= One way Anova, Kruskal-Wallis test, \$ = Mann-Whitney test. \* Significant.


**Table 2.** Spearman's correlation analysis of adipose gene expression of *SRA1* with TLRs, their signaling mediators, and IRFs in the total study population. **Adipose Marker** *r p* **Value** TLR2 0.218 0.036 \*

**Table 2.** Spearman's correlation analysis of adipose gene expression of *SRA1* with TLRs, their sig-

**All participants (N = 108)**

*Cells* **2022**, *11*, x FOR PEER REVIEW 7 of 19

tives. Immunohistochemistry analysis showed that TLR4 (Figure 3A,B), IRAK1 (Figure 36

performed immunohistochemistry analysis on TLR4, IRAK1 and NF-kB as representa-

script levels, in adipose tissue, are positively correlated with TLRs and their signaling molecules.

Furthermore, to confirm the expression of protein levels in the adipose tissue, we

(r<sup>2</sup> = 0.379; *p* = 0.0085; Figure 5C).

4A,B) and NF-kB (Figure 5A,B) were significantly upregulated in individuals with obesity. Our protein data show that SRA1 positively correlated with TLR4 protein (r<sup>2</sup> = 0.623; *p* = 0.0002; Figure 3C). IRAK1 protein (r<sup>2</sup> = 0.0649; *p* < 0.0001; Figure 4C) and NF-kB protein

**Figure 3.** Adipose TLR4 protein expression in adipose tissue and its correlation with SRA1 protein expression. TLR4 protein expression was determined in adipose tissue samples from lean (normal weight; NW) and obese individuals (n = 8 for each group) using by immunohistochemistry (IHC) as described in Materials and Methods. IHC staining intensity was expressed as arbitrary units (AU) and the data (mean ± SEM) were compared between NW and obese populations using unpaired *t*test and *p* < 0.05 was considered significant. (**A**) The representative IHC images are shown for NW and obese individuals. (**B**) IHC staining intensity data (AU). (**C**) Correlation between TLR4 protein expression with SRA1. **Figure 3.** Adipose TLR4 protein expression in adipose tissue and its correlation with SRA1 protein expression. TLR4 protein expression was determined in adipose tissue samples from lean (normal weight; NW) and obese individuals (n = 8 for each group) using by immunohistochemistry (IHC) as described in Materials and Methods. IHC staining intensity was expressed as arbitrary units (AU) and the data (mean ± SEM) were compared between NW and obese populations using unpaired *t*-test and *p* < 0.05 was considered significant. (**A**) The representative IHC images are shown for NW and obese individuals. (**B**) IHC staining intensity data (AU). (**C**) Correlation between TLR4 protein expression with SRA1.

**Figure 4.** Adipose IRAK1 protein expression in adipose tissue and its correlation with SRA1 protein expression. IRAK1 protein expression was determined in adipose tissue samples from lean (normal weight; NW) and obese individuals (n = 8 for each group) using by immunohistochemistry (IHC) as described in Materials and Methods. IHC staining intensity was expressed as arbitrary units (AU) and the data (mean ± SEM) were compared between NW and obese populations using unpaired *t*test and *p* < 0.05 was considered significant. (**A**) The representative IHC images are shown for NW and obese individuals. (**B**) IHC staining intensity data (AU). (**C**) Correlation between IRAK1 protein expression with SRA1. **Figure 4.** Adipose IRAK1 protein expression in adipose tissue and its correlation with SRA1 protein expression. IRAK1 protein expression was determined in adipose tissue samples from lean (normal weight; NW) and obese individuals (n = 8 for each group) using by immunohistochemistry (IHC) as described in Materials and Methods. IHC staining intensity was expressed as arbitrary units (AU) and the data (mean ± SEM) were compared between NW and obese populations using unpaired *t*-test and *p* < 0.05 was considered significant. (**A**) The representative IHC images are shown for NW and obese individuals. (**B**) IHC staining intensity data (AU). (**C**) Correlation between IRAK1 protein expression with SRA1.

**Figure 5.** Adipose NF-kB protein expression in adipose tissue and its correlation with SRA1 protein expression. NF-kB protein expression was determined in adipose tissue samples from lean (normal weight; NW) and obese individuals (n= 8 for each group) using by immunohistochemistry (IHC) as described in Materials and Methods. IHC staining intensity was expressed as arbitrary units (AU) and the data (mean ± SEM) were compared between NW and obese populations using unpaired *t*test and *p* < 0.05 was considered significant. (**A**) The representative IHC images are shown for NW and obese individuals. (**B**) IHC staining intensity data (AU). (**C**) Correlation between NF-kB protein expression with SRA1. **Figure 5.** Adipose NF-kB protein expression in adipose tissue and its correlation with SRA1 protein expression. NF-kB protein expression was determined in adipose tissue samples from lean (normal weight; NW) and obese individuals (n= 8 for each group) using by immunohistochemistry (IHC) as described in Materials and Methods. IHC staining intensity was expressed as arbitrary units (AU) and the data (mean ± SEM) were compared between NW and obese populations using unpaired *t*-test and *p* < 0.05 was considered significant. (**A**) The representative IHC images are shown for NW and obese individuals. (**B**) IHC staining intensity data (AU). (**C**) Correlation between NF-kB protein expression with SRA1.

#### *3.2. Association of Adipose SRA1 Expression (mRNA) with TLRs, Their Signaling Mediators, 3.2. Association of Adipose SRA1 Expression (mRNA) with TLRs, Their Signaling Mediators, and IRFs in Individuals Classified as Those with NW, Overweight, and Obesity*

*and IRFs in Individuals Classified as Those with NW, Overweight, and Obesity* Regarding the association between adipose *SRA1* gene expression and meta-inflammatory markers studied (Table 3), we found that *SRA1* correlated with TLR2 (r = 0.317, *p* Regarding the association between adipose *SRA1* gene expression and meta-inflammatory markers studied (Table 3), we found that *SRA1* correlated with TLR2 (r = 0.317, *p* = 0.017), TLR3 (r = 0.531, *p* < 0.0001), TLR4 (r = 0.311, *p* = 0.022), TLR7 (r = 0.305, *p* = 0.015),

Heat map shown in Figure 6)

= 0.017), TLR3 (r = 0.531, *p* < 0.0001), TLR4 (r = 0.311, *p* = 0.022), TLR7 (r = 0.305, *p* = 0.015), TLR9 (r = 0.374, *p* = 0.002), MyD88 (r = 0.324, *p* = 0.010), IRAK1 (r = 0.255, *p* = 0.044), NF-κB

individuals with obesity. *SRA1* associated inversely with TLR9 expression (r = −0.489, *p* = 0.005) only in overweight group. In NW participants, *SRA1* was associated with MyD88 (r = 648, *p* = 0.043), IRF3 (r = 0.0857, *p* = 0.014), and IRF5 (r = 0.929, *p* = 0.003) expression.

TLR9 (r = 0.374, *p* = 0.002), MyD88 (r = 0.324, *p* = 0.010), IRAK1 (r = 0.255, *p* = 0.044), NF-κB (r = 0.454, *p* < 0.001), IRF3 (r = 0.290, *p* = 0.030) and IRF5 (r = 0.321, *p* = 0.010) expression in individuals with obesity. *SRA1* associated inversely with TLR9 expression (r = −0.489, *p* = 0.005) only in overweight group. In NW participants, *SRA1* was associated with MyD88 (r = 648, *p* = 0.043), IRF3 (r = 0.0857, *p* = 0.014), and IRF5 (r = 0.929, *p* = 0.003) expression. Heat map shown in Figure 6. *Cells* **2022**, *11*, x FOR PEER REVIEW 11 of 19 **Table 3.** Spearman´s correlation analysis of adipose gene expression of *SRA1* with TLRs, their signaling

**Table 3.** Spearman's correlation analysis of adipose gene expression of *SRA1* with TLRs, their signaling mediators, and IRFs in individuals differing by obesity levels. **Obesity Level Normal Weight Overweight Obese (N = 12) (N = 32) (N = 64)**


mediators, and IRFs in individuals differing by obesity levels.

*p* ≤ 0.05 \*, *p* < 0.01 \*\*, *p* < 0.001 \*\*\*, *p* < 0.0001 \*\*\*\*. *p* ≤ 0.05 \*, *p* ˂ 0.01 \*\*, *p* ˂ 0.001 \*\*\*, *p* ˂ 0.0001 \*\*\*\*.

**Figure 6.** Heat map of the correlation of SRA1 expression with TLRs and their signaling molecules in adipose tissues obtained from NW, Overweight and Obese. *p* ≤ 0.05 \*, *p* ˂ 0.01 \*\*, *p* ˂ 0.001 \*\*\*, *p* ˂ 0.0001 \*\*\*\*. **Figure 6.** Heat map of the correlation of SRA1 expression with TLRs and their signaling molecules in adipose tissues obtained from NW, Overweight and Obese. *p* ≤ 0.05 \*, *p* < 0.01 \*\*, *p* < 0.001 \*\*\*, *p* < 0.0001 \*\*\*\*.

## *3.3. Association of Adipose SRA1 Expression (mRNA) with TLRs, Their Signaling Mediators, and IRFs in Individuals with/without T2D*

In the analysis whether diabetic status affected associations between adipose *SRA1* expression and other markers of inflammation, we found that in people with T2D, *SRA1* expression was associated with TLR3 (r = 0.555, *p* < 0.0001), TLR4 (r = 0.302, *p* = 0.044), TLR7 (r = 0.292, *p* = 0.040), TLR9 (r = 0.398, *p* = 0.003), NF-κB (r = 0.381, *p* = 0.005, MyD88 (r = 0.311, *p* = 0.030), TRAF6 (r = 0.286, *p* = 0.038), and IRF5 (r = 0.288, *p* = 0.043) expression (Table 4; Figure 7). On the other hand, in participants without T2D, *SRA1* was inversely associated with TLR9 (r = −0.290, *p* = 0.034) and TRAF6 (r = −0.318, *p* = 0.019) expression (Table 4).

**Table 4.** Spearman's correlation analysis of adipose gene expression of *SRA1* with TLRs, their signaling mediators, and IRFs in individuals with/without T2D.


*p* ≤ 0.05 \*, *p* < 0.01 \*\*, *p* < 0.0001 \*\*\*\*.

*Mediators, and IRFs*

predictors of adipose SRA1 expression.

**Figure 7.** SRA1 gene expression is correlated with TLR3, TLR4, TLR7, TLR9, Myd88, TRAF6, NF-kB or IRF5 in the adipose tissues obtained from individuals with T2D (**A**–**G**). **Figure 7.** SRA1 gene expression is correlated with TLR3, TLR4, TLR7, TLR9, Myd88, TRAF6, NF-kB or IRF5 in the adipose tissues obtained from individuals with T2D (**A**–**G**).

*3.4. Analysis of the Independent Associations between SRA1 and TLRs, Their Signaling* 

In order to determine independent associations, the markers showing significant as-

and IRAK1 independently predicted the adipose SRA1 expression. In participants with T2D (N = 53), TLR3 and TLR9 were identified as the independent predictors of adipose *SRA1 expression,* but not in participants without T2D (N = 55). Regression analysis stratified by obesity status revealed the independent association of adipose SRA1 expression with MyD88 in NW (N = 12), and with TLR9 in overweight (N = 32) participants. In individuals with obesity (N = 64), TLR3, TLR7 and IRAK1 were detected as the independent

## *3.4. Analysis of the Independent Associations between SRA1 and TLRs, Their Signaling Mediators, and IRFs*

In order to determine independent associations, the markers showing significant associations with adipose *SRA1* expression were further assessed by multivariable stepwise linear regression analysis (Table 5). We found that in the total population (N = 108), TLR3 and IRAK1 independently predicted the adipose SRA1 expression. In participants with T2D (N = 53), TLR3 and TLR9 were identified as the independent predictors of adipose *SRA1* expression, but not in participants without T2D (N = 55). Regression analysis stratified by obesity status revealed the independent association of adipose SRA1 expression with MyD88 in NW (N = 12), and with TLR9 in overweight (N = 32) participants. In individuals with obesity (N = 64), TLR3, TLR7 and IRAK1 were detected as the independent predictors of adipose SRA1 expression.


**Table 5.** The multiple linear regression analysis.

Multiple linear regression analysis is performed to identify TLRs and their signaling molecules associated with SRA1 as predictor variables.

## **4. Discussion**

There is a relative lack of data highlighting the role and significance of long noncoding RNAs in metabolic inflammation and insulin resistance. In the present study, we show that the adipose tissue SRA1 protein expression is significantly increased in individuals with obesity, as compared to their normal weight counterparts, regardless of T2D status. While *SRA1* mRNA expression is significantly elevated, only in people with obesity people without T2D as compared to normal-weight counterparts. We further demonstrate that adipose *SRA1* gene expression is associated with the expression of TLRs, their signaling mediators and IRFs including TLR2, TLR3, TLR4, TLR7, TLR9, MyD88, IRAK1, NF-κB, IRF3, and IRF5. There is a general consensus that obesity is associated with low-grade chronic inflammation, marked by the abnormal production of pro- and anti-inflammatory cytokines and adipokines by the expanding white adipose tissue, which influences changes in adipose tissue expression of the innate immune receptors such as TLRs. The growing evidence now supports that TLRs play a role in inducing a systemic acute phase response characterized by chronic inflammation and oxidative stress [1–4]. Apart from PAMPs, FFAs also act as TLR agonists which consolidates emerging role of TLRs as metabolic sensors and as receptors of immune-metabolic significance [20]. Regarding nutrient sensing mechanism and ensuing inflammatory responses, TLR signaling is initiated in the TIR domain and the TLR-downstream signaling is propagated through the pathways dependent or independent of MyD88 adaptor protein [16,17]. MyD88-dependent signaling is induced after TLR-ligand engagement and TLR dimerization, leading to recruitment of MyD88 and IRAKs to the

TIR domain [54,55]. IRAKs are known as the death domain-containing serine/threonine kinases and adapter proteins which play key roles in signaling pathways of the IL-1 family receptors and TLRs. IRAK1 is activated after phosphorylation by IRAK4 and associates with TRAF6. The IRAK1/TRAF6 complex then engages with the TGFβ-activated kinase (TAK)-1 and TAB-1/2 adapter proteins to yield a macromolecular complex [56]. IRAK1, following its phosphorylation, disengages from the signaling complex and TAK1 activates the inhibitor of NF-κB kinase alpha/beta (IKKα/β), which leads to the phosphorylation, ubiquitination, and degradation of IκBα to allow nuclear translocation of the p65 NF-κB complexes [57]. In parallel, TAK1 phosphorylates MAPKs such as MKK4, MKK3, or MKK6 which, in turn, activate the ERK, JNK, p38 MAPK, NF-κB, and IRFs cascades. Activation of these signaling pathways leads to the increased expression of inflammatory markers including cytokines, chemokines, and adhesion molecules [16–19,58].

TLRs, especially, the TLR4 and TLR2 have emerged as key players in metabolic inflammation by nutrient sensing of LPS as well as sFFAs, both of which are abundantly found in individuals with obesity/T2D [35]. Increased expression of TLR2 and TLR4 has been found in patients with T2D, highlighting their associations with the pathogenesis of diabetes [24,28]. We previously showed that the elevated adipose gene expression of TLR4, TLR2, and MyD88 in people with obesity/T2D was associated with the IRAK1 gene expression [59]. Overall, several studies have reported the altered expression of TLRs (TLRs 1/2, TLRs 4-10) in people with obesity and/or T2D [25–27].

LncRNAs are emerging as key players in inflammatory and innate immune signaling cascades and are involved in pre- and post-transcriptional gene regulation [60,61]. Many lncRNAs are known to be differentially expressed in different tissues from individuals with obesity. SRA1 expression is more notable in the tissues or organs that have high energy demand such as heart, adipose tissue, skeletal muscle, and liver [36,62,63]. Furthermore, in these tissues/organs, SRA1 regulates several pathophysiological processes such as myocyte/adipocyte differentiation, hepatic steatosis, stem cell function, steroidogenesis, mammary gland development, and tumorigenesis [64]. Increased *SRA1* expression has been reported, in order, in adipocyte fractions from white adipose tissue, brown adipose tissue, and in preadipocytes [40,41]. *SRA1* is involved in regulation of adipocyte differentiation as well as in glucose homeostasis and insulin sensitivity of adipocytes [37,40,41].

Using a mouse model, Liu et al. demonstrated that *SRA1* knockout mice, compared to wild-type controls, had reduced body weight, an increased percentage of lean mass, and less fat percentage, less epididymal/subcutaneous white fat mass, and less liver mass [40]. Two different studies using this mouse model present congruent data showing better insulin sensitivity, less proclivity for obesity development under high-fat diet feeding, reduced hepatic steatosis, and improved glucose tolerance together with less inflammation and lower plasma TNFα levels as well as reduced expression of inflammatory genes (*Tnfα*, and *Ccl2*) in the white fat [40,41]. We also reported that the adipose *SRA1* expression was higher in non-diabetic persons with obesity compared with non-diabetic lean participants, and that the changes associated directly with BMI, PBF, fasting serum insulin, HOMA-IR, certain inflammatory markers but inversely with HbA1c level [42]. The results from previous studies collectively suggest that SRA1 may play a role in the adipose tissue function, pathobiology, as well as in inflammatory cascades implicated with insulin resistance, presenting it as a new target for therapeutic intervention of metabolic inflammation and insulin resistance.

Our data further show significant association of *SRA1* expression with that of TLR3, TLR4, TLR7, TLR9, NF-κB, IRF5, MyD88, and TRAF6 in the adipose tissue. Interestingly, TLR9 and TRAF6 associated differentially with the adipose *SRA1* expression depending on the diabetic status: TLR9/TRAF6 associated positively with *SRA1* expression in individuals with T2D but associated negatively with *SRA1* expression in individuals without T2D. We speculate that the negative association between these factors and adipose *SRA1* expression in the absence of diabetes factor imparts tenacity to the positive relationship of these factors

with *SRA1* expression in the cohort with diabetes. These observations need to be further validated in larger cohorts with more diverse populations.

Our data further show that TLR3/TLR9 expression was associated independently with the expression of *SRA1* in individuals with T2D, while TLR3, TLR7 and IRAK1 were identified as the independent predictors of adipose *SRA1* expression in individuals with obesity. Taken together, TLR3 remains the independent predictor of adipose *SRA1* expression in settings of both obesity and T2D. Overall, these data support the plausibility of adipose *SRA1* expression to be considered as a novel, surrogate biomarker of adipose inflammation in obesity/T2D. However, caution will be needed for interpreting results of this primarily correlative study and further investigations will be required to determine: (i) which immunometabolic insults such as adipocytokines, chemokines, as well as glucolipotoxic and oxidative stresses may lead to the induction or upregulation of adipose *SRA1* expression, (ii) how does the *SRA1* expression differ between adipocytes and stromal vascular fraction, especially in monocytes/macrophages, and (iii) how will *SRA1* gene silencing and overexpression in adipocytes and monocytes/macrophages affect the insulinstimulated glucose uptake as a measure of cellular function modulation.

## **5. Conclusions**

In conclusion, our data show that the AT *SRA1* expression correlates with TLRs-3,4,7,9, MyD88, NF-κB, and IRF5 expression in individuals with T2D and with TLR2, IRAK1, and IRF3 expression in individuals with obesity. AT-*SRA1* expression was independently predicted by TLR3/TLR7 and IRAK1 in those with obesity and by TLR3/TLR9 in individuals with T2D. An association between the increased *SRA1* expression in the adipose tissue and the markers of immune signaling derangement in this compartment suggest that SRA1 molecules may have significance as new surrogate biomarkers of metabolic inflammation.

**Supplementary Materials:** The following supporting information can be downloaded at: https:// www.mdpi.com/article/10.3390/cells11244007/s1, Figure S1: H&E staining of human adipose tissue samples; Table S1: Demographic and clinical characteristics of study population; Table S2: List of TaqMan gene expression assays.

**Author Contributions:** Conceptualization, S.K. and R.A.; data curation, S.K., R.T. and T.J.; formal analysis, H.A., S.S. (Sardar Sindhu), R.T., T.J., A.A.-S., S.S. (Steve Shenouda), H.A.K., F.A.-M., J.T. and R.A.; funding acquisition, F.A.-R. and R.A.; investigation, S.K., H.A., S.S. (Sardar Sindhu), H.A.K., J.T. and R.A.; methodology, S.K., T.J. and J.T.; resources, R.A.; writing—original draft, H.A. and S.S. (Sardar Sindhu); writing—review and editing, S.K., A.A.-S., S.S. (Steve Shenouda), F.A.-R., H.A.K., F.A.-M., J.T. and R.A. All authors have read and agreed to the published version of the manuscript.

**Funding:** This study was supported by Kuwait Foundation for Advancement of Sciences (KFAS) (Grant #: RA-2010-003; RA AM 2016-007).

**Institutional Review Board Statement:** The study was conducted according to the guidelines of the Declaration of Helsinki, reviewed and approved by the ethics committee of Dasman Diabetes Institute, Kuwait (No.: RA-2010-003).

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

**Data Availability Statement:** The data presented in this study are available on request from the corresponding author.

**Acknowledgments:** The authors also thank Fahad Al-Ghamlas for help with patient recruitment.

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

#### **References**

1. Calder, P.C.; Ahluwalia, N.; Brouns, F.; Buetler, T.; Clement, K.; Cunningham, K.; Esposito, K.; Jönsson, L.S.; Kolb, H.; Lansink, M.; et al. Dietary factors and low-grade inflammation in relation to overweight and obesity. *Br. J. Nutr.* **2011**, *106* (Suppl. S3), S5–S78. [CrossRef] [PubMed]


*Article*

## **Physical Exercise A**ff**ects Adipose Tissue Profile and Prevents Arterial Thrombosis in** *BDNF* **Val66Met Mice**

**Leonardo Sandrini 1,2 , Alessandro Ieraci <sup>2</sup> , Patrizia Amadio <sup>1</sup> , Marta Zarà 1 , Nico Mitro <sup>2</sup> , Francis S. Lee <sup>3</sup> , Elena Tremoli <sup>1</sup> and Silvia Stella Barbieri 1,\***


Received: 15 July 2019; Accepted: 10 August 2019; Published: 11 August 2019

**Abstract:** Adipose tissue accumulation is an independent and modifiable risk factor for cardiovascular disease (CVD). The recent CVD European Guidelines strongly recommend regular physical exercise (PE) as a management strategy for prevention and treatment of CVD associated with metabolic disorders and obesity. Although mutations as well as common genetic variants, including the *brain-derived neurotrophic factor (BDNF)* Val66Met polymorphism, are associated with increased body weight, eating and neuropsychiatric disorders, and myocardial infarction, the effect of this polymorphism on adipose tissue accumulation and regulation as well as its relation to obesity/thrombosis remains to be elucidated. Here, we showed that white adipose tissue (WAT) of humanized knock-in BDNFVal66Met (BDNFMet/Met) mice is characterized by an altered morphology and an enhanced inflammatory profile compared to wild-type BDNFVal/Val. Four weeks of voluntary PE restored the adipocyte size distribution, counteracted the inflammatory profile of adipose tissue, and prevented the prothrombotic phenotype displayed, per se, by BDNFMet/Met mice. C3H10T1/2 cells treated with the Pro-BDNFMet peptide well recapitulated the gene alterations observed in BDNFMet/Met WAT mice. In conclusion, these data indicate the strong impact of lifestyle, in particular of the beneficial effect of PE, on the management of arterial thrombosis and inflammation associated with obesity in relation to the specific BDNF Val66Met mutation.

**Keywords:** BDNF; Val66Met polymorphism; adipose tissue; adipogenesis; arterial thrombosis; physical exercise

## **1. Introduction**

Despite the huge growth in knowledge and advances in the prevention and treatment of cardiovascular disease (CVD), this pathology is still the leading cause of morbidity and mortality in the world and is predicted to reach 23.3 million by 2030 [1]. It is well known that an important modifiable risk factor for CVD mortality and morbidity is represented by excessive weight [2], and several follow-up studies demonstrated that a body mass index (BMI) >25 (>75th percentile based on percentile curves of BMI in the US reference population) is associated with a higher mortality rate [3,4]. Excessive body weight may influence CVD through its effect on risk factors such as hypertension, glucose intolerance, and dyslipidemia and may contribute through not already identified mechanisms [5]. In overweight and obese patients, adipose tissue accumulation is associated with a low-grade inflammatory profile and a higher secretion of cytokines and chemokines in the circulation compared to normal weight people [6]. The resulting subclinical inflammation is associated, among others, with hypercoagulability and increased thrombotic risk due to the enhanced platelet and leukocyte numbers and reactivities [7–9].

International guidelines, including 2016 European Guidelines on CVD prevention in clinical practice [10], strongly recommended regular physical exercise (PE) as management for the prevention and treatment of CVD, in particular when related to obesity and metabolic disorders. Regular PE reduces adipose-derived systemic inflammation, improves endothelial function, decreases platelet and leukocyte activation, and halts the progression of coronary stenosis in both obese and normal-weight individuals [8,11–14].

Starting from the discovery that several rare forms of obesity, called monogenic obesity, result from a mutation in single genes primarily located in the leptin–melanocortin pathway [15,16], recent evidence has identified additional selected genes associated with obesity, providing that the genetic background can play a pivotal role in causing or triggering susceptibility to the pathology when associated with environmental factors such as overeating and PE reduction [17,18]. Of note, *brain-derived neurotrophic factor* (*BDNF*) is included among these genes. Genome-wide association studies (GWAS) have shown a strong association between the *BDNF* locus and anorexia nervosa, bulimia nervosa [19], or obesity [20,21]. Indeed, it is well known that *BDNF* plays an important role in energy metabolism food intake and weight control [22,23].

In this context, the common human *BDNF* Val66Met variant through reduction of the activity-dependent secretion and signaling of mature BDNF, is associated not only to neuro-psychiatric disorders [24] and CVD [25] but also to eating disorders and obesity in humans [26–30]. Interestingly, a knock-in mouse carrying the human *BDNF* Val66Met polymorphism has a significantly higher body weight than wild-type littermates [31], associated with a proinflammatory and prothrombotic phenotype [25]. The frequency of the Met allele has a wide range of values: in Asians, Met allele frequency is nearly 50% heterozygous, while is about 20%–30% homozygous [32,33]. In the Caucasian population the Met allele is less frequent, with a frequency of 20%–30% heterozygous and only about 4% homozygous [33,34].

The aim of the present study was to investigate the relationship between the *BDNF* Val66Met polymorphism, obesity, and thrombosis, by analyzing the adipose tissue profile in BDNFMet/Met mice, and to evaluate the ability of PE to affect adipose tissue and reduce the prothrombotic phenotype in *BDNF* Val66Met knock-in mice. Finally, in vitro studies were performed to investigate the functional relevance of *BDNF* Val66Met polymorphism on adipogenesis.

## **2. Materials and Methods**

## *2.1. Mice*

All experiments were performed in adult (3–4 months old) wild-type BDNFVal/Val and BDNFMet/Met littermate mice, generated by Chen Z-Y et al. [31]. Only male mice were used to avoid the potential impact of hormones involved in the ovarian cycle in adipose tissue present in female mice. All experiments were approved by the National Ministry of Health-University of Milan Committee and of DGSA (12/2015 and 349/2015). Surgical procedures were performed in mice anesthetized with ketamine chlorhydrate (75 mg/kg; Intervet, Segrate, Milan, Italy) and medetomidine (1 mg/kg; Virbac, Milan, Italy). Mice were housed under standard conditions (20–22 ◦C, 12 h light/dark cycle, light on at 7 a.m.) with water and food ad libitum. All efforts were made to minimize animal distress and to reduce the numbers of animals used in this study.

## *2.2. Voluntary Physical Exercise (PE) Protocol*

Mice underwent voluntary PE protocol as previously described [35,36]. Briefly, BDNFVal/Val and BDNFMet/Met mice were weighed and allocated randomly into running (RUN) and control (sedentary, SED) groups in cages equipped with or without running wheels, respectively, for 4 weeks with free access to food and water. Four sedentary control mice were housed in a standard polypropylene mice cage. Four runner mice were housed in standard polypropylene rat cages, with free access to two running wheels (12 cm diameter and 5.5 cm width). The greater dimensions of cages for runner

mice were necessary for an adequate setup of running wheels. Running wheels were connected to an electronic counter, and the total distance ran was recorded daily. The average distance ran by a single mouse was calculated by dividing by 2 the total distance recorded per wheel (two running wheels × cage × four mice). The average distance ran by a single mouse, in our model, was comparable with the average distance reported by others [35–38].

## *2.3. Arterial Thrombosis Model*

Experimental arterial thrombosis was induced as previously described [39]. Briefly, the left carotid artery of anesthetized mice was freely dissected, and a flow probe (model 0.7 VB, Transonic System, Ithaca, NY, USA) connected to a transonic flowmeter (TransonicT106) was used to measure blood flow. When blood flow was constant for 7 min (at least 0.8 mL/s), a strip of filter paper (Whatman N◦1) soaked with FeCl<sup>3</sup> (7% solution; Sigma-Aldrich, Saint Louis, MO, USA) was applied for 3 min, and the flow was recorded for 30 min. An occlusion was considered to be total and stable when the flow was reduced by >90% from baseline until the 30 min observation time.

## *2.4. Whole Blood Counts*

Blood was collected by cardiac venipuncture into 3.8% sodium citrate (1:10 vol:vol) from anesthetized mice, and white blood cells and platelets were counted optically.

## *2.5. Platelet–Leukocyte Aggregate Analysis*

Platelet/leukocyte aggregates were analyzed as previously described [40]. Briefly, citrated blood was stimulated with 5 µM ADP (Sigma-Aldrich, Saint Louis, MO, USA) for 5 min, red blood cells were lysed by FACS Lysing solution, and samples were stained with anti-CD45 and anti-CD41 and analyzed by flow FACS "Novocyte 3000". A minimum of 5000 events were collected in the anti-CD45<sup>+</sup> gate.

## *2.6. Cell Culture, Treatment, and Di*ff*erentiation*

The C3H10T1/2 cell line has been used to evaluate the effect of different compounds on adipogenesis processes, as previously shown [41–43]. C3H10T1/2 cells (ThermoFisher Scientific, Paislay, Scotland, UK) were cultured in DMEM medium supplemented with 100 U/mL penicillin (Gibco, Rodano, Milan, Italy), 100 µg/mL streptomycin (Gibco, Rodano, Milan, Italy) and 10% FBS at 37 ◦C in 5% CO2/95% air atmosphere. Cells were plated in 6-well plates at a concentration of 3.5 <sup>×</sup> <sup>10</sup><sup>4</sup> cells/mL, and when they reached 80% confluence (day –2), they were treated with 10 ng/mL of ProBDNFVal or ProBDNFMet synthetic peptide (Alomone Labs, Jerusalem, Israel) [44–46] to simulate the kinetics of *BDNF* expression occurring in physiological conditions during adipogenesis [47]. Forty-eight hours later (day 0), cells were treated with adipogenic commitment mix (5 µg/mL insulin, 2 µg/mL dexamethasone, 0.5 mM IBMX, and 5 µM rosiglitazone; all from Cayman Chemical, Arcore, Italy). Insulin (5 µg/mL) was added at days 3, 5, and 7, and complete differentiation of the cells was reached at day 9.

## *2.7. Adipogenesis Evaluation by Flow Cytometry and Oil-Red-O*

After ProBDNFVal or ProBDNFMet treatment, C3H10T1/2 cells were analyzed during adipogenesis by flow cytometry, as previously described [48]. Briefly, at days 3, 5, and 9, cells were harvested in ice-cold PBS, analyzed by flow cytometry, and, according to granularity (SSC-H), were divided into four categories that correlated with the increased level of cell lipid accumulation after adipogenic commitment. In particular, noninduced cells were detected in the R1 gate, while cells with increasing granularity were identified in the regions from R2 to R4.

Oil-Red-O staining was performed as already described [49] on day 9. Lipid content was quantified as absorbance at a wavelength of 518 nm using a Tecan Infinite M1000 plate reader spectrophotometer (TECAN, Männedorf, Switzerland).

## *2.8. Quantitative Real-Time Polymerase Chain Reaction (RT-qPCR)*

Total RNA was isolated from mouse adipose tissue or C3H10T1/2 cells with TRIzol Reagent (Sigma-Aldrich, Saint Louis, MO, USA) and a Direct-zol RNA extraction kit (Zymo Research, Irvine, CA, USA) according to the manufacturer's instructions. One microgram of RNA was reverse-transcribed using an iScript Advanced cDNA Synthesis Kit (Bio-Rad Laboratories, Segrate, Milan, Italy).

Samples of cDNA were incubated in 15 µL Luna® Universal qPCR mix containing the specific primers and fluorescent dye SYBR Green (New England Biolabs, Pero, Milan, Italy). RT-qPCR was carried out in triplicate for each sample on the CFX Connect real-time System (Bio-Rad Laboratories, Segrate, Milan, Italy) as previously described [39]. Gene expression was analyzed using parameters available in CFX Manager Software 3.1 (Bio-Rad Laboratories, Segrate, Milan, Italy). qPCR was then carried out using the primer sequences shown in Table S1. In particular, the expression of genes relevant in adipogenesis, inflammation, and the *BDNF* pathway were assessed (*Ppar*γ*, C*/*ebp-*α *and C*/*ebp-*β*, Adipoq, Fabp4, Adra2a, Il-6, Mcp-1, Tnf-*α*, Tgf-*β*, Pai-1, Tf, CD163, CD80, Sorl1, Sirt1, Bdnf, TrkB* full and *TrkB-T1*).

## *2.9. Adipose Tissue Histology and Quantification of Adipocyte Size and Number*

Immunocytochemistry and the analysis of adipocytes were performed in inguinal (ingWAT) and epididymal (epiWAT) white adipose tissue. Tissues were fixed overnight in 4% formalin, embedded in paraffin, cut at 5 µm, and mounted on polarized slides [50]. Five sections at three different levels along the whole length of epiWAT for each animal were analyzed. In particular, the mean values for each group were obtained from a total of 90 sections (5 sections × 3 points × 6 animals/group). The tissue contiguous to the epididymis were excluded from the analyses since its structure is different from that of general adipose tissue [51].

The number and size of adipocytes were evaluated in hematoxylin and eosin stained sections by counting five 5× microscopic fields for each tissue sample using the ImageJ-Macro Adipocytes Tool. Images were taken with a Zeiss Axioskop (Carl Zeiss, Milan, Italy) equipped with an intensified charge-coupled device (CCD) camera system (Photometrics, Tucson, AZ, USA).

## *2.10. Statistical Analysis*

Statistical analyses were performed with GraphPad Prism 7.0 and SAS versus 9.4 software (SASA Institute, Cary, NC, USA). Data were analyzed by Student's t-test, two-way or three-way ANOVA with or without repeated measures for main effects of genotype and treatment or time and stimuli, as reported in every graph, followed by a Bonferroni post hoc analysis as appropriate. When data distribution was not normal, the variables were included in the analyses after logarithmic transformation. Values of *p* < 0.05 were considered statistically significant. Data are expressed as mean ± SEM.

## **3. Results**

## *3.1. Characterization of the White Adipose Tissue Depots in BDNFMet*/*Met Mice*

As previously shown, BDNFMet/Met mice have a significantly greater body weight compared to wild-type BDNFVal/Val littermates (Figure 1A). In addition, we observed that the percentage of both inguinal white adipose tissue (ingWAT) and epididymal white adipose tissue (epiWAT) on total body weight were significantly enhanced in BDNFMet/Met mice compared to BDNFVal/Val (Figure 1B,C).

*Cells* **<sup>2019</sup>**, *<sup>8</sup>*, 875 *Cells* **2019**, *8*, x 5 of 17

**Figure 1.** Characterization of white adipose tissue depots in BDNFVal/Val and BDNFMet/Met mice. (**A**) Body weight, percentage of (**B**) inguinal (ingWAT) and (**C**) epidydimal (epiWAT) white adipose tissue on total mouse body weight. (i) Representative hematoxylin and eosin (H&E) staining images and (ii) analysis of the frequency distribution of adipocyte sizes in (**D**) ingWAT and (**E**) epiWAT. Size bar: 100 µm. Black arrow: large adipocytes, green arrow: medium adipocytes, and red arrow: small adipocytes. Data are expressed as mean ± SEM. *n* = 6 mice/group. (**A**–**C**) Student's t-test and (**D**,**E**) **Figure 1.** Characterization of white adipose tissue depots in BDNFVal/Val and BDNFMet/Met mice. (**A**) Body weight, percentage of (**B**) inguinal (ingWAT) and (**C**) epidydimal (epiWAT) white adipose tissue on total mouse body weight. (i) Representative hematoxylin and eosin (H&E) staining images and (ii) analysis of the frequency distribution of adipocyte sizes in (**D**) ingWAT and (**E**) epiWAT. Size bar: 100 µm. Black arrow: large adipocytes, green arrow: medium adipocytes, and red arrow: small adipocytes. Data are expressed as mean ± SEM. *n* = 6 mice/group. (**A**–**C**) Student's t-test and (**D**,**E**) two-way ANOVA followed by Bonferroni post hoc analysis. \* *p* < 0.05, \*\* *p* < 0.01.

The histological examination of adipose depots revealed no difference in the frequency distribution of adipocyte sizes in ingWAT, while the BDNFMet/Met mice showed enrichment in small-The histological examination of adipose depots revealed no difference in the frequency distribution of adipocyte sizes in ingWAT, while the BDNFMet/Met mice showed enrichment in small-size and a reduction in middle-size adipocytes in the epiWAT when compared to BDNFVal/Val (Figure 1D,E).

two-way ANOVA followed by Bonferroni post hoc analysis. \* *p* < 0.05, \*\* *p* < 0.01.

size and a reduction in middle-size adipocytes in the epiWAT when compared to BDNFVal/Val (Figure 1D,E). Then, the molecular signatures underlying the distinct morphological features of the epiWAT were investigated. Mutant mice had significantly lower levels of *Pparγ, C/ebp-α* and *C/ebp-β* genes involved in the adipogenic program, as well as *Adipoq*, but a similar expression of *Fabp4* compared to BDNFVal/Val mice (Figure 2A). Interestingly, the *BDNF* Val66Met polymorphism affected also the Then, the molecular signatures underlying the distinct morphological features of the epiWAT were investigated. Mutant mice had significantly lower levels of *Ppar*γ*, C*/*ebp-*α and *C*/*ebp-*β genes involved in the adipogenic program, as well as *Adipoq*, but a similar expression of *Fabp4* compared to BDNFVal/Val mice (Figure 2A). Interestingly, the *BDNF* Val66Met polymorphism affected also the expression of *Adra2a*, *Sirt1*, and *Sorl1*, genes involved in both adipose tissue energy balance and adipocyte morphology (Figure 2A–C).

expression of *Adra2a*, *Sirt1*, and *Sorl1*, genes involved in both adipose tissue energy balance and adipocyte morphology (Figure 2A–C). In addition, a significant increase in the expression of Il-6, Tnf-α, Tgf-β, Mcp-1, and Pai-1 in BDNFMet/Met mice compared to BDNFVal/Val was found, whereas similar levels of TF between the two groups were found (Figure 2B). The enhanced inflammatory profile of BDNFMet/Met epiWAT was In addition, a significant increase in the expression of Il-6, Tnf-α, Tgf-β, Mcp-1, and Pai-1 in BDNFMet/Met mice compared to BDNFVal/Val was found, whereas similar levels of TF between the two groups were found (Figure 2B). The enhanced inflammatory profile of BDNFMet/Met epiWAT was associated with a greater expression of CD80, an M1 inflammatory macrophage marker, and with a reduction of CD163, an alternatively activated M2 macrophage marker (Figure 2B).

associated with a greater expression of CD80, an M1 inflammatory macrophage marker, and with a reduction of CD163, an alternatively activated M2 macrophage marker (Figure 2B). Finally, BDNFMet/Met mice had a higher *BDNF* mRNA level in epiWAT, whereas no differences in the expression of both *TrkB*-full length and the truncated isoform *TrkB-T1* were found (Figure 2C).

Finally, BDNFMet/Met mice had a higher *BDNF* mRNA level in epiWAT, whereas no differences in

the expression of both *TrkB*-full length and the truncated isoform *TrkB-T1* were found (Figure 2C).

*Cells* **2019**, *8*, x 6 of 17

**Figure 2.** Gene expression profile of epidydimal white adipose tissue (epiWAT) in BDNFVal/Val and BDNFMet/Met mice. mRNA levels of genes related to (**A**) adipogenesis, (**B**) inflammation, and (**C**) BDNF/TrkB pathway in epidydimal white adipose tissue (epiWAT) of BDNFVal/Val and BDNFMet/Met mice. Data are expressed as mean ± SEM. *n* = 6 mice/group. Student's t-test. \* *p* < 0.05, \*\* *p* < 0.01, and \*\*\* *p* < 0.005. **Figure 2.** Gene expression profile of epidydimal white adipose tissue (epiWAT) in BDNFVal/Val and BDNFMet/Met mice. mRNA levels of genes related to (**A**) adipogenesis, (**B**) inflammation, and (**C**) BDNF/TrkB pathway in epidydimal white adipose tissue (epiWAT) of BDNFVal/Val and BDNFMet/Met mice. Data are expressed as mean ± SEM. *n* = 6 mice/group. Student's t-test. \* *p* < 0.05, \*\* *p* < 0.01, and \*\*\* *p* < 0.005.

## *3.2. Evaluation of the Role of Mutant BDNF Val66Met Protein on Adipogenesis*

(day 9) (Figure 3A and Figure S3).

*3.2. Evaluation of the Role of Mutant BDNF Val66Met Protein on Adipogenesis*  Next, in vitro studies were performed to investigate the role of the BDNF Val66Met protein on adipogenesis. Pre-confluent C3H10Ts1/2 murine mesenchymal stem cells were exposed to ProBDNFVal or to ProBDNFMet synthetic peptides before inducing the adipocyte differentiation Next, in vitro studies were performed to investigate the role of the BDNF Val66Met protein on adipogenesis. Pre-confluent C3H10Ts1/2 murine mesenchymal stem cells were exposed to ProBDNFVal or to ProBDNFMet synthetic peptides before inducing the adipocyte differentiation program. Synthetic peptide treatment did not affect cell number and morphology (Figure S1).

program. Synthetic peptide treatment did not affect cell number and morphology (Figure S1). Notably, gene expression analysis at late (day 9) stages of differentiation showed that pretreatment with the peptide carrying the Met mutation determined a significant down-regulation of adipogenic genes, including *Pparγ*, *C/ebpα* and *C/ebpβ* mRNA levels (Figure 3A). In addition, ProBDNFMet treatment decreased the percentage of cells with low granularity (noninduced; R1) and increased those with high granularity (R4) both at 3 and 9 d post-induction (Figure 3B and Figure S2). However, at day 9, as provided by the oil-red-O staining, a similar accumulation of lipid droplets Notably, gene expression analysis at late (day 9) stages of differentiation showed that pretreatment with the peptide carrying the Met mutation determined a significant down-regulation of adipogenic genes, including *Ppar*γ, *C*/*ebp*α and *C*/*ebp*β mRNA levels (Figure 3A). In addition, ProBDNFMet treatment decreased the percentage of cells with low granularity (noninduced; R1) and increased those with high granularity (R4) both at 3 and 9 d post-induction (Figure 3B and Figure S2). However, at day 9, as provided by the oil-red-O staining, a similar accumulation of lipid droplets was detected in both samples (Figure 3C).

was detected in both samples (Figure 3C). In this experimental condition, among the genes that were previously modulated in epiWAT of BDNFMet/Met mice, only *Sorl1* was enhanced by ProBDNFMet treatment at late stages of differentiation In this experimental condition, among the genes that were previously modulated in epiWAT of BDNFMet/Met mice, only *Sorl1* was enhanced by ProBDNFMet treatment at late stages of differentiation (day 9) (Figure 3A and Figure S3).

*Cells* **2019**, *8*, 875 *Cells* **2019**, *8*, x 7 of 17

**Figure 3.** Effect of proBDNFMet on adipogenic differentiation of C3H10T1/2 cells. (**A**) mRNA levels of (i) *Pparγ*, (ii) *C/ebp-α*, (iii) *C/ebp-β,* and (iv) *Sorl1*. (**B**) Percentage of different cell populations based on their granularity profile analyzed by flow cytometry (R1: noninduced, R2-R3: growing granularity, and R4: high granularity) at day 3 (D3), day 5 (D5), and day 9 (D9) of differentiation, and (**C**) Oil-Red-O staining absorbance measurement in C3H10T1/2 cells. Data are expressed as mean ± SEM. *n* = 5 independent experiments/group. (**A**) Two-way ANOVA followed by Bonferroni post hoc analysis. **Figure 3.** Effect of proBDNFMet on adipogenic differentiation of C3H10T1/2 cells. (**A**) mRNA levels of (i) *Ppar*γ, (ii) *C*/*ebp-*α, (iii) *C*/*ebp-*β*,* and (iv) *Sorl1*. (**B**) Percentage of different cell populations based on their granularity profile analyzed by flow cytometry (R1: noninduced, R2-R3: growing granularity, and R4: high granularity) at day 3 (D3), day 5 (D5), and day 9 (D9) of differentiation, and (**C**) Oil-Red-O staining absorbance measurement in C3H10T1/2 cells. Data are expressed as mean ± SEM. *n* = 5 independent experiments/group. (**A**) Two-way ANOVA followed by Bonferroni post hoc analysis. (**B**,**C**) Student's t-test. \* *p* < 0.05, \*\* *p* < 0.01, \*\*\* *p* < 0.005, and \*\*\*\* *p* < 0.001.

## *3.3. E*ff*ect of Physical Exercise (PE) on Adipose Tissue Phenotype of BDNF Val66Met Mice*

(**B**,**C**) Student's t-test. \* *p* < 0.05, \*\* *p* < 0.01, \*\*\* *p* < 0.005, and \*\*\*\* *p* < 0.001.

BDNFMet/Met mice showed an opposite trend, even if less marked (Figure 4B).

*3.3. Effect of Physical Exercise (PE) on Adipose Tissue Phenotype of BDNF Val66Met Mice*  According to international cardiovascular guidelines [10] that recommend regular PE as management for the prevention and treatment of CVD, we evaluated the potential beneficial effect of According to international cardiovascular guidelines [10] that recommend regular PE as management for the prevention and treatment of CVD, we evaluated the potential beneficial effect of PE on adipose tissue and on prothrombotic phenotypes in *BDNF* Val66Met knock-in mice.

PE on adipose tissue and on prothrombotic phenotypes in *BDNF* Val66Met knock-in mice. BDNFVal/Val and BDNFMet/Met mice underwent 4 weeks of voluntary PE in cages equipped with a running wheel. As previously reported [35], no difference in the daily running distance was found between BDNFVal/Val and BDNFMet/Met mice (BDNFVal/Val: 6.676 ± 0.720 Km/d and BDNFMet/Met 6.657 ± 0.602 Km/d; *p* = 0.9837) in our experimental setting. In addition, we showed that PE did not affect the percentage of ingWAT and epiWAT on the total body weight in both BDNFVal/Val and BDNFMet/Met mice, compared to sedentary mice, whereas the morphology of adipose depots was modified as BDNFVal/Val and BDNFMet/Met mice underwent 4 weeks of voluntary PE in cages equipped with a running wheel. As previously reported [35], no difference in the daily running distance was foundbetween BDNFVal/Val and BDNFMet/Met mice (BDNFVal/Val: 6.676 <sup>±</sup> 0.720 Km/d and BDNFMet/Met 6.657 <sup>±</sup> 0.602 Km/d; *<sup>p</sup>* <sup>=</sup> 0.9837) in our experimental setting. In addition, we showed that PE did not affectthe percentage of ingWAT and epiWAT on the total body weight in both BDNFVal/Val and BDNFMet/Met mice, compared to sedentary mice, whereas the morphology of adipose depots was modified as provided by histological analyses (Figure 4).

provided by histological analyses (Figure 4). PE induced a change in the profile of the frequency distribution of adipocyte sizes in the ingWAT of both genotypes; however, this effect was more evident in BDNFVal/Val than in BDNFMet/Met mice PE induced a change in the profile of the frequency distribution of adipocyte sizes in the ingWAT of both genotypes; however, this effect was more evident in BDNFVal/Val than in BDNFMet/Met mice (Figure 4A).

(Figure 4A). Interestingly, in the epiWAT, BDNFVal/Val running mice displayed a significant enrichment in small-size adipocytes and a reduction in medium-size ones compared to sedentary mice, whereas Interestingly, in the epiWAT, BDNFVal/Val running mice displayed a significant enrichment in small-size adipocytes and a reduction in medium-size ones compared to sedentary mice, whereasBDNFMet/Met mice showed an opposite trend, even if less marked (Figure 4B).

after PE (Figure S4).

**Figure 4.** Impact of voluntary physical exercise (PE) on epiWAT morphology. (**A**) Inguinal (ingWAT) and (**B**) epidydimal (epiWAT) white adipose tissue on total mouse body weight. (i) Representative hematoxylin and eosin (H&E) staining images and (ii) analysis of the frequency distribution of adipocyte sizes in (**A**) ingWAT and (**B**) epiWAT. Size bar: 100 µm. Black arrow: large adipocytes, green arrow: medium adipocytes, and red arrow: small adipocytes. Data are expressed as mean ± SEM. *n* = 6 mice/group. Two-way ANOVA followed by Bonferroni post hoc analysis. \* *p* < 0.05, \*\* *p* < **Figure 4.** Impact of voluntary physical exercise (PE) on epiWAT morphology. (**A**) Inguinal (ingWAT) and (**B**) epidydimal (epiWAT) white adipose tissue on total mouse body weight. (i) Representative hematoxylin and eosin (H&E) staining images and (ii) analysis of the frequency distribution of adipocyte sizes in (**A**) ingWAT and (**B**) epiWAT. Size bar: 100 µm. Black arrow: large adipocytes, green arrow: medium adipocytes, and red arrow: small adipocytes. Data are expressed as mean ± SEM. *n* = 6 mice/group. Two-way ANOVA followed by Bonferroni post hoc analysis. \* *p* < 0.05, \*\* *p* < 0.01, and \*\*\*\* *p* < 0.001.

0.01, and \*\*\*\* *p* < 0.001. Notably, PE strongly influenced the gene expression profile of epiWAT. In particular, in BDNFVal/Val,4 weeks of PE enhanced mRNA levels of *Adipoq*, whereas it did not modify the expression of genes involved in the adipogenic program (Figures 5A and S4) and in inflammation compared to the sedentary mice. In BDNFMet/Met mice, PE was not sufficient to affect the expression of adipogenic genes, but it was sufficient to improve the inflammatory profile, decreasing the expression of Il-6, Tnf-α, Tgf-β, Mcp-1, and Pai-1 , and to switch M1/M2 macrophage polarization, reducing the Notably, PE strongly influenced the gene expression profile of epiWAT. In particular, in BDNFVal/Val , 4 weeks of PE enhanced mRNA levels of *Adipoq*, whereas it did not modify the expression of genes involved in the adipogenic program (Figure 5A and Figure S4) and in inflammation compared to the sedentary mice. In BDNFMet/Met mice, PE was not sufficient to affect the expression of adipogenic genes, but it was sufficient to improve the inflammatory profile, decreasing the expression of Il-6, Tnf-α, Tgf-β, Mcp-1, and Pai-1, and to switch M1/M2 macrophage polarization, reducing the expression of CD80 and increasing the expression of CD163, (Figure 5B).

expression of CD80 and increasing the expression of CD163, (Figure 5B). In addition, the expression of *Sorl1* was markedly reduced by PE in both BDNFVal/Val and BDNFMet/Met mice, whereas *Adra2a* and *Sirt1* were only slightly, but not significantly, decreased in In addition, the expression of *Sorl1* was markedly reduced by PE in both BDNFVal/Val and BDNFMet/Met mice, whereas *Adra2a* and *Sirt1* were only slightly, but not significantly, decreased in BDNFMet/Met running mice (Figure 5C and Figure S4).

BDNFMet/Met running mice (Figures 5C and S4). Conversely, PE modulated the *BDNF* expression in the two groups of mice. In particular, *BDNF* mRNA levels increased in BDNFVal/Val running mice and reduced in BDNFMet/Met running mice when compared to their respective sedentary controls (Figure 5C). Of note, the expression of both *TrkB* full length and the *TrkB-T1* isoform were slightly, but not significantly, increased in both groups of mice Conversely, PE modulated the *BDNF* expression in the two groups of mice. In particular, *BDNF* mRNA levels increased in BDNFVal/Val running mice and reduced in BDNFMet/Met running mice when compared to their respective sedentary controls (Figure 5C). Of note, the expression of both *TrkB* full length and the *TrkB-T1* isoform were slightly, but not significantly, increased in both groups of mice after PE (Figure S4).

**Figure 5.** Impact of voluntary physical exercise (PE) on the gene expression profile of adipose tissue isolated from BDNFVal/Val and BDNFMet/Met mice. (**A**) Adipogenesis, (**B**) inflammation, and (**C**) BDNF/TrkB pathway related to mRNA levels in epiWAT of sedentary and running BDNFVal/Val and **Figure 5.** Impact of voluntary physical exercise (PE) on the gene expression profile of adipose tissue isolated from BDNFVal/Val and BDNFMet/Met mice. (**A**) Adipogenesis, (**B**) inflammation, and (**C**) BDNF/TrkB pathway related to mRNA levels in epiWAT of sedentary and running BDNFVal/Val and BDNFMet/Met mice. Data are expressed as mean <sup>±</sup> SEM. *<sup>n</sup>* <sup>=</sup> 6 mice/group. Two-way ANOVA followed by Bonferroni post hoc analysis. \* *p* < 0.05, \*\* *p* < 0.01, \*\*\* *p* < 0.005, and \*\*\*\* *p* < 0.001.

BDNFMet/Met mice. Data are expressed as mean ± SEM. *n* = 6 mice/group. Two-way ANOVA followed

#### by Bonferroni post hoc analysis. \* *p* < 0.05, \*\* *p* < 0.01, \*\*\* *p* < 0.005, and \*\*\*\* *p* < 0.001. *3.4. E*ff*ect of Physical Exercise (PE) on the Pro-Thrombotic Phenotype in BDNFMet*/*Met Mice*

*3.4. Effect of Physical Exercise (PE) on the Pro-Thrombotic Phenotype in BDNFMet/Met Mice*  Finally, we investigated the ability of 4 weeks of PE to improve the prothrombotic phenotype Finally, we investigated the ability of 4 weeks of PE to improve the prothrombotic phenotype already observed in BDNFMet/Met [25], in terms of platelet and leukocyte aggregates and FeCl3-induced arterial thrombosis.

already observed in BDNFMet/Met [25], in terms of platelet and leukocyte aggregates and FeCl3-induced arterial thrombosis. As previously shown, in the BDNFMet/Met mice there was a higher number of circulating blood cells, a higher platelet activation state, and enhanced arterial thrombosis [25]. PE restored the physiological number of platelets and leukocytes, and the natural percentage of platelet/leukocyte As previously shown, in the BDNFMet/Met mice there was a higher number of circulating blood cells, a higher platelet activation state, and enhanced arterial thrombosis [25]. PE restored the physiological number of platelets and leukocytes, and the natural percentage of platelet/leukocyte aggregates in response to ADP in BDNFMet/Met mice, without affecting significantly these parameters in BDNFVal/Val mice (Figure 6A–C).

aggregates in response to ADP in BDNFMet/Met mice, without affecting significantly these parameters in BDNFVal/Val mice (Figure 6A–C). Application of FeCl3 to the carotid artery reduced the blood flow in all BDNFMet/Met sedentary mice, leading to a stable occlusion in 100% of mice, whereas only a slight reduction was observed in BNDFVal/Val mice. Of note, PE ameliorated arterial thrombosis, preventing completely the occlusion of the carotid artery in BDNFMet/Met mouse group (Figure 6D). In addition, no statistical differences were observed among sedentary BNDFVal/Val mice and running BNDFVal/Val and/or running BDNFMet/Met mice in terms of carotid artery occlusion (Figure 6D). In line with these data, total occlusion (flow reduction Application of FeCl<sup>3</sup> to the carotid artery reduced the blood flow in all BDNFMet/Met sedentary mice, leading to a stable occlusion in 100% of mice, whereas only a slight reduction was observed in BNDFVal/Val mice. Of note, PE ameliorated arterial thrombosis, preventing completely the occlusion of the carotid artery in BDNFMet/Met mouse group (Figure 6D). In addition, no statistical differences were observed among sedentary BNDFVal/Val mice and running BNDFVal/Val and/or running BDNFMet/Met mice in terms of carotid artery occlusion (Figure 6D). In line with these data, total occlusion (flow reduction >90%) was reached only in sedentary BDNFMet/Met mice after an average time of 15 min (Figure 6E).

>90%) was reached only in sedentary BDNFMet/Met mice after an average time of 15 min (Figure 6E). Overall, these data show that a paradigm of 4 weeks of voluntary PE is able to prevent the Overall, these data show that a paradigm of 4 weeks of voluntary PE is able to prevent the prothrombotic phenotype of BDNFMet/Met mice.

prothrombotic phenotype of BDNFMet/Met mice.

*Cells* **2019**, *8*, x 10 of 17

**Figure 6.** Effect of voluntary physical exercise (PE) on the prothrombotic phenotype of BDNFVal/Val and BDNFMet/Met mice. Numbers of circulating (**A**) platelets and (**B**) leukocytes. (**C**) Percentage of platelet/leukocytes in whole blood analyzed by flow cytometry. Arterial thrombosis induced by topical application of FeCl3 to the carotid artery: (**D**) blood flow and (**E**) time to occlusion measured in sedentary and running BDNFVal/Val and BDNFMet/Met mice. *n* = 6 mice/group. (**A**–**C** and **E**) Two-way ANOVA followed by Bonferroni post hoc analysis. (**D**) Three-way ANOVA with repeated measures followed by Bonferroni post hoc analysis. \*\* *p* < 0.01, \*\*\* *p* < 0.005. **Figure 6.** Effect of voluntary physical exercise (PE) on the prothrombotic phenotype of BDNFVal/Val and BDNFMet/Met mice. Numbers of circulating (**A**) platelets and (**B**) leukocytes. (**C**) Percentage of platelet/leukocytes in whole blood analyzed by flow cytometry. Arterial thrombosis induced by topical application of FeCl<sup>3</sup> to the carotid artery: (**D**) blood flow and (**E**) time to occlusion measured in sedentary and running BDNFVal/Val and BDNFMet/Met mice. *n* = 6 mice/group. (**A**–**C** and **E**) Two-way ANOVA followed by Bonferroni post hoc analysis. (**D**) Three-way ANOVA with repeated measures followed by Bonferroni post hoc analysis. \*\* *p* < 0.01, \*\*\* *p* < 0.005.

#### **4. Discussion 4. Discussion**

Although mutations, as well as genetic variants, including *BDNF* Val66Met polymorphism, have been associated with increased body weight and eating disorders in both human and animal models [19–23,31,35,52–55], the factors and mechanisms involved in the development of obesity in presence of the *BDNF* Met homozygosity remain to be elucidated. It is only known that *BDNF*-to-*TrkB* signaling is an important downstream target of MC4R-mediated signaling involved in the regulation Although mutations, as well as genetic variants, including *BDNF* Val66Met polymorphism, have been associated with increased body weight and eating disorders in both human and animal models [19–23,31,35,52–55], the factors and mechanisms involved in the development of obesity in presence of the *BDNF* Met homozygosity remain to be elucidated. It is only known that *BDNF*-to-*TrkB* signaling is an important downstream target of MC4R-mediated signaling involved in the regulation of energy balance and food intake [55–57].

of energy balance and food intake [55–57]. Using a knock-in *BDNF* Val66Met mouse model, here we confirmed that BDNFMet/Met mice had a higher body weight when compared to BDNFVal/Val [31], and we showed that this increase was related to the enhanced percentage of epiWAT and ingWAT. In particular, adipocytes from epiWAT of mutant mice had a different size distribution, with an enrichment in the percentage of small-sized adipocytes. The presence of small adipocytes in epiWAT of BDNFMet/Met might trace back to hyperplasia or expansion of the small cell population, which are mechanisms of defense that the adipose tissue can undergo in obesity after a threshold of hypertrophy is reached [58–60]. This hypothesis is also supported by the higher expression of *Adra2a* and *Sorl1* found in epiWAT of BDNFMet/Met. Indeed, overexpression of *Adra2a* in animal models has been associated with adipose tissue hyperplasia [61]. In addition, it is well known that the activation of *Adra2a* has an antilipolytic effect, and the increased alpha/beta adrenoreceptor ratio as well as the gain of function mutations of *Adra2* have been associated with obesity in humans [62–64]. Similarly, upregulation of the expression of *Sorl1*, which encodes for the protein Sorla, has been related to reduced lipolytic activity in adipocytes [65], and GWAS analyses have associated *Sorl1* with obesity in humans and in mouse models [21,66], suggesting its key role in metabolic diseases. Using a knock-in *BDNF* Val66Met mouse model, here we confirmed that BDNFMet/Met mice had a higher body weight when compared to BDNFVal/Val [31], and we showed that this increase was related to the enhanced percentage of epiWAT and ingWAT. In particular, adipocytes from epiWAT of mutant mice had a different size distribution, with an enrichment in the percentage of small-sized adipocytes. The presence of small adipocytes in epiWAT of BDNFMet/Met might trace back to hyperplasia or expansion of the small cell population, which are mechanisms of defense that the adipose tissue can undergo in obesity after a threshold of hypertrophy is reached [58–60]. This hypothesis is also supported by the higher expression of *Adra2a* and *Sorl1* found in epiWAT of BDNFMet/Met. Indeed, overexpression of *Adra2a* in animal models has been associated with adipose tissue hyperplasia [61]. In addition, it is well known that the activation of *Adra2a* has an antilipolytic effect, and the increased alpha/beta adrenoreceptor ratio as well as the gain of function mutations of *Adra2* have been associated with obesity in humans [62–64]. Similarly, upregulation of the expression of *Sorl1*, which encodes for the protein Sorla, has been related to reduced lipolytic activity in adipocytes [65], and GWAS analyses have associated *Sorl1* with obesity in humans and in mouse models [21,66], suggesting its key role in metabolic diseases.

The adipose tissue accumulation found in BDNFMet/Met mice was accompanied by a higher expression of the M1 proinflammatory marker CD80, of the monocyte chemoattractant protein-1 (Mcp-1) and of the mediators of inflammation such as Pai-1, Tnf-alpha, and Il-6, which is in line with

The adipose tissue accumulation found in BDNFMet/Met mice was accompanied by a higher expression of the M1 proinflammatory marker CD80, of the monocyte chemoattractant protein-1 (Mcp-1) and of the mediators of inflammation such as Pai-1, Tnf-alpha, and Il-6, which is in line with the well-established paradigm that overweight and obesity are related to adipose tissue inflammation [67]. In addition, the higher levels of these inflammatory transcripts, concomitant with the lower expression of *Adipoq* measured in the epiWAT of BDNFMet/Met mice and the higher number of circulating leukocytes and platelets as well as their activation state, might well summarize the relationship between adipose tissue inflammation and thrombosis. Indeed, the inflammatory profile of adipose tissue in obese subjects as well as the increased presence of these proteins in the circulation have a direct role in the onset and progression of the pathology [68–70], enhancing platelet activation and ability of leukocytes to produce, in turn, inflammatory factors such as Il-6, Tnf, and Cox-2 [9,68,71–74]. All these findings thoroughly summarize data obtained in human adipose tissue samples. Indeed, a positive correlation between proinflammatory cytokines, including IL-6, TNF-α and MCP-1, and adipocyte size was found. Interestingly, the small adipocytes expressed anti-inflammatory factors such as IL-10 and IL-8 [75].

Of note, the reduced levels of *Ppar*γ along with those of adiponectin found in *BDNF* mutant mice might also contribute to the observed adipose tissue inflammation. It is well known that *PPAR*γ, alongside the role of master regulator of adipogenesis, is also involved in the regulation of adipose tissue inflammation. In particular, it was demonstrated that *PPAR*γ downregulates inflammatory adipokines in WAT. Specifically, *PPAR*γ activation downregulates the expression of inflammatory markers such as MCP-1 and TNFα and, thus, reduces inflammation in activated macrophages [56,76–78]. Moreover, *PPAR*γ activation induces adiponectin expression, thus further contributing to the reduction of chronic inflammation [79].

Remarkably, *BDNF* expression was markedly greater in epiWAT of mutant mice, supporting our hypothesis that the *BDNF* Val66Met polymorphism contributes to adipose tissue pathophysiology.

Indeed, studies performed using *BDNF*-(si)RNA-mediated knockdown in the 3T3 cell line showed a reduced adipogenic differentiation ability, supporting the hypothesis that *BDNF* expression is of functional relevance for adipogenesis. In addition, it was reported that *BDNF* expression is dramatically downregulated during adipocyte differentiation, and mature adipocytes only marginally contribute to the production of BDNF in the adipose tissue [80].

Interestingly, we showed that the treatment of C3H10T1/2 cells with Pro-BDNFMet before cell commitment well recapitulated the expression profile of genes that were found altered in the epiWAT of mutant mice. Pro-BDNFMet reduced *Ppar*γ and upregulated *Sorl1* expression, and it increased the percentage of mature adipocytes evaluated in the flow cytometry analysis, suggesting a direct role of the *BDNF* Val66Met polymorphism in the regulation of adipogenesis. However, Pro-BDNFMet was not able to affect *Adipoq* and *Adra2a* as well as *Pai-1* expression, leading us to hypothesize a more complex process that may involve the fraction stromal vascular cells. Indeed, it is suggested that mesenchymal progenitor/stem cells, preadipocytes, endothelial cells, pericytes, T cells, and macrophages, and not mature adipocytes, are the main source of adipokines and PAI-1 in adipose tissue. Of note, the stromal vascular fraction in adipose tissue increases with an increasing degree of obesity [81].

Adipose tissue accumulation represents an independent and modifiable risk factor for CVD [5], and regular PE was recently recognized and strongly recommended as a valuable management strategy for the prevention and treatment of CVD and metabolic disorders from the European Guidelines of cardiology [10,82].

In the present study, we provide evidence that, in mutant BDNFMet/Met mice, four weeks of PE was sufficient to change epiWAT morphology and the inflammatory profile with a concomitant reversion of the prothrombotic phenotype. In particular, the change in adipose tissue morphology observed in BDNFMet/Met running mice was accompanied with a reduction in *Sorl1* and *Adra2a* expression, thus suggesting that PE might improve the metabolic profile of mutant mice, ultimately affecting lipolysis [65,83,84].

The beneficial effect of PE has been provided in animal studies and human trials, showing an impact on both systemic [14,85] and specific reduction of visceral fat mass [86,87], protecting against chronic inflammation-associated disease [88]. Several mechanisms have been proposed to explain the beneficial anti-inflammatory effect of PE. By affecting AMPK and PGC-1α pathways, PE decreases mitochondrial dysfunction and reduces oxidative stress [89,90], with the consequent reduction of proinflammatory adipokines released from the visceral fat mass. Moreover, PE increased production of anti-inflammatory molecules from skeletal muscle and leukocytes [91]. PE decreases Toll-like receptors on monocytes and macrophages, thus preventing their infiltration into adipose tissue and inducing the M1 to M2 macrophage switching to limit macrophage M1 polarization [88].

In line with this evidence, we showed that PE in BDNFMet/Met mice reduced the levels of inflammation mediators, induced a switch in macrophage polarization, and decreased the number of circulating leukocytes and platelets, modifications that, in turn, occur to improve the prothrombotic phenotype observed in mutant mice. Interestingly, for the first time, we provide evidence that PE influenced differently the expression of *BDNF* in the two genotypes, increasing and decreasing its levels in BDNFVal/Val and BDNFMet/Met, respectively. These results might be related to the intrinsic adipose tissue morphology of BDNFVal/Val and BDNFMet/Met mice, suggesting a strong relationship between adipocyte dimension and *BDNF* levels. In fact, the great number of small adipocytes was associated with high levels of *BDNF* (e.g., sedentary BDNFMet/Met and running BDNFVal/Val), and conversely, low levels of transcript were measured in epiWAT when the mean adipocyte dimension was higher (e.g., sedentary BDNFVal/Val and running BDNFMet/Met). The different involvement of the stromal vascular cell fraction in sustaining the adipocyte turnover, as well as the potential contribution of the peripheral nervous system, might explain the different mRNA levels of *BDNF* detected in our experimental setting [92–94]. In this regard, the inability of PE to enhance *BDNF* transcripts in the central nervous system of mutant mice [35] might have important consequences on the levels of BDNF in the peripheral nervous system, thus affecting their levels in epiWAT. Interestingly, it is worth mentioning that, contrary to data presented here related to CVD, the *BDNF* Val66Met polymorphism impairs the beneficial neurobiological changes induced by physical exercise in mice [35].

## **5. Conclusions**

Cardiovascular disease still represents the first cause of mortality worldwide, and obesity is a well-known modifiable risk factor for this pathology. Of note, PE is highly recommended to manage the prevention and treatment of CVD and obesity, showing beneficial cardiometabolic effects.

In human subjects, the *BDNF* Val66Met polymorphism is known to be related to adipose tissue accumulation and cardiovascular risk.

Interestingly, our in vitro data well support the role of Pro-BDNFMet in adipogenesis, in line with data obtained in the BDNFMet/Met WAT mice.

Taking advantage of a mouse model carrying the human *BDNF* Val66Met polymorphism, we showed that 4 weeks of voluntary physical exercise was sufficient to induce positive morphological changes and reduce the inflammatory profile of the adipose tissue.

These beneficial effects might be the bases of the observed reduction in the prothrombotic phenotype detected in this animal model. Future studies are required to assess this relationship.

These data indicate the strong impact of lifestyle, in particular the beneficial effect of PE, on the management of arterial thrombosis and obesity-associated inflammation in relation to genetic mutations that predisposes one, per se, to these pathologies. Nevertheless, human studies need to support these results.

**Supplementary Materials:** The following are available online at http://www.mdpi.com/2073-4409/8/8/875/s1, Figure S1: Cell number and morphology are not altered by ProBDNFVal and ProBDNFMet treatment. Figure S2: Representative flow cytometry graphs showing gate selected for cell granularity analyses at day 3, 5 and 9. Figure S3: Evaluation of the functional relevance of BDNF Val66Met protein on C3H10T1/2 cells adipogenic differentiation. Figure S4: Impact of voluntary physical exercise on gene expression profile of adipose tissue. Table S1: Primers sequences of the analyzed genes. Table S2: F and P values referred to each graph analyzed by Two-way ANOVA or Three-way ANOVA.

**Author Contributions:** Conceptualization, S.S.B.; formal analysis and data curation, L.S., A.I., P.A., S.S.B.; investigation, L.S., A.I., P.A.; original draft preparation, L.S. and S.S.B.; contributed to the discussion on the results from a biological point of view, review and editing, A.I., P.A., M.Z., N.M., F.S.L. and E.T.; supervision and funding acquisition, S.S.B. All authors read and approved the final manuscript.

**Funding:** This research was funded by the Italian Ministry of Health (Ricerca Corrente) grants number BIO35-2015: 2622789; BIO37-2016: 2613074; BIO37-2017: 2631213; MPP1.2A-2018: 2634597. Co-funding provided by the contribution of the Italian "5 × 1000" tax (2014, 2015 and 2016).

**Acknowledgments:** L.S. is supported by the 32nd cycle Ph.D. program in "Scienze farmacologiche sperimentali e cliniche", Università degli Studi di Milano. The authors would like to thank Alessandra Giannopulo Nicolini for the technical support.

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

## **References**


© 2019 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 (http://creativecommons.org/licenses/by/4.0/).

*Article*

## **Melatonin Supplementation Attenuates the Pro-Inflammatory Adipokines Expression in Visceral Fat from Obese Mice Induced by A High-Fat Diet**

**Talita da Silva Mendes de Farias, Regislane Ino da Paixao, Maysa Mariana Cruz, Roberta Dourado Cavalcante da Cunha de Sa, Jussara de Jesus Simão, Vitor Jaco Antraco and Maria Isabel Cardoso Alonso-Vale \***

Department of Biological Sciences, Institute of Environmental Sciences, Chemical and Pharmaceutical, Federal University of Sao Paulo, Diadema 09913-130, Brazil

**\*** Correspondence: alonsovale@gmail.com

Received: 13 July 2019; Accepted: 5 September 2019; Published: 6 September 2019

**Abstract:** Obesity is defined as a condition of abnormal or excessive fat accumulation in white adipose tissue that results from the exacerbated consumption of calories associated with low energy expenditure. Fat accumulation in both adipose tissue and other organs contributes to a systemic inflammation leading to the development of metabolic disorders such as type 2 diabetes, hypertension, and dyslipidemia. Melatonin is a potent antioxidant and improves inflammatory processes and energy metabolism. Using male mice fed a high-fat diet (HFD—59% fat from lard and soybean oil; 9:1) as an obesity model, we investigated the effects of melatonin supplementation on the prevention of obesity-associated complications through an analysis of plasma biochemical profile, body and fat depots mass, adipocytes size and inflammatory cytokines expression in epididymal (EPI) adipose depot. Melatonin prevented a gain of body weight and fat depot mass as well as adipocyte hypertrophy. Melatonin also reversed the increase of total cholesterol, triglycerides and LDL-cholesterol. In addition, this neurohormone was effective in completely decreasing the inflammatory cytokines leptin and resistin in plasma. In the EPI depot, melatonin reversed the increase of leptin, Il-6, *Mcp-1* and *Tnf-*α triggered by obesity. These data allow us to infer that melatonin presents an anti-obesity effect since it acts to prevent the progression of pro-inflammatory markers in the epididymal adipose tissue together with a reduction in adiposity.

**Keywords:** adipose tissue; leptin; adiponectin; Il-6; Mcp-1; Tnf-α

## **1. Introduction**

Obesity results from an imbalance between energy consumption and energy expenditure, promoting an abnormal or excessive accumulation of fat in various regions of the body. Associated with the fat increase, a chronic low-grade inflammation that contributes to systemic metabolic disorders such as dislipidemia, hypertension, nonalcoholic fatty liver diseases, steatohepatitis, cardiovascular diseases, type-2 diabetes and even some cancers has been observed.

Despite the fact that the molecular mechanisms that associate obesity with higher incidences of these diseases are not yet fully defined, evidence indicates that white adipose tissue(s) is one of the first tissues to develop inflammatory responses in obesity conditions, as evidenced by the activation of classical proinflammatory pathways, exacerbated infiltration of macrophages, neutrophils and lymphocytes and a variety of pro-inflammatory mediators secretion [1–3]. It is now well established that white adipose tissue (WAT) is not only involved in energy storage but also functions as an endocrine organ that secretes various bioactive substances or adipo (cyto)kines such as leptin, adiponectin, tumor necrosis factor-alfa (TNF-α), interleukin-6 (IL-6), resistin and macrophage chemoattractant protein 1 (MCP-1), which are known to be involved in a wide range of physiological processes [4]. These adipokines play an important role in the pathophysiological link between increased adiposity and cardiometabolic alterations [4,5]. Indeed, it is also now established that an imbalance of pro- and anti-inflammatory adipokines secretion by WAT due to the expansion of fat mass in obesity exerts potential effects on obesity-linked metabolic disorders [6,7]. Since obesity-associated increases in these adipokines present a great contribution to the development of a dysfunctional WAT, characterized by the infiltration of pro-inflamatory immune cells in this tissue and unresolved inflammation, in addition to inappropriate extracellular matrix remodeling and impaired angiogenesis [8], all of which lead to the development of chronic low grade inflammation [9].

High levels of leptin resulting from leptin resistance and reduced adiponectin levels have been associated with obesity pathophysiology disorders [10]. On the other hand, insulin resistance has been associated with leptin resistance and a reduction of plasma adiponectin, which is reverted by the simultaneous administration of leptin and adiponectin [11,12]. Indeed, a dysregulated expression of these (and others) adipokines by WAT that occurs in obesity is one of the most important components that predisposes one to insulin resistance and that is predictive of the development of metabolic syndrome [13,14]. This occurs because a lot of adipokines interact with the insulin pathway and interfere in glucose and lipid metabolism.

Evidence in the literature indicates that melatonin has a modulatory effect on energy metabolism [15–17], insulin secretion and insulin action, as well as the glucose end lipid metabolism of adipose tissue from rats [18–21]. Therefore, we hypothesized that the melatonin used as a therapeutic strategy could be used as a way to improve the low-grade inflammation observed in obesity conditions. According to some studies, the use of melatonin, a neurohormone produced by the pineal gland only in the night phase and which is responsible for the synchronization of innumerable physiological effects, is related to beneficial effects on the control of obesity and its complications [22,23]. Additionally, chronobiological melatonin aspects and their interrelationship with cytokines produced by adipocytes such as leptin and adiponectin have been evaluated [10,24] and promising results in the prevention and control of complications caused by obesity have been suggested.

Though some studies have evaluated the effects of melatonin and/or pinealectomy on the repercussions of adiponectin and leptin gene expression [19,24], there are a lack of studies related with the role of melatonin on the expression of these and other adipokines involved in the inflammatory response and insulin signaling in WAT by fat depots. Thus, we herein investigate whether melatonin supplementation could prevent the characteristic increase of pro-inflammatory adipokines produced by epididymal WAT during the development of obesity in mice.

## **2. Materials and Methods**

## *2.1. Animals and Melatonin Supplementation*

The study was performed according to protocols approved by the Ethics Committee of the Federal University of São Paulo (CEUA 5998280515). Eight-week-old male C57BL/6 mice obtained from the Center for Development of Experimental Models (CEDEME), Federal University of São Paulo, were housed at 3 mice per cage in a room with light–dark cycle (12-h light, 12-h dark cycle, lights on at 0600) and temperature of 24 ± 1 ◦C. Mice were divided into three groups: (a) Control (low fat) diet (CO), (b) high-fat diet (Obese), and (c) high-fat diet supplemented with melatonin 1 mg/kg (Obese + Mel). The CO diet contained 76% carbohydrate, 15% protein and 9% fat, and the high-fat diet (HFD) contained 26% carbohydrate, 15% protein, and 59% fat, in % kcal. Lard and Soybean oil (9:1) was used as fat source The detailed composition of the diet and energy distribution was provided in our previous study [25].

During obesity induction, the animals were supplemented with melatonin (1 mg/kg) [26] in drinking water during the dark phase, daily, for 10 weeks. Body weight and food intake were measured weekly. After 10 weeks of the experimental protocol, 12-hour fasted mice were killed by cervical dislocation, which occurred between 9am and 11am, after isoflurane anesthesia. Blood samples were centrifuged at 1500 rpm for 20 min at 4 ◦C, and serum were stored at −80 ◦C. Adipose depots were collected and weighted, and epididymal adipose fat (EPI) was processed as described below.

## *2.2. Glucose and Insulin Tolerance Tests*

An oral glucose tolerance test (oGTT) and an insulin tolerance test (ITT) were evaluated after a 6-hour fast. For oGTT analysis, we administrated by gavage a 20% glucose 20% solution (1 g/kg b.w.). The blood glucose measurements were performed at 0, 15, 30, 45, 60 or 90 min. For ITT, animals were injejected intraperitoneally. with insulin (Humulin R, Lilly, 0.75 UI/kg b.w.), and glucose measurements were performed at 0, 10, 20, 30, 40, 50 or 60 min after injection. In both tests, blood samples were collected from the tail vein. This method was not stressful, as indicated by the low basal levels of the stress hormone corticosterone. oGTT and ITT were determined by using a glucometer (One Touch Ultra, Johnson and Johnson, New Brunswick, NJ, USA). The assays were always performed in all groups concomitantly in order to avoid any interference in the obtained results.

## *2.3. Adipocyte Isolation*

Adipocyte isolation was performed as previously described [27] with slight modifications [28]. Briefly, Epi fat pads were diced in small fragments in a flask containing 4 mL of DMEM supplemented with HEPES (20 mM), glucose (5 mM), bovine serum albumin (BSA, 1%), and collagenase type II (1 mg/mL) at pH 7.4 and incubated for approximately 40 min at 37 ◦C in an orbital shaker. Isolated adipocytes were filtered through a plastic mesh (150 µm) and washed three times in a fresh buffer without collagenase. After washing and brief spinning, the medium was thoroughly aspirated, and adipocytes were harvested. Aliquots of isolated adipocytes suspensions were placed in a microscope slide, and 6 fields were photographed under an optical microscope (×100 magnification) coupled to a microscope camera (AxioCam ERc5s; Zeiss, Oberkochen, Germany), and mean adipocyte volume (4/3 × π × r 3 ) was determined by measuring 100 cells using AxioVision LE64 software.

### *2.4. Blood Measurements*

Triacylglycerol (TG) [29], fasting glucose, total cholesterol (TC), LDL-cholesterol [30], and HDL-cholesterol levels [31] were determined by colorimetric assays (Labtest Diagnostics, Lagoa Santa, MG, Brazil).

## *2.5. RNA Extraction and Quantitative Real-Time Polymerase Chain Reaction (qPCR)*

Total RNA was extracted from an EPI depot, reverse transcribed, and destined for quantitative PCR analysis as previously described [25]. An analysis of real-time PCR data was performed using the 2−∆∆<sup>C</sup> <sup>T</sup> method [32]. Data are expressed as the ratio between the expression of the target gene and housekeeping gene (18S gene). Primers used are presented: *Adipoq* (5'-3'sense: GCAGAGATGGCACTCCTGGA; 5'-3'antisense: CCCTTCAGCTCCTGTCATTCC), *Tnf-*α (5'-3'sense: CCCTCACACTCAGATCATCTTCT; 5'-3'antisense: GCTACGACGTGGGCTACAG), *Il-6* (5'-3'sense: TTCTCTGGGAAATCGTGGAAA; 5'-3'antisense: TCAGAATTGCCATTGCACAAC), *Lep* (5'-3'sense: CATCTGCTGGCCTTCTCCAA; 5'-3'antisense: ATCCAGGCTCTCTGGCTTCTG), *Mcp-1* (5'-3'sense: GCCCCACTCACCTGCTGCTACT; 5'-3'antisense: CCTGCTGCTGGTGATCCTCTTGT) and *18S* (5'-3'sense: GGCCGTTCTTAGTTGGTGGAGCG; 5'-3'antisense: CTGAACGCCACTTGTCCCTC).

## *2.6. Adipokine Measurements*

Lysates from the EPI depot and peripheral blood were used to perform the ELISA test. The concentrations of the adipokines IL-6, resistin, adiponectin and leptin were determined using specific commercially available DuoSet ELISA kits according to instructions supplied by the manufacturer (R&D Systems, Minneapolis, MN, USA; Catalog numbers DY406, DY1069, DY1119, DY 498, respectively). The concentrations of the cytokines were expressed in ng per 100 mg of tissue or in ng/ml, as indicated. ng/ml, as indicated. *2.7. Statistical Analysis*

*Cells* **2019**, *8*, x FOR PEER REVIEW 4 of 13

respectively). The concentrations of the cytokines were expressed in ng per 100 mg of tissue or in

#### *2.7. Statistical Analysis* Data are presented as mean ± SEM. A one-way ANOVA and a Tukey post-test were used for the

Data are presented as mean ± SEM. A one-way ANOVA and a Tukey post-test were used for the comparison between groups. GraphPad Prism 5.0 software (GraphPad Software, Inc., San Diego, CA, USA) was used for analysis. The level of significance was set at *p* < 0.05. comparison between groups. GraphPad Prism 5.0 software (GraphPad Software, Inc., San Diego, CA, USA) was used for analysis. The level of significance was set at *p* < 0.05.

## **3. Results**

**3. Results**

#### *3.1. Melatonin Supplementation Decreased Body Mass (BM), Adipose Depots Mass, Adipocytes Hypertrophy and Blood Biochemical Parameters Triggered by HFD-Induced Obesity 3.1. Melatonin Supplementation Decreased Body Mass (BM), Adipose Depots Mass, Adipocytes Hypertrophy and Blood Biochemical Parameters Triggered by HFD-Induced Obesity*

After 10 weeks of diet-induced obesity (DIO), it was found that the HFD was efficient in increasing the body mass of the animals (between approximately two- and five-fold). Concerning food, calories, and fat intake, as compared to CO diet, mice fed with the HFD presented a reduction (by 44%, *p* < 0.05) in food intake but an increase (by three-fold, *p* < 0.05) in fat intake, whereas a slight reduction was observed in calorie (by 20%) and water (by 22%) intake. However, mice that received the HFD associated with melatonin supplementation presented a lower body mass gain (between approximately three- and eight-fold compared to the control group). Though the mice supplemented with melatonin gained less body mass, when the food intake was mesured, both the Obese and Obese + Mel groups presented the same pattern of food, calories and fat consumption (Figure 1A and Table 1). After 10 weeks of diet-induced obesity (DIO), it was found that the HFD was efficient in increasing the body mass of the animals (between approximately two- and five-fold). Concerning food, calories, and fat intake, as compared to CO diet, mice fed with the HFD presented a reduction (by 44%, *p* < 0.05) in food intake but an increase (by three-fold, *p* < 0.05) in fat intake, whereas a slight reduction was observed in calorie (by 20%) and water (by 22%) intake. However, mice that received the HFD associated with melatonin supplementation presented a lower body mass gain (between approximately three- and eight-fold compared to the control group). Though the mice supplemented with melatonin gained less body mass, when the food intake was mesured, both the Obese and Obese + Mel groups presented the same pattern of food, calories and fat consumption (Figure 1A and Table 1).

**Figure 1.** Effects of a high-fat diet (HFD) and melatonin supplementation (Mel, 1 mg/kg b.w., diluted in drinking water, daily, for 10 weeks) on body weight and adipocytes size. (**A**) Mice body mass gain at the end of experimental protocol. (**B**) Volume (in picoliters) of epididymal (EPI) isolated adipocytes. (**C**) Isolated EPI adipocytes photographed under optic microscope (×100 magnification). Adipocyte volume (4/3 × π × r<sup>3</sup> ) was determined by measuring 100 cells per animal (six fields for each slide). Results were analyzed by a one-way ANOVA and a Tukey post-test. Values are mean ± SEM (Control *n* = 21; Obese *n* = **Figure 1.** Effects of a high-fat diet (HFD) and melatonin supplementation (Mel, 1 mg/kg b.w., diluted in drinking water, daily, for 10 weeks) on body weight and adipocytes size. (**A**) Mice body mass gain at the end of experimental protocol. (**B**) Volume (in picoliters) of epididymal (EPI) isolated adipocytes. (**C**) Isolated EPI adipocytes photographed under optic microscope (×100 magnification). Adipocyte volume (4/3 × π × r 3 ) was determined by measuring 100 cells per animal (six fields for each slide). Results were analyzed by a one-way ANOVA and a Tukey post-test. Values are mean ± SEM (Control *n* = 21; Obese *n* = 17; Obese + Mel *n* = 20). \* *p* < 0.05 vs. control; # *p* < 0.05 vs. obese.

17; Obese + Mel *n* = 20). \**p* < 0.05 vs. control; #*p* < 0.05 vs. obese.


**Table 1.** Body mass (BM), food intake, organ weights and blood biochemical parameters after 10 weeks of a high-fat diet feeding and melatonin suplementation in mice.

Results were analyzed by a one-way ANOVA and a Tukey post-test. Values are mean ± SEM. Control *n* = 20; Obese *n* = 17; Obese + Mel *n* = 21 to body mass, food, and fat intake and depots weight; Control *n* = 10; Obese *n* = 9; Obese + Mel *n* = 9 to blood biochemical parameters). \**p* < 0.05 vs. Control; #*p* < 0.05 vs. Obese.

Corroborating the lower body mass gain, we observed that the adipocyte size from the visceral (epididymal—EPI) region of the Obese + Mel group was 42% smaller than the obese group, thus preventing the hypertrophy triggered by the HFD (Figure 1B,C). In the same way, a significant reduction was observed in the EPI and inguinal (ING) depots mass from the animals supplemented with melatonin. The retroperitoneal (RP) and the brown fat (BAT) depots' mass did not present a significant reduction (Table 1).

The analysis of glucose and lipids serum concentrations indicated that the HFD significantly increased fasting glucose (21%), triglycerides (55%), total-cholesterol (52%) and LDL-cholesterol (60%) in the Obese group when compared to the Control group. This increase was prevented in animals that were supplemented with melatonin (Obese + Mel group), since the fasting glucose levels remained similar to the control group, and the serum levels of triglycerides, total cholesterol and LDL-cholesterol were reduced (a reduction of 26%, 21%, and 23%, respectively, in relation to the obese group). There were no significant differences in serum HDL-cholesterol between the groups (Table 1).

## *3.2. Melatonin Supplementation Did Not Alter Glycemic Curve after Glucose and Insulin Tolerance Test (GTT and ITT Test)*

After glucose load, both groups receiving the HFD (Obese and Obese + Mel) presented higher blood glucose levels (Figure 2A). The HFD groups showed lower responsiveness to insulin compared to the control group (Figure 2B). Melatonin supplementation did not alter both the oGTT and ITT tests—that is, it did not prevent the development of insulin intolerance or the response to oral glucose loading.

**GTT**

expression (Figure 3E).

**0**

10). \**p* < 0.05 vs. Control; #*p* < 0.05 vs. Obese.

**ITT**

**Figure 2.** Effects of the high-fat diet (HFD) and melatonin supplementation (Mel, 1 mg/kg b.w., diluted in drinking water, daily, for 10 weeks) on glucose and insulin tolerance tests (GTT and ITT, respectively) in mice. (**A**) GTT or glucose concentration versus time after administration of glucose (2 g/kg b.w.); (**B**) ITT or glucose decay curve versus time after insulin administration (0.75 mU/g b.w.). **Figure 2.** Effects of the high-fat diet (HFD) and melatonin supplementation (Mel, 1 mg/kg b.w., diluted in drinking water, daily, for 10 weeks) on glucose and insulin tolerance tests (GTT and ITT, respectively) in mice. (**A**) GTT or glucose concentration versus time after administration of glucose (2 g/kg b.w.); (**B**) ITT or glucose decay curve versus time after insulin administration (0.75 mU/g b.w.). Values are mean ± SEM (Control *n* = 11, Obese *n* = 7; Obese + Mel *n* = 11). \* *p* < 0.05 versus Control. **Figure 2.** Effects of the high-fat diet (HFD) and melatonin supplementation (Mel, 1 mg/kg b.w., diluted in drinking water, daily, for 10 weeks) on glucose and insulin tolerance tests (GTT and ITT, respectively) in mice. (**A**) GTT or glucose concentration versus time after administration of glucose (2 g/kg b.w.); (**B**) ITT or glucose decay curve versus time after insulin administration (0.75 mU/g b.w.). Values are mean ± SEM (Control *n* = 11, Obese *n* = 7; Obese + Mel *n* = 11). \**p* < 0.05 versus Control.

#### Values are mean ± SEM (Control *n* = 11, Obese *n* = 7; Obese + Mel *n* = 11). \**p* < 0.05 versus Control. *3.3. Melatonin Supplementation Decreased the Gene Expression of Inflammatory Cytokines on EPI Depot 3.3. Melatonin Supplementation Decreased the Gene Expression of Inflammatory Cytokines on EPI Depot.*

*3.3. Melatonin Supplementation Decreased the Gene Expression of Inflammatory Cytokines on EPI Depot.* Whereas obesity is accompanied by a low-grade systemic inflammation, we evaluated the gene expression of the main adipokines produced in the visceral adipose depot from mice under an obesity condition. It was found that the HFD significantly increased the gene expression of *Lep, Il-6, Mcp-1* and *Tnf-α* ((Figure 3A–D) compared to the Control group. Melatonin was able to prevent some of these effects, since the Obese + Mel group presented a significant reduction in the expression of *Lep* (50%) and *Mcp-*1 (55.8%) in relation to the Obese group. Moreover, the expressions of *Il-6* and *Tnf-α*  in the EPI depot of mice supplemented with melatonin were partially prevented (44.6% and 44.8%, respectively), when compared to the Obese group. No change was observed in the *Adipoq* gene Whereas obesity is accompanied by a low-grade systemic inflammation, we evaluated the gene expression of the main adipokines produced in the visceral adipose depot from mice under an obesity condition. It was found that the HFD significantly increased the gene expression of *Lep, Il-6, Mcp-1* and *Tnf-*α ((Figure 3A–D) compared to the Control group. Melatonin was able to prevent some of these effects, since the Obese + Mel group presented a significant reduction in the expression of *Lep* (50%) and *Mcp-*1 (55.8%) in relation to the Obese group. Moreover, the expressions of *Il-6* and *Tnf-*α in the EPI depot of mice supplemented with melatonin were partially prevented (44.6% and 44.8%, respectively), when compared to the Obese group. No change was observed in the *Adipoq* gene expression (Figure 3E). Whereas obesity is accompanied by a low-grade systemic inflammation, we evaluated the gene expression of the main adipokines produced in the visceral adipose depot from mice under an obesity condition. It was found that the HFD significantly increased the gene expression of *Lep, Il-6, Mcp-1* and *Tnf-α* ((Figure 3A–D) compared to the Control group. Melatonin was able to prevent some of these effects, since the Obese + Mel group presented a significant reduction in the expression of *Lep* (50%) and *Mcp-*1 (55.8%) in relation to the Obese group. Moreover, the expressions of *Il-6* and *Tnf-α*  in the EPI depot of mice supplemented with melatonin were partially prevented (44.6% and 44.8%, respectively), when compared to the Obese group. No change was observed in the *Adipoq* gene expression (Figure 3E).

**Control Obese Obese+Mel Control Obese Obese+Mel Figure 3.** Effects of the high-fat diet (HFD) and melatonin supplementation (Mel, 1 mg/kg b.w., diluted in drinking water, daily, for 10 weeks) on mRNA levels of genes related to inflammation in epididymal (EPI) WAT from mice. (**A**) mRNA levels of *Lep*; (**B**) mRNA levels of *Il-6*; (**C**) mRNA levels of *Mcp-1*; (**D**) mRNA levels of *Tnf-α*; (**E**)mRNA levels of *Adipoq*. Results were analyzed by a one-way **Figure 3.** Effects of the high-fat diet (HFD) and melatonin supplementation (Mel, 1 mg/kg b.w., diluted in drinking water, daily, for 10 weeks) on mRNA levels of genes related to inflammation in epididymal (EPI) WAT from mice. (**A**) mRNA levels of *Lep*; (**B**) mRNA levels of *Il-6*; (**C**) mRNA levels of *Mcp-1*; (**D**) mRNA levels of *Tnf-α*; (**E**)mRNA levels of *Adipoq*. Results were analyzed by a one-way ANOVA and a Tukey post-test. Values are mean ± SEM (Control *n* = 9; Obese *n* = 8; Obese + Mel *n* = 10). \**p* < 0.05 vs. Control; #*p* < 0.05 vs. Obese. **Figure 3.** Effects of the high-fat diet (HFD) and melatonin supplementation (Mel, 1 mg/kg b.w., diluted in drinking water, daily, for 10 weeks) on mRNA levels of genes related to inflammation in epididymal (EPI) WAT from mice. (**A**) mRNA levels of *Lep*; (**B**) mRNA levels of *Il-6*; (**C**) mRNA levels of *Mcp-1*; (**D**) mRNA levels of *Tnf-*α; (**E**) mRNA levels of *Adipoq*. Results were analyzed by a one-way ANOVA and a Tukey post-test. Values are mean ± SEM (Control *n* = 9; Obese *n* = 8; Obese + Mel *n* = 10). \* *p* < 0.05 vs. Control; # *p* < 0.05 vs. Obese.

ANOVA and a Tukey post-test. Values are mean ± SEM (Control *n* = 9; Obese *n* = 8; Obese + Mel *n* =

**0.0**

*Depot and Peripheral Bood.*

#### *3.4. Melatonin Supplementation Reduced the Protein Expression of Inflammatory Cytokines on EPI Depot and Peripheral Bood* supplementation reduced these levels by 28% (Obese vs Obese + Mel group, *p* < 0.05, Figure 4E). Melatonin supplementation also prevented the increase of IL-6 protein in the EPI depot

In plasma, we observed an even more pronounced effect, where the Obese group showed an increase in plasma leptin of approximately three-fold compared to the Control group, and melatonin

*Cells* **2019**, *8*, x FOR PEER REVIEW 7 of 13

The protein expression analysis by ELISA corroborated the data presented by the gene expression analysis. *Lep* expression in the EPI depot was significantly increased in the Obese group compared to the control group (82%, *p* < 0.05). In contrast, the Obese + Mel group showed a reduction by 30% (*p* < 0.05) compared to the Obese group (Figure 4A). Thus, melatonin supplementation

*3.4. Melatonin Supplementation Reduced the Protein Expression of Inflammatory Cytokines on EPI*

The protein expression analysis by ELISA corroborated the data presented by the gene expression analysis. *Lep* expression in the EPI depot was significantly increased in the Obese group compared to the control group (82%, *p* < 0.05). In contrast, the Obese + Mel group showed a reduction by 30% (*p* < 0.05) compared to the Obese group (Figure 4A). Thus, melatonin supplementation partially prevented this increase, indicating its important action in reducing these adipokine levels. In plasma, we observed an even more pronounced effect, where the Obese group showed an increase in plasma leptin of approximately three-fold compared to the Control group, and melatonin supplementation reduced these levels by 28% (Obese vs Obese + Mel group, *p* < 0.05, Figure 4E). triggered by the HFD since the Obese + Mel group showed a reduction of 51% (*p* < 0.05) when compared to the Obese group (Figure 4B). No statistical differences in adiponectin and resistin protein levels were observed in this adipose depot (Figure 4C,D). However, adiponectin expression showed a tendency (*p* = 0.0629) to increase in the Obese + Mel group (39% increase compared to the Obese group). Finally, the HFD increased the resistin expression in plasma by 34%. Melatonin supplementation partially prevented this effect, since the Obese + Mel group presented an increase of only 11% compared to the Control group (Figure 4E).

**Cytokines in epididymaI WAT**

**Figure 4.** Effects of the high-fat diet (HFD) and melatonin supplementation (Mel, 1 mg/kg b.w., diluted in drinking water, daily, for 10 weeks) on ELISA analysis of cytokine levels in lysates from epididymal (EPI) WAT and peripheral blood from mice. (**A**) Leptin expression on the EPI depot, (**B**) **Figure 4.** Effects of the high-fat diet (HFD) and melatonin supplementation (Mel, 1 mg/kg b.w., diluted in drinking water, daily, for 10 weeks) on ELISA analysis of cytokine levels in lysates from epididymal (EPI) WAT and peripheral blood from mice. (**A**) Leptin expression on the EPI depot, (**B**) Il-6 expression on the EPI depot, (**C**) adiponectin expression on the EPI depot, (**D**) resistin expression on the EPI depot, (**E**) leptin expression on peripheral blood, and (**F**) resistin expression on peripheral blood. Results were analyzed by a one-way ANOVA and a Tukey post-test. Values are mean ± SEM (Control *n* = 10; Obese *n* = 9; Obese + Mel n = 11). \* *p* < 0.05 vs. Control; # *p* < 0.05 vs. Obese.

**0**

**Control Obese Obese+Mel**

**Control Obese Obese+Mel**

**0**

Melatonin supplementation also prevented the increase of IL-6 protein in the EPI depot triggered by the HFD since the Obese + Mel group showed a reduction of 51% (*p* < 0.05) when compared to the Obese group (Figure 4B). No statistical differences in adiponectin and resistin protein levels were observed in this adipose depot (Figure 4C,D). However, adiponectin expression showed a tendency (*p* = 0.0629) to increase in the Obese + Mel group (39% increase compared to the Obese group).

Finally, the HFD increased the resistin expression in plasma by 34%. Melatonin supplementation partially prevented this effect, since the Obese + Mel group presented an increase of only 11% compared to the Control group (Figure 4E).

## **4. Discussion**

In this study, we evaluated the effects of melatonin on WAT inflammatory aspects in obese mice induced by an HFD. Initially, we verified that the experimental model adopted for DIO was efficient, since we observed that the animals fed an HFD showed a greater gain of body mass (with greater adiposity) and an increase in fasting glucose, triglycerides, total cholesterol and LDL-cholesterol in plasma, glucose intolerance and insulin resistance, corroborating the results obtained previously by our group [25]. Melatonin supplementation was effective in preventing most of the alterations triggered by the HFD, significantly hampering the gain of body mass and preventing the dyslipidemia progression.

There are, in the literature, some works showing the effect of melatonin decreasing body weight, and this hormone has therefore been considered as a possible therapeutic agent against obesity [33,34]. Herein, using a daily melatonin dose of 1 mg/kg (close to the endogenous physiological level), it was observed that the Obese + Mel group had an attenuation of their body mass gain in relation to the non-supplemented group. Taken together with a reduction in both the EPI and ING depots, melatonin suplemmentation performed a protective effect on the development of obesity and adiposity in animals fed an HFD.

Similarly to the findings reported here, using HFD-induced obese Wistar rats and supplementing them with melatonin (25 ug/ml) for 11 weeks, Rios Lugo et al. [34], observed a decrease in the body mass gain of these animals. Favero et al. [33], using a leptin-deficient (ob/ob) mouse supplemented with melatonin in drinking water (100 mg/kg) for eight weeks, also observed a decrease of approximately 5% in total body mass and a decrease in the weight of visceral and subcutaneous fat depots (53% and 41%, respectively) in animals supplemented with melatonin.

In contrast, melatonin supplementation in humans (3 mg daily for three months) did not promote changes in the body mass gain or loss of these individuals [35]. In the same way, Nduhirabandi et al. [36], using obese rats treated with melatonin (4 mg/kg), did not observe any decrease in the body mass despite having presented a cardioprotective effect. These controversial data may be due to the different doses and methods of melatonin administration.

It is believed that melatonin's effect on reducing body mass observed in rodents is probably mediated by central and peripheral target tissues, which leads to the synchronization of circadian rhythms and improved glucose uptake acting directly in adipocytes [24,34]. Here, we showed that dietary-induced obese animals supplemented with melatonin presented a significant reduction in fasting glycemia, indicating again the beneficial action of melatonin on glucose homeostasis, Corroborating these data, other studies employing pinealectomized animals have presented a significant reduction in the expression of the glucose transporter (GLUT 4), as well as glucose intolerance and insulin resistance, which were reverted by melatonin treatment [16,37]. The direct action of melatonin on the adipocytes and myocytes metabolism has been reported. In isolated adipocytes from the inguinal fat of rats, melatonin inhibited isoproterenol-stimulated lipolysis through the inhibition of the cAMP-PKA pathway [38]. In C2C12 skeletal muscle cells, melatonin activates the IRS-1 insulin receptor, thus stimulating GLUT4 expression [39], and these data corroborate a reduced glucose uptake in the skeletal muscle from melatonin receptor-1 knockout mice [40].

It is known that dyslipidemia triggered by obesity plays an important role in the development and worsening of the inflammatory state [41]. Here, we have shown that melatonin supplementation was effective in partially preventing the increase in total cholesterol, LDL-cholesterol and triglycerides characteristic of DIO. Previous studies have already shown this action of melatonin on lipid homeostasis [42–44]. Wistar rats induced to diabetes by streptozotocin and treated with melatonin (10 mg/kg and 20 mg/kg) i.p. for two weeks also showed significant reductions in TG, TC and LDL-cholesterol levels [45]. However, in another study [46], C57Bl/6 mice induced to obesity by an HFD and treated with melatonin (10 mg/kg) for 12 weeks presented a reduction in body mass and LDL levels but not in TG levels. Considering melatonin supplementation in humans, it was observed that after three months of treatment, the daily use of melatonin (3 mg) was effective in significantly lowering TC and TG levels [35]. It is important to note that dyslipidemia is frequently observed in obese and/or diabetic individuals and that high plasma concentrations of total cholesterol and LDL-cholesterol are associated with an increased risk of cardiovascular disease [47,48]. On the other hand, the effects of melatonin on the cardiovascular system are well known. The removal of circulating melatonin causes hypertension in rats, and melatonin replacement prevents or reduces this effect [49,50]. Thus, the attenuation of serum triglyceride levels, total cholesterol and LDL-cholesterol reported here in animals treated with melatonin suggests a role for melatonin in the atherosclerosis prevention, one of the main complications of obesity.

There is a positive correlation between the increase in visceral WAT and hypertension, dyslipidemia, fasting glucose, non-alcoholic steatosis, age, and gender [51,52]. Thus, visceral obesity leads to increased risk of insulin resistance and cardiovascular disease. This fat depot displays a higher production of inflammatory cytokines in obese individuals [53,54].

We observed that the melatonin supplementation reduced the mass of the EPI depot by preventing cell hypertrophy because the adipocyte volume was reduced by 42%. Favero et al. [33], through histological and morphometric observations in WAT of animals supplemented with melatonin, observed that adipose depots from non-obese animals are composed of smaller and regular adipocytes, whereas in obese animals the adipocytes are larger and have a wide lipid droplet with presence of inflammatory infiltrate, macrophages and monocytes with degranulation signs which characterize the inflammatory tissue state (wherein there is a greater adipokines pro-inflammatory expression). Taken together, we suggest that melatonin contributes to the prevention of the inflammatory process in the visceral WAT (here represented by the EPI depot) because prevented the adipocytes hypertrophy.

As expected, we observed a significant increase in the expression of genes encoding proinflammatory cytokines, such as *IL-6, Lep, Mcp-1*, and *Tnf-*α in the EPI depot of animals with DIO. Melatonin fully reversed the increase of *Lep* and *Mcp-1* and partially reversed *Il-6* and *Tnf-*α gene expression. Melatonin supplementation also prevented an increase of leptin and IL-6 protein expression in the EPI depot triggered by the HFD. Moreover, melatonin was effective in decreasing the levels of leptin and resistin in plasma of animals induced to obesity by an HFD.

In the subcutaneous adipose tissue of ob/ob mice, immunofluorescence analyses also revealed melatonin's effects in reducing the expression of TNF-α, resistin, and visfatin, as well as in increasing the expression of adiponectin and its receptors [33]. In the liver of obese mice, Sun et al. [46] found that melatonin treatment also reduced the expression of proinflammatory markers *Tnf-*α*, Il-1*β, and *Il-6*. The same downregulation of these markers was observed in the liver of senescence accelerated prone male (SAMP8) mice [55]. In Wistar rats induced to obesity by an HFD, the plasma analyses demonstrated that melatonin attenuated the increase of the leptin observed in obese animals [34].

It is important to emphasize that the increase of these cytokines, such as leptin and IL-6, establishes a pathophysiological link between increased adiposity and obesity-associated cardiovascular disease, insulin resistance, type 2 diabetes, hypertension, and dyslipidemia [4,5,9]. On the other hand, obesity due to the ingestion of an HFD may be a consequence of a desynchronization in the biological rhythms of important metabolic processes [56,57] Consequently, the obese phenotype may be originated from this circadian desynchronization. Corroborating this, the clock genes and adipocytokines show circadian rhythmicity. The dysfunction of these genes is involved in the alteration of these adipokines during the development of obesity. Desynchronization between the central and peripheral clocks by an altered diet composition can lead to the uncoupling of peripheral clocks from the central pacemaker and to the development of metabolic disorders, leading to obesity. While CLOCK expression levels are increased with HDF-induced obesity, peroxisome proliferator-activated receptor (PPAR) alpha increases the

transcriptional level of brain and muscle ARNT-like 1 (BMAL1) in obese subjects. Consequently, the disruption of clock genes results in dyslipidemia, insulin resistance and obesity [58]. Since melatonin is recognized by its important synchronization of diurnal and circadian rhythms [59]**,** the supplementation with melatonin, respecting its physiological pattern of secretion (exclusively at night as performed in this work), is an important synchronizer to break changes triggered by diet [59] and could prevent changes observed in the obese phenotype.

Other assays need to be performed so that we can conclude in more detail how melatonin acts to reduce proinflammatory cytokines in adipose tissue. However, based on the facts that chronic inflammation in visceral WAT is one of the first steps in triggering obesity-associated diseases and that supplementation with melatonin for 10 weeks revealed significant effects in reducing body mass gain, adiposity and visceral adipocytes hypertrophy, plasma lipids and fasting glucose, as well as the expression of proinflammatory markers in visceral adipose tissue, we can infer here that melatonin must be considered to be a reliable therapeutic agent for the treatment of obesity.

**Author Contributions:** T.d.S.M.d.F.: Concept/design, design of experiments, acquisition of data, data analysis/interpretation, drafting and revision of the manuscript. R.I.d.P.: Concept/design, acquisition of data, revision of the manuscript. M.M.C.: Acquisition of data, revision of the manuscript. R.D.C.d.C.d.S.: Acquisition of data, revision of the manuscript. J.d.J.S.: Acquisition of data, revision of the manuscript. V.J.A.: Acquisition of data, revision of the manuscript. M.I.C.A.V.: Concept/design, data analysis/interpretation, revision of the manuscript, approval of the article.

**Funding:** This work was supported by a grant from FAPESP (2018/05485-6), TSMF is the recipient of Post-doc fellowship from FAPESP (2015/03554-2 and 2016/07638-9).

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

## **References**


© 2019 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 (http://creativecommons.org/licenses/by/4.0/).

*Article*

## **Obesity as an Inflammatory Agent Can Cause Cellular Changes in Human Milk due to the Actions of the Adipokines Leptin and Adiponectin**

**Tassiane C. Morais 1,2 , Luiz C. de Abreu 1,2,3,\* , Ocilma B. de Quental <sup>2</sup> , Rafael S. Pessoa 2,4 , Mahmi Fujimori 2,4, Blanca E. G. Daboin <sup>2</sup> , Eduardo L. França 2,4 and Adenilda C. Honorio-França 2,4,\***


Received: 21 March 2019; Accepted: 7 May 2019; Published: 29 May 2019

**Abstract:** Adiponectin and leptin play roles in the hunger response, and they can induce the inflammatory process as the initial mechanism of the innate immune response. It is possible for alterations in the levels of these adipokines to compromise the functional activity of human colostrum phagocytes. Therefore, the objective of this study is to analyze the effects of adiponectin and leptin on colostrum mononuclear (MN) cells. Colostrum was collected from 80 healthy donors, who were divided into two groups: the control group and the high body mass index (BMI) group. MN cells were used to analyze phagocytosis by flow cytometry, and reactive oxygen species (ROS), intracellular calcium, and apoptosis were assessed by fluorimetry using a microplate reader. Adipokines restored the levels of phagocytosis to the high BMI group (*p* < 0.05), with a mechanism that is action-dependent on the release of ROS and intracellular calcium. However, adiponectin and leptin simultaneously contributed to better microbicidal activity, thus reflecting an increase in the apoptosis level (*p* < 0.05) in the high BMI group. Probably, the maintenance of the balance between adiponectin and leptin levels enhances the protection and decreases the indices of neonatal infection in the breastfeeding infants of women with high BMI values. Therefore, policies that support pre-gestational weight control should be encouraged.

**Keywords:** adiponectin; body mass index; colostrum; leptin; phagocytes; obesity; overweight; oxidative stress

## **1. Introduction**

Obesity is considered a complex, recurrent, and progressive chronic disease. It can be characterized as an inflammatory condition involving elevated oxidative stress, and it is associated with other diseases [1,2] due to the risk of developing comorbidities such as asthma, musculoskeletal and sleep disorders, diabetes mellitus type 2, liver and kidney dysfunction, cardiovascular diseases, infertility, and cancer [1,3,4].

It is important to highlight that women are more affected by obesity than men, which has the potential to cause an impact on the health of future generations [5,6], whether through metabolic changes transmitted to the fetus during pregnancy [6] or through the changes in an infant's nutritional programming. Having adequate nutrition in the early stages of life is essential for the development of a child's metabolic programming, with human milk being the most recommended food for newborns, as it provides inclusive protection against metabolic changes associated with various conditions such as obesity and diabetes. In this way, breastfeeding represents an important pathway that impacts on the reduction of excess weight in both the mother and the child [7–12], because infants who are breastfed and mothers who breastfeed have lower rates of obesity [7,10].

For the mother, breastfeeding facilitates postpartum weight loss; it is positively associated with the lean mass index and inversely associated with visceral fat thickness [9,12]. Thus, this is a way to improve a woman's health after pregnancy, as it may help her to return to a normal metabolic profile and to lose the weight she gained during pregnancy. In infants, breastfeeding may protect against obesity through the components of human milk and behaviors related to infant feeding [8]. In any case, the programing of satiety control is one of the main breastfeeding mechanisms that helps to control obesity [12]. However, in the literature, the number of studies in this area is still scarce, so the mechanisms involved in this process have not yet been totally described.

Studies have demonstrated that metabolic changes due to maternal obesity can lead to changes in the constituents of colostrum and human milk, thereby modifying the different concentrations of hormones regulating appetite and metabolism, such as adiponectin and leptin, through breastfeeding [13–17]. In previously published results from this research group, it was found that the colostrum adiponectin concentration was 8.61 and 13.82 ng/mL, and the leptin concentration was 0.19 and 0.32 mg/dL in normal weight and obese groups, respectively. Furthermore, it was shown that changes in maternal serum constituents caused by maternal obesity are not necessarily reflected in colostrum constituents, especially the adiponectin levels, which were negatively correlated in colostrum and serum [15].

Adiponectin and leptin are mainly secreted by the adipose tissue and inhibit feeding by different mechanisms, acting on their respective receptors (AdipoRs and LepRs) located in neurons in the hypothalamus [18]. These adipokines levels are important indicators for the development of obesity and metabolic syndrome, since there is a reduction in the endogenous concentrations of adiponectin and an increase in leptin levels in overweight and obese individuals [19,20].

It is interesting to note that the actions of adiponectin and leptin are not restricted only to the hunger response. They also have binding sites in the cells of the immune system and activate monocytes/macrophages, triggering the inflammatory process in order to eliminate invading microorganisms. This process can also undergo alterations due to obesity and metabolic syndrome. It is emphasized that these adipokines usually act differently during the inflammatory process; adiponectin acts in a more anti-inflammatory manner, whereas leptin is more proinflammatory [3,21–25].

Several studies in the literature have reported that having a high maternal body mass index (BMI) is correlated with alterations in the constituents of colostrum and human milk [15,26–29]. However, the impact of maternal BMI on the functional activity of colostrum phagocytes is not entirely clear.

It is emphasized that the mother, through breastfeeding, transmits colostrum phagocytes to the infant, which are activated in the presence of microorganisms. So, the colostrum phagocytes encompass the invading particles, and during the process of phagocytosis, microbicidal activity develops via an oxidative burst, and consequently, there is an increase in the release of reactive oxygen species (ROS). Thus, colostrum phagocytes represent an additional mechanism of protection for the baby in case of neonatal infections until the infant's immune system is developed [30–32].

In this study, we investigate the actions of adiponectin and leptin on human colostrum phagocytes. It is possible that changes in adiponectin and leptin in human colostrum due to maternal overweightness can lead to alterations in the responses of mononuclear phagocytes in human colostrum and compromise this fundamental mechanism of infant protection, thus increasing the risk of neonatal infection. Therefore, the aim of this study is to analyze the effects of exogenous adiponectin and leptin and their repercussions on human colostrum mononuclear phagocytes as a function of maternal body mass index (BMI).

## **2. Materials and Methods**

## *2.1. Design and Sample*

A cross-sectional study was carried out with the participation of 80 clinically healthy colostrum donors enrolled at the Hospital of the University of São Paulo, SP, Brazil, in 2017.The women were divided into two groups according to their pre-maternal BMI: normal BMI (18.5–24.9 kg/m<sup>2</sup> ) and high BMI (≥30.0 kg/m<sup>2</sup> ).

The inclusion criteria of the study were as follows: aged from 18 to 35 years; pre-gestational weight known or measured until the end of the 13th gestational week; gestational age at delivery between 37 and 416/<sup>7</sup> weeks; negative serological reactions for hepatitis, HIV, and syphilis; prenatal and non-food restrictions; and informed consent form signed. The exclusion criteria were as follows: gestational diabetes; twin pregnancy; fetal malformations; and delivery before the 36th week of gestation.

The study was approved by the Institutional Committee for Ethics in Research of the Hospital of the University of São Paulo (HU/USP) (CAAE 46643515.0.3001.0076), and all the subjects gave informed written consent before entering the experimental protocol.

## *2.2. Obtaining Colostrum and Cell Separation*

About 5 mL of colostrum was collected manually from both breasts of each woman into sterile plastic tubes between 48–72 h postpartum. Colostrum was collected between feeding intervals, in the daytime period (between 10:00 and 12:00). The samples were stored at –80 ◦C until analysis. The experimental assays were carried out in 2017 and 2018.

The samples were thawed and then centrifuged for 10 min (160× *g*, 4 ◦C), which separated colostrum into three different phases: cell pellet, an intermediate aqueous phase, and an upper fat layer. The upper fat layer and the aqueous supernatant were discarded, and the cell pellet was separated using the Ficoll–Paque concentration gradient (Pharmacia, Uppsala, Sweden). The cells were resuspended in Medium 199 (Gibco, Grand Island, NE, USA) at a concentration of 1 <sup>×</sup> <sup>10</sup><sup>6</sup> cells/mL [30–32] and immediately used in the assays.

## *2.3. Treatment of Mononuclear Cells with Adipokines and Zymosan*

The activation of mononuclear (MN) cells was performed by incubation with Zymosan in the presence and absence of the exogenous adipokines human adiponectin (Sigma, St Louis, MO, USA) and human leptin (Thermo Fisher, Carlsbad, CA, USA), each at a concentration of 100 ng/mL. The concentrations were in accordance with data from the scientific literature [25], and preliminary pilot tests were conducted to standardize the concentrations used.

The MN cells were incubated with Zymosan (for 2 h at 37 ◦C under gentle shaking) and treated with 199 medium (negative control), adiponectin, leptin, and adiponectin+leptin.

The phagocytosis assays were performed with Zymosan pHrodo Green™ (Thermo Fisher, Carlsbad, CA, USA), because it emits green fluorescence in the presence of an acidic pH during the phagocytosis process. Free radical release, apoptosis, and intracellular calcium assays were performed with Zymosan (Sigma, St Louis, MO, USA) without conjugated fluorochrome to avoid interference in the fluorescence intensity of the reagents used in each assay.

## *2.4. Phagocytosis Assays*

The phagocytosis assay using Zymosan pHrodo Green™ (Thermo Fisher, Carlsbad, CA, USA) did not require wash steps and quencher dyes. So, after the incubation period, 10,000 cells were analyzed by flow cytometry using FACSCalibur™ (BD Biosciences, San Jose, CA, USA) with excitation/emission maxima of 509/533 nm. The results were expressed by the Phagocytosis Index (%). The experiments were performed in duplicate.

## *2.5. Tests for the Analysis of Free Radicals*

The cells were incubated with Zymosan in the presence of 5 µ/mL of dihydrorhodamine 123 (DHR123) (Sigma, St Louis, MO, USA). The intensity of fluorescence emitted was proportional to the amount of reactive oxygen species released [33]. The Fluoroskan Ascent FL™ plate reader (Thermo Scientific, Vantaa, Finland) was used, with the 485-nm excitation and 538-nm emission filters. The results were expressed as the DHR123 mean fluorescence intensity. The experiments were performed in duplicate.

## *2.6. Intracellular Calcium Assay*

The cells were incubated with Zymosan in the presence of the 5-µL Fluo-3 AM solution (Sigma, St Louis, MO, USA). The cells were washed and resuspended in HBSS (Hank's Balanced Salt Solution) containing bovine serum albumin (BSA). The fluorescence intensity was measured by a Fluoroskan Ascent FL™ Microplate reader (Thermo Scientific, Vantaa, Finland) using 485-nm excitation and 538-nm emission filters. The rate of intracellular Ca2<sup>+</sup> release was expressed as the mean fluorescence intensity of Fluo-3. The experiments were performed in duplicate.

## *2.7. Apoptosis Assay*

Cells undergoing apoptosis were detected using FITC Annexin V (BD Biosciences, Erembodegem, Belgium). The fluorescence intensity was obtained with the Fluoroskan Ascent FL™ Microplate reader (Thermo Scientific, Vantaa, Finland) using 485-nm excitation and 538-nm emission filters. The results were expressed as Apoptosis Index values (%). The experiments were performed in duplicate.

## *2.8. Statistical Analysis*

Statistical analyses were performed with BioEstat® version 5.0 software (Mamirauá Institute, Belém, Brazil). The results are presented as mean (± standard deviation). The D'Agostino normality test and variance analysis (ANOVA) were used, followed by Tukey's test. Significant differences were considered when *p* < 0.05, and the power of the test for the sample size used was 80%.

### **3. Results**

The women who had a high pre-pregnancy BMI also presented higher BMI values at the time of delivery. The other maternal/infant parameters did not differ between groups (Table 1).

**Table 1.** Maternal and neonate characteristics according to maternal pre-gestational body mass index (BMI) (normal or high BMI).


Maternal and neonatal data are shown as mean (±SD) or number (%). They were assessed by ANOVA and Tukey's test. # Statistical difference among the normal and high BMI groups (*p* < 0.05).

ANOVA and Tukey's test.

hormones within groups (*p* < 0.05).

Delivery BMI (kg/m<sup>2</sup>

(BMI) (normal or high BMI).

Maternal pre-gestational BMI (kg/m<sup>2</sup>

Pre-gestational excess weight caused a reduction in the phagocytosis index of the human colostrum mononuclear cells (*p* < 0.05). However, the addition of adiponectin and leptin (100 ng/mL) increased the percentage of phagocytosis in the group with high BMI (*p* < 0.05) (Figure 1). Pre-gestational excess weight caused a reduction in the phagocytosis index of the human colostrum mononuclear cells (*p* < 0.05). However, the addition of adiponectin and leptin (100 ng/mL) increased the percentage of phagocytosis in the group with high BMI (*p* < 0.05) (Figure 1).

*Cells* **2019**, *8*, x FOR PEER REVIEW 5 of 12

**Table 1.** Maternal and neonate characteristics according to maternal pre-gestational body mass index

Age (years) 26.02 (±5.43) 25.60 (±4.88) Diabetes or gestational diabetes (%) 00 (0.00%) 00 (0.00%)

Gestational weight gain 10.52 (±4.08) 9.11 (±3.24) Gestational age at delivery(weeks) 38.77 (±1.10) 38.70 (±1.04) Infant sex—female (%) 23 (57.50%) 21 (52.50%) Birth weight (g) 3263.00 (±430.31) 3367.375(±495.50) Birth height (cm) 47.31 (±2.84) 47.58 (±2.69)

(18.5–24.9 kg/m<sup>2</sup>

) 25.72 (±2.32) 34.77 (±3.78)

)

) 21.51 (±2.32) 31.05 (±3.77)

#Statistical difference among the normal and high BMI groups (*p* < 0.05).

**High BMI (n = 40)** (≥30 kg/m<sup>2</sup>

)

#

#

**Maternal and child characteristics Normal BMI (n = 40)**

**Figure 1.** Effects of adiponectin and leptin on phagocytosis by colostrum mononuclear (MN) cells from women with normal BMI and high BMI values. Phagocytosis Index (%) values were determined by the assessment of pHRodo™ Green zymosan in MN cells from human colostrum. The results are presented as mean ± SD (n = 10 per group). They were assessed by ANOVA and Tukey's Test. \* Statistical difference between colostrum cells incubated with 119 medium and **Figure 1.** Effects of adiponectin and leptin on phagocytosis by colostrum mononuclear (MN) cells from women with normal BMI and high BMI values. Phagocytosis Index (%) values were determined by the assessment of pHRodo™ Green zymosan in MN cells from human colostrum. The results are presented as mean ± SD (n = 10 per group). They were assessed by ANOVA and Tukey's Test. \* Statistical difference between colostrum cells incubated with 119 medium and hormones within groups (*p* < 0.05). # Statistical difference among groups with the same treatment and samples (*p* < 0.05).
