The global prevalence of overweight and obesity is increasing continuously, affecting more than 2.6 billion individuals, which is approximately 38% of the world population. Projections estimate that, by 2035, more than 50% of the global population could be classified as overweight or obese.¹ Obesity increases the risk for several chronic diseases, including cardiovascular disease, diabetes mellitus, chronic kidney disease, several types of cancer, and various musculoskeletal disorders.² Experts predict that meeting the global aim to reduce the rate of obesity by half by 2025 is nearly impossible. While some countries have made progress in reducing the prevalence of obesity, no country has successfully reversed its obesity epidemic.^3-5 Several genome-wide association studies (GWAS) have found an association between Src homology 2 B adaptor protein 1 (SH2B1) single nucleotide polymorphisms (SNPs) and common obesity.^6,7 SH2B1 is a member of the SH2B protein family of adaptor proteins that play significant roles in various signaling pathways, including leptin and insulin signaling. Initial identification of SH2B1 as a Janus kinase 2 (JAK2)-interacting protein through a yeast two-hybrid screen revealed that SH2B1 enhanced JAK2 activation after binding of leptin to its receptor.^8-10 SH2B1 was found to enhance insulin sensitivity by binding to the insulin receptor.¹¹ Animal models of SH2B1 null mice exhibit severe hyperphagia, hyperleptinemia, hyperglycemia, and hyperinsulinemia, with elevated expression of orexigenic neuropeptide Y (NPY) and agouti-related peptide (AgRP).¹² Restoration of SH2B1 in the hypothalamus of SH2B1 null mice improved hyperphagia, obesity, hyperglycemia, and glucose intolerance, suggesting that the effects of this gene on obesity are mediated through the central nervous system (CNS). Furthermore, the overexpression of SH2B1, specifically in neurons, resulted in dose-dependent protection against leptin resistance and obesity induced by a high-fat diet.¹³ Patients with the SH2B1 220 kb deletion gain weight rapidly, exhibit hyperphagia, and have disproportionately elevated fasting plasma insulin levels compared to age- and obesity-matched controls and patients with other obesity syndromes.¹⁴ These findings highlight the importance of SH2B1 as a susceptibility gene for common obesity. The current study aimed to study the association between the common missense variant of SH2B1, namely rs7498665, with common obesity and elucidate its interactions with actionable lifestyle and behavioral factors in an Israeli population.

Participants

A total of 3,070 Israeli adults (≥18 years) with a mean age of 55.21±14.31 years was included in this study. Participants were genotyped for SH2B1 rs7498665 and completed an online lifestyle and eating habits questionnaire between December 21, 2021 and October 1, 2022. Information from an anonymous genetic database listed in the Israeli registry database (#700068969) of Lev Hai Genetics LTD–MyGenes was analyzed. Ethical approval for this study was received by the Helsinki Committee of Ariel University (#AU-HEA-RB-20220214). Written informed consent by the patients was waived due to a retrospective nature of our study. Exclusion criteria were individuals younger than 18 years, those with a medically identified genetic disorder (self-reported), or those missing genetic or anthropometric data (n=88).

Anthropometric and lifestyle variables and fasting glucose level

Participants completed an online questionnaire to report their physical activity (PA), drinking habits, eating habits, and daily stress levels. PA was assessed by asking questions such as “Are you physically active?” with possible answers of yes or no, “How many days a week do you engage in physical activity?” with answers in the form of the number of physically active days, and “What is the duration of each physical activity you engage in?” with possible answers of 30, 60, or more than 60 minutes. PA was considered at least 150 minutes of any PA per week. Regular wine consumption was defined as 1 to 3 drinks (1 drink=5 ounces) weekly. Anthropometric measurements were self-reported, with weight measured in kilograms and height measured in centimeters. Body mass index (BMI) was calculated as the ratio of weight to height squared (kg/m²). Individuals with a BMI ≥30 kg/m² were classified as obese, while those with a BMI <30 kg/m² were classified as non-obese. Individuals with a BMI ≥25 kg/m² were classified with overweight and obesity, while those with a BMI <25 kg/m² were classified as normal weight according to BMI cut-off points.¹⁵ The eating habits questionnaire contained multiple statements pertaining to eating habits, such as “I consume large portions” and “I eat rapidly,” and study participants were instructed to evaluate each statement on a Likert scale (ranging from 0 never to 4 always). Stress levels of the study participants were classified as either low or high based on their self-reported assessments, with the measurement of perceived stress levels ranging on a Likert scale from 0 (never) to 4 (always). For analytical purposes, we dichotomized this variable into ‘never/rarely’ and ‘frequently/always.’ Participants were classified as type 1 diabetes mellitus (T1DM) or type 2 diabetes mellitus (T2DM) based on physician diagnoses and was self-reported. Plasma glucose levels data were from health maintenance organization (HMO) standard blood tests results reported by the participants. HMO standard blood glucose levels were measured in peripheral venous plasma collected after at least 8 hours of fasting.

Selection of SNPs and Hardy-Weinberg equilibrium calculations

The process of selecting SH2B1 SNPs and ensuring they were in Hardy-Weinberg equilibrium (HWE) involved prioritizing rs7498665 of the SH2B1 gene, which has previously been linked to obesity in other GWAS.^6,7 Selection was based on a minor allele frequency >0.01 in our sample and its inclusion in the validated catalog of published GWAS. The SNP was assessed for HWE using a 1 degree of freedom χ² test.

Statistical analysis

A priori power analysis was carried out using G*Power 3.1.9.7 software¹⁶ to ascertain the minimum sample size required to detect the association between obesity and common SNPs across two groups (odds ratio [OR], 1.5 and a power of 0.95; α=0.05); the minimum calculated sample size was 503. Continuous descriptive characteristics of the study participants are presented as mean±standard deviation. Dichotomized characteristic variables were compared between participants classified as obese and those classified as non-obese and between participants classified as overweight and obese compared to normal weight subjects using the chi-square statistical test. For continuous variables that were not normally distributed, the Mann-Whitney test for independence was performed. Binary logistic regression was employed to elucidate the association between the SNP of interest as a predictor of the relative likelihood of obesity or overweight or obesity using recessive, dominant, additive, and codominant genetic models. Regressions were adjusted for probable confounders (sex, age, T1DM, and T2DM). To determine the OR of gene-environment interactions, logistic regression with interaction was conducted with a model that included independent variables of age, sex, T2DM, T1DM, wine consumption, PA, and genotype. PA duration was calculated by summing weekly duration of total PA for each participant. We calculated a composite eating habits score (EHS) for each participant as the sum of the individual scores generated from a Likert scale questionnaire. Stress level was categorized as “lower” if the participant reported experiencing no or low to medium stress, whereas stress level was classified as “high” if participants reported high or very high levels of stress. To further classify the participants based on their stress levels and eating habits, we generated a composite variable named “EHS+stress.” Participants who were classified as ‘high stress’ and who had an EHS score above the median were categorized as EHS ≥median+high stress. Statistical analyses were performed using SPSS version 29.0 for Windows (IBM Co.).

Population characteristics

The study involved 3,070 adult participants who were grouped according to BMI status. As shown in Table 1, there was a significant difference in sex, weight, BMI, T2DM, PA, EHS, EHS+stress, wine consumption, and glucose levels between participants with obesity (BMI ≥30 kg/m²) and without obesity (BMI <30 kg/m²). There was a significant difference in sex, weight, height, BMI, smoking status, PA, EHS, EHS+stress, T2DM, and glucose levels between participants with overweight or obesity (BMI ≥25 kg/m²) compared to normal weight participants (BMI <25 kg/m²). The non-obese (BMI <30 kg/m²) and normal weight (BMI <25 kg/m²) groups had significantly (P<0.001) higher PA status (at least 150 minutes/week) and significantly lower mean EHS than the group with obesity (BMI ≥30 kg/m²) and the group with overweight or obesity (BMI ≥25 kg/m²) (P<0.001). Furthermore, non-obese participants (BMI <30 kg/m²) showed a greater tendency to consume 1 to 3 drinks of wine per week regularly than the group with obesity.

SH2B1 rs7498665 association with obesity risk

SH2B1 rs7498665 polymorphism genotype frequency was 56.8% for the homozygous reference allele (AA), 36.4% for the heterozygous (AG) allele, and 6.2% for the homozygous polymorphic alteration allele (GG). The associations of SH2B1 rs7498665 with obesity (BMI ≥30 kg/m²) and overweight and obesity (BMI ≥25 kg/m²) classifications were analyzed using dominant, recessive, additive, and codominant genetic models. Significant associations were found between SH2B1 rs7498665 and increased risk of obesity in the recessive (OR, 1.36; 95% confidence interval [CI], 1.01 to 1.85; P=0.048), additive (OR, 1.14; 95% CI, 1.14 to 1.28; P=0.029), and codominant (OR, 1.41; 95% CI, 1.03 to 1.92; P=0.026) genetic models, with 0.8 BMI increment for GG genotype carriers compared to the wild type (AA) (P=0.048). Furthermore, a significant association was found between SH2B1 rs7498665 and increased risk of BMI ≥25 kg/m² using the recessive (OR, 1.90; 95% CI, 1.04 to 3.47; P=0.035), additive (OR, 1.24; 95% CI, 1.02 to 1.51; P=0.030), and codominant genetic models (OR, 2.00; 95% CI, 1.09 to 3.67; P=0.026) (Table 2), with approximately 1 BMI increment for GG genotype carriers compared to the wild type (P=0.036).

SH2B1 rs7498665 interaction with lifestyle factors

Significant gene-environment interactions were found for SH2B1 rs7498665 with PA, EHS, stress level, and wine consumption. Regular PA of at least 150 minutes/week and consuming wine in moderation (1 to 3 drinks per week) were associated with significant attenuation of obesity risk (OR, 0.35; 95% CI, 0.22 to 0.55; P<0.001; and OR, 0.71; 95% CI, 0.61 to 0.84; P=0.017, respectively). PA also attenuated overweight and obesity conferred by SH2B1 rs7498665 (OR, 0.56; 95% CI, 0.35 to 0.89; P<0.010). Conversely, SH2B1 rs7498665 individuals with EHS score ≥median or elevated stress level had a higher obesity risk (OR, 1.32; 95% CI, 1.14 to 1.52; P<0.001; and OR, 1.19; 95% CI, 1.01 to 1.42; P=0.035, respectively). Furthermore, participants with combined EHS ≥median and high stress levels were at significantly higher risk for obesity compared to other participants (OR, 3.63; 95% CI, 1.68 to 7.85; P=0.001). The same significant gene x life and behavior interactions were found for the overweight and obesity categories (BMI ≥25 kg/m²) compared to the normal weight category (OR, 1.59; 95% CI, 1.22 to 2.11; P<0.001; OR, 1.43; 95% CI, 1.06 to 1.94; P=0.024; and OR, 2.10; 95% CI, 1.35 to 3.29; P<0.001) for EHS scores ≥median, high stress levels, and EHS ≥median and high stress levels, respectively. Regular wine consumption did not interact with genotype to affect BMI ≥25 kg/m² (Table 3).

Impact of lifestyle factors on obesity in carriers of two risk alleles

PA, wine consumption, and EHS ≥median+high stress levels interacted with SH2B1 rs7498665 in homozygous risk allele carriers to modify the risk of obesity. Both PA (>150 minutes/week) and wine consumption (1 to 3 drinks/week) were significantly associated with reduced obesity risk (OR, 0.18; 95% CI, 0.04 to 0.71; P=0.010; and OR, 0.40; 95% CI, 0.19 to 0.86; P=0.018, respectively). However, this protective effect of wine consumption was not observed among participants in the BMI ≥25 kg/m² category (OR, 0.79; 95% CI, 0.51 to 1.22; P=0.289). High EHS ≥median combined with a high stress level were significantly associated with increased obesity risk (OR, 3.63; 95% CI, 1.52 to 8.54; P=0.004).

For the overweight and obesity category (BMI ≥25 kg/m²), PA was also significantly associated with a reduced obesity risk (OR, 0.40; 95% CI, 0.22 to 0.73; P=0.003). EHS ≥median+high stress levels were positively associated with overweight and obesity risk (OR, 2.07; 95% CI, 1.19 to 3.60; P=0.014) (Table 4).

A significantly higher obesity frequency (BMI ≥30 kg/m²) was found in homozygous risk allele carriers with EHS ≥above the median (71.6%) than in the normal population and in heterozygous risk allele carriers with EHS ≥median (59.2%, P=0.014). Additionally, obesity frequency was significantly elevated in normal and ≥one risk gene carriers with EHS ≥the median compared to their counterparts with EHS P=0.015, respectively) (Fig. 1A).

Regardless of genotype, participants with EHS ≥median and high stress level had a significantly higher frequency of obesity (BMI ≥30 kg/m²) than other participants (61.7% vs. 55%, P=0.003). A significantly higher frequency of obesity was found among homozygous risk allele carriers with EHS ≥median and high stress levels (81.8%) compared to normal+heterozygous risk allele carriers with EHS ≥median and high stress levels (60.1%, P=0.004). Obesity frequency was significantly elevated in normal and ≥one risk gene carriers with EHS ≥median with high stress levels compared to their counterparts with EHS P=0.024; and 81.1% vs. 57.5%, P=0.004, respectively) (Fig. 1B).

Blood glucose levels

A fasting blood glucose level ≥90 mg/dL interacted significantly with the homozygous SH2B1 rs7498665 genotype carriers to elevate the risk of overweight and obesity (BMI ≥25 kg/m²) (OR, 5.82; 95% CI, 1.04 to 3.26; P=0.040), independent of age, sex, T1DM, and T2DM. Frequency of overweight and obesity was significantly higher among homozygous risk allele carriers with a blood glucose level ≥90 mg/dL than in normal+heterozygous risk allele carriers with a blood glucose level ≥90 mg/dL (97.9% vs. 85.7%, P=0.02). Overweight and obesity frequency were also significantly elevated within each genotype group for those who had a blood glucose level ≥90 mg/dL compared to <90 mg/dL (85.7% vs. 81.9%, P<0.01, respectively for normal+heterozygous risk allele carriers; and 97.7% vs 91.9%, P=0.01, respectively for homozygous risk allele carriers) (Fig. 1C).

SH2B1 rs7498665 (A484T) has been identified in GWAS as a novel susceptibility SNP for complex obesity⁶ and was previously related to higher body weight in a Caucasian female twin population.¹⁷ Our results in the Israeli population support the association between SH2B1 rs7498665 and obesity, consistent with the findings of a Belgian study that the rs7498665 minor allele increased the risk of obesity by 26%¹⁸ and another study that reported an association between SNPs in SH2B1 and a higher genetic risk of obesity in children and adolescents (P=0.04).¹⁹ Conversely, another study found no significant impact of the SH2B1 rs7498665 variant on weight loss-related phenotypes in children following a lifestyle intervention.²⁰ Our findings suggest that the effect of these SNPs might be more pronounced or possibly different in adults than children or dependent on the type of lifestyle intervention. Furthermore, we demonstrated that carriers of the rs7498665 SNP were significantly predisposed to being overweight (BMI ≥25 kg/m²) as well as obese (BMI ≥30 kg/m²), indicating that SH2B1 rs7498665 has a strong effect on the propensity to gain weight. The precise causative mechanism by which the missense variant rs7498665 predisposes to obesity is currently unknown, though SH2B1 plays a known role in energy homeostasis.²¹ Interestingly, SH2B1 is highly expressed in the brain (particularly in the hypothalamus), consistent with an important role in CNS processes involving weight regulation.⁶ The present study investigated the effects of lifestyle and behavioral factor interactions with SH2B1 rs7498665 on obesity risk predisposition, including PA, wine consumption, stress levels, and eating habits. Our findings show that at least 150 minutes/week of PA and consuming wine in moderation (1 to 3 drinks/day) attenuated the increased obesity risk associated with SH2B1 rs7498665. Engaging in PA for at least 150 minutes/week interacted with SH2B1 rs7498665 to reduce the risk of obesity and overweight and of obesity by 65% and 44%, respectively. Engaging in PA for at least 150 minutes/week is in line with the PA guidelines for Americans.²² Similarly, other studies have shown that interactions between SH2B1 rs7498665 and PA alter obesity risk.^23,24 However, these studies were performed mainly in European cohorts and may have been population-specific interactions. The mechanism behind the observed effects of PA on attenuating the risk of obesity in SH2B1 rs7498665 carriers may involve improved brain insulin sensitivity, which can mediate the beneficial peripheral effects of exercise, such as healthier body fat distribution and reduced hunger perception. This is supported by evidence that PA can improve insulin sensitivity in the brain and peripheral tissues, as well as modulate the expression of genes involved in energy metabolism and appetite regulation.²⁵ Additionally, SH2B1 null mice, which exhibit elevated expression of orexigenic neuropeptides NPY and AgRP, display severe hyperphagia, hyper-leptinemia, hyperglycemia, and hyperinsulinemia, suggesting a role for SH2B1 in energy homeostasis.¹²

Our findings indicate that moderate wine consumption may have a protective effect against obesity among carriers of the SH2B1 rs7498665 G allele. Specifically, GG genotype carriers who consumed wine in moderation had a significantly reduced obesity risk (60% reduction) compared to their GG genotype counterparts who did not consume any wine. Resveratrol, a major antioxidative polyphenol found in wine, may modulate obesity-related metabolic pathways by mimicking the effects of caloric restriction and improving exercise performance, which results in insulin sensitivity.²⁶ Additionally, resveratrol has been found to decrease adipogenesis and increase lipid mobilization in adipose tissue, leading to a reduction in body fat.²⁷ Given these beneficial effects, resveratrol has emerged as a promising anti-obesity bioactive agent.²⁸ Several studies have suggested that alcohol consumption, even moderate consumption, may elevate the probability of several diseases.²⁹ Therefore, wine consumption should be confined to low-to-moderate levels (1 to 2 drinks/day or approximately 150 to 300 mL/day). This amount of wine consumption follows the Mediterranean diet recommendations, which is considered favorable for health.³⁰

High levels of stress and poor eating habits were found to exacerbate the risk of obesity among SH2B1 rs7498665 risk allele carriers (for carriers of at least 1 risk allele, P=0.035 and P<0.001, respectively). Our findings indicate that stress alongside maladaptive eating behaviors can exacerbate the predisposition of SH2B1 rs7498665 carriers to obesity, particularly among those carrying two risk alleles (P=0.004). This underscores the need to consider the interaction between genetic factors with behavioral factors such as stress and eating habits when devising personalized obesity prevention and treatment plans. Recent studies highlight the complex interplay between stress and obesity, which involves a range of behavioral, physiological, and psychological factors.³¹ Understanding these mechanisms is crucial for developing targeted interventions to reduce the impact of stress on obesity and promote healthy weight management. Stress has been found to disrupt cognitive processes such as executive function and self-regulation, which may lead to overeating and consumption of high-calorie, high-fat, or high-sugar foods.^32,33 Furthermore, stress can result in decreased PA and sleep duration, which contribute to obesity by altering energy balance.³⁴ Physiological mechanisms that underlie the link between stress and obesity are alterations in the hypothalamic-pituitary-adrenal axis, reward processing in the brain,³⁵ and potentially the gut microbiome.³⁶ Stress may also stimulate the production of biochemical hormones and peptides such as leptin, ghrelin, and NPY, which have been linked to appetite regulation and energy metabolism.³⁷ Moreover, the state of obesity itself can be a source of stress due to weight stigma, which can exacerbate the psychological burden associated with obesity and lead to further weight gain.^37,38 Incorporating stress reduction and management strategies into treatment strategies may prove especially beneficial for individuals with a higher genetic predisposition for obesity, as our results suggest a more pronounced impact of stress on obesity in SH2B1 rs7498665 carriers.

We found a marked difference in obesity frequency between SH2B1 rs7498665 homozygous carriers with EHS ≥median, which highlights the crucial role of eating habits in modulating the impact of genetic predisposition on obesity. This observation emphasizes the importance of targeted dietary interventions and healthy eating habits, particularly for individuals with a higher genetic risk. By addressing modifiable environmental factors, such as eating habits, we can potentially mitigate the risk of obesity in those carrying risk alleles. SH2B1 is known to play a critical role in controlling human food intake and body weight, as well as insulin signaling.¹² Mutations in the SH2B1 gene in patients with severe obesity have been found to be associated with hyperphagia and childhood-onset obesity, as well as a spectrum of behavioral abnormalities, suggesting a link between SH2B1 and maladaptive human behavior.³⁹ Furthermore, although commonly independently investigated and treated, central brain neural circuits that regulate feeding behavior overlap reward circuitry both anatomically and functionally.⁴⁰ Thus, we propose that the interaction between SH2B1 rs7498665 and high maladaptive behavior could have similar underlying mechanisms.

We found that individuals with two SH2B1 rs7498665 alleles who had a blood glucose level ≥90 mg/dL were more likely to be classified as overweight and obesity (BMI ≥25 kg/m²) than wild type carriers and those with only one risk allele who had a glucose level ≥90 mg/dL compared to their counterparts who had a glucose level <90 mg/dL. We chose to set the glucose level of 90 mg/dL as a threshold, as this is indicative of pre-diabetes. Based on our research results, carriers of SH2B1 rs7498665 homozygous risk alleles need to be particularly mindful of their glucose levels due to the significant role of SH2B1 in glucose metabolism and obesity. This finding underscores a critical window of opportunity for intervention and management to prevent progression to T2DM. The SH2B1 gene is known to play a crucial role in leptin and insulin signaling. SH2B1 deficiency results in severe leptin and insulin resistance, which leads to T2DM in mice.¹² SH2B1 promotes insulin signaling by enhancing insulin receptor catalytic activity and protecting against de-phosphorylation of insulin receptor substrates (IRS) proteins. Additionally, SH2B1 promotes pancreatic β-cell expansion and insulin secretion to counteract insulin resistance in obesity, and SH2B1 plays a crucial role in maintaining normal energy and glucose homeostasis.^41,42 This indicates the importance for individuals who carry the rs7498665 allele to monitor and manage their glucose levels. How SH2B1 rs7498665 interacts with glucose to alter obesity risk is unknown; however, a study of 18,014 middle-aged Danes also found a strong association between SH2B1 rs7498665 and increased risk of T2DM after adjusting for age, sex, and BMI.⁴³

While our study provides valuable insights into the associations between SH2B1 rs7498665, lifestyle factors, overweight and obesity, several limitations must be acknowledged. The cross-sectional design of our research limited our ability to establish causality. Additionally, the assessment of PA was based solely on the duration of exercise, without considering the intensity or type of PA. This approach might not fully capture the actual energy expenditure or the varied effects of different physical activities on health outcomes. Our study population was limited to Israeli individuals, which limits the generalizability of our findings to other ethnicities or populations. Another limitation pertains to the methods of data collection; our reliance on self-reported data collected through an online questionnaire involves potential for recall bias and self-reporting inaccuracies. Furthermore, the availability of glucose level only limited the scope of our biomedical analysis. Nevertheless, our sample size was substantial at 3,070 participants. This large number of participants enhanced our statistical power and may have helped mitigate some of the limitations associated with our research design and methods.

In summary, SH2B1 rs7498665 significantly elevates the risk of obesity. Actionable lifestyle and behavioral factors significantly modulate the rs7498665 genetic predisposition to obesity; PA and moderate wine consumption attenuate the risk, while high stress levels, EHS, and glucose levels increase the obesity risk among carriers. Our findings suggest that modification of lifestyle factors, maintenance of optimal blood glucose levels, and behavioral adaptations may help carriers of SH2B1 rs7498665 manage overweight and obesity more effectively.

Ruth Birk is a scientific consultant of MyGenes. Danyel Chermon has no conflicts of interest to declare.

We would like to thank Lev Hai Genetics LTD–MyGenes for the data analyzed in this study.

Study concept and design: DC and RB; acquisition of data: DC and RB; analysis and interpretation of data: DC and RB; drafting of the manuscript: DC and RB; critical revision of the manuscript: DC and RB; statistical analysis: DC and RB; administrative, technical, or material support: DC and RB; and study supervision: RB.