Do We Need a New Unifying Theory for Obesity?

Despite unprecedented public awareness of the adverse health consequences associated with obesity, its global prevalence continues to rise; 14% of the world population was obese in 2020, and this fraction is estimated to grow to 24% by 2035 (1). Approximately 42% of the US adult population have been classified as obese (body mass index > 30 kilograms/square meter of body surface area) (2). Why is obesity increasing worldwide?

It is widely believed that obesity results from a caloric imbalance when calories ingested exceeds energy expended—the so-called energy balance model (EBM). Despite the demonstration that glucagon-like peptide-1 receptor agonists such as semaglutide and liraglutide decrease weight by reducing appetite, NHANES data convincingly demonstrated that caloric intake has not increased from 1999 to 2017 (2), nor has physical activity decreased (3), despite the incidence of obesity increasing from ∼30% to ∼42%. The EBM holds that the nature of the calories is less important than the number of calories consumed. We do not understand yet why central control of appetite and satiety is dysregulated in people with obesity or why weight lost returns when calories are no longer restricted or when patients cease taking glucagon-like peptide-1 mimics.

The carbohydrate–insulin model (CIM) focuses on the nature of calories consumed and their effects on insulin (4). Rapidly metabolizable carbohydrates prevalent in ultraprocessed foods disproportionately trigger insulin response, increasing the storage of excess energy into adipose tissues. The CIM focuses on the effects of diet on the endocrine system and fuel partitioning in peripheral tissues such muscle and adipose. In the CIM, the nature of calories consumed and how this affects insulin signaling is emphasized over the number of calories consumed. Notably, the obesity pandemic exploded in the late 1970s when dietary guidelines shifted to emphasize reduced fat consumption, resulting in their replacement with sugars in processed foods.

A recent perspective discussed the strengths and weaknesses of the EBM and CIM as models for increased obesity and for homeostatic regulation of energy balance, pointing out that neither was sufficient to explain the rise in obesity or to fully model homeostatic control of weight (5). The energy reduction–oxidation (REDOX) model recognizes that the biochemical transformation of fuel molecules such as glucose and fats results in the production of small amounts of reactive oxygen species (ROS) that can serve as intracellular signals of fuel availability (6). The presence and removal of these fuel-derived ROS were proposed to underlie a signaling network responsive to the availability of intracellular fuel by modulating insulin release from beta cells, triglyceride storage in adipocytes, hepatic gluconeogenesis, and satiety via the central nervous system (6). The REDOX model can intersect with the CIM via its effects on insulin secretion and the EBM by modulating food consumption. In principle, mitigating the excess ROS production resulting from overconsumption of fuel could be counterbalanced by increasing ROS removal.

The obesogen model (OBS) was based on the observation that exposure to certain chemicals could lead to weight gain in exposed individuals at any life stage (reviewed in 7). These obesogens can act by disrupting endocrine signaling, acting as ligands for nuclear hormone receptors, disrupting cellular signaling pathways controlling adipocyte commitment, differentiation, function, size, and number, together with insulin levels, food intake, and metabolic rate (7). At least 50 obesogens were associated with increased fat accumulation in vivo, but the strongest support for the OBS model comes from pharmaceuticals with weight gain side effects (7). Prenatal obesogen exposure can lead to lifelong, irreversible derangement of metabolic function such that the exposed individuals and their descendants are predisposed to become obese when dietary fat is increased (7).

So, which model can correctly guide efforts to prevent the development of obesity? The value of any obesity model lies in its ability to predict physiology, response to perturbations, the concordance between animal experiments and human observations, and the robustness of the mechanisms proposed to explain biological observations. When considering the predictive value of the various obesity models, I am reminded of the parable of the blind men and the elephant, where each tried to determine what type of animal an elephant was by feeling only 1 part of its body. Individual observations failed to correctly predict the elephant's nature because they lacked the context of the entire animal, despite that each prediction matched the observation. Each obesity model has merit and is supported by extensive data; however, none adequately explains, or provides, a viable strategy for controlling the obesity pandemic.

Heindel and colleagues attempted to develop a unified theory of obesity (8). Recognizing the scientific support for each of the obesity models, they proposed a framework integrating these into a plausible whole. They posited that chemical obesogens (including added sugars) in a diet rich in packaged, ultraprocessed foods influence food consumption and insulin signaling by direct action on signaling pathways (OBS) or via interactions with the REDOX system. This modulates energy intake via the central nervous system (EBM) and energy partitioning in the peripheral tissues via insulin (CIM). This OBS/REDOX model offers a mechanistic underpinning for the rise in obesity and redox signaling via fuel metabolites could modulate energy homeostasis, and its ability to be perturbed by the fuels in ultraprocessed foods and the obesogens they contain. Generation of elevated ROS levels by excess fuel and/or via the effects of obesogens on REDOX and critical pathways in energy balance, insulin signaling and metabolic function offers a framework in which disturbances at various points will elicit predictable results on energy balance and weight gain. OBS/REDOX also offers a possible explanation for developmental programming of obesity susceptibility. OBS/REDOX is testable by modulating the amounts of dietary simple carbohydrates and obesogens and/or by increasing ROS removal and observing effects on metabolism.

Ultimately, there may not be a single cause for the rise in obesity and, like the disparate parts of the elephant, each model has something to contribute. The integrated OBS/REDOX model is worth further exploration, considering the high societal cost of obesity and the even higher cost of reducing it with the current generation of effective, but costly, pharmaceutical interventions.

Abbreviations

CIM

carbohydrate–insulin model

EBM

energy balance model

OBS

obesogen model

REDOX

energy reduction–oxidation

ROS

reactive oxygen species

Funding

Work in the Blumberg laboratory is supported by grants from the US National Institutes of Health, R01ES023316 and R01ES031139.