Abstract

Introduction: Bile acid signaling has been suggested to promote BAT activity in various experimental models.brown adipose tissue However, little is known if and how physiologic bile acid metabolism is linked to BAT function in humans.Here Bile acids we investigated the association between BAT activity and circulating bile acid concentrations in lean and obese Thermogenesis individuals.

CYP8B1 Methods: BAT18F-fluorodeoxyglucose uptake was measured after a standardized cooling protocol by positron emission tomography/computed tomography. Cold-induced thermogenesis was assessed by indirect calorimetry. Fasting bile acid concentrations were determined by high performance liquid chromatography-high-resolution mass spectrometry.

Results: In a cohort of 24 BAT-negative and 20 BAT-positive individuals matched by age, sex, and body mass index, circulating bile acid levels were similar between groups except for higher ursodeoxycholic acid and a trend towards a lower 12α-OH/non-12α-OH bile acid ratio in lean participants with active BAT compared to those without. Moreover, the 12α-OH/non-12α-OH ratio, a marker of CYP8B1 activity, correlated negatively with BAT volume and activity.

Conclusion: Fasting concentrations of major bile acids are not associated with cold-induced BAT activity in humans. However, the inverse association between BAT activity and 12α-OH/non-12α-OH ratio may suggest CYP8B1 as a potential new target in BAT function and warrants additional investigation.

1. Introduction

Due to its energy dissipating qualities, brown adipose tissue (BAT) has emerged as a promising therapeutic target for the treatment of obesity and related sequelae. Activation of BAT in humans results in increased energy expenditure, improves insulin sensitivity and lipid metabolism (Chondronikola et al., 2016; Iwen et al., 2017; van Marken Lichtenbelt et al., 2009). BAT recruitment by means of chronic cold exposure also reduced body fat mass (Yoneshiro et al., 2013). However, little is known about the endogenous regulation of BAT activity. Recently, bile acid metabolism has been linked to BAT function. It has been suggested that bile acids could exert beneficial metabolic effects through the G protein-coupled bile acid receptor TGR5 or the nuclear farnesoid x receptor (FXR) with variable affinities of different bile acid species for these receptors (Chen et al., 2011; de Aguiar Vallim et al., 2013; Shapiro et al., 2018). TGR5 is expressed on adipocytes and increases thermogenesis by interacting with the thyroid hormone-converting enzyme type II iodothyronine deiodinase (Watanabe et al., 2006). Intestinal FXR agonism promoted adipose tissue browning in mice and reduced insulin resistance and obesity (Fang et al., 2015). Cholic acid administration in mice induced expression of uncoupling protein-1 in BAT, increased energy turnover and reduced adiposity (Zietak and Kozak, 2016). In humans, short-term treatment with the primary bile acid chenodeoxycholic acid resulted in enhanced BAT activity and energy expenditure (Broeders et al., 2015). A study in mice reported that cold exposure leads to increased bile acid concentrations which mediate an expansion of BAT activity (Worthmann et al., 2017).

Previous studies reported that obesity is itself associated with an increased bile acid synthesis, especially increased 12α-OH bile acids (cholic acid and deoxycholic acid) through CYP8B1, as well as an impaired hepatic sinusoidal transport (Haeusler et al, 2013, 2016). Thus obesity-induced alterations in bile acid concentrations could serve as a negative feedback mechanism by increasing BAT activity.Given these observations, we aimed to establish the association between endogenous circulating bile acids and cold-induced BAT activity in a population of both lean and obese but otherwise healthy adults.

2. Methods
2.1. Population

We included participants of two prospective studies investigating BAT activity in healthy subjects who are lean or obese (cohort 1, NCT02381483) as well as subjects with morbid obesity (cohort 2, NCT03168009). Inclusion criteria included age between 18 and 50 years and a body mass index either between 18.5 and 25 kg/m2 (lean) or 30 and 55 kg/m2 (obese). The exclusion criteria included endocrinological disease except substituted hypothyroidism, chronic kidney disease, chronic liver disease, malignancies or autoimmune disease requiring systemic immune-modulatory treatment. For the determination of bile acid concentrations, we selected a subgroup of the participants of the two studies. With the aim of an equal representation of lean and obese participants in this study, westratified the participants in a lean (BMI < 25 kg/m2) and an overweight (BMI > 30 kg/m2) group.A total of 31 lean and 57 obese individuals were available from two previous studies (cohort 1 and cohort 2) in whom cold-induced BAT activity had been characterized by positron emission tomography and computed tomography Environmental antibiotic (PET/CT). Within each group, we matched BAT positive (BAT pos) with BAT negative (BAT neg) participants using propensity score matching to achieve an even age and sex distribution. This resulted in 45 participants (lean: 10 BAT neg/10 BAT pos; obese: 15 BATneg/10 BAT pos).

2.2. Study procedures

After an overnight fast (>10 h), the patients arrived at the metabolic unit of the Clinical Division of Endocrinology and Metabolism at Vienna General Hospital where their height was measured, and body composition was analyzed on a scale incorporating bio-electrical impedance analysis (seca mBCA 515, seca GmbH & Co. KG, Hamburg, Germany). After 30 min of rest in supine position, resting energy expenditure was measured by indirect calorimetry (Quark RMR, COSMED srl., Rome, Italy). Baseline EDTA plasma samples were drawn from an indwelling catheter in an antecubital vein. Then the patients were fitted with a water-perfused cooling vest.The temperature was set above their respective shivering threshold determined by surface electromyography (EMG Quattro, OT Bioeletronica, Torino, Italy). For this purpose, surface electrodes were placed above the major pectoral muscle. The electromyographical signal was assessed in realtime for signs of shivering such as sudden non-voluntary increases in amplitude either continuously or in bursts. From 60min to 90min of cold exposure, a second indirect calorimetry was performed to assess cold-induced non-shivering thermogenesis (CIT) which is expressed as percentage increase in energy expenditure. Next, 2.5 Megabecquerel of 18F-fluorodeoxyglucose (18FFDG) per kilogram of body weight were administered intravenously and cold exposure was continued until 150min. After cold exposure, a second blood draw was performed for the determination of norepinephrine levels in plasma. All laboratory analyses but the bile acid determination were performed using diagnostic assays implemented at the institution ’s Department of Laboratory Medicine.

2.3. Determination of plasma bile acid concentrations

Bile acids (cholic acid, deoxycholic acid, chenodeoxycholic acid, lithocholic acid, ursodeoxycholic acid) were determined as unconjugated acids and as taurine and glycine conjugates using an Orbitrapmethod (Q-Exactive, Thermo Fisher Scientific) in full-scan mode and high resolution (Amplatz et al., 2017). High-performance liquid chromatography was performed on a reversed-phase (C18) column using a methanol/water gradient for chromatographic solution of isobaric bile acids. Quantitation was performed by the use of deuterated internal standards and correlation of peak area ratios in linear regression (Amplatz et al., 2017). Composite bile acids were calculated as follows: total bile acids: sum of all bile acid species independent of conjugation status; primary bile acids: sum of cholicacid and chenodeoxycholicacid including their glycine and taurine conjugates; secondary bile acids: sum of deoxycholic acid, lithocholic acid, and ursodeoxycholic acid including their glycine and taurine conjugates; 12-α hydroxylated bile acids: sum of cholic acid and deoxycholic acid including their glycine and taurine conjugates. Lithocholic acid was only detected in 4 participants and thus was not analyzed individually.

2.4. PET/CT imaging and image analysis

60 min after the application of the radiotracer, a combined PET/CT acquisition was started on a Siemens Biograph 64 True Point scanner (Siemens Healthineers, Erlangen, Germany). First, a low dose CT scan (120 kV, 50 mAs) was performed for attenuation and scatter correction as well as for precise anatomical localization, followed by PET acquisition in 3D mode of 3 min/bed position. Both, PET and CT were acquired from the base of the skull to mid-thigh. The images were analyzed using Hermes Hybrid 3D Viewer (Hermes Medical Solutions, Stockholm, Sweden). BAT was delineated using a double-threshold model using lean body mass adjusted SUV thresholds as proposed in the BARCIST Criteria (Chen et al., 2016). For the delineation of BAT, the regions of interest were delineated in the axial fused images. Two experienced nuclear medicine physicians visually inspected each slice for potential spillover from adjacent non-fat tissues such as muscles, glands, or lymph nodes. Those with detectable active BAT had BAT volumes ranging from 19 to 452 ml and are referred to as BAT positive.

2.5. Statistics

Between-group comparisons were analyzed using Student ’s independent samples t-test or Mann-Whitney-U test, as appropriate. Correlation analysis was performed using Spearman ’s rank correlation coefficient. Due to the retrospective and hypothesis-generating design, no adjustment for multiplicity was performed. The analyses were performed using the statistical software package SPSS 25 (IBM Corp., Armonk, NY, USA). Graphs were built using GraphPad Prism (GraphPad Software, La Jolla, California USA).

3. Results
3.1. Metabolic parameters did not differ between age-, sexand BMImatched individuals with and without active BAT

Of the 45 participants included in this study, one was excluded because of fasting bile acid levels above the commonly accepted upper normal limit for healthy adults of 10 μmol/L. This resulted in a total cohort of 20 subjects with (BAT pos) and 24 without (BAT neg) active BAT. By design of the study, the BAT neg and BAT pos participants did not differ by age, sex, or body composition (Table 1). Important metabolic parameters selleck kinase inhibitor such as blood glucose, cholesterol and triglycerides were similar in both groups. Only the liver enzyme aspartate amino
transferase was significantly lower in BAT pos compared to BAT neg (Table 1).

3.2. BATpos and BATneg participants had similar bile acids profiles

Given the accumulating preclinical evidence that bile acid signaling may affect BAT activity in vivo, we investigated if fasting plasma bile acid concentrations were associated with the presence of active BAT in humans. Cold-induced BAT activity was assessed by 18F-FDG-PET/CT in the entire study cohort 150 min after mild cold exposure using a personalized cooling protocol. While BATneg subjects had no significant 18F-FDG fat uptake after cold exposure, BAT pos subjects had strong 18FFDG uptake in the typical cervical and thoracic BAT regions following cold exposure (Fig. 1A–D). As expected, cold-induced thermogenesis (CIT) was markedly higher in BATpos compared to BATneg subjects (Fig. 1E). To exclude an insufficient noradrenergic response in those without detectable BAT, we measured circulating norepinephrine levels before and after cold exposure. The increase of norepinephrine concentrations was similar in BAT pos and BAT neg participants (Fig. 1F). Accordingly, the temperatures of the cooling vests did not differ between the two groups (Suppl. Fig. 1).We found no statistical differences in total bile acids or distinct bile acid groups between BAT pos and BAT neg subjects (Table 2). To consider potential effects of leanness versus obesity on bile acid profiles we performed subgroup analysis. In general, bile acid concentrations did not differ between BAT pos and BAT neg lean and obese participants,respectively. However, ursodeoxycholicacid (UDCA) levels were higher in lean BAT pos than in lean BAT neg participants whereas the ratio of 12αOH/non-12α-OH bile acids was almost 2-fold lower (p = 0.096) in lean BATpos versus BATneg subjects with no difference between these two groups in obese participants (Table 2).

3.3. The 12“-OH/non-12“-OH ratio is associated with BAT activity markers

Next, we studied a potential association between BAT function and bile acid profiles. CIT, a functional read-out of BAT activity, was not associated with any bile acids (Table 3). However, BAT volume and SUVmean were inversely correlated with 12α-OH bile acids including DCA and particularly the ratio of 12α-OH/non-12α-OH bile acids (Table 3). Weaker associations were found between BAT volume and SUVmean with secondary as well as with the ratio of primary to secondary bile acids (Table 3).

4. Discussion

An increasing body of evidence suggests that bile acids are critical regulators of energy metabolism pathways including thermogenic processes in adipocytes (Broeders et al., 2015; Fang et al., 2015; Velazquez-Villegas et al., 2018; Watanabe et al., 2006; Zietak and Kozak, 2016). The emerging importance of BAT as an energy dissipating metabolic organ has spurred efforts to elucidate if BAT activation confers health benefits in humans and how energy metabolism may differ between individuals with and without active BAT (Becher et al., 2021; Chondronikola et al., 2016; Hanssen et al., 2015; Kulterer et al., 2020; Mihalopoulos et al., 2020; van der Lans et al., 2013; Yoneshiro et al., 2013). However, the role of physiologic bile acid concentrations has not been studied in this context. Here we investigated the relationship between endogenous circulating bile acids and the presence as well as the activity of BAT in lean and obese humans.

Bile acid concentrations did not differ significantly between individuals with and without active BAT irrespective of their weight status, except for higher total UDCA levels in lean BATpos compared to lean BAT neg participants (Table 2). UDCA is, compared to other bile acids,both aweak FXR and TGR5 agonist (Iguchi et al., 2011; Lefebvre et al., 2009). In mice, however,pharmacological doses were shown to improve systemic insulin sensitivity by alleviating endoplasmic reticulum stress (Ozcan et al., 2006). Hypothalamic and adipocyte endoplasmic reticulum stress are suggested to directly impair BAT function (Contreras et al., 2017; Yuliana et al., 2019). Still, compared to other primary or secondary bile acids species, fasting UDCA levels are very low. As UDCA is formed by intestinal dehydrolyzation of CDCA by gut microbiota, fasting UDCA levels might thus rather act as a marker of the gut microbiome than a direct effector in metabolic health (Ridlon et al., 2006). The gut microbiota has been shown to influence BAT activity through short chain fatty acids or the regulation of systemic inflammation (Li et al., 2019; Su rez-Zamorano et al., 2015´(a) ).

Fig. 1. Representative pairs of PET images at thermoneutrality (left) and after cold exposure (right) of A) a lean BATpos participant, B) a lean BATneg participant, C) an obese BATpos participant, and D) an obese BATneg participant. 18F-FDG uptake in BAT can only be detected after cold exposure in subjects A and C. It is primarily located in the supraclavicular and cervical region and shows a higher distribution in the lean subject where it extends in the deep axillar fat as well as into the thoracic paraspinal fat. Cold-induced thermogenesis (CIT) was however, significantly higher in BATpos Hepatitis B chronic participants (E). Participants with detectable BAT had a similar cold-induced increase in norepinephrine levels (Δnorepinephrine) compared to BAT negative subjects (F).

In addition, the 12“-OH/non-12“-OH bile acid ratio was almost twice as high in lean BAT neg compared to BAT pos participants with a trend for statistical significance whereas no such difference was seen in individuals with obesity (Table 2). In line with this finding, the imagingderived BAT markers, BAT volume and BAT SUVmean were inversely associated with the 12α-OH/non-12α-OH bile acid ratio and total 12αOH bile acids levels including DCA (Table 3). This is an interesting observation given that 12α-OH bile acids are generally increased in obesity, diabetes and non-alcoholic steatohepatitis and have been suggested to promote hepatic steatosis and fibrosis (Brufau et al., 2010; Haeusler et al, 2013, 2016; Lee et al., 2020; Puri et al., 2018; Wei et al., 2020; Xie et al., 2021). The enzyme responsible for the production of 12α-OH bile acids (mainly CA) versus non-12α-OH bile acids (mainly CDCA) is the sterol 12α hydroxylase CYP8B1. Hence, CYP8B1 enzymatic activity is reflected by 12α-OH/non-12α-OH bile acid ratio which has also been linked to obesity and insulin resistance in humans (Haeusler et al, 2013, 2016). Ablation of CYP8B1 impaired diet-induced weight gain, hepatic steatosis and insulin resistance in rodents and improved glucose metabolism due to higher intestinal fatty acid-induced GLP-1 secretion (Bertaggia et al., 2017; Kaur et al., 2015). CYP8B1 deletion also resulted in an increase in thermogenic markers and mitochondrial density in white adipose tissue accompanied by increased energy expenditure in diet-induced obese mice (Axling et al., 2020). Thus, the observed inverse association between BAT activity and 12α-OH/non-12α-OH bile acid ratio in our study could point towards a functional role of CYP8B1 in human BAT.However, we found no association between physiological concentrations of total bile acids or individual bile acid species with BAT activity, except for the low abundant bile acid UDCA. Most of the current knowledge about the interaction between bile acids and BAT stems from studies conducted in mice. Whereas accumulating evidence suggests a role of bile acid metabolism in murine BAT activity, this has not yet been proven in humans. Hence, additional translational studies are warranted to elucidate a potential contribution of endogenous bile acids towards human BAT physiology.

A limitation of this study is that we could only analyze fasting bile acid concentrations. A study on postprandial bile acid kinetics found that total bile acid concentrations after a mixed-meal exhibited a 3.3fold mean increase with a high inter-individual variation which remained constant for 6 h. A shift towards increased circulating conjugated bile acids was observed and the authors noted that the postprandial area under the curve did not correlate with fasting values (Fiamoncini et al., 2017). We can therefore not speculate if different postprandial bile acid responses may be associated with BAT function. While fasting CDCA levels were not correlated with BAT activity in our study, an interventional study in lean healthy women found that the acute administration of CDCA resulted in increased BAT activity as assessed by PET/CT imaging (Broeders et al., 2015). This was achieved by acutely raising CDCA levels to 10 μmol/L which is approximately eight-fold the median total bile acid concentration and more than 200-fold the median CDCA concentrations observed in our lean cohort (Broeders et al., 2015). Hence even if the differences in physiologic bile acid levels between BAT pos and BAT neg subjects appear rather modest, pharmacological bile acid concentrations may have a significant impact on BAT function in humans.In summary, this is the first study investigating the association between endogenous circulating bile acid concentrations and BAT activity in humans. While total bile acids and other major bile acid groups did not differ between participants with and without active BAT, the 12αOH/non-12α-OH ratio, a marker of CYP8B1 activity, was inversely associated with BAT activity and showed a tendency towards lower values in lean BAT pos compared to BAT neg. CYP8B1, previously linked to metabolic perturbation in rodents and humans, could serve as a potential new BAT target that could be investigated in more detail in future.