Childhood obesity often persists into adulthood (1), bringing with it a host of metabolic abnormalities (2). Compared to people who develop obesity as adults, those with childhood-onset obesity (CO) face a heightened risk of type 2 diabetes (3–11). While the mechanisms behind this increased risk remain unclear, chronic, low-grade inflammation—implicated in both insulin resistance and β-cell dysfunction (12)—may play a key role. Inflammation also contributes to arterial stiffness, a marker of subclinical cardiovascular disease (13). Increased arterial stiffness often precedes the development of insulin resistance and type 2 diabetes (14, 15) and is linked to the microvascular and macrovascular complications of type 2 diabetes (16).

Adipose tissue serves as a central hub linking inflammation to cardiometabolic disease. As both an immunological and endocrine organ, it harbors diverse immune cells and secretes inflammatory factors that can act locally or systemically. During obesity development in mice, adipose tissue immune cell and secretory profiles shift from anti-inflammatory to pro-inflammatory (17). Children with obesity already exhibit macrophage infiltration in adipose tissue and elevated circulating inflammatory markers, such as C-reactive protein (18). Whether their adipose tissue and systemic inflammation worsen as they enter adulthood remains unknown.

Our group has compared adipose tissue macrophage (ATM) populations between female bariatric surgery patients with CO and adult-onset obesity (AO), matched for type 2 diabetes status. We found that the age of obesity onset did not affect M1-like ‘pro-inflammatory’ and M2-like ‘anti-inflammatory’ macrophage content in abdominal subcutaneous adipose tissue (SAT) and visceral adipose tissue (VAT) (19). However, we do not know if such similarities manifest in adults with milder obesity before the onset of comorbidities or if they extend to other immune cells and circulating adipokines.

Furthermore, we do not fully understand whether conventional weight loss treatments can effectively target inflammation in people with CO and AO. While some pro-inflammatory adipose tissue immune cells decrease after bariatric surgery (20, 21), this is not always the case after lifestyle interventions—despite cardiometabolic improvements (22, 23). We suspect that the age of obesity onset may contribute to this variability.

In this preliminary, hypothesis-generating study, we aimed to examine the effect of age of obesity onset (CO vs. AO) on abdominal and femoral SAT macrophage and T-cell populations (primary outcomes), as well as circulating adipokines and subclinical cardiometabolic risk factors (hyperinsulinemia and arterial stiffness) before and after moderate (~10%) weight loss.

This study was motivated by the increased risk of type 2 diabetes associated with CO and is the first of its kind to examine the effects of age of obesity onset on regional SAT immune cells and systemic inflammation before and after weight loss. Contrary to our expectations, we did not find a more pro-inflammatory SAT immune cell profile in females with CO than in those with AO. Instead, we found a slightly higher percentage of M1-like macrophages in the abdominal SAT of females with AO, and no other differences in macrophage or T-cell percentages between groups. Fasting plasma adipokine and insulin concentrations, as well as arterial stiffness, were also not significantly different between groups. Moderate (~10%) weight loss similarly reduced arterial stiffness and plasma concentrations of insulin, leptin, and PAI-1 in females with CO and AO. However, SAT immune cell percentages only shifted in females with AO. After weight loss, their M1-like macrophages in abdominal SAT decreased, while their CD3⁺CD4⁺ T cells in abdominal and femoral SAT increased.

Consistent with our baseline findings on M1-like macrophages, females with AO exhibit greater IL-6 gene expression in abdominal SAT than do those with CO (42). A mouse study by Wernstedt Asterholm et al. (43) prompts us to question whether increased abdominal SAT pro-inflammatory macrophages in AO represent a temporary adaptive response to recent weight gain rather than a driver of chronic inflammation. Wernstedt Asterholm et al. (43) found that, during the early stages of obesity development in mice, adipose tissue inflammation is crucial for proper extracellular matrix remodeling and angiogenesis, processes that promote adipogenesis. They postulated that by promoting healthy tissue expansion, acute adipose tissue inflammation can resolve before becoming chronic (43). While no studies have compared adipogenic capacity between people with CO and AO, our lab has provided evidence suggesting increased lipogenic capacity in females with AO (42).

Although it is appealing to view the increased abdominal SAT proinflammatory macrophages in AO as an adaptive response to recent weight gain, we must consider that adult obesity typically develops over several months or even years in humans. This raises questions about what constitutes an ‘acute’ inflammatory response to weight gain in humans and how long healthy adipose tissue remodeling can persist before chronic, low-grade inflammation takes over. Short-term overfeeding studies in healthy adults show that after a ~3-kg weight gain, the total number of macrophages in abdominal SAT remains stable. However, the M1/M2 macrophage ratio increases along with markers of extracellular matrix remodeling and angiogenesis (44, 45). What remains unclear is whether these changes after weight gain are tempered once weight stabilizes over the long term. In our study, participants were weight-stable for at least three months, yet we still observed slight differences in abdominal SAT M1-like macrophages between those with recent weight gain (AO) and those with CO. Studies comparing bariatric surgery patients with CO and AO indicate that by the time they have severe obesity with comorbidities, their ATM profiles converge (19). The adiposity trajectories leading to this convergence remain to be studied.

Even though our findings suggest that SAT immune cells and circulating adipokines may not explain the increased type 2 diabetes risk associated with CO, they do not discount the role of inflammation entirely. While the degree of SAT and systemic inflammation is one consideration for cardiometabolic risk, the duration of exposure to inflammation is another. Still, this reasoning does not align with our subclinical findings. If those with CO experience chronic, low-grade inflammation for longer, why don’t they exhibit worse arterial stiffness than those with AO? Differences in other triggers of arterial stiffness, such as oxidative stress, endothelial dysfunction, or structural changes in the artery, could provide an explanation.

Average plasma insulin concentration was ~100 pmol/L greater in females with CO than in those with AO, and although this difference was not statistically significant, it may still be clinically relevant. Hyperinsulinemia, a well-established risk factor for type 2 diabetes, remains difficult to define due to variability in insulin assays. However, the mean plasma insulin concentrations of ~300 pmol/L in our CO group and ~200 pmol/L in our AO group are notably higher than the ~60 pmol/L we measured in a small sample of lean females using the same methodology. Plasma insulin concentrations reflect the balance between insulin secretion and clearance. An increase in the former and a decrease in the latter can lead to fasting and postprandial hyperinsulinemia in people with obesity (46).

Drawing from current evidence, we suspect that compensatory hyperinsulinemia may be more common in AO, whereas hypersecretion-induced hyperinsulinemia may be more common in CO. Studies using the hyperinsulinemic-euglycemic clamp technique have found worse insulin resistance in people with AO compared to those with CO (47, 48). Moreover, at the same level of glycemia and insulin sensitivity, children with obesity have hyperresponsive β-cells compared to adults with obesity (49–51). This hyperresponsiveness could lead to accelerated β-cell exhaustion and an earlier onset of type 2 diabetes in adults with CO (52). Studies comparing insulin kinetics and multi-organ insulin sensitivity between adults with CO and AO over time are required to test this hypothesis.

Many cardiometabolic risk factors improve following lifestyle weight loss interventions (53–55), but whether changes in adipose tissue immune cell profiles drive these improvements remains unclear. Collectively, human studies suggest that total abdominal SAT macrophage content increases or remains stable in the early phases of rapid or gradual diet and/or exercise-induced weight loss (< ~10%), despite improvements in cardiometabolic risk factors (22, 23, 55–59). However, with continued weight loss (> ~15%), total abdominal SAT macrophage content decreases (55, 56, 60). There is also evidence that exercise, with or without weight loss, shifts abdominal SAT macrophages toward an anti-inflammatory M2-like phenotype (57, 61). In our study, only females with AO experienced a decrease in the proportion of abdominal SAT M1-like macrophages after weight loss, eliminating the baseline difference between those with AO and CO. Therefore, in some people, a 10% weight loss achieved through diet and exercise may reduce pro-inflammatory macrophages, with baseline levels possibly playing a role.

In our study, the percentages of abdominal and femoral SAT CD3⁺CD8⁺ T cells (pro-inflammatory) did not change after weight loss, regardless of obesity onset. In contrast, the proportion of CD3⁺CD4⁺ T cells increased in both SAT regions in females with AO, but not in those with CO. Our findings in females with AO align with those of Kratz et al., who examined changes in abdominal SAT T-cell populations after a 1-year lifestyle intervention (~7% weight loss) in people with obesity and type 2 diabetes (23). Given their participants were older (~50 years), it is plausible that a higher proportion had AO.

SAT Treg cells, an anti-inflammatory subset of CD3⁺CD4⁺ T cells, are both reduced and less functional in adults with obesity (62). Due to chronic antigen stimulation, they adopt an exhausted phenotype marked by impaired activity and a diminished ability to proliferate (62). Interestingly, Cottom et al. found that adipose tissue Treg cells remain low and maintain an exhausted phenotype after weight loss in obese mice (63). If the increase in CD3⁺CD4⁺ T cells in people with AO is due to a rise in Treg cells, it might suggest that their baseline Treg cells—even if similar in quantity to the CO group—are less exhausted. Conversely, in those with CO, 10% weight loss may not be enough to recover Treg cell exhaustion. Further phenotyping of CD3⁺CD4⁺ T cells in people with CO and AO, both before and after weight loss, is needed to explore this possibility.

For people with obesity engaged in lifestyle interventions, an initial weight loss target of 5-10% is considered clinically significant. We found that after 10% weight loss, females with CO and AO improved their plasma insulin concentration and arterial stiffness. However, our findings, along with those of others (22, 55), suggest that 10% weight loss may not be sufficient to induce positive changes in certain aspects of SAT biology—at least in some people. Consistent with other studies (55, 64), we found that circulating concentrations of only select adipokines, leptin and PAI-1, declined in both groups after moderate weight loss. We also found that adipocyte size decreased and adiponectin secreted from SAT increased in both groups, whereas SAT immune cell profiles only changed in those with AO.

While these results may suggest that changes in SAT immune cell profiles are not required for short-term improvements in cardiometabolic risk factors, we wonder whether they are necessary to sustain improvements over the long-term. Since immune cell infiltration is one of the first changes with weight gain, could it be the last to resolve with weight loss before lasting metabolic health is restored? In mice with diet-induced obesity, a proinflammatory immune cell profile in adipose tissue persisted after weight loss and, along with glucose tolerance, worsened after weight regain (63). The authors speculated that a ‘memory-like immunological imprinting’ may contribute to the exacerbated metabolic dysfunction associated with weight regain (63). Interestingly, after a 10% diet-induced weight loss in humans, the lower the reduction in gene expression of immune cell integrins in abdominal SAT, the greater the weight regain over the following nine months (65). These findings may be particularly relevant to those with CO, whose SAT immune cell profile did not change after 10% weight loss and whose hyperplastic abdominal SAT may predispose them to weight regain (66). Longitudinal studies are required to determine whether those with CO need greater weight loss or alternative interventions compared to those with AO to prevent weight regain and sustain metabolic improvements. GLP-1 receptor agonists may offer one such alternative, as they can reduce appetite and—at least in preclinical models—modulate adipose tissue inflammation (67).

We acknowledge that the M1/M2 classification may oversimplify the inflammatory and functional diversity of ATM. Measuring cytokine production from isolated macrophage populations would have provided greater insight into their inflammatory phenotype. We were also unable to assess CD3⁺CD4⁺ T-cell subtypes, which limits our understanding of how they contribute to the SAT inflammatory milieux. Moreover, our analysis was limited to macrophages and T cells in SAT, preventing a comprehensive understanding of the adipose tissue immune landscape. Our findings are derived from a small sample of healthy females with obesity, which limits generalizability to other populations, including males and people with varying degrees of cardiometabolic risk.

A major strength of our study is the use of the gold standard technique, flow cytometry, to quantify immune cell percentages in two SAT regions. SAT immune cell percentages quantified by flow cytometry demonstrate good test-retest reliability (68), which is crucial for our pre-post design. Another strength is that we verified the age of obesity onset through photographic evidence and body rating scales, ensuring accurate classification of participants. Furthermore, our narrow eligibility criteria resulted in groups well-matched for age, body composition, and body fat distribution, enhancing our ability to isolate the effect of age of obesity onset on SAT immune cell profiles.

Our findings provide preliminary evidence that the age of obesity onset influences changes in SAT macrophages and T cells following moderate weight loss. Unexpectedly, females with AO had slightly more proinflammatory macrophages at baseline in abdominal SAT compared to those with CO, though this difference diminished after weight loss. In contrast, T-cell populations in both abdominal and femoral SAT were unaffected by the age of obesity onset at baseline; however, CD3⁺CD4⁺ T cells increased after weight loss only in those with AO. Our results suggest that SAT immune cell profiles may not fully account for the elevated type 2 diabetes risk in people with CO or the short-term cardiometabolic benefits of weight loss. However, the long-term implications of SAT immune changes—or lack thereof—after weight loss remain unclear. Future research should determine if sustained metabolic improvements in people with CO and AO require specific changes in SAT immune cell populations and whether targeted interventions are needed to support these changes.