The proposed systemic thermogenic metabolites succinate and 12,13-diHOME are inversely associated with adiposity and related metabolic traits: evidence from a large human cross-sectional study

Obesity is associated with the development of type 2 diabetes and cardiovascular disease. The rising prevalence of obesity has focused attention on the discovery of safe and effective therapeutics promoting weight-loss and/or a healthy metabolic profile. Mammals possess two adipocyte types. White adipocytes function to store excess energy whereas brown adipocytes dissipate energy as heat and serve to maintain body temperature via non-shivering thermogenesis. In addition to thermogenesis, activation of brown adipose tissue (BAT) in rodents was shown to ameliorate hyperlipidaemia, promote glucose tolerance and insulin sensitivity, and protect against obesity and atherosclerosis [1‐3]. In adult humans BAT activity was also shown to be inversely correlated with total and visceral adiposity and fasting plasma glucose [4, 5]. Chronic cold exposure and β3-adrenergic agonist treatment further led to BAT activation concomitant with an improvement in insulin sensitivity and/or a reduction in body fat mass [6‐10]. These findings highlight that pharmacological induction of BAT thermogenesis in humans may ameliorate obesity and related metabolic disorders. However, pharmacologic targeting of β3 adrenoceptors has thus far demonstrated limited clinical efficacy and/or unacceptable side-effects. Hence, discovering novel BAT activation pathways is an important research focus.

Succinate is a tricarboxylic acid cycle (TCA) intermediate which can also exit mitochondria and function in the circulation. Succinate was recently identified as a systemic signal promoting BAT thermogenesis in rodents [11]. Specifically, circulating succinate was shown to increase following cold exposure in response to muscle shivering and selectively taken up by BAT leading to acute activation of thermogenesis. Pharmacological elevation of plasma succinate drove BAT thermogenesis with resultant protection against diet-induced obesity and glucose intolerance. Using a global lipidomic analysis, another landmark study identified 12,13-dihydroxy-9Z-octadecenoic acid (12,13-diHOME) as a lipid species increased in the circulation of humans and mice following cold exposure [12]. Injection of 12,13-diHOME in rodents acutely activated BAT NEFA uptake leading to enhanced cold tolerance. Chronic treatment of diet-induced obese mice with 12,13-diHOME led to a reduction in circulating triacylglycerol. In a follow-up study, plasma 12,13-diHOME levels were also shown to increase in response to an acute bout of exercise in humans and mice [13]. Acute treatment of mice with 12,13-diHOME led to increased skeletal muscle NEFA uptake and oxidation. Prompted by these findings, we set out to investigate the relevance of plasma succinate and 12,13-diHOME in the regulation of energy metabolism in a large cohort of healthy volunteers.

Our study provides novel insights, as well as confirming previous cross-sectional data from small cohorts, about the association of plasma succinate and 12,13-diHOME with adiposity and metabolic traits in humans. Our data highlight a potential role for both metabolites in the regulation of energy balance, systemic metabolism and BAT activation in humans, thereby extending the results of recent landmark studies demonstrating that circulating succinate and 12,13-diHOME promote brown adipocyte thermogenesis and protect against metabolic disorders in rodents [11, 12]. Nonetheless, despite—to our knowledge—this being the largest human series to date to directly examine the associations of these metabolites with body composition and plasma biochemistry, our study has limitations. Namely, we did not investigate the effects of pharmacological supplementation with succinate or 12,13-diHOME on adiposity, metabolic profile or BAT activation; nor did we measure BAT activity or systemic markers thereof (e.g. FGF21) in our study participants. As such our results should be interpreted with caution. Additionally, physical activity was determined using a questionnaire rather than accelerometer devices which are the gold standard. Finally, we did not have information as to whether study participants had been physically active in the few hours prior to blood extraction. This may be another limitation of our study given that an acute spurt of exercise has been shown to increase both plasma succinate and 12,13-diHOME levels in humans [13, 23]. However, all blood sampling was undertaken in the fasted state in the early morning hence it is unlikely that this was the case. Furthermore, any noise introduced in our dataset by physical activity prior to blood sampling is likely to have diluted rather than strengthened the observed associations due to ‘measurement error’.

In addition to acting as a TCA intermediate, succinate is used as a food additive and dietary supplement. Succinate might also be released into the circulation by gut commensal bacteria which produce large amounts of this metabolite through fermentation of dietary fibre [25]. The association between circulating succinate levels and cardiometabolic disease was previously investigated only in small series. In a cohort of 91 participants, plasma succinate was higher in obese than in lean individuals and those with type 2 diabetes (n = 20) vs BMI-matched normoglycaemic participants [26]. Similarly, van Diepen et al found elevated plasma succinate levels in 58 participants with type 2 diabetes compared with 76 healthy control participants [27]. In contrast, another study found no difference in fasting and postprandial serum succinate between 23 participants with type 2 diabetes and 15 control participants [28]. Similarly, no differences in serum succinate were detected between individuals with (n = 172) and without (n = 61) coronary artery stenosis [29]. Herein, we extend these findings by demonstrating an inverse association between plasma succinate and DXA-determined total and visceral adiposity, as well as a positive association with HDL-cholesterol, NEFA, glycerol and lactate levels. The positive association between plasma succinate and NEFA, as well as glycerol concentrations, was unexpected given that succinate acting via succinate receptor 1 (SUCNR1) was previously shown to inhibit adipocyte lipolysis [30]. This finding is likely to reflect increased succinate production from TCA metabolism by oxidation of carbons derived from NEFA. A similar mechanism is also likely to explain the positive relationship between systemic succinate and lactate levels as glucose was shown to feed the TCA cycle via circulating lactate [31]. We were unable to examine the association between plasma succinate levels and type 2 diabetes status as our study population was composed solely of healthy participants. Of note, with the exception of total fat mass percentage, the associations between circulating succinate and reduced adiposity, as well as increased HDL-cholesterol, became non-significant after adjustment for physical activity. These results suggest that, in addition to potentially having direct effects on systemic metabolism, plasma succinate may also serve as a surrogate marker of cardiorespiratory fitness; indeed exercise is known to lead to elevated plasma succinate through increased muscle contraction [32].

12,13-diHOME is produced by oxidation of linoleic acid, an essential fatty acid which is abundant in many nuts and fatty seeds, followed by hydrolysis which is catalysed by soluble epoxide hydrolases. The latter are encoded by four genes, EPHX1-4. Like succinate, gut commensal bacteria might also contribute to plasma 12,13-diHOME levels [33]. A potential link between 12,13-diHOME and metabolic traits was first reported by Schuchardt et al [34] who showed that the serum concentration of this lipokine was lower in 20 normolipidaemic vs 20 hyperlipidaemic middle-aged men. Subsequently, plasma 12,13-diHOME levels were shown to be strongly and negatively correlated with BMI, HOMA-IR and fasting plasma triacylglycerols in 55 predominantly female participants, including ten with type 2 diabetes [12]. Associations between 12,13-diHOME and BMI and triacylglycerols were confirmed in a follow-up study comprising 39 healthy individuals which also demonstrated a correlation between 12,13-diHOME and cardiovascular fitness [13]. Our study broadly confirms these findings. However, in our cohort the associations between systemic 12,13-diHOME levels and anthropometric and metabolic traits were weaker than those previously reported. Also, contrary to these earlier studies, which found no significant associations between 12,13-diHOME and HDL-cholesterol and/or fasting glucose, these two metabolites exhibited statistically robust relationships with 12,13-diHOME in the OBB. These differences are probably due to the large size and composition of our study population which comprised healthy individuals with a more narrow age and BMI range than those reported in earlier studies. Due to the relatively large size of our study cohort we were additionally able to show that circulating 12,13-diHOME levels were higher in women than men. The sexual dimorphism in systemic 12,13-diHOME concentration had not been observed in earlier studies [12, 13]. We also detected strong positive associations between plasma 12,13-diHOME and NEFA, as well as glycerol levels. We interpret these data as being consistent with the release of 12,13-diHOME from adipocytes into the systemic circulation consequent to fasting-induced lipolysis. In this respect, Lynes et al detected a high concentration of 12,13-diHOME in mouse white adipose tissue and, whilst the concentration of this lipokine has not been determined in human fat, based on RNAseq data from GTEx (https://​gtexportal.​org) EPHX1 and 2 are both highly expressed in this tissue. Plasma 12,13-diHOME also correlated positively with circulating β-OH butyrate levels, a result in keeping with elevated fasting-induced lipolysis providing increased NEFA substrates for hepatic ketogenesis. An acute bout of moderate exercise has been shown to cause a pronounced increase in circulating 12,13-diHOME levels in humans. Active male participants also had significantly higher baseline 12,13-diHOME levels compared with their sedentary counterparts [13]. In this regard, it is noteworthy that the associations of circulating 12,13-diHOME levels with body composition and metabolic profile, whilst attenuated, remained significant after adjustment for physical activity. These data suggest that this lipokine has direct effects on systemic metabolism, possibly by promoting BAT activity.

MR analyses did not support a causal influence of increased obesity in reducing succinate or 12,13-diHOME plasma levels. Whilst these results may be due to insufficient statistical power (the OBB dataset with available metabolomic data was small [n = 2248]), any undetected effects are likely to be very small. Together with the observational data presented herein, these findings suggest that higher circulating levels of succinate and 12,13-diHOME may be causally linked to, rather than being a consequence of, reduced fat mass accumulation. Unfortunately, we were unable to directly test this hypothesis using MR owing to the lack of available genetic instruments associated with plasma succinate or 12,13-diHOME levels. Similarly, we discovered no evidence that a genetically determined abnormal metabolic profile modulates systemic levels of succinate or 12,13-diHOME although, in the case of raised fasting glucose, insulin and HOMA-IR, this may have been due to the lack of sufficient genetic instruments and low statistical power. On the other hand we did find that a genetic predisposition to upper-body fat distribution was associated with higher circulating levels of both metabolites. We speculate that this is a consequence of the increased spontaneous lipolysis associated with central obesity [35].

In conclusion, we provide evidence that circulating succinate and 12,13-diHOME may play a role in the regulation of energy balance and systemic metabolism in humans. Further studies involving direct assessment of physical and BAT activity, dietary supplementation and bidirectional MR are necessary to examine the potential health benefits of these proposed thermogenic metabolites.