Novel observations emerged from this study that are relevant to weight regain in humans, but are not likely to emerge from clinical studies. While no intervention prevented weight regain following weight loss, RS and exercise, alone and in combination, attenuated the rate of weight regain and lowered the overall defended body weight. RS and exercise reduced the energy gap between the biological drive to eat and total energy requirements by different mechanisms and at different stages of relapse. The effect of exercise on food intake was at the initial transition to relapse, whereas the effects of RS on food intake occurred during the subsequent weeks. A more subtle effect of their combination on food intake persisted for four weeks after the initial transition from the weight maintenance phase to relapse. Surprisingly, the weight gain prevented by exercise was primarily lean mass, as exercise increased the number of adipocytes in subcutaneous adipose tissue and diverted excess nutrients to this depot. Conversely, RS consumption attenuated regain of fat mass and promoted lean mass accrual when combined with exercise. The distinct mechanisms of these two interventions were clearly complementary in this metabolic context. These observations suggest that RS and exercise, alone or in combination, may facilitate weight maintenance after weight loss by reducing the energy gap between energy intake and expenditure. Their combination may be particularly beneficial during transient excursions from the weight maintenance diet that involve high fat overfeeding.
The magnitude of overfeeding on day 1 of relapse provides insight about the biological pressures to cease or suspend a calorie-restricted, weight maintenance diet. This energy imbalance was roughly twice as large relapsing on a HC/HF diet in comparison to our previous relapse studies with a LF diet [10
]. Because of the well-known properties of HC/HF diets [34
], it was not surprising that the level of overfeeding was greater in the present study than those employing a LF diet. However, the HC/HF diet also greatly reduced the beneficial effects exercise on weight regain. On a LF diet, treadmill exercise reduced the energy gap by ~80% on day 1 of relapse primarily by reducing the magnitude of overfeeding, and this affect on food intake persisted for 8 weeks [12
]. Moreover, exercise attenuated fat mass accrual, particularly in abdominal depots, during relapse on a LF diet. The observations in the present study suggest that reverting to unrestricted HC/HF feeding, even for brief periods, may substantially minimize the magnitude and persistence of the beneficial effects of exercise on weight regain and body composition.
In contrast to exercise, the effects of RS on food intake were delayed in initiation but more sustained into the relapse period. This delay may reflect the time for the consumption of RS to affect gut microbiota and the production of short chain fatty acids. The diminution of the RS effect on energy intake occurred as the lost weight was regained, which may reflect a progressive impairment in RS fermentation as the animals relapsed to obesity. In ad libitum
fed animals, the beneficial effects of RS are dependent upon RS fermentation in the lower gut, and obesity is accompanied by a lack of RS fermentation and no effect of RS on body weight or adiposity [29
]. The mechanisms of this obesity-associated lack of RS fermentation are not well understood, but the expectation is that this impairment would return as excess body weight returned. If these observations translate to the human condition, the implication is that RS may be effective for facilitating maintenance of a reduced weight rather than promoting weight loss, per se
RS consumption has been shown to reduce post-prandial excursions in blood glucose and insulin [22
], and these effects translate to the overfeeding event experienced on the first day of relapse. This effect of RS has been attributed to the combination of delayed digestion and fermentation to short chain fatty acids which may improve whole body insulin sensitivity. Insulin and glucose are known to induce the lipogenic pathways [40
], and the effects of RS consumption on insulin and glucose are accompanied by reduced expression of adipose tissue glucose transporters (GLUT4) and fatty acid synthase (FAS) and suppression of de novo
]. These effects of RS may have contributed to its impact on body composition during relapse in this metabolic context. However, RS, when combined with exercise, prevented the increase CORT excretion that occurred with weight regain, which may also have contributed the preferential deposition of lean mass in this group.
In previous studies with this model, a population of small, presumably new, adipocytes (<20 μm) appears in abdominal adipose depots early in relapse, increasing the total number of adipocytes [10
]. The increase in adipocyte number persists to the end of relapse [11
]. Preferential trafficking of dietary fat and greater retention of lipid from de novo
lipogenesis in these small adipocytes would not only facilitate clearance of ingested nutrients, but would also contribute to further expansion of the fat depot causing weight gain which surpasses the original, obese weight. Exercise prevents the increase in adipocyte number that results from HF feeding [42
] and weight regain on a LFD [12
]. Observations in the present study are distinct in that exercise had no effect on cell number in abdominal depots, but instead increased cell number in subcutaneous adipose tissue. Generation of new adipocytes in subcutaneous adipose and the high insulin response in this group may explain the preferential deposition of ingested nutrients in this depot.
It should be noted that, due to its resistance to digestion and passage through the large bowel, the caloric availability of RS is lower than that for digestible starches. Most digestible carbohydrates provide 4kCal/g whereas RS provides around 3 kCal of available energy per g. This difference between the RS and C diets may also be a contributing factor to the lower rate of weight regain and the lower overall defended body weight in the RS rats relative to C rats.
The magnitude of weight regain in this study is greater than that usually observed in clinical studies. Our studies in rodent models reflect the biological pressures to gain weight after weight loss. Unlike rats, humans are subjected to additional psychosocial pressures which motivate them, for reasons unrelated to energy homeostasis, to keep excess weight off. For weight reduced humans, this clearly pits cognitive motivations against the biological drive to regain, with outcomes that can vary greatly in the degree of success. The high recidivism rates observed in obesity therapeutics should provide some warning as to how powerful these biological pressures are. Moreover, clinical studies rarely, if ever, monitor the relapse period in subjects that return to obesogenic dietary habits or follow subjects for more than 5 years [43
], and they are limited by a substantial number of drop-outs [44
]. There is reason suspect that those subjects with the greatest weight gain during follow-up are those who are most reluctant to continue their participation in clinical studies. Regardless, the trajectory of percent weight lost that is regained over 9 weeks in the present study and in our previous studies of weight regain [8
] reflect those that occur in humans over 3 to 5 years [1
], and, similar to these models, as many as 50% of individuals participating in clinical studies end up surpassing their original weight [45
]. Ultimately, the RS diet and exercise still lowered the rate of weight regain and reduced the overall weight regain in this rat model of obesity.
While the EX/SED arm of the present study was not in the initial study design, we have included this data because of the potential implications to the human condition. The inability to sustain a program of regular exercise is a substantial problem in clinical weight loss studies [17
] and is a common occurrence in the real world setting. In our LF relapse studies with this model, compliance scores declined sharply at the transition from weight maintenance to relapse [12
]. However, they did not progressively decline with weight regain nor were we required to remove any animals from the exercise program. Studies in similar formerly-obese models show that volitional activity also declines after weight loss [15
]. These studies suggest that there may be a biological contribution to this noncompliance issue and that it may be exacerbated during relapse on a HC/HF diet.
Because the EX/SED arm was selected based upon compliance scores, rather than randomized, it is unclear if these animals had a pre-existing metabolic disposition, experienced an injury during relapse, or experienced particularly detrimental metabolic effects during the early stages of weight regain. While the EX/SED rats appeared normal at the end of weight maintenance, their metabolic health declined severely by the end of the study. While some beneficial effect of RS was apparent in the protein disappearance and energy balance measures by day 35, it is unclear if the decline in metabolic health occurred early in relapse or after the cessation of exercise. The exercise regimen employed in this study undoubtedly imposed a chronic metabolic stress, and obesity, HF feeding, and overfeeding have been shown to impair the response to chronic stress [46
]. Exercise cessation may have imparted an additional metabolic insult, as it has known negative consequences on energy balance, weight gain, and metabolic health [49
]. The impaired ability to respond to these compounding metabolic challenges is likely to have contributed to the detrimental effects of exercise and exercise cessation on body composition. Given the frequent problem with sustained compliance in humans, studies clarifying the impact of exercise cessation after weight loss and during weight regain on a HC/HF diet, beyond the preliminary data reported here, are certainly warranted.