIT was associated with a decreased protein and fat oxidation and an increase in carbohydrate oxidation. Because the net energy balance favored energy gain, when coupled with diminished fat oxidation, the decreased resting metabolic rate promoted a disproportionate accumulation of body fat. A more recent report by Leiter et al. (1995) illustrates that the weight gain associated with IIT continues to increase up to 9 y after the onset of therapy. A novel observation in these studies is that waist circumference was higher in the IIT group. Because abdominal obesity is more strongly associated with the risk for cardiovascular disease (Kannel and McGee 1979, Kissebah 1982, Ward 1994), such an observation is particularly noteworthy and is not without concern. Thus, these recent observations suggest that although IIT may have a beneficial clinical effect on selected secondary complications of diabetes including cardiovascular abnormalities, increased body fatness secondary to altered nutritional energetics is a repeatable yet unwanted sequela of IIT.

In the present paradigm, improved insulin secretion was associated with a normalization of growth and a proportionate gain in body fat mass. Although the present studies did not include direct comparisons to intensive insulin therapy, subsequent investigations have demonstrated a 19% greater body fat in diabetic rats treated with multiple daily subcutaneous insulin injections, a magnitude similar to DCCT results (Tobin and Marchello 1996). These studies illustrated that islet transplantation was associated with 17% less body fat than that seen in IIT-treated rats, and transplant recipients were not significantly different than controls. Thus, data are accumulating that indicate islet transplantation is better than IIT at preventing the disproportionate body fat gain associated with strict glycemic control achieved through exogenous insulin administration.

The present study also illustrated a gain in body protein in islet-transplanted rats, to a level not different than control rats. A single mechanism of action that describes the effect of insulin on proteolysis or proteogenesis remains to be clearly elucidated. Decreased lean body mass in diabetes may be due to a decreased number and translational efficiency of ribosomes (Jefferson et al. 1977, Morgan et al. 1971), as well as alterations in peptide chain elongation/termination (Peavy et al. 1978). Several studies additionally suggest that these effects may be modulated in part by modifications in insulin like growth factor-1 (IGF-1). Streptozotocin diabetic rats that are insulin deficient lack IGF-1, and growth retardation in IDDM infants has been ascribed to a lack of proper insulinization (Froesch et al. 1990). Recent studies further suggest that protein nutrition, insulin and growth may be modulated via IGF-1 (Lemozy et al. 1994, Straus 1994). In an animal model of noninsulin-dependent diabetes mellitus (NIDDM), Tse et al. (1995) demonstrated that dietary protein restriction attenuates islet insulin secretion, without altering hyperglycemia or hyperinsulinemia, yet attenuates weight gain in obese Zucker rats. Although NIDDM is not usually associated with a primary insulin secretory deficiency, both IDDM and NIDDM engender similar cellular insulin deficiencies. The influence of islet transplantation in IDDM on IGF-1 is not known. However, if insulin deficiency at the cellular level results in similar modulations of IGF-1 in IDDM or NIDDM, such a relationship may warrant further investigation.

In summary, we have demonstrated that islet transplantation in female diabetic rats is associated with proportionate gains in body fat and lean body mass. A portion of this nutritional effect is due to a normalization in energy efficiency that occurs concomitant with a reduction in energy intake. These data indicate that islet transplantation is not associated with a disproportionate gain in body fat observed in clinical IIT trials and may be particularly noteworthy for female islet transplant recipients; reticence towards insulin therapy manifests itself to a greater extent in females than males. Thus, the present studies offer an optimistic outlook for the continued development of more physiological insulin delivery strategies because islet transplantation results in proportionate gains in protein and fat mass. Such results may prove to enhance the clinical attractiveness of islet transplantation for patient populations who wish to avoid the excessive weight gain associated with intensive insulin administration via subcutaneous injection.

ACKNOWLEDGMENT