Maternal obesity and programming of metabolic syndrome in the offspring: searching for mechanisms in the adipocyte progenitor pool

The prevalence of obesity in Western nations was 13% in the late twentieth century [1], climbing to 35% over the last two decades. Today, obesity is one of the most significant public health problems, affecting 650 million individuals globally [2], a crisis driven by multifaceted environmental, behavioral, and socioeconomic factors. Socioeconomic disparities in obesity burden exist within high income countries, while globalization of the food and beverage industry has led to a double burden of obesity and malnutrition in the global south. Building more resilient populations and mitigating the burgeoning rise in health care costs are incumbent upon addressing the obesity crisis.

The age-of-onset of obesity has declined in parallel with its increasing prevalence. Over the past several decades, the rise in obesity has occurred more rapidly in children and adolescents than in adults. A study published in the Lancet showed that between 1975 and 2016, obesity among children and adolescents increased globally, with an increase of 30–50% per decade in high income countries and an increase of 400% in parts of Africa [3]. More recently, obesity rates among youth appear to have plateaued [4, 5], albeit at high levels (18–20%) in some high-income regions including North America [6, 7]. Currently, it is unclear how future trends of childhood obesity will be impacted by the COVID-19 pandemic, which resulted in a doubling of the rates of BMI increase in children, with the most drastic changes observed in 5–11-year-olds [8, 9]. Obesity at younger ages increases the duration over which the obesity burden is carried throughout the life course. A meta-analysis published by Simmonds et al. reported that 55% of obese children become obese as adolescents, while 80% of adolescents that are obese remain obese in adulthood [10]. Data collected in the USA show that obesity at younger ages increases years of lost life across the lifespan [11]. Earlier onset of obesity-associated morbidity due to higher rates of childhood obesity will come with high economic costs for future generations with impacts to productivity in the working age population and greater burden on health care systems.

Consequent to the declining age-of-onset, a greater number of women are overweight or obese at reproductive age. In 2019, 29% of US women were obese immediately prior to pregnancy, an increase of 11% since 2016 [12]. Maternal obesity has become one of the most common complications of pregnancy and is a major risk factor for gestational diabetes (GDM), both of which increase the risk for adverse perinatal outcomes and later-life chronic disease [13‐16]. Mothers that are obese or diabetic during pregnancy are more likely to have babies that develop obesity, often before adulthood [17‐19]. Therefore, this transmission of obesity across generations may be amplifying its prevalence and contributing to its declining age of onset.

Obesity is a central pathophysiological mechanism in development of the metabolic syndrome (MetS), a clustering of risk factors for cardiovascular disease (CVD) including dyslipidemia, hypertension, and insulin resistance [20, 21]. According to NHANES data (2011–2016), 20% of 20–34-year-olds and 49% of individuals over 60 in the USA have MetS [22]. The presence of MetS doubles the risk for CVD and increases the risk for T2DM by 6-fold [23]. Results from the Global Burden of Disease study showed CVD to account for 41% of deaths and 34% of disability-adjusted life years related to high BMI [24]. Among adults with T2DM, more than 80% of individuals were overweight or obese [25], suggesting that the prevalence of T2DM is largely driven by obesity. Therefore, obesity and MetS are major drivers of CVD and T2DM, two main causes of mortality and morbidity in populations across the globe.

The childhood obesity crisis has resulted in an earlier onset of MetS. Among children and adolescents, the prevalence of MetS is reported to fall between 2.5 and 3.5% in Canada and between 4.2 and 9.2% in the USA [26]. At least one component of MetS is present in one third of youth in North America [26], with abdominal obesity being the most common. According to NHANES data (2015-2018), obesity is a common co-morbidity in children and adolescents with hypertension, which occurred in 4.6% of children between the ages of 8–12 and in 3.7% of adolescents between the ages of 13–17 [27]. This trend of obesity-associated cardiometabolic disorders at younger ages is predicted to reverse the steady rise in life expectancy that has occurred over the past two centuries, as the burden of morbidity will span further across the life course in future generations [28]. Given the rise in obesity at reproductive age and associated metabolic complications of pregnancy, it is plausible that transmission of obesity across generations due to fetal overnutrition has played a key role in the emergence of pediatric MetS. Several studies have shown that development of obesity and MetS before adulthood is more common in babies born to mothers with metabolic disease during pregnancy [14‐16]. Dabelea et al. showed that over a 10-year period, the increase in prevalence of T2DM in Pima children was almost entirely accounted for by exposure to maternal diabetes during pregnancy [29]. Therefore, preventative measures targeted to metabolic health during pregnancy are critical in efforts to mitigate the global epidemic of MetS and its declining age of onset.

Obesity is characterized by the excessive accumulation of white adipose tissue (WAT), which is distributed throughout the body in anatomically distinct depots that fulfill related but unique functional roles [30]. Regardless of anatomic location, the principal role of WAT is to serve as a “metabolic sink” that stores energy in the form of triglycerides packaged in lipid droplets of adipocytes, the functional units of WAT. Additionally, WAT synthesizes and secretes adipokines that play a key role in the regulation of systemic metabolism through paracrine and endocrine actions. The first adipokine to be discovered was leptin, which is produced by adipocytes and functions in the regulation of hunger and satiety through crosstalk with neuronal targets [31]. Acting on its cognate receptors in the hypothalamus, leptin stimulates hunger in the fasted state and dampens it in the fed state. Adiponectin is specifically expressed and secreted from adipocytes, predominately in subcutaneous (SAT) depots, and enters the circulation to act distally on the liver and muscle where it enhances sensitivity to insulin, increases fatty acid oxidation, and decreases influx of fatty acids [32]. Thus, WAT orchestrates responses to nutritional and environmental cues for fine control of systemic lipid metabolism and energy balance

The primary anatomical divisions of WAT are the subcutaneous (SAT) and visceral adipose tissue (VAT) depots. The majority of WAT volume is localized to the SAT [33], located just beneath the skin and is the principal site of energy storage. The SAT is considered a protective adipose depot as its expansion is associated with a favorable metabolic phenotype [30]. The VAT, located primarily in the abdominal cavity surrounding the organs, comprises only 10–20% and 5–8% of total body fat in men and women, respectively [33]. Accumulation of fat in the abdominal region and VAT compartments is a pattern of obesity termed central or android obesity that is associated with higher cardiometabolic risk and is characteristic of males [34]. Waist-to-hip circumference, a proxy measurement of central obesity, is a better predictor of T2DM and CVD than BMI [35, 36]. The protection of females from cardiometabolic disease relative to males during the reproductive years is at least in part attributable to their tendency to accumulate fat in the gluteo-femoral region and SAT depots. This protection is lost after menopause when the distribution of body fat shifts to a male-like pattern with higher expansion of VAT [37‐39]. The favorable influence of SAT on metabolic homeostasis has been experimentally demonstrated by the improvement in insulin sensitivity, plasma lipid profiles, and hepatic lipid content after autologous transplantation of SAT into VAT depots in rodent models of diet-induced obesity [40, 41]. Overall, there is abundant evidence showing that distribution across anatomical depots, rather than total body fat mass, is the important factor driving the relationship between obesity and risk for MetS.

Under obesogenic conditions, WAT expands to accommodate the energy surplus by two mechanisms: hypertrophy of existing adipocytes and hyperplastic expansion in which new adipocytes are generated by recruiting a resident population of adipose-derived stem cells (ASC) for differentiation. Adipogenesis generates new adipocytes through sequential differentiation steps, in which multipotent mesenchymal stem cells commit to the adipocyte lineage and subsequently become preadipocytes, followed by terminal differentiation to form a mature lipid-storing adipocyte. While the transcriptional program that controls lineage specification and differentiation is well-defined, much remains to be learned with respect to environmental cues that govern the behavior of progenitors within their native microenvironment.

It is thought that proliferation and differentiation of ASC are stimulated through paracrine feedback when hypertrophic growth has reached a defined limit of capacity. Pro-adipogenic signals released from engorged adipocytes may be triggered by mechanical strain [42], local hypoxia [43], or hypertrophy-induced cell death [44]. Recently, Li et al. demonstrated that adipocytes from obese or insulin-resistant humans were prone to undergoing premature senescence [45], raising the possibility of a mechanism common to adipocytes that have either reached a critical size or the end of their lifespan. While the signals emanating from adipose tissue to recruit the ASC pool for differentiation remain unresolved, current evidence demonstrates that adipose expansion is associated with a shift in the adipose secretome towards higher release of pro-inflammatory mediators [46‐48]. Interestingly, a recent study showed an increase in differentiation of ASC exposed to the secretome of adipose explants from high fat-fed mice [46]. Adipocyte-specific dominant-negative mutation of the pro-inflammatory cytokine, TNFα, prevented diet-induced WAT expansion, demonstrating that adipocyte inflammation is required for adaptive WAT remodeling [49]. It is also possible that non-adipocyte cells within the adipose microenvironment mediate crosstalk with the ASC pool. Indeed, a number of studies demonstrate that the pro-inflammatory environment of expanding adipose tissue is in part attributable to macrophages that surround and eliminate dead adipocytes [49]. Work by Lee et al. demonstrated that macrophages recruited to dead adipocytes secret a chemotactic signal that activates local progenitors [50]. The above evidence suggests that the ASC pool senses and responds to a local and transient inflammatory environment, contributing to adaptive adipose tissue remodeling that replaces dead or senescent adipocytes. Redox signaling plays a key role in mediating cellular responses to inflammation and recent work from our laboratory reveals a novel role for reactive oxygen species (ROS) in stimulating adipogenesis [51]. Thus, ROS may serve as second messengers in the transduction of signals from the local environment to control behavior of the progenitor pool. New techniques that preserve cell-to-cell and cell-to-matrix interactions within the adipose microenvironment will shed light on signaling pathways that govern the response of ASC to the energy buffering needs of WAT.

Recruitment of the progenitor pool allows WAT depots to meet increased requirements for lipid storage in the face of a prolonged positive energy balance. The adipogenic response to lipid overload is depot specific. Traditional thought held that stimulation of adipogenesis in response to an obesogenic diet occurs primarily in the SAT. This supposition was based on studies measuring adipogenic potential in ASC isolated from different adipose depots, cultured, and exposed to a differentiation cocktail in vitro. This experimental approach consistently demonstrates higher intrinsic adipogenic potential in ASC isolated from SAT compared to those isolated from VAT [52, 53]. Further, a study by Tchoukalova et al. revealed lower adipogenic capacity in ASC isolated from abdominal SAT compared to those isolated from the femoral depot, suggesting variability across subcutaneous depots [54]. However, newer in vivo approaches such as pulse-chase transgenic lineage tracing have called into question the assumption that intrinsic adipogenic potential in vitro reflects adipogenic responses of ASC in their native environment. In mouse models of high fat feeding, proliferation of ASC occurs rapidly in the first 12 h in both VAT and SAT but thereafter is observed in only VAT [55]. By 8 weeks, proliferating ASC undergo differentiation in VAT, as indicated by an increase in BrdU incorporation into mature adipocytes and decrease in proliferating ASC [55]. Using a pulse-chase system of labeling pre-existing adipocytes, Wang et al. demonstrated that after 35 days of high fat feeding, adipocyte hypertrophy was apparent in both SAT and epididymal VAT, while after 56 days of the obesogenic diet new adipocytes were generated in epididymal fat only [56]. Likewise, another model of inducible adipocyte tagging that distinguishes new from pre-existing adipocytes showed that diet-induced adipogenesis occurred predominately in VAT [55]. Interestingly, the same group subsequently published a study revealing that diet-induced adipogenesis exclusively in VAT is a male-specific response. In females, diet-induced proliferation and differentiation of ASC occurred in both SAT and VAT, leading to a similar increase in mass of these two depots [57]. Ovariectomy in female mice recapitulated the VAT-dominant adipogenic response observed in males [57]. The default exclusion of females in pre-clinical studies has obscured our understanding of how sex influences WAT responses in obesity. Sexual dimorphism in behavior of the adipocyte progenitor pool may be a key contributor to sex differences in patterns of body fat distribution and risk for cardiometabolic disease.

Gene expression signatures and differentiation potential of ASC vary across adipose depots, suggesting that cell-autonomous properties contribute to depot-specific responses to lipid storage needs [53]. A role for the microenvironment in regulating the adipocyte progenitor pool has also been demonstrated. While low adipogenic potential is characteristic of ASC isolated from VAT and cultured in a monolayer [53], adipogenic capacity was similar between VAT and SAT-derived ASC differentiated in 3D culture that recapitulates the cell-to-cell and cell-to-matrix interactions of the adipose microenvironment [58]. Jeffrey et al. found that ASC isolated from SAT of male mice differentiate readily when transplanted in VAT of male recipients exposed to an obesogenic stimulus, while the adipogenic response was lower when ASC were isolated from VAT and transplanted into SAT [57]. Therefore, inhibitory or stimulatory signals originating from the microenvironment may be more important in regulating adipogenic responses than cell-intrinsic differentiation potential. Females are less susceptible to high fat diet-induced adipocyte apoptosis and inflammation [59], which occurs predominately in the VAT of obese males [60, 61], possibly explaining sex and depot-specific differences in adipogenic responses.

The “adipose tissue expandability hypothesis” posits that a failure of SAT to meet energy buffering needs of the body triggers the onset of obesity-induced cardiometabolic disease (Fig. 1). An overreliance on hypertrophic growth results in engorged adipocytes that become hypoxic, fibrotic, and inflamed. Dysfunctional adipocytes in obesity-associated adiposopathy or “sick fat” lose their ability to buffer energy and become resistant to the anti-lipolytic effects of insulin, leading to lipid spillover into the circulation, increased lipid storage in visceral depots, and deposition in ectopic organs such as the liver and arteries [62, 63]. Adiposopathy is characterized by an imbalance in lipid handling with increased lipogenesis and lipolysis and decreased fatty acid oxidation, as well as dysregulated adipokine secretion. Retained expansion capacity of SAT is thought to distinguish those with unhealthy obesity from the metabolically healthy obese (MHO), a term used to describe individuals with a BMI that falls in the obese category (BMI ≥ 30 kg/m²) but absent of other components of MetS (e.g., insulin resistance, hypertension) [64‐66]. Overexpression of adiponectin in the ob/ob mouse model of obesity exacerbates total weight gain, primarily as a consequence of increased SAT volume, decreases adipocyte diameter in SAT, and normalizes insulin sensitivity and hepatic lipid content [67]. Insulin sensitivity is maintained in mice that develop obesity as a consequence of collagen VI deficiency that results in unrestricted WAT expansion [68]. This study not only demonstrates the adaptive role of WAT expansion in obesity but also highlights that the extracellular scaffold imposes a barrier to WAT expansion that may play a role in impaired expandability when hypertrophic adipocytes become fibrotic. These transgenic models provide strong evidence supporting a critical role for adipose plasticity in determining metabolic health in obesity. In fact, weight gain is a common side-effect of thiazolidinediones (TZD), a class of anti-diabetic drugs that act as synthetic agonists of PPARg and improve insulin sensitivity in part through stimulating adipogenesis in SAT [69, 70]. Thus, the critical pathophysiological factor in obesity-associated cardiometabolic risk is not the absolute amount of WAT but rather its capacity to carry out its critical role in regulating systemic lipid homeostasis.

The capacity for hyperplastic expansion in WAT is limited by the availability of adipocyte progenitors and their differentiation potential. The theory that adiposopathy in obesity is triggered by a deficiency in the capacity for adipogenesis is supported primarily by studies using cultured ASC isolated from obese humans or rodent models of obesity. A few studies have demonstrated impaired in vitro proliferation and differentiation capacity of ASC isolated from the stromal vascular fraction of obese humans [71, 72]. In a mouse model of diet-induced obesity, reversing the impairment in differentiation potential of ASC from inguinal SAT prevented the development of insulin resistance and hepatic lipid accumulation [73]. An older study showing reduced preadipocytes in the stromal vascular fraction of obese patients is often cited as evidence for impaired differentiation capacity in obesity [74]. However, in this study, preadipocytes in the stromal vascular fraction were identified solely on the basis of staining for the fatty acid binding protein, aP2 [74, 75]. More recent single cell RNAseq studies have lent support to the adipose tissue expandability hypothesis by showing diet-induced obesity in mice to cause shifts toward less differentiated and pro-inflammatory subpopulations of adipose progenitors [76, 77].

Another line of evidence supporting a role for impaired adipogenesis as a mechanism underlying adiposopathy comes from studies in aged humans and rodents. Aging is accompanied by a decrease in WAT volume and a redistribution of fat in favor of VAT depots. Thus, adiposopathy likely contributes to age-related increases in the risk for cardiometabolic disease. Several studies have demonstrated blunted differentiation capacity in ASC isolated from adipose depots of aged humans or rodents [78‐80]. A study using single cell RNAseq in the SAT of mice showed that aging leads to depletion in populations of mesenchymal stem cells, committed ASC and preadipocytes [81]. Cellular senescence in aging is a proposed mechanism for the exhaustion of adipocyte progenitors and their blunted differentiation capacity. Interestingly, Li et al. showed premature senescence in freshly isolated adipocytes from insulin resistant obese humans [45], and other groups have observed markers of senescence in cultures of ASC isolated from obese rodents and humans [79, 82, 83]. Therefore, in both obesity and aging, cellular senescence may be either a cause or consequence of chronic inflammation in WAT, a hallmark of adiposopathy. Findings from our laboratory show that the effect of ROS on adipogenesis is non-linear, such that low levels of ROS potentiate adipogenesis, while high levels impair differentiation [51]. Thus, signals emanating from an adipose microenvironment that has progressed to a state of chronic inflammation and redox imbalance may instigate the transition from adaptive WAT remodeling to adiposopathy. Adipose tissue inflammation in obesity is not only attributable to excessively hypertrophied adipocytes but also to an increase in the presence of macrophages due to proliferation of resident macrophages and infiltration of monocytes, as well as a polarization toward the pro-inflammatory M2 subtype [84, 85]. Exposure of ASC to media enriched with M2 macrophages inhibits their differentiation [86]. Recent studies using single cell RNAseq reveal heterogeneity within the APC pool with subpopulations differing in proliferative and adipogenic potential [77, 87, 88]. Pro-inflammatory, anti-adipogenic ASC have been found to appear in higher abundance in SAT of aged mice [81, 88, 89] Interestingly, subpopulations of ASC give rise to adipocytes that differ in metabolic properties such as lipolytic activity and fatty acid uptake [87]. Therefore, phenotypic changes in the progenitor pool may negatively influence capacity for differentiation and self-renewal as well as adipocyte function. A study by Tang et al. demonstrated that prolonged treatment of mice with TZD resulted in a reduced differentiation capacity of isolated ASC and a decrease in the population of ASC without a change in the rate of apoptosis [90]. This finding suggests that the progenitor pool and its ability to self-renew may be finite. More studies are required to better understand the molecular and pathophysiological mechanisms responsible for reduced WAT plasticity and its role in the development of adiposopathy.