**Changes in Trimethylamine-N-oxide Levels in Obese Patients following Laparoscopic Roux-en-Y Gastric Bypass or Sleeve Gastrectomy in a Korean Obesity Surgical Treatment Study (KOBESS)**


**Abstract:** Trimethylamine N-oxide (TMAO), a gut microbe-dependent metabolite, has been implicated as a novel risk factor for cardiovascular events related to obesity and type 2 diabetes mellitus (T2DM). The aim of the study was to test the hypothesis if TMAO is associated with the reduction of cardiovascular disease in the Korean obese patients who underwent bariatric surgery. From a subgroup of a multicenter, nonrandomized, controlled trial, titled KOBESS, 38 obese patients, 18 with and 20 without T2DM, who underwent Roux-en-Y gastric bypass (RYGB) or sleeve gastrectomy (SG) were investigated. Bariatric surgery is indicated for Korean patients with a body mass index (BMI) <sup>≥</sup> 35 kg/m2 or for Korean patients with a BMI <sup>≥</sup> 30 kg/m2 who have comorbidities. Serum levels of TMAO and its precursors, betaine, carnitine, and choline were measured before and six months after bariatric surgery. The levels of TMAO and its precursors did not differ between obese patients with T2DM and non-T2DM at baseline. However, TMAO increased more than twofold in patients with T2DM after RYGB surgery, but not in patients without T2DM. Choline levels were decreased by half in all patients after RYGB. In patients with T2DM who underwent SG, TMAO, betaine, and carnitine levels did not change after the surgery. Furthermore, in obese patients who underwent bariatric surgery, increased TMAO levels were associated with both T2DM and RYGB, while reduced choline levels were associated with RYGB. These associations need to be further elucidated in follow-up studies to gain further insights into the relationship between TMAO levels and bariatric surgery outcomes.

**Keywords:** trimethylamine N-oxide; cardiovascular disease; obesity; bariatric surgery; diabetes mellitus

#### **1. Introduction**

Obesity is an ever-growing disease that is strongly associated with metabolic syndrome, characterized by insulin resistance, hyperglycemia, hyperlipidemia, and hypertension. Patients with obesity and diabetes mellitus (DM) are at an increased risk of

**Citation:** Lee, S.J.; Park, Y.S.; Kim, Y.-J.; Han, S.-U.; Hwang, G.-S.; Han, Y.; Heo, Y.; Ha, E.; Ha, T.K. Changes in Trimethylamine-N-oxide Levels in Obese Patients following Laparoscopic Roux-en-Y Gastric Bypass or Sleeve Gastrectomy in a Korean Obesity Surgical Treatment Study (KOBESS). *J. Clin. Med.* **2021**, *10*, 5091. https://doi.org/ 10.3390/jcm10215091

Academic Editor: David Benaiges Boix

Received: 13 September 2021 Accepted: 28 October 2021 Published: 29 October 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

cardiovascular morbidity and mortality. Several clinical trials have shown the significant effects of bariatric surgery, including weight loss, improvements in serum glucose control, and reduced risk of cardiovascular diseases [1,2].

Trimethylamine N-oxide (TMAO), a gut microbe-dependent metabolite, is a small organic compound derived from dietary choline, betaine, and L-carnitine through metabolic processes of gut microbiota and subsequently by hepatic flavin monooxygenases [3–5]. Evidence suggests that TMAO induces platelet hyperactivity and thrombosis, thereby increasing the atherosclerotic burden [6]. These findings are replicated in other clinical studies that showed an association between elevated TMAO levels and an increased risk of atherosclerosis and cardiovascular disease (CVD) [7,8]. Moreover, prospective cohort studies have shown that increased TMAO levels could predict an elevated risk of major adverse events, such as myocardial infarction, stroke, or death [8–10]. In addition, increased levels of TMAO are strongly associated with obesity and DM [11,12]. Recent observational studies have reported TMAO levels to be elevated after bariatric surgery [13,14].

As expected, bariatric surgery changes the composition of the gut microbiome owing to the characteristics of the surgical procedures involving the reconstruction of the small intestine [15]. However, TMAO levels were reported to be increased in Norwegians after bariatric surgery [13]. Given the well-established beneficial effects of bariatric surgery on attenuating the risks of CVDs, increased level of TMAO, a molecule that has been suggested as a risk factor for CVD, after bariatric surgery is conflicting and contradictory. To the best of our knowledge, as of today, the impact of bariatric surgery on TMAO change in Asians has not been reported.

Thus, in this prospective study, we investigated the levels of TMAO and its precursors to elucidate the association between TMAO and the risk of CVD after bariatric surgery. We subdivided patients in this study according to the presence or absence of DM and the types of bariatric surgery.

#### **2. Material and Methods**

#### *2.1. Patients and Study Design*

The current study is part of a clinical trial entitled Korean Obesity Surgical Treatment Study (KOBESS), registered at www.ClinicalTrials.gov, accessed on 10 September 2021 (NCT03100292) [16]. KOBESS is a prospective, multicenter, nonrandomized, controlled study of obese Korean patients who underwent primary sleeve gastrectomy (SG) or Rouxen-Y gastric bypass (RYGB). All patients were recruited between August 2016 and April 2019. Patients with a body mass index (BMI) ≥ 35 kg/m2 or a BMI 30.0–34.9 kg/m2 and obesity-related comorbidities, such as DM, hypertension, or hyperlipidemia, were considered eligible for KOBESS. The protocol was approved by the Institutional Review Board of each clinical center (approval number for the coordinating investigator: 000000 2016-06-015), and informed written consent was obtained from all participants. Patients who had complete results for laboratory tests conducted at baseline and six months after surgery were enrolled. Patients with serum samples of less than 50 μL, which was considered insufficient volume for analysis, were excluded from the study. Of 64 KOBESS patients, 38 obese patients, 18 with and 20 without type 2 DM (T2DM), were enrolled and investigated in the current study (Supplementary Figure S1). Seventeen patients (7 with T2DM and 10 without T2DM) underwent RYGB, and 21 patients (11 with T2DM and 10 without T2DM) underwent SG.

#### *2.2. Anthropometric and Laboratory Assessments*

Patients were assessed for anthropometry and blood chemistry at baseline two weeks before surgery. Follow-up examinations were performed six months after surgery. Anthropometric data and laboratory test results were recorded for each of the 38 patients six months after bariatric surgery. Height, weight, sex, systolic and diastolic blood pressure (mmHg), and BMI (kg/m2) were measured and recorded. Fasting blood samples were collected using a standard venipuncture and stored at –70 ◦C. Laboratory tests, including complete blood count, fasting plasma glucose, glycosylated hemoglobin (HbA1c), lipid profile (triglycerides, total cholesterol, high-density lipoprotein cholesterol [HDL-C], low-density lipoprotein cholesterol [LDL-C]), liver panel [aspartate aminotransferase (AST), alanine aminotransferase (ALT), γ-glutamyl transpeptidase (GTP), alkaline phosphatase (ALP)], renal panel [creatinine, blood urea nitrogen (BUN), uric acid (mg/dL)], ferritin, iron, vitamin B, and folate were performed. The diagnosis of T2DM at baseline was defined according to a previous diagnosis; HbA1c ≥ 6.5%, fasting serum glucose when fasting for more than 8 h ≥ 126 mg/dL, or serum glucose after 75 g oral glucose tolerance test ≥ 200 mg/dL.

#### *2.3. Metabolomic Analysis*

For targeted quantitative analysis, we performed ultra-high-performance liquid chromatography/triple quadrupole mass spectrometry (UPLC/TQ-MS) analysis. Prior to analysis, 20 μL of serum sample was extracted using 80 μL of methanol, and the aqueous supernatant was diluted with 20% acetonitrile (*v*/*v*) containing 5 ng/mL betaine-d11, an internal standard. UPLC/TQ-MS analysis was performed on an Agilent 1290 Infinity LC and an Agilent 6495 triple quadrupole MS system equipped with an Agilent Jet Stream electrospray ionization source (Agilent Technologies, USA). Chromatographic separation was carried out on an Acquity UPLC BEH amide column (2.1 mm × 50 mm, 1.7 μm; Waters) with a binary gradient system comprising 10 mM ammonium formate in water (solvent A) and acetonitrile (solvent B). The linear gradient elution was as follows: 0–1.0 min, 85% B; 1.0–2.5 min, 85–40% B; 2.5–3.0 min, 40% B; 3.0–3.1 min, 40–85% B; 3.1–5.1 min, 85% B. Quantification was performed in the multiple reaction monitoring mode using MS operation in positive ionization mode. Mass Hunter Workstation (Ver B.06.00, Agilent Technologies, USA) software was used for data acquisition and analysis. Metabolite analysis was performed on serum samples collected at baseline and six months after bariatric surgery from 37 patients. Data from one patient were excluded because of measurement failure.

#### *2.4. Statistical Analysis*

All data are expressed as the mean ± standard deviation (SD). A paired *t*-test or independent *t*-test was used to assess the difference in each variable between baseline and six months after surgery. A significance level of 0.05 was used. Statistical analyses were performed using commercial software packages (SPSS version 19; IBM, Chicago, IL, USA), and *p*-values less than 0.05 were considered to be statistically significant.

#### **3. Results**

#### *3.1. Clinical Characteristics*

As a subgroup of a prospective multicenter clinical trial, a total of 38 obese patients who underwent both bariatric surgery and laboratory tests conducted at baseline and six months after surgery were enrolled (Supplementary Figure S1). The baseline characteristics of the two groups divided by the prevalence of T2DM (18 patients with T2DM and 20 without T2DM) are presented in Table 1. At baseline, the mean age in the T2DM group was 8 years older (*p* = 0.034) and the HbA1c level was 2.2% higher than in the non-T2DM group (*p* < 0.001). In the T2DM group, BMI was slightly higher than that of non-T2DM (37.8 ± 5.9 kg/m<sup>2</sup> vs. 40.1 ± 6.4 kg/m2), but there was no significant difference. There were no differences in sex, type of surgery, systolic blood pressure (SBP), diastolic blood pressure (DBP), total cholesterol, and triglyceride levels between the two groups. In the T2DM group, 7 (38.9%) patients underwent RYGB, and 11 (61.1%) underwent SG. In the non-T2DM group, 10 (50.0%) patients underwent RYGB, and 10 (50.0%) underwent SG. All surgical procedures were performed without any significant postoperative complications, and all patients received recommendations regarding dietary habits and lifestyle modifications and micronutrient supplementation (vitamin D, vitamin B12, multivitamin, calcium) as required.


**Table 1.** Baseline characteristics of patients with type 2 diabetes mellitus (T2DM) and non-T2DM.

Data are presented as the mean ± standard deviation (SD) or number (n). T2DM, type 2 diabetes mellitus; BMI, body mass index; RYGB, Roux-en-Y gastric bypass; SG, sleeve gastrectomy, HbA1c, glycosylated hemoglobin; SBP, systolic blood pressure; DBP, diastolic blood pressure.

Table 2 shows BMI and changes in serum biochemical indices six months after surgery for two groups. After bariatric surgery, BMI decreased significantly by more than 10 kg/m<sup>2</sup> in both groups (both *p* < 0.001). In addition, both SBP and DBP decreased in the two groups, but the decrease was not significant. In patients with T2DM, HbA1c decreased significantly from 7.8% to 6.2% and mean value of serum glucose markedly decreased from 144 mg/dL to 105 mg/dL. Even in non-T2DM patients, mean value of serum glucose decreased from 102 mg/dL to 94 mg/dL (*p* = 0.003). AST, ALT and HDL-C were significantly improved in both groups (*p* < 0.05 for all). In addition, in non-T2DM patients, GTP, total cholesterol and triglyceride decreased significantly. Levels of BUN and uric acid of both groups remained unchanged. The level of ferritin decreased in both groups, and VitB12 decreased in T2DM after bariatric surgery.

#### *3.2. Changes in Metabolites*

Table 3 shows the serum levels of betaine, carnitine, choline, and TMAO at baseline and six months after bariatric surgery (RYGB and SG). Betaine, carnitine, and choline are the precursors of TMAO. Serum levels of TMAO and its precursors did not change at six months after surgery. (*p* > 0.05). When stratified by the presence and absence of T2DM, serum levels of TMAO, although not statistically significant (2.2 ± 1.6 vs. 4.9 ± 5.9 μM, *p* = 0.072), appeared increased in patients with T2DM (Figure 1 and Supplementary Table S1). The levels of betaine, carnitine, and choline remained unchanged after surgery in both the T2DM and non-T2DM groups.

We stratified the patients according to the type of surgery, RYGB and SG (Table 4). We found no difference in baseline levels of TMAO and its precursors. After surgery, TMAO, betaine, and carnitine levels did not significantly change in both the T2DM and non-T2DM groups. Intriguingly, we observed significantly decreased levels of choline (3.2 ± 2.1 μM at baseline, 1.7 ± 1.4 μM after six months, *p* = 0.004) after RYBG, but not after SG (*p* = 0.850).

Based on the increased level of TMAO in patients with T2DM, we stratified patients with T2DM and non-T2DM according to the type of bariatric surgery: T2DM with RYGB (n = 7), T2DM with SG (n = 11), non-T2DM with RYGB (n = 10), and non-T2DM with SG (n = 9). We also analyzed the levels of the metabolites (Table 5). Betaine and carnitine levels were not affected by the presence of T2DM or the type of surgery. Contrary to the levels of betaine and carnitine, those of choline and TMAO appeared to be influenced by the presence of T2DM or the type of surgery. In T2DM patients who underwent RYGB, the level of TMAO increased more than twofold (2.3 ± 1.5 to 5.5 ± 3.1 μM, *p* = 0.043). In contrast, in patients with non-T2DM, TMAO levels did not change in either the RYGB or SG groups. These results suggest a possible association of TMAO with T2DM and RYGB. We also observed that the level of choline was associated with RYGB. The level of choline

was attenuated in non-T2DM patients who underwent RYGB (3.2 ± 2.0 to 2.1 ± 1.5 μM, *p* = 0.036). The mean level of choline, although not significant, decreased from 3.3 ± 2.5 to 1.2 ± 1.2 μM (*p* = 0.091) in T2DM patients who underwent RYGB.

**Table 2.** Serum indices at baseline and six months after bariatric surgery in patients with T2DM and non-T2DM.


Data are presented as the mean ± standard deviation. T2DM, type 2 diabetes mellitus; BMI, body mass index; SBP, systolic blood pressure; DBP, diastolic blood pressure; HbA1c, glycosylated hemoglobin; AST, aspartate aminotransferase; ALT, alanine aminotransferase; GTP, γ-glutamyl transpeptidase; ALP, alkaline phosphatase; HDL-C, high-density lipoprotein cholesterol; LDL-C, low-density lipoprotein cholesterol.



Data are presented as the mean ± standard deviation. Δ, difference between baseline and six months after surgery; TMAO, trimethylamine-N-oxide.

**Figure 1.** Serum levels of metabolites at baseline (black bars) and six months after bariatric surgery (gray bars) in (**A**) Type 2 diabetes mellitus (T2DM) and (**B**) non-T2DM patients. Abbreviation: T2DM, type 2 diabetes mellitus; TMAO. Trimethylamine-N-oxide.

**Table 4.** Serum indices at baseline and after six months in subjects who underwent Roux-en-Y gastric bypass (RYGB) and sleeve gastrectomy (SG).


*\** Metabolite data from 20 patients. § Comparison of baseline values between the two groups. Data are presented as the mean ± standard deviation or number (n). RYGB, Roux-en-Y gastric bypass; SG, sleeve gastrectomy; T2DM, type 2 diabetes mellitus; BMI, body mass index; SBP, systolic blood pressure; DBP, diastolic blood pressure; HbA1c, glycosylated hemoglobin; TMAO, trimethylamine-N-oxide.


**Table 5.** Changes in metabolites according to diabetes and surgery type.

Data are presented as the mean ± standard deviation or number (n). RYGB, Roux-en-Y gastric bypass; SG, sleeve gastrectomy; T2DM, type 2 diabetes mellitus; TMAO, trimethylamine-N-oxide.

#### **4. Discussion**

In our study of obese patients who underwent bariatric surgery, we observed that the serum levels of TMAO increased substantially in T2DM patients who underwent RYGB, while other precursor metabolites, betaine and carnitine, were not altered. We also observed that the level of choline decreased significantly in all patients who underwent RYGB. Meanwhile, there was no significant change in metabolites including TAMO in patients with SG regardless of diabetes.

In the current study, we observed that TMAO levels were particularly associated with T2DM and RYGB. Previous studies have reported that TMAO levels after bariatric surgery, particularly RYGB, are increased [13,14], but not after vertical banded gastroplasty [17]. Trøseid et al. reported that plasma levels of TMAO more than doubled compared to the preoperative level in 27 obese patients one year after RYGB (4.4 μM vs. 10.5 μM, *p* < 0.001) [13]. Tremaroli et al. observed increased levels of TMAO only in patients who underwent RYGB but not in those that underwent vertical-banded gastroplasty, nine years after bariatric surgery [18]. In addition, they did not observe differences in other metabolites, carnitine and betaine, between the control and surgery groups.

The mechanism underlying the increase in TMAO levels after RYGB surgery remains to be elucidated. One possibility is that this may be due to adaptive shifts in the gut microbiota. Studies have indicated that bariatric surgery produces a specific shift in the gut microbiota that persists for up to a decade after surgery and is different from the shifts related to dietary intervention for weight loss [15,18,19]. Li et al. observed a major shift in the gut phyla towards higher concentrations of Proteobacteria (52-fold), lower concentrations of Firmicutes (4.5-fold), and Bacteroidetes (twofold) in a non-obese RYGB rat model compared with sham-operated rats [20]. Tremaroli et al. suggested that the increased level of TMAO after bypass surgery might be the consequence of less anaerobic metabolism in the intestine after bypass surgery, a hypothesis that is supported by the broad increase in facultative anaerobes in the intestine after RYGB [18]. Supporting this hypothesis, a recent study indicated that Proteobacteria is the most important bacteria as it encodes the *cutC* gene that codes for choline to trimethylamine (TMA)-mediating enzyme, choline TMA-lyase [21]. Thus, adaptive shifts in the gut microbiota of RYGB surgery may be responsible for the increased level of TMAO. In the current study, however, contrary to the more than doubled level in T2DM patients who underwent RYGB, we did not observe any change in TMAO levels in non-T2DM patients who underwent RYGB. This result clearly does not support the hypothesis that adaptive shifts in the gut microbiota are responsible for the increase in TMAO levels after RYGB surgery.

Another possible explanation for the increased TMAO level after RYGB surgery is that flavin-containing monooxygenase 3 (FMO3), the hepatic enzyme that produces TMAO, might be responsible for the increase in TMAO levels. Higher levels of TMAO are associated with T2DM [11,22]. A recent meta-analysis demonstrated a positive dose-dependent association between TMAO levels and an increased risk for T2DM [23]. The study indicated that the odds ratio for DM prevalence increased by 54% per 5 μM increment in serum TMAO levels. An interesting recent study showed that FMO3 is suppressed by insulin and increases in obese/insulin-resistant humans [24]. More intriguingly, a study reported

that RYGB corrects fasting hyperinsulinemia in patients with T2DM [25]. Based on the above-referenced studies, decreased production of insulin during fasting after RYGB could contribute to the increased activity of FMO3, leading to increased levels of TMAO.

In addition to the increased level of TMAO in T2DM patients who underwent RYGB, we also observed that reduced choline levels were associated with RYGB. The decreased choline level is likely due to the rearrangement of the anatomy of the intestine rather than the presence of T2DM, since choline levels were significantly reduced in all patients who underwent RYGB. In addition, we did not observe any change in choline levels in patients who underwent SG, in which the anatomy of the intestine remains intact. Adaptive shifts in gut microbiota towards higher concentrations of bacteria, such as Proteobacteria, that actively convert choline into TMA could explain the decreased choline level after RYGB.

This study is limited to the insufficient power of statistical significance. However, insufficient power of statistical significance does not exclude the possible effect of TMAO on cardiovascular disease. The aim of the study was to test the hypothesis if TMAO is associated with the reduction of cardiovascular disease in the Korean obese patients who underwent bariatric surgery. The prevalence of obesity in Korea steadily increased and the incidence of metabolic disease also concomitantly increased. Bariatric surgery has been reimbursed by National Health Insurance Service in Korea since 2019. The indication of bariatric surgery in Korea is different from that in Western countries. Bariatric surgery is indicated for Korean patients with BMI ≥ 35kg/m2 or for Korean patients with BMI ≥ 30kg/m2 who have comorbidities (T2DM, hypertension, dyslipidemia, obstructive sleep apnea etc.). Therefore, these data cannot be applied to the patients in Western countries. Asian people have similar characteristics and comorbidity. Based on this preliminary study, a large clinical study is underway to determine if TMAO can be a risk factor for cardiovascular disease (Clinical trials No. NCT04554758).

To the best of our knowledge, this is the first prospective study to reveal that the increase in TMAO levels after RYGB is associated with T2DM. The current study is a relatively short-term study of six months. Given the conflicting facts that high TMAO levels are implicated in CVD, while RYGB surgery reduces the risk of CVD, further prospective long-term studies are necessary to gain further insights into the relationship between TMAO levels and bariatric surgery outcomes.

**Supplementary Materials:** The following are available online at https://www.mdpi.com/article/ 10.3390/jcm10215091/s1, Figure S1: Flow chart of the KOBESS study. Table S1: The changes in metabolites at baseline and six months in subjects with T2DM and non-T2DM.

**Author Contributions:** Conceptualization: S.J.L., E.H. and T.K.H.; Investigation: Y.S.P., Y.-J.K., S.-U.H., Y.H. (Yeyoung Han), G.-S.H., Y.H. (Yoonseok Heo) and T.K.H.; Data Curation: Y.S.P., Y.-J.K., S.-U.H., Y.H. (Yeyoung Han), G.-S.H., Y.H. (Yoonseok Heo) and T.K.H.; Formal Analysis: S.J.L.; Validation: S.J.L.; Writing—Original Draft Preparation: S.J.L. and E.H.; Writing-Review and Editing: E.H. and T.K.H. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by the Korean Health Technology R&D Project [HC15C1322], Ministry of Health &Welfare, Republic of Korea (for YH), and the Basic Science Research Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Education [2018R1D1A1B07045737].

**Institutional Review Board Statement:** This study was approved by the Institutional Review Board of each participating hospital.

**Informed Consent Statement:** Informed consent was obtained from all subjects involved in the study.

**Data Availability Statement:** The data described in the manuscript, code book, and analytic code will be made available upon request pending application and approval.

**Conflicts of Interest:** The authors declare no conflict of interest.

### **References**


### *Review* **Impact of Bariatric Surgery on Adipose Tissue Biology**

**Óscar Osorio-Conles 1,2, Josep Vidal 1,2,3,\* and Ana de Hollanda 2,3,4,\***


**Abstract:** Bariatric surgery (BS) procedures are actually the most effective intervention to help subjects with severe obesity achieve significant and sustained weight loss. White adipose tissue (WAT) is increasingly recognized as the largest endocrine organ. Unhealthy WAT expansion through adipocyte hypertrophy has pleiotropic effects on adipocyte function and promotes obesity-associated metabolic complications. WAT dysfunction in obesity encompasses an altered adipokine secretome, unresolved inflammation, dysregulated autophagy, inappropriate extracellular matrix remodeling and insufficient angiogenic potential. In the last 10 years, accumulating evidence suggests that BS can improve the WAT function beyond reducing the fat depot sizes. The causal relationships between improved WAT function and the health benefits of BS merits further investigation. This review summarizes the current knowledge on the short-, medium- and long-term outcomes of BS on the WAT composition and function.

**Keywords:** bariatric surgery; adipose tissue; obesity; subcutaneous adipose tissue; visceral adipose tissue; cytokines; adipokines; adipocyte

#### **1. Introduction**

White adipose tissue (WAT) has evolved to become the largest endocrine organ. Its plasticity in response to excess or deficit of nutrients is crucial to maintain metabolic health. The remodeling and expansion capacity of adipose tissue implies the orchestrated response of adipocytes, immune cells, endothelial cells, fibroblasts, the extracellular matrix, and its secretome (cytokines, hormones, microRNAs) as mediators of crosstalk between the main organs involved in metabolic health. Dysfunctional expansion of adipose tissue emerges as a key determinant of obesity-related complications. WAT expansion beyond the subcutaneous adipose tissue (SAT) capacity leads to visceral adipose tissue (VAT) expansion and ectopic fat deposition in other tissues, which are major contributors to cardiovascular disease and metabolic risk above body mass index (BMI) [1]. The precise mechanism leading to impaired adipose tissue expandability are not fully understood. Bariatric surgery (BS) currently results in weight loss and better control of comorbid obesity conditions than medical therapy. BS is also associated with a reduced risk of mortality and of some types of cancer [2]. Currently, Roux-en-Y gastric bypass (RYGBP), sleeve gastrectomy (SG), and biliopancreatic diversion (BPD) are the main surgical techniques used worldwide [3].

This review aims to delve into the biology of adipose tissue in the context of obesity and its changes after BS.

**Citation:** Osorio-Conles, Ó.; Vidal, J.; de Hollanda, A. Impact of Bariatric Surgery on Adipose Tissue Biology. *J. Clin. Med.* **2021**, *10*, 5516. https:// doi.org/10.3390/jcm10235516

Academic Editor: David Benaiges Boix

Received: 27 September 2021 Accepted: 22 November 2021 Published: 25 November 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

#### **2. Obesity-Related White Adipose Tissue Dysfunction**

To identify what hypothetical benefits BS has on adipose tissue biology, we need to cite first the most consensed features of obesity-related WAT dysfunction: an altered adipokine secretome [4,5], unresolved inflammation [6,7], inappropriate extracellular matrix (ECM) remodeling, and insufficient angiogenic potential [8]. The causal order in this context is not completely known; however, hypertrophic adipocytes seem more prone to this scenario as they reach the diffusional limit of oxygen, resulting in persistent hypoxia and ultimately leading to unhealthy WAT tissue expansion [8]. Given its role in WAT remodeling, some authors add autophagy dysregulation to this context [9,10]. Among these features, inflammation-related phenomena, i.e., impaired adipokine and cytokine secretion, have been undoubtedly the most exhaustively studied and tracked parameters during the postsurgical follow-up period after BS.

Obese WAT is characterized by macrophage infiltration, a condition considered as both the cause and consequence of its immune response, which leads to chronic inflammation [11,12]. Obesity-related accumulation of adipose tissue macrophages (ATMs) has been clearly demonstrated in multiple studies [13–15] and the majority of such ATMs accumulate in omental rather than subcutaneous depots [14–16]. Thus, while a small number of macrophages, preferentially localized near blood vessels and dispersed among mature adipocytes are found in lean WAT, subjects with severe obesity show a higher abundance of infiltrating macrophages forming crown-like structures (CLS) around single adipocytes [13]. Such macrophages predominantly present the M1 pro-inflammatory phenotype and promote inflammation by releasing tumor necrosis factor alpha (TNF-α) and interleukin 6 (IL-6), thus contributing to insulin resistance. Alternatively activated, M2-like macrophages play a role in WAT expansion, thermoregulation, antigen presentation, and iron homeostasis [17]. In lean humans, the number of M2 ATMs predominates, secreting anti-inflammatory cytokines and utilizing oxidative metabolism to maintain WAT homeostasis. During obesity development, their proportion compared to M1 ATMs decreases and both populations may adopt a glycolytic metabolism [18]. Once the WAT healthy growth capacity is exceeded, the production of specific adipokines and cytokines by adipocytes and ATMs is compromised and can affect other organ systems.

Although the secretome of many other pro- [19–21] and anti-inflammatory [19,22,23] mediators have been found altered in the context of obesity, those described here are the most comprehensively evaluated, and their postsurgical modulation at different follow-up times is summarized below.

#### *2.1. Adipose Tissue-Derived Cytokines*

Although the systemic impact of WAT cytokine production in the context of obesity and diabetes has been recently called into question [24], it has been consistently shown that BS-induced weight loss progressively decreases the infiltration of macrophages and WAT secretion of pro-inflammatory molecules [25]. Such cytokines can be both released by WAT-resident immune cells or directly from adipocytes. Below, we describe the most well-studied in the context of obesity.

#### 2.1.1. Pro-Inflammatory

TNF-α is a 17 kDa pro-inflammatory cytokine that can be secreted from mature adipocytes but is predominantly produced within WAT stromovascular fraction, including preadipocytes, endothelial cells, smooth muscle cells, fibroblasts, leukocytes, and macrophages [26–28]. The latter is thought to be the major responsible for the elevated expression during obesity [29].

IL-1β is a major 17.5 kDa pro-inflammatory cytokine secreted mostly by macrophages [30], and its release from WAT nonfat cells is augmented during obesity [31]. With other inflammatory mediators, its production is greater in the visceral than in the subcutaneous depot [32]. Increased circulating IL-1β levels have been associated with the risk of developing type 2 diabetes [33], inasmuch as IL-1β contributes to inhibiting β-cell function and destroying β-cell mass [34,35] and impairs adipocyte insulin signaling [36].

Adipocytes, fibroblasts, and endothelial and immune cells secrete IL-6, which induces fever and liver production of the acute phase reactants, and also mediates chronic inflammatory responses [37]. Both adipocytes and macrophages are responsible for its overexpression in WAT during obesity [38,39]. Visceral rather than subcutaneous depot seems to be the main source of circulating IL-6 levels [40].

Different cell types express IL-8, such as monocytes, macrophages, fibroblasts, endothelial cells, and adipocytes [41,42]. IL-8 acts as a chemokine, attracting leukocytes [39]; as a pro-angiogenic factor [43]; and an amplifier of inflammation [44]. Secretion of IL-8 from WAT is increased [45], mainly in the visceral depot [37] during obesity, and is associated with insulin resistance [42].

IL-18 is another pro-inflammatory cytokine, produced by both hematopoietic cells and non-hematopoietic cells, which has been found to be increased in obesity [46,47] and associated with the metabolic syndrome independently of obesity and insulin resistance [48].

Chemoattractant chemokine ligand 2 (CCL2), also referred to as MCP-1, is a chemoattractant cytokine produced by, among others, myeloid cells [49] and adipocytes [50]. The latter enhances MCP-1 secretion during obesity [51], recruiting and activating macrophages through the MCP-1/IL-1β/CXCL12 signaling pathway [52]. Nevertheless, WAT expansion augmentation does not influence circulating MCP-1 levels [51].

Three isoforms of transforming growth factor-beta (TGF-β) have been identified in mammals, which are produced by all-white blood cells lineages and, to a lesser extent, by mature adipocytes [53]. Despite originally being thought to have overlapping functions, isoform-specific knockout mouse models revealed non-redundant phenotypes [54–56], TGF-1β being the predominant and most important isoform [57]. TGF-1β release from WAT is enhanced in obesity and in response to insulin and inhibitors of TNF-α and IL-β and correlates BMI and adiposity [53].

#### 2.1.2. Anti-Inflammatory

Secreted by Th2 T-cells, M2 macrophages, and adipocytes [39,58], IL-10 is an antiinflammatory cytokine that suppress macrophage activation [59], which has been inversely associated with BMI and body fat content [60].

Primarily secreted from mast cells and eosinophils, cytokines IL-4 and IL-13 are closely related, where the former stimulates the production of the latter [61], both sharing similar anti-inflammatory functions and receptor complexes [62]. The presence of these cytokines promotes alternative activation of macrophages into M2 cells and inhibits M1-like classical activation [63]. Both IL-4 [64] and IL-13 [65] serum concentrations are increased in obesity. Moreover, recent research showed a role for IL-4 in promoting adipocyte thermogenic capacity [66] and lipolysis [67] through hormone-sensitive lipase (HSL) modulation [68].

#### *2.2. Adipose Tissue-Derived Hormones*

#### 2.2.1. Pro-Inflammatory

Mostly produced by adipocytes, leptin is a highly conserved 167 kDa peptide. It is secreted proportionally to the amount of adiposity [69,70]. Leptin acts to reduce the food intake at the level of the hypothalamus and fat stores at the level of the adipocyte [71], as well as promoting pro-inflammatory cytokine production by immune cells [38,58].

Resistin, traditionally considered a WAT-specific secretory factor, is a 12.5 kDa hormone which acts as a modulator of body cholesterol trafficking, increasing low-density lipoprotein (LDL)-cholesterol and degrading liver LDL receptors, thus contributing to atherosclerosis pathogenesis. Within WAT, resistin promotes pro-inflammatory cytokine production through the resistin receptor and is found to be increased in obesity [72]. Nevertheless, mounting evidence reveals inconsistencies between resistin's role in rodents and humans, and its relationship with insulin resistance in humans is still controversial [73], with arguments existing both for [74–76] and against [77–79] this association.

Visfatin is another proinflammatory adipokine that plays a role in insulin sensitivity and whose production is increased in obesity and correlates with visceral adiposity [80].

#### 2.2.2. Anti-Inflammatory

Adiponectin is secreted from WAT as an oligomer of varying sizes in an inversely proportional manner to the degree of visceral adiposity [81]. Adiponectin plays an antiinflammatory role and promotes insulin sensitivity by increasing fatty acid oxidation, thus regulating lipoprotein metabolism and inhibiting hepatic glucose production. The adiponectin-leptin ratio is considered a biomarker of inflammation in WAT [82,83].

Predominantly expressed in the visceral depot [84], omentin -34 kDa- is an antiinflammatory adipokine with insulin-sensitizing effects whose levels are decreased in obesity and diabetes [85] and inversely correlated BMI [86]. The role of other molecules such as apelin, vaspin, and RBP4 in inflammation is less clear.

#### *2.3. Extracellular Matrix Remodeling and Fibrosis*

WAT is a highly dynamic organ, as it is responsible for storing and releasing energy in response to nutrient excess or shortage. As WAT expands (by adipocyte enlargement hypertrophy; and preadipocyte recruitment—hyperplasia), ECM is remodeled to accommodate healthy WAT expansion. Like in other organs, sustained WAT inflammation can trigger aberrant ECM deposition leading to WAT fibrosis. Profibrotic mediators such as TGF-β or connective tissue growth factor (CTGF) participate in this pathway [87]. When WAT becomes fibrotic, ECM stiffness impedes healthy remodeling, causing the tissue to be metabolically dysfunctional, displaying, e.g., adipocyte death, decreased lipolysis, and disrupted cell–cell interactions [87,88]. Thus, inflammation can disarrange the tight balance between ECM composition, extracellular metalloproteinases (MMPs), and their inhibitors (TIMPs) [89]. Data from three independent studies carried out by Karine Clément's group in BS subjects identified the degree of fibrosis in SAT as a predictor for poorer weight loss response after BS [90–92]. In this context, HIF1α has been proposed to link the hypoxic milieu to fibrosis and inflammation [93]. Certainly, accumulating evidence demands further research on the relationship between multiple ECM components and adipocyte function in the context of obesity [88,94–97] and diabetes [98–100], and possible associations with BS outcomes should be explored in depth.

#### *2.4. Basal and Stimulated Lipolysis*

Obesity is associated with an increase in basal lipolysis and impaired insulin ability to suppress the FFA outflow [101,102]. Antagonistically, plasma catecholamines are important stimulators of lipolysis via adrenergic receptors, particularly through beta-1 (ADRB1) and beta-3 adrenergic receptors (ADRB3) in human WAT [103], and catecholamine-stimulated lipolysis has also been found to be impaired in obesity [104]. Although the classical notion of 'catecholamine resistance' in obesity seems to receive little attention today, some authors recommend its revisitation [105].

More than two decades ago, Kaartinen et al. found a good correlation between the fat cell size and response to isoproterenol in isolated SAT adipocytes from subjects with obesity undergoing BS [106]. Interestingly, after substantial BS-induced weight loss, the lipolytic effect of isoproterenol stimulation of adrenergic receptors was higher than lean controls, despite no difference in receptor density between groups. Similar results have been reported after short-term nutrition interventions [107,108]. Fasting FFA circulating levels are other relevant measures of basal lipolysis, though they are not only dependent on WAT lipolysis but also on clearance by muscle and the liver.

#### *2.5. Angiogenesis*

Adipogenesis and angiogenesis are tightly related processes during 'healthy' WAT expansion since adipocyte differentiation trigger blood vessel formation [109,110], and in turn, WAT endothelial cells promote preadipocyte differentiation [111]. Vascular endothelial

growth factor (VEGF)-A, highly expressed in WAT, plays a capital role in angiogenesis, and its expression is raised during adipogenesis [112,113]. Besides the family of VEGF factors, the angiopoietin (ANGPT) family is also involved in vascular remodeling, maturation, and stabilization [114]. ANGPT-2, expressed in WAT endothelial cells, is considered a proangiogenic factor. Although its overexpression in mice improved the metabolic status [115], its role in the angiogenic process has not yet been elucidated. Platelet endothelial cell adhesion molecule-1 (PECAM-1/CD31) plays a role as an adhesion and signaling molecule with several roles in vascular and inflammatory processes, and its levels are increased in young men with severe obesity [116].

#### *2.6. Autophagy*

Autophagy, the cellular mechanism that promotes cell survival during nutrient depletion, may also be relevant under basal or nutrient excess conditions. During nutrient depletion, autophagy can provide essential components for energy production and biosynthesis. In circumstances of nutrient excess, autophagy plays an important role in eliminating unfolded proteins and toxic aggregates and facilitating endoplasmic reticulum homeostasis [117]. In this regard, liver autophagy has been the subject of extensive research [118], while WAT autophagy has been receiving growing attention in recent years and is now considered a key regulator of adipogenesis [9] with intricate implications in ECM remodeling and inflammation [10]. Although some authors have reported attenuated WAT autophagy in obesity [119], not all studies could confirm the sense in which obesity and/or metabolic disruption is related to WAT autophagy alterations [10], and most studies point to overactivation of WAT autophagy in obesity [120–122] and diabetes [123,124]. However, several considerations should be taken into account [10]. Since WAT autophagy can be regarded as a protective mechanism to avoid WAT maladaptation to nutritional stress, this may explain enhanced autophagy despite the increased inflammation in dysfunctional WAT. In addition, autophagy has different functions depending upon the cell type; thus, WAT cell heterogeneity should be taken into consideration. Finally, the varied technical approaches used to measure autophagy and the different depots analyzed could explain conflicting results among these studies. All of this together calls for much more research into the relationship between autophagy, obesity, and BS outcomes.

#### **3. Bariatric Surgery—Related Changes in White Adipose Tissue Biology**

Since there is no standardization and the definition of short-, mid- and long-term terminologies can vary among published reports [125], from here on, the current knowledge on this topic is summarized across five follow-up time points commonly used to report BS outcomes: ≤3 months (3 m), 6 m, 1 year (1 y)—all often considered to be short-term; ≥2 y <5 y—referred to as medium-term; and >5 y—frequently regarded as long-term post-surgery. All bariatric interventions considered in Table 1 consisted of SG, RYGB, or BPD.



#### *J. Clin. Med.* **2021** , *10*, 5516

**Table 1.** *Cont*.


#### *J. Clin. Med.* **2021** , *10*, 5516

**Table 1.**

*Cont*.

subcutaneous

 adipose tissue; VAT, visceral adipose tissue; MNC, mononuclear

 cells. \* after significant weight loss, collection time not reported.

#### *3.1. Short Term*

During the first year, coinciding with the rapid weight loss phase after BS, both SAT [126–130] and VAT [126–130,134,135] depots progressively reduce their size, and this is accompanied by a reduction in the area of subcutaneous [136–138] and visceral [138] adipocytes, respectively. A large adipocyte size was independently associated with a lower incidence of insulin resistance 6 months after RYGBP [136].

In the very short term after BS (≤3 months), Cancello et al. showed a significant decrease of total ATMs (HAM56+ cells) in SAT after RYGB [13]. These results were confirmed in another study from the same group, wherein CD40<sup>+</sup> cells (M1-like) were also found to be decreased and CD206<sup>+</sup> and CD163+ cells (M2-like) increased 3 months after RYGB [16]. This was accompanied by a reversion to the lean WAT profile, with CLS remission and ATMs again located near blood vessels [13].

Such early changes in the WAT cellular composition seem to alter the production of some cytokines, while others generate conflicting results between studies or do not seem to be modulated in the short term after BS. Thus, among the proinflammatory cytokines, MCP-1 was found to be concomitantly decreased during this period [13,150,165,166], while TGF-β or IL-1β seem to decrease only at 1 year after BS [142,147,156]. Reports on IL-6 production give conflicting results at 3 and 6 months but agree on a consistent decrease 1 year after surgery [139,142–144,146,150,152,156,158–160,168]. Similarly, reduced circulating levels of IL-18 were found 1 y post-BS [141] and after massive BS-induced weight loss, irrespective of the time elapsed since surgery [195,196].

In contrast, there is less consensus about TNF-α and IL-8, which have been found in different studies to both be increased [140,155,156,197], decreased [141–144,163,164], or unchanged [146–152,157] during this period. Similarly, BS-related outcomes on antiinflammatory cytokine production have yielded highly contradictory results between studies during the short-term follow-up period, as is the case with IL-4 [64,140,167], IL-10 [141,147,162,164,167,168], and IL-13 [144,167]. Interestingly, circulating omentin levels decrease as early as 24 h post-BS, before any fat mass loss, and maintained for1y[172].

Inasmuch as surgical weight loss predominantly reduces the body fat content, it is understandable that leptin levels were found to be consistently reduced following BS [140,141,144,148–150,156,158,162,163,168,170]. The leptin levels were also reduced after the novel endovascular bariatric procedure [198]. Nevertheless, systemic leptin levels are not directly related to the amount of body weight or fat loss, since early reductions of adiposity more dramatically reduce leptin levels than later periods of weight loss [162,170]. Again, there is a lack of consensus regarding the short-term effect of BS on resistin levels, given several studies have found it to be decreased [141,142,145,158,168] or unchanged [150,171]. In the case of visfatin, Lima et al. showed unaltered levels throughout the first year after BS [150].

Despite some conflicting reports in the very short term [146,147,149,150,156], circulating adiponectin levels appear to be consistently increased 1 year after BS [139,141,142, 150,152,156,158,168]. For its part, omentin was found to be increased as early as 24 h after BPD [172], and such a change is maintained for up to 1 year [172,173]. Apelin, a multifaceted biomarker [174], and vaspin, an insulin-sensitizing adipokine [175], are less investigated adipokines that showed a short-term reduction after BS. Regarding RBP-4, most studies reported a decrease in the circulating [177,178] or SAT mRNA [176] levels early after BS.

One study performed by Chabot and collaborators showed no resolution of SAT fibrosis 6 months after BS and suggested a transient association between SAT fibrosis and insulin resistance in humans with obesity [180]. Similarly, Katsogiannos et al. did not find significant differences in either the basal or stimulated lipolysis rate in SAT adipocytes at 1 and 6 months after BS but reported a decrease in isoproterenol-stimulated lipolysis at 6 versus 1 month after BS [181]. Conversely, insulin-suppressed free fatty acid (FFA) release has been found to be enhanced at 4 months [137], 7 months [183], and 1 year after RYGBP [101]. While some authors found increased FFA levels in the early months after BS [101,148,181], others reported no differences in this period [148,184–186].

García de la Torre et al. found higher VEGF-A levels in obese women undergoing BS compared to lean controls, and such levels significantly decreased 1 y after surgery, irrespective of the surgical procedure performed [187]. At this same follow-up period, another recent study showed, in addition to VEGF-A, lower levels of several angiogenesis biomarkers such as angiopoietin 2 (ANGPT-2), follistatin, hepatocyte growth factor (HGF), and the platelet endothelial cell adhesion molecule (PECAM-1) in patients who underwent SG or laparoscopic adjustable gastric banding (LAGB) [188].

Finally, Soussi et al. found attenuated WAT autophagy in obesity, and pre- versus post-BS comparisons indicated ameliorated adipocyte autophagic clearance in all patients within 3 to 12 months after the intervention, although at different degrees because of the large time-frame in post-surgery sample collection [119].

#### *3.2. Medium Term*

Two years after surgery, both visceral and subcutaneous depots maintain reduced sizes [126,131,132] as does the abdominal subcutaneous fat cell volume [132]. There is much less data available on circulating parameters beyond 1 y after BS. While IL-6 levels are consistently found reduced2y[145,153], 3 y [154], and 4 y after BS [169], reports on TNF-a continue to report conflicting data [145,154]. Although reports on IL-10 also seem quite inconsistent, some authors find that, after a temporary rise in the short term, its levels return to baseline values at2y[145], or even continue falling at4y[169].

BS outcomes on leptin and adiponectin levels seem much more solid. Circulating leptin has been repeatedly found to be reduced at 2 [145,153], 3 [154], and 4 y [169], and such reductions seem to be mainly attributed to early changes in WAT. Conversely, adiponectin levels continue to progressively rise in the medium term [145,153,154]. Only one report seems to oppose this view, a contradiction that could arise from the limited number of subjects and the variety of surgical techniques included in the study [169].

Beyond the short-term inconsistencies mentioned above, a single study showed that circulating resistin, after an early decline, recovered baseline levels 2 y after gastric bypass [145]. Finally, the RBP-4 levels were found still lowered 24 months after BS. Such changes were more pronounced in the subgroup without metabolic syndrome and correlated with reductions in the waist and visceral fat diameter [179].

Despite negative results reported by Katsogiannos et al. in the short term in a mixedsex cohort [181], Löfgren and collaborators found reduced basal and stimulated lipolysis rates at 2 y after BS exclusively in females [182], where differences in the basal rates remained only significant when lipolysis was expressed per cell surface area. In another study, the glycerol release in women who underwent RYGBP was found to be decreased postsurgically at 2 y and then increased dramatically to similar levels observed before surgery at 5 y [133]. Similarly, Manco et al. found reduced FFA levels in normoglucosetolerant obese women 3 years after BPD [154]. Finally, insulin-mediated suppression of FFA outflow has been found to be enhanced 3 years after RYGBP [102].

#### *3.3. Long Term*

Studies on long-term outcomes after BS are restricted almost exclusively to weight-loss parameters. Thus, a recent meta-analysis at 10 or more years after all bariatric procedures reported weighted means of 56.7% excess weight loss (EWL) after GB, 45.9%EWL after LAGB, 74.1%EWL after BPD and 58.3%EWL after SG [199]. The same study reported a 48.9%EWL and 22.2%TWL 20 y after LAGB. Very similar results were previously reported by the same group at 15 y after LAGB [200]. A lower incidence [201] and greater remission [202] of T2DM have also been reported in the long term; reductions in all-cause, cardiovascular, and T2DM mortality have also been found [203]. Nevertheless, the potential impact of body fat loss on these metabolic outcomes deserves further investigation since some variables

appear to be more weight-dependent, while others seem to be more adiposity-dependent from the medium-term [204].

Regarding the outcomes in WAT exclusively, we only have evidence from a single study carried out in women by Hoffstedt and collaborators at the long-term follow-up [133]. The authors reported decreased amounts of estimated SAT and VAT at 2 and 5 y and diminished SAT cell volume and increased adiponectin levels at 5 y post-BS. This study also found augmented basal glycerol release from isolated SAT adipocytes at 5 y, despite not finding changes in fasting plasma levels.

#### *3.4. Summary of BS Outcomes on WAT*

In summary, after bariatric surgery, SAT and VAT reduce their size progressively during the weight-loss phases. M1-like decrease and M2-like ATMs increase early after surgery; however, there are no data beyond the short term after BS.

Most pro-inflammatory cytokines begin to decrease early after surgery and continue to decline in the medium- and long-term. However, TGF-B or IL1B decrease only after one year of BS. There are controversial data on short-term TNFα and IL-8 levels after surgery as well as in anti-inflammatory cytokine levels in the short- and medium-term after surgery. Leptin levels drop rapidly soon after BS and then continue to decline during the follow-up; conversely, adiponectin and omentin levels rise after surgery. Resistin and visfatin dynamics show less agreement.

Regarding fibrosis, only one study reported no changes at short-term. Gender differences seem to affect basal and stimulated rates of lipolysis, which have been found decreased only in females at mid-term after BS. For its part, insulin inhibition of lipolysis was found consistently enhanced at medium- and long-term after surgery. Finally, autophagy increases and several angiogenesis-related molecules decrease at short-term, although there is a lack of reports on longer follow-up periods.

#### **4. Other Proposed Novel Mechanisms for WAT Improvement after BS**

Mitochondrial function and biogenesis have been found to be impaired in obesity, and T2DM and BS may attenuate mitochondrial damage in adipocytes. Thus, Varela-Rodríguez and colleagues reported an increased mitochondrial density and coverage, together with enhanced mitochondrial function at both the gene and protein level in abdominal SAT after RYGB- or SG-induced weight loss in a reduced cohort of patients [205]. In the very short term after RYGB, an induction of genes involved in mitochondrial biogenesis was found in SAT [206]. Similarly, increased SAT expression of transcripts related to the oxidative phosphorylation (OXPHOS) pathway has been shown 3 m [207] and 1 y after BS [137,208]. More recently, Van der Kolk et al. confirmed these findings in abdominal SAT in the short and medium term after RYGB, while opposite results were found after a low-calorie diet [209], suggesting a BS-specific effect. The authors also showed an induction of the tricarboxylic acid (TCA) cycle and fatty acid oxidation 2 y after surgery.

Beiging is the process through which WAT can change its phenotype to a brown-like adipose tissue known as beige/brite adipose tissue. Accumulating evidence from human and rodent studies in the last years suggest that RYGB predominantly enhances beige thermogenesis, while SG seems to promote brown adipose tissue thermogenesis [210]. A role for bile acids and the gut microbiome has been proposed in these mechanisms. Moreover, such thermogenic effects could depend on the fat content of the postoperative diet.

Beyond the fat mass loss and biological pathways discussed above, other mechanisms could contribute to the improvement of WAT dysfunction after BS. Thus, Frikke-Schmidt and colleagues recently summarized other potential pathways affecting the WAT function that can play a role after BS [105]. Bile acids, whose levels are persistently found to be increased after surgery [211], may improve adipocyte function acting upon the FXR receptor [212]. Another hypothesized mechanism implies gut microbiome composition. It has been found that the composition of the bacteria in the gut change after BS, and accumulating evidence shows how the bacterial composition can modulate the host immune cell

population. Thus, there is the possibility that changes in the gut microbiome initiated by BS can significantly impact the WAT metabolic function by modulating immune-resident cells in WAT [213,214]. Finally, as BS has demonstrated effects on the central regulation of metabolism, potential changes in neural enervation to the WAT may mediate physiological changes after BS, as results in mice suggest [215].

Lastly, recent studies have indicated significant alterations in the expression of several putative adipose tissue-derived microRNAs (miRNAs) after BS [216]. Thus, 1 y after BS, Sangiao-Alvarellos et al. reported raised circulating levels of miR-221 and miR-222 [190], and as early as 21 days post-RYGB, Atkin et al. found modulated levels of seven miRNAs (miR-7-5p, let-7f-5p, miR-15b-5p, let-7i-5p, miR-320c, miR-205-5p, and miR-335-5p) [189], mostly related with diabetes and insulin resistance pathways. Regarding SAT expression, Ortega and collaborators identified 12 modulated miRNAs 2 y after BS, some of them previously found raised in mature adipocytes after inflammatory stimulation (such as miR-146b, miR-376c, and again miR-221) [194], while another study from the same group found significant modifications in 15 mature miRNAs, mostly related to cell cycle, metabolism, and inflammation pathways, in women who underwent RYGB [191].

#### **5. Future Perspectives**

Better understanding of the fascinating biology of WAT following BS deserves further investigation. Evaluation of the modifications of WAT biology associated not only with time elapsed after surgery but also with the amount of weight loss is a priority. Studies should help us better understand the relationship between shrinkage in WAT volume and improved WAT function with the health benefits of BS. The health burden associated with a particular BMI in subjects with a weight-reduced state following BS appears to be eased as compared to that in subjects with comparable BMI that have not undergone BS. Thus, future studies should help disentangle how BS helps restore the crosstalk between the different components of the WAT as well as the crosstalk between WAT and other organs.

**Author Contributions:** Conceptualization, Ó.O.-C., J.V. and A.d.H.; writing—original-draft preparation, Ó.O.-C.; writing—review and editing, Ó.O.-C., J.V. and A.d.H.; supervision, J.V. and A.d.H. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**

