Statistical difference among groups with the same treatment and sample (*p* < 0.05).

*Cells* **2019**, *8*, x FOR PEER REVIEW 6 of 12

**\*#**

**\***

**Figure 3.** The effects of adiponectin and leptin on the release of intracellular calcium by colostrum mononuclear phagocytes from women with normal BMI and high BMI values incubated with Zymosan A from *Saccharomyces cerevisiae*. Intracellular calcium release was performed with Fluo-3AM (n = 10). The results are presented as the 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 **Figure 3.** The effects of adiponectin and leptin on the release of intracellular calcium by colostrum mononuclear phagocytes from women with normal BMI and high BMI values incubated with Zymosan A from *Saccharomyces cerevisiae*. Intracellular calcium release was performed with Fluo-3AM (n = 10). The results are presented as the 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 sample (*p* < 0.05).

treatment and sample (*p* < 0.05). Irrespective of the BMI status of the women tested, their colostrum cells showed similar apoptosis rates when treated with adiponectin or leptin compared to untreated cells. When the colostrum cells from higher BMI group were treated with both hormones, they exhibited higher Irrespective of the BMI status of the women tested, their colostrum cells showed similar apoptosis rates when treated with adiponectin or leptin compared to untreated cells. When the colostrum cells from higher BMI group were treated with both hormones, they exhibited higher apoptosis rates than in cells from normal BMI group (Figure 4). *Cells* **2019**, *8*, x FOR PEER REVIEW 7 of 12

**Figure 4.** The effects of adiponectin and leptin on the apoptosis index of colostrum mononuclear phagocytes from women with normal BMI and high BMI values incubated with Zymosan A from *Saccharomyces cerevisiae*. The apoptosis assay was performed by Annexin V-fluorescein isothiocyanate (FITC) staining (n = 10). The results are presented as mean ± SD. 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 **Figure 4.** The effects of adiponectin and leptin on the apoptosis index of colostrum mononuclear phagocytes from women with normal BMI and high BMI values incubated with Zymosan A from *Saccharomyces cerevisiae*. The apoptosis assay was performed by Annexin V-fluorescein isothiocyanate (FITC) staining (n = 10). The results are presented as mean ± SD. 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 sample (*p* < 0.05).

child health, as well as controlling weight in overweight and obese individuals [7,8,12].

phagocytosis index values and ROS release in the presence of Zymosan.

Obesity is characterized by an increase in adipose tissue, which is considered a large organ with plastic and endocrine properties [34]. Adipocytes secrete inflammatory cytokines and hormones called adipokines that have complex interactions with each other and are altered in overweight and obese individuals. In obese individuals, an increase in proinflammatory adipokine levels and a

In obese mothers, a state of chronic inflammation can trigger an exaggerated inflammatory response in the placenta, leading to the accumulation of macrophages and the production of proinflammatory mediators [38]. However, a different mechanism that leads to the production and regulation of inflammatory mediators probably operates in colostrum from obese mothers, which exhibits immunological components that protect newborns without triggering inflammatory processes [15]. This suggests that breastfeeding may represent a strategy for enhance maternal and

The findings of the present study indicate that as an inflammatory agent, obesity alters phagocytosis and reactive oxygen species release in colostrum phagocytes. In addition, colostrum cells from obese mothers showed differences in functional activity in the presence of adiponectin

Colostral macrophages produce reactive oxygen species (ROS) and show phagocytic activity [30–32]. Reactive oxygen species (ROS) release is one of the last events in the course of the innate immune response and plays an essential role in the death process of invasive microorganism [39]. In this study, colostrum phagocytes from women with high BMI values were associated with lower

Several studies have reported that maternal obesity causes alterations in the components of colostrum and human milk [15,26–29]. Some authors have shown that obesity is associated with a decreased macrophage phagocytic index [40,41], whereas others suggest an increase in phagocytes

Interestingly, the treatment of colostrum cells with adiponectin and leptin restored the phagocytic capacity of colostrum cells from obese mothers to similar phagocytic index values to normal weight mothers. However, ROS release only showed higher levels when the cells were

## **4. Discussion**

Obesity is characterized by an increase in adipose tissue, which is considered a large organ with plastic and endocrine properties [34]. Adipocytes secrete inflammatory cytokines and hormones called adipokines that have complex interactions with each other and are altered in overweight and obese individuals. In obese individuals, an increase in proinflammatory adipokine levels and a reduction in anti-inflammatory adipokines occurs, promoting a state of chronic inflammation [35]. The most abundant adipokines secreted by adipocytes are adiponectin and leptin [36,37].

In obese mothers, a state of chronic inflammation can trigger an exaggerated inflammatory response in the placenta, leading to the accumulation of macrophages and the production of proinflammatory mediators [38]. However, a different mechanism that leads to the production and regulation of inflammatory mediators probably operates in colostrum from obese mothers, which exhibits immunological components that protect newborns without triggering inflammatory processes [15]. This suggests that breastfeeding may represent a strategy for enhance maternal and child health, as well as controlling weight in overweight and obese individuals [7,8,12].

The findings of the present study indicate that as an inflammatory agent, obesity alters phagocytosis and reactive oxygen species release in colostrum phagocytes. In addition, colostrum cells from obese mothers showed differences in functional activity in the presence of adiponectin and leptin.

Colostral macrophages produce reactive oxygen species (ROS) and show phagocytic activity [30–32]. Reactive oxygen species (ROS) release is one of the last events in the course of the innate immune response and plays an essential role in the death process of invasive microorganism [39]. In this study, colostrum phagocytes from women with high BMI values were associated with lower phagocytosis index values and ROS release in the presence of Zymosan.

Several studies have reported that maternal obesity causes alterations in the components of colostrum and human milk [15,26–29]. Some authors have shown that obesity is associated with a decreased macrophage phagocytic index [40,41], whereas others suggest an increase in phagocytes and oxidative stress in these cells [42].

Interestingly, the treatment of colostrum cells with adiponectin and leptin restored the phagocytic capacity of colostrum cells from obese mothers to similar phagocytic index values to normal weight mothers. However, ROS release only showed higher levels when the cells were treated with leptin.

Leptin also plays an important role in the action of phagocytes, and can be used as a potent phagocytosis-inducing agent [43]. Immunomodulatory effects have been attributed to this hormone, which represents a link between nutritional status and neuroendocrine and immunological functions [44] because of its dual action as both a hormone and cytokine [45] that can control macrophage activation [46]. Its mechanism of action involves the activation of the JAK/STAT signaling pathways, which induce phagocytosis, oxidative burst, and the increased secretion of proinflammatory cytokines [22].

Studies have shown a significant relationship between high levels of leptin and the inflammatory state present in obesity [47,48]. Increased ROS production modifies the response of intracellular Ca2<sup>+</sup> during oxidative metabolism [49]. It is important to note that although leptin is related to ROS elevation, when macrophages are in a hyperinflammatory state, they may alter the oxidative burst [21]. In this study, it is possible that the alteration in ROS by colostrum cells treated with leptin may have been associated with the reduction of intracellular calcium released by the cells from the high BMI group.

In contrast to leptin, although the presence of adiponectin in colostrum cells from obese mothers increased phagocytosis and intracellular calcium release, it did not alter the ROS release. Adiponectin, despite being known for its anti-inflammatory action, is derived from its ability to stimulate the expression of macrophage markers of the M2 phenotype. This adipokine also causes stimuli in macrophages [50,51]. Adiponectin promotes macrophage activation, inducing a proinflammatory response that resembles M1 more than M2. It induces a limited program of inflammatory activation that likely desensitizes these cells to future proinflammatory stimuli [50]. Cot/tpl2 also play an important role in the production of inflammatory mediators upon the stimulation of macrophages with

adiponectin. Furthermore, activation of this axis plays an important role in the M1 proinflammatory program induced by adiponectin in macrophages [52].

Through phosphatidyl inositol 3-kinase (PI3K)/Akt, adiponectin induces the activation of the enzyme IkB kinase. This enzyme is responsible for the degradation of inhibitory protein kB and thus the consequent activation of NF-kB, which will promote the release of proinflammatory cytokines in the nucleus and trigger the inflammatory response [51].

It must be highlighted that intracellular calcium levels are fundamental for the phagocytosis process and to control the subsequent steps involved in phagosome maturation [53]. Here, colostrum cells from mothers with high BMI values treated with adiponectin exhibited higher concentrations of intracellular calcium, suggesting that it plays an important role in its mechanism of action. Intracellular calcium is responsible for the control of various cellular processes including proliferation, differentiation, and cell death [54]. Alterations in the intracellular Ca2<sup>+</sup> influx by human cells may cause cell damage and have been associated with apoptosis [55].

Considering that leptin was able to activate cellular oxidative mechanisms and that adiponectin increased the calcium influx, we suggest that treatment with adiponectin plus leptin induces a more efficient phagocytic response with an association mechanism that is dependent on ROS and intracellular calcium, which culminated in increased rates of apoptosis. Possibly, this association potentiated the microbicidal activity of these phagocytes, and thus led to the consequent increase in the levels of apoptosis in the high BMI group.

According to the scientific literature, adipokines can enhance the cellular response, possibly through activation of the classical NF-kB pathway, which regulates the genes responsible for the production of most ROS in the cells [56].

From the literature, it is already known that the ratio between leptin and adiponectin is an important parameter that can be used to indicate the risk of developing obesity and even metabolic syndrome [19,20].

However, in this study, the data suggest that the actions of adiponectin and leptin on the functional activity of colostrum cells occur by different mechanisms. Adiponectin is probably associated with microbicidal activity, which is essential for the immunological response. Leptin induces high levels of ROS, which, although important for microbicidal mechanisms, may be responsible for the inflammatory processes generated by the activation of macrophages. Thus, their association is a considerable defense strategy, because in addition to inducing apoptosis, it reduced oxidative stress to levels similar to those of the non-obese group, and consequently may have controlled inflammatory processes. The maintenance of the balance of adiponectin and leptin concentrations is essential for the adequate function of mononuclear phagocytes and is extremely important for the immunologic and metabolic programming that breastfeeding can confer to the child.

This study used a cross-sectional design that brings limitations to the research. It would be interesting to develop a prospective cohort study to evaluate the effect of adipokines on human colostrum phagocytes and on the health outcomes of infants during the childhood. Such research could add important contributions that may allow the development of new strategies against the obesity epidemic, for example, through the encouragement of maternal weight control or initiatives that use colostrum and breast milk as an intervention strategy. In this sense, breastfeeding could be used at strategic hours which take into account the serum fluctuations of adipokines. Thus, we could create new alternatives with potential impacts against possible inflammatory processes caused by obesity as well as reduce the prevalence of childhood infections.

## **5. Conclusions**

Pre-gestational maternal overweightness and obesity alter the functional activity of human colostrum phagocytes. In this study, the action of adiponectin on colostrum phagocytes did not alter the level of reactive oxygen species and increased the release of intracellular calcium in obese women. In contrast, leptin increased the level of reactive oxygen species and reduced the intracellular calcium

concentration. The association of adiponectin with leptin was shown to be essential to restore the level of reactive oxygen species and maintain the intracellular calcium level. It also contributed to better microbicidal activity, thus reflecting an increase in the apoptosis level.

These data reinforce the importance of maternal weight control from the pre-gestational period to ensure a balance among the endogenous concentrations of adipokines and to provide to the infant, through breastfeeding, colostrum mononuclear cells that are capable of eliciting a more effective response against childhood infection. Moreover, these data support the importance of breastfeeding in the context of obesity to control the inflammatory process and reduce childhood infections.

**Author Contributions:** Conceptualization, A.C.H.-F.; Data curation, T.C.M., O.B.d.Q., R.S.P. and M.F.; Formal analysis, T.C.M., B.E.G.D. and E.L.F.; Funding acquisition, L.C.d.A. and A.C.H.-F.; Investigation, R.S.P. and M.F.; Methodology, O.B.d.Q., M.F. and B.E.G.D.; Project administration, E.L.F.; Supervision, L.C.d.A., E.L.F. and A.C.H.-F.; Writing—original draft, T.C.M., E.L.F. and A.C.H.-F.; Writing—review & editing, L.C.d.A. and B.E.G.D.

**Funding:** Funding for this research was provided by Fundação de Amparo a Pesquisa de São Paulo (FAPESP N◦ : 2015/19922-0; N◦ : 2015/01051-3) and Conselho Nacional de DesenvolvimentoCientífico e Tecnológico (CNPq-N◦ : 303983/2016-7; N◦ : 403383/2016-1).

**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/).

*Review*

## **Associations between Fatty Acid-Binding Protein 4–A Proinflammatory Adipokine and Insulin Resistance, Gestational and Type 2 Diabetes Mellitus**

**Marcin Trojnar <sup>1</sup> , Jolanta Patro-Małysza <sup>2</sup> , Zaneta Kimber-Trojnar ˙ 2,\* , Bozena Leszczy ´nska-Gorzelak ˙ <sup>2</sup> and Jerzy Mosiewicz <sup>1</sup>**


Received: 4 February 2019; Accepted: 3 March 2019; Published: 8 March 2019

**Abstract:** There is ample scientific evidence to suggest a link between the fatty acid-binding protein 4 (FABP4) and insulin resistance, gestational (GDM), and type 2 (T2DM) diabetes mellitus. This novel proinflammatory adipokine is engaged in the regulation of lipid metabolism at the cellular level. The molecule takes part in lipid oxidation, the regulation of transcription as well as the synthesis of membranes. An involvement of FABP4 in the pathogenesis of obesity and insulin resistance seems to be mediated via FABP4-dependent peroxisome proliferator-activated receptor γ (PPARγ) inhibition. A considerable number of studies have shown that plasma concentrations of FABP4 is increased in obesity and T2DM, and that circulating FABP4 levels are correlated with certain clinical parameters, such as body mass index, insulin resistance, and dyslipidemia. Since plasma-circulating FABP4 has the potential to modulate the function of several types of cells, it appears to be of extreme interest to try to develop potential therapeutic strategies targeting the pathogenesis of metabolic diseases in this respect. In this manuscript, representing a detailed review of the literature on FABP4 and the abovementioned metabolic disorders, various mechanisms of the interaction of FABP4 with insulin signaling pathways are thoroughly discussed. Clinical aspects of insulin resistance in diabetic patients, including women diagnosed with GDM, are analyzed as well.

**Keywords:** adipose tissue; fatty acid-binding protein 4; proinflammatory adipokine; insulin resistance; gestational diabetes mellitus; type 2 diabetes mellitus

## **1. Introduction**

Type 2 diabetes mellitus (T2DM) represents a common metabolic disorder that is characterized by chronic hyperglycemia. For more than half a century, the link between insulin resistance and T2DM has been well recognized. Insulin resistance is not only the most powerful predictor of future development of T2DM, but it is also a therapeutic target. On the other hand, gestational diabetes mellitus (GDM) is one of the most common metabolic disorders of pregnancy and its incidence has considerably increased by 10–100% in the last 20 years [1]. It should be emphasized that women with a previous history of GDM have a significantly increased risk of developing T2DM, obesity, and cardiovascular diseases in the future [2–5]. Women who had prior GDM are nearly eight times more likely to develop future T2DM compared with those with normal glucose tolerance during their pregnancy [6]. Up to one-third of women with T2DM have been previously diagnosed with GDM [7,8]. The identification of women with GDM who are at high risk of developing subsequent diseases offers a remarkable opportunity to alter their future health [1,9].

There is ample evidence to suggest a link between fatty acid-binding protein 4 (FABP4) and insulin resistance, GDM, and T2DM.

## **2. Fatty Acid-Binding Protein 4**

FABP4, also referred to in the literature as adipocyte fatty acid-binding protein (AFABP), is a relatively novel adipokine [10], which belongs to the calycin protein superfamily. This protein has also been termed adipocyte P2 (aP2) since there is high sequence similarity (67%) with the myelin P2 protein (M-FABP/FABP8) [11]. FABP4 is highly expressed in adipocytes and represents approximately 1% of all soluble proteins in adipose tissue [11]. FABP4 is able to reversibly bind to hydrophobic ligands, such as saturated and unsaturated long-chain fatty acids, eicosanoids, and other lipids. Accordingly, it takes part in the regulation of lipid trafficking and responses at the cellular level [12–15]. FABPs, a family of intracellular lipid chaperones, are engaged in the transport of fatty acids to specific organelles in the cell, including mitochondria, peroxisomes, the nucleus, and the endoplasmic reticulum [12,16]. Therefore, FABPs play a significant role in lipid oxidation, lipid-mediated transcriptional regulation, and the signaling, trafficking, and synthesis of membranes. In addition, FABPs are also engaged in the regulation of the enzymatic activity and storage of lipid droplets in the cytoplasm [17], the conversion of fatty acids to eicosanoids, and the stabilization of leukotrienes [18].

The human FABP4 consists of 132 amino acids. Its molecular mass has been assessed at 14.6 kDa. FABP4 expression markedly increases at the time of adipocyte differentiation [12]. Due to the abovementioned observation, this molecule has been suggested as an adipocyte differentiation marker [19]. FABP4 expression is also enhanced during differentiation from monocytes to macrophages. A wide spectrum of different proinflammatory factors modify and control the expression of FABP4 in these cells [20]. In macrophages, FABP4 stimulates the foam cell formation. Foam cell formation, which is believed to be mediated by modified low density lipoproteins (LDLs), often occurs in the presence of increased concentrations of insulin and glucose. These increased concentrations are characteristic of the insulin resistance associated with diabetes, obesity, and the metabolic syndrome [21]. Siersbaek et al. [22] found that peroxisome proliferator-activated receptor γ (PPARγ) and CCAAT/enhancer-binding protein (C/EBP) regulate the expression of most genes associated with adipogenesis. PPARγ not only promotes the proliferation and differentiation of adipocytes, but also confers insulin sensitivity to adipocytes [23]. An increase in insulin sensitivity can in turn promote the expression of the PPARγ gene in the adipose tissue, thereby positively accelerating the differentiation of adipocytes. FABP4 expression is controlled at the transcriptional level by PPARγ and C/EBP [12]. However, the aforementioned transcriptional regulators seem to be dysregulated in T2DM [24,25].

Other actions encompass modifications of the inflammatory response mediated by activation of the IKK-NF-κB and c-Jun N-terminal kinase (JNK)-activator protein-1 (AP-1) pathways [20]. In addition, FABP4 enhances the hydrolytic activity of hormone-sensitive lipase (HSL) [26].

FABP4 as an adipokine may influence insulin sensitivity. On the other hand, the expression of FABP4 is highly induced during adipocyte differentiation and transcriptionally controlled by PPARγ agonists, fatty acids, dexamethasone, and insulin [27]. Insulin downregulates only microvesicle-free-mediated and microvesicle-secreted FABP4. However, the release of FABP4 via adipocyte-derived microvesicles is a small fraction and conveys a minor activity [27].

Furthermore, FABP4-/- animals exhibit a defect in β-adrenergic stimulated insulin secretion, even under lean conditions, suggesting an effect on β cell function, another cell type that does not express FABP4. This is complemented by human studies showing that higher serum FABP4 levels correlate with a higher insulin response index in T2DM patients, and a higher insulinogenic index in non-diabetics [28]. In vivo, FABP4 levels are suppressed under refeeding or insulin [28].

Various biological effects of exogenous FABP4 have been demonstrated in different types of cells; FABP4 has been reported to enhance hepatic glucose production in vivo and in vitro and to increase the proliferation of vascular smooth muscle cells and glucose-stimulated insulin secretion in pancreatic β cells [29]. Furthermore, glucose oxidation and glycolysis are inhibited and glucose uptake as well as

its utilization in the muscles and liver are considerably limited [30–32]. Moreover, FABP4 inhibits the activation of the insulin-signaling pathway, resulting in decreased activation of the endothelial nitric oxide synthase (eNOS) in vascular endothelial cells and nitric oxide production, inducing endothelial dysfunction [33].

## **3. Relationship between FABP4 and PPAR**γ

An involvement of FABP4 in the pathogenesis of obesity and insulin resistance might be mediated via FABP4-dependent PPARγ inhibition [34]. FABP4 triggers the ubiquitination and subsequent proteasomal degradation of PPARγ, a crucial regulator of adipogenesis and insulin responsiveness [35]. Importantly, FABP4-null mouse preadipocytes as well as macrophages exhibited increased expression of PPARγ, and complementation of FABP4 in the macrophages reversed the increase in PPARγ expression. The FABP4-null preadipocytes exhibited a remarkably enhanced adipogenesis compared with wild-type cells, indicating that FABP4 regulates adipogenesis by downregulating PPARγ [35].

The FABP4 level was higher and PPARγ level was lower in human visceral fat and mouse epididymal fat compared with their subcutaneous fat. Furthermore, FABP4 was higher in the adipose tissues of obese diabetic individuals compared with healthy ones. Suppression of PPARγ by FABP4 in visceral fat may explain the reported role of FABP4 in the development of obesity-related morbidities, including insulin resistance, diabetes, and atherosclerosis [35]. The most recent studies also suggest that the FABP4 inhibitor-BMS309403 significantly improves insulin sensitivity in the ob/ob mice, and the FABP4 inhibitor reduces plasma triacylglycerol levels [10].

PPARγ is activated by natural or synthetic agonists, such as the antidiabetic, thiazolidinedione (TZD) [36]. The disruption of PPARγ specifically in myeloid cells also predisposes mice to the development of diet-induced obesity, insulin resistance, and glucose intolerance, whereas activation of PPARγ within macrophages promotes lipid efflux, thereby stabilizing atherosclerotic lesions [37,38].

### **4. Relationship between FABP4 and Diseases of Civilization**

It has been reported that increased circulating FABP4 levels are associated with obesity, insulin resistance, T2DM, cardiovascular disorders, arterial hypertension, cardiac dysfunction, kidney damage, fatty liver disease, and atherosclerosis [10,29,31,32,39–42]. Increased FABP4 concentrations were found in obese subjects and were positively correlated with waist circumference, blood pressure, and insulin resistance [40]. A 10-year prospective study also documented that high FABP4 levels independently predicted the development of T2DM [31]. An increased concentration of FABP4 was an independent biomarker of the development of metabolic syndrome in a five-year perspective in a Chinese population [41]. Also, Stejskal et al. [43] found that serum FABP4 levels might be a significant predictor of metabolic syndrome in a Caucasian population. Cabré et al. [44] reported that higher FABP4 plasma concentrations were associated with the early presence of metabolic syndrome components, along with inflammation and oxidation markers in T2D subjects.

Circulating FABP4 is not only a potent biomarker, but, as an adipokine, it also plays an important role in the development of metabolic syndrome and cardiovascular diseases [29]. Furthermore, FABP4 could be a treatment target in T2DM [45]. A small molecular ligand for FABP4 that blocks the binding of endogenous ligands may be developed into a drug for the treatment of T2DM [46].

## **5. Associations of FABP4 with Adipogenesis and Inflammation**

FABP4 plays a crucial role in the regulation of lipid-mediated actions, such as the initialization of inflammation and oxidative stress processes [33]. The expression of this adipogenic protein can be induced by vascular endothelial growth factor (VEGF) signaling [47]. FABP4 increases the ciatriphosphate-binding cassette A1 pathway [32]. Emerging data suggests an essential involvement of FABP4 in endothelial dysfunction [47]. Exogenous FABP4 interferes with insulin stimulated production of nitric oxide in endothelial cells [48]. Cabré et al. [44] observed that increased FABP4 concentrations were associated with excessive oxidative stress and inflammatory markers in diabetes.

FABP4 negatively regulates PPARγ in macrophages and adipocytes, affecting adipocyte differentiation. Higher levels of FABP4 and lower levels of PPARγ in visceral adipose tissue, when compared with subcutaneous adipose tissue, suggest a causative link between FABP4 and the metabolic syndrome. It may also explain certain morphological and functional differences between the adipocytes found in these two types of fat tissue. Visceral preadipocytes are known to proliferate and differentiate into mature adipocytes less actively than those from subcutaneous fat. Visceral adipose tissue secretes more proinflammatory cytokines [35].

It has also been postulated that the transcription factor forkhead box protein O1 (FOXO1) is involved in lipid metabolism. The available scientific evidence suggests that metformin may have a protective effect against lipid accumulation in macrophages as it decreases FABP4 expression at the mRNA level. The exact mechanism is not yet fully understood; it may deal with decreases in transcription or by promotion of mRNA degradation. For this reason, some authors suggest that metformin should also be perceived as a therapeutic agent for the prevention and targeting of atherosclerosis in metabolic syndrome [49].

FABP4 contributes to the accumulation of short-chain free fatty acids and suppresses the activity of relevant proteins in the phosphatidylinositol 30 -kinase (PI3K)-AKT signal pathway. Accordingly, FABP4 inhibits the presence of glucose oxidation and glycolysis and decreases the uptake and utilization of glucose in human organs, such as in muscles and the liver [50]. It is known that FABP4 can bind various intracellular fatty acids and probably mediates intracellular lipid trafficking between cellular compartments [51]. It might also modulate the availability and composition of fatty acids in muscles and adipose tissues [52]. In the myocytes and adipose tissue in mice, improved glucose homeostasis after the ablation of FABP4 was documented [53].

Increased FABP4 production could contribute to macrophage activity, possibly by activation of inflammatory pathways, resulting in inflammation [42]. An alternative hypothesis, more in line with the published evidence, is that the elevated circulating FABP4 reflects the increased cellular production in both adipocytes and macrophages in response to greater lipid availability, with increased Kupffer cell production of FABP4 triggering an increased inflammatory response [42,54].

FABP4 plays a crucial role in mediating the endoplasmic reticulum stress observed in macrophages upon lipotoxic signal exposure, which contributes to atherosclerosis, inflammation, and perhaps plaque vulnerability [55]. High levels of FABP4 may contribute to adverse prognosis via the induction of endoplasmic reticulum stress in macrophages and upregulation of pro-inflammatory cytokine production. Indeed, FABP4 modulates inflammatory responses in macrophages through a positive feedback loop involving c-Jun NH2-terminal kinases and activator protein-1 [42,56]. FABP4 is also implicated in modulating the eicosanoid balance by affecting both cyclooxygenase 2 (COX2) activity and leukotriene A4 (LTA4) stability, and upregulates uncoupling protein 2 (UCP2); all these processes influence macrophage function and adipose tissue inflammation. FABP4 can also interact with HSL and Janus kinase 2 (JAK2) [57] (Figure 1). Two different mechanisms of the latter action have been proposed. The first one includes direct protein–protein interactions between FABP4 and HSL and JAK2 [58,59]. FABP4 and HSL may also interact indirectly via two protein kinases A and G pathways [29].

**Figure 1.** Relationship between fatty acid-binding protein 4 and pathophysiology of type 2 diabetes mellitus. COX2- cyclooxygenase-2; eNOS- endothelial nitric oxide synthase; FA- fatty acid; HSLhormone sensitive lipase; JNK AP1-Jun N-terminal kinase-activator protein 1; LTA4-leukotriene A4; NF-κB-nuclear factor-kappa B; PGE2-prostaglandin E2; PPARγ–peroxisome proliferator-activated receptor γ; UCP2-uncoupling protein 2.

Several proinflammatory stimuli have been reported to upregulate FABP4 expression in macrophages, including oxidized low-density lipoproteins, toll-like receptor (TLR) agonists, and PPARγ agonists [60–63]. In this respect, lipopolysaccharide (LPS), a TLR4 ligand, stimulates FABP4 expression and the activated FABP4, reciprocally, enhances the LPS-TLR4 signaling-evoked JNK inflammatory pathways [56]. In addition, prolonged hyperglycemia has been shown to induce FABP4 expression in mesangial cells and trigger the release of proinflammatory cytokines [60,64] (Figure 2).

**Figure 2.** Fatty acid-binding protein 4 inducing factors. TLR- toll-like receptor; LPS- lipopolysaccharide; FABP4-fatty acid-binding protein 4.

### **6. Relationship between FABP4 and Insulin Resistance**

FABP4 is a cytoplasmic fatty acid chaperone clearly engaged in the onset of insulin resistance [35]. Studies in animal models suggest that FABP4 is important for glucose homeostasis [50]. Deletion of the FABP4 gene protected mice against insulin resistance as well as hyperinsulinemia associated with both diet-induced obesity and genetic obesity [45,50,53,54,65,66]. A reduced ability of adipocytes to take up and retain free fatty acids, leading to ectopic lipid accumulation, and abnormalities in the release of adipokines by adipocytes are critical factors for insulin resistance and the development of T2DM [66].

FABP4 was also detected in apoptotic granulosa cells in atretic antral follicles of the mouse ovary, which suggests a potential link to polycystic ovary syndrome (PCOS), which is known to be frequently associated with insulin resistance [11]. Expression of FABP4 mRNA in isolated granulosa cells was found to be higher in patients diagnosed with PCOS than in controls [67].

It has been proven that FABP4 is negatively correlated with the glucose-disposal rate (GDR) [68]. In non-DM subjects, serum FABP concentrations were negatively correlated with the mean rate of glucose infusion during the last 30 min of the clamp test, which reflects insulin sensitivity [69]. Nakamura et al. [35] also showed that circulating FABP4 concentrations were negatively correlated with GDR, which is a marker of insulin resistance in skeletal muscles in individuals with T2DM. On the contrary, FABP4 concentration was positively related to the insulinogenic index in non-diabetic participants. Wu et al. [70] reported that circulating FABP4 concentrations were correlated with glucose-stimulated insulin secretion in healthy controls.

It has been suggested that insulinotropic potential of FABP4 is similar to the effects of glucagon-like peptide-1 (GLP-1) [35]. To maintain glucose homeostasis, FABP4 may stimulate β cells and alter insulin secretion. Moreover, FABP4 presented a positive relationship with insulin secretion at an early stage in the non-diabetic group, which may be due to the fact that insulin secretion is damaged relatively early in T2DM.

Nakamura et al. [35] found the most pronounced negative correlation between FABP4 and GDR when compared to some other markers of insulin resistance or body composition in T2DM. FABP4 has been reported to have a negative correlation with GDR in type 1 diabetes mellitus, T2DM, and the controls of Asian Americans [71]. FABP4 represents an important molecule dealing with insulin resistance in T2DM.

## **7. Relationship between FABP4 and Gestational Diabetes Mellitus**

FABP4 and leptin are known to be involved in the pathophysiology of GDM and its long-term post-partum complications. Placental and non-placental origins of these adipokines are likely to contribute to insulin resistance and β-cell dysfunction [72].

Previous studies found that the serum FABP4 concentrations were significantly increased in women diagnosed with GDM when compared to the control group [65,73,74]. Zhang et al. [65] found that the fasting insulin and age adjusted FABP4 concentrations were significantly higher in the GDM group compared with the normal glucose tolerance participants in the mid and late stages of pregnancy. Furthermore, there was a significant increase in FABP4 from the second to third trimester in women with GDM [65]. Maternal FABP4 concentrations were obviously upregulated in the first trimester in women who later developed GDM [50]. Also, Li et al. [30] observed that FABP4 levels in the GDM group were significantly higher than those of controls, and noted that FABP4 was an independent risk factor for increased insulin resistance during pregnancy. Nevertheless, Ortega-Senovilla et al. [74] did not find any differences in FABP4 levels between the GDM and normal glucose tolerance groups when the FABP4 values were corrected with insulin. Meanwhile, the cited authors revealed significant correlations between maternal blood FABP4 levels and pre-pregnancy BMI values in both control and GDM pregnant patients [74].

Li et al. [32] explained the high circulating FABP4 levels in the maternal serum of pregnant GDM women by its additional release from placenta and adipocytes. Expression of FABP4 mRNA in the placenta and decidua of pregnant women with GDM is greater than that in normal organs [32]. Circulating FABP4 is associated with lipolysis and may aggravate insulin resistance compared to normal physiological insulin resistance during pregnancy [54].

Moreover, candidates for the placental hormones to induce FABP4 overexpression in the placenta and decidua in GDM include human placental lactogen and progesterone [32] as well as the combination of estrogen and progesterone [75]. Their concentrations increase continuously until term and may be associated with an increased insulin resistance along with the progressing gestational age [43,48]. Elevated placental hormones present in serum in GDM may increase the expression of FABP4 mRNA in adipocytes. The synergistic effects of FABP4 from the placenta and adipocytes can affect both metabolic and inflammatory pathways via adipocytes. These actions may play crucial roles in the development of insulin resistance and T2DM in the future lives of post-partum women [32].

On the other hand, FABP4 as expressed in human placental trophoblasts is a key regulator of trophoblastic lipid transport and accumulation during placental development. It has also been reported that maternal FABP4 levels are elevated in preeclampsia, even before the clinical onset of the disease [11]. The level of the second trimester plasma FABP4 in the preeclampsia GDM group was significantly higher than that of the patients with only GDM [76]. Besides, Wotherspoon et al. [77] revealed that increased second-trimester FABP4 levels independently predicted pre-eclampsia in women with type 1 diabetes mellitus.

The serum FABP4 levels were also associated with overweight in GDM patients. Ning et al. [11] concluded that serum FABP4 may be a potential biomarker in GDM diagnosis and is associated with overweight, insulin resistance, and TNF-α in GDM patients.

Our previous study showed that the serum FABP4 levels were significantly higher in the GDM group in the early puerperium in comparison with healthy mothers and women with excessive gestational weight gain [2]. Based on our findings and previous studies, it appears that increased circulating FABP4 concentrations can persist in GDM patients after delivery and might contribute to the increased risk of T2DM and metabolic syndrome. On the other hand, evaluation of FABP4 may be used as a predictive marker for mothers with a history of GDM [2].

In a follow-up study of women six years after GDM, Svensson et al. [78] identified three factors—all related to the adipose tissue and body composition—that, independently of body mass index (BMI) and ethnicity, may increase the risk of progression to T2DM following GDM. These factors included increased serum FABP4 levels, weight gain following index pregnancy, and a lower proportion of fat-free mass. High BMI and abdominal fat distribution were also associated with the development of T2DM after GDM [78].

## **8. Relationship between FABP4 and T2DM and its Complications**

A 10-year prospective study proved that high levels of FABP4 at baseline independently predicted the development of T2DM [33]. Serum FABP4 concentrations have been reported to be associated with inadequate glucose control in T2DM [33]. Increased FABP4 levels are also linked to the early presence of metabolic syndrome components, as well as inflammation and oxidation markers in T2DM subjects [44]. Inflammation and oxidative stress are suggested to play a role in the pathogenesis of complications of T2DM [10,79,80].

A previous study showed that the concentrations of FABP4 were negatively associated with endothelial function in T2DM [81]. Another study identified a correlation between FABP4 and impaired endothelial function in diabetes, which lead to an increased cardiovascular risk [82]. Furthermore, the prognostic value of FABP4 for kidney damage [83] and cardiac contractile dysfunction [84] in patients with T2DM have also been proposed. In addition, previous studies suggest that increased FABP4 synthesis in atherosclerotic plaques is associated with disease severity [85,86]. Holm et al. [87] reported that FABP4 was linked to atherogenesis, plaque instability, and adverse outcomes in patients with carotid atherosclerosis and acute ischemic stroke. Two other studies have shown the association of enhanced FABP4 expression within human carotid atherosclerotic lesions with poor prognosis [85,86]. There is available data to suggest that higher levels of FABP4 are also associated with elevated cardiovascular diseases mortality among men with T2DM [42,88].

#### **9. Relationship between FABP4 and Diabetic Retinopathy**

FABP4 shows clinically relevant potential as a novel predictor of diabetic retinopathy (DR). According to some authors, strict glycemic control and more frequent retinal examination should be recommended for the T2DM patients with the detected highest quartile range of FABP4 [10].

The impact of FABP4 on atherosclerosis seems to be attributed to its action in macrophages [89]. Circulating FABP4 induces insulin resistance, which is an independent biomarker of proliferative retinopathy [90]. Lipopolysaccharides (LPS) stimulate FABP4 transcription through JNK, which in turn induces c-Jun recruitment to a highly conserved activator protein-1 recognition site within the proximal region of the FABP4 promoter [56]. LPS-binding protein is involved in the immune response triggered by inflammatory injury characteristic of DR. Moreover, FABP4 is an obligatory mediator coupling toxic lipids (i.e., saturated fatty acids) to endoplasmic reticulum stress in macrophages in vitro and in vivo [55]. Interestingly, endoplasmic reticulum stress represents an initial event in retina pathogenesis in diabetes [10,91].

## **10. Relationship between FABP4 and Diabetic Nephropathy**

In the past decades, several biomarkers have emerged for the detection of early diabetic nephropathy (DN) besides the glomerular filtration rate (GFR) and urine albumin-to-creatinine ratio (UACR). Among them, FABP4 has attracted increased attention [92]. FABP4 was also reported at increased concentrations in nondiabetic as well as T2DM patients with end-stage renal disease [93]. Yeung et al. [94] reported that serum levels of FABP4 had a significantly inverse relationship with the estimated GFR (eGFR) and was independently associated with macrovascular complications and DN staging classified by albuminuria.

Cabré et al. [73] reported that FABP4 was independently associated with eGFR in T2DM patients with eGFR <sup>≥</sup> 60 mL/min/1.73 m<sup>2</sup> . Ni X et al. [92] found that serum FABP4 along with UACR or a panel of biomarkers might be more sensitive for the detection of early DN. Serum FABP4 had an inverse correlation with GFR and could be an independent predictor for early DN [92].

The mechanisms behind the elevation of FABP4 in patients with diabetic kidney disease are not yet fully understood. It is known that FABP4 is abundantly expressed in adipocytes, macrophages, and endothelial cells [92]. Firstly, it is suggested that, during the early stage of DN, the accumulation of active macrophages is more evident in the kidney because of the increased oxidative stress and chronic inflammation, which consequently induces greater expression of serum FABP4 [44,94]. Secondly, damage to glomeruli and tubulointerstitium might result in both decreased glomerular filtration and increased tubular reabsorption, leading to an increase in FABP4 in the circulation [83]. Okazaki et al. [16] reported that urinary excretion of FABP4 was associated with the progression of proteinuria and renal dysfunction in healthy subjects. The cited authors suggested that the urinary FABP4 reflects damage of the glomerular with the hypothesis proposed by Tanaka et al. [13] that the main source of the urinary FABP4 is derived from ectopic expression of glomerular FABP4 rather than increased adiposity and that locally increased FABP4 in the glomerulus affects renal dysfunction.

Toruner et al. [93] found that serum FABP4 levels were independently and positively associated with the albumin excretion rate in patients with T2DM, suggesting an involvement of the increased serum FABP4 levels in the occurrence and development of microalbuminuria among patients with T2DM. Furthermore, researchers from Hong Kong [94] also documented that among patients with diabetes mellitus, serum FABP4 levels were shown to be independently associated with the severity of nephropathy. Their findings raised the possibility that FABP4 might be used as a serum biomarker for stratifying nephropathy stages in patients with T2DM. The serum FABP4 level can be used as an indicator of microalbuminuria not only in diabetic patients with early stage disease, but also for hyperglycemic individuals before the onset of diabetes [95].

## **11. Relationship between FABP4 and Non-Alcoholic Fatty Liver Disease**

There are multiple reports available that clearly link metabolic disorders, including insulin resistance and T2DM, to non-alcoholic fatty liver disease (NAFLD), which nowadays represents the most frequent liver disease worldwide. Its incidence has been estimated at 20–30% in the populations of Western countries. Approximately 70% of T2DM and obese subjects present some extent of NAFLD. This abnormal lipid accumulation in the liver is considered a significant causative factor of cancerogenesis resulting in hepatocellular carcinoma (HCC) [96].

At present, HCC represents the third leading cause of cancer-related mortality. Interestingly, except for the well-known risk factors of malignancy, such as viral hepatitis and alcohol, a higher incidence of HCC was observed in T2DM patients. In the analysis, obese subjects and patients with metabolic syndrome were also at a higher risk of developing cancer [96]. Additionally, HCC treatment outcomes of all those patients appeared worse with T2DM, representing a prognostic predictor of the increased risk of mortality [96].

In a study by Thompson et al., the FABP4 expression was indeed proven to be significantly increased in animal models of obesity promoted HCC [97]. These findings are consistent with the previous report, which concluded that PPARγ is considered a tumor suppressor gene, whereas FABP4 plays a role in tumorigenesis [35].

## **12. Conclusions**

FABP4, an intracellular lipid chaperone, is highly expressed in adipocytes and macrophages [98]. Abnormalities in the level of FABP4 have been correlated with the development of adiposity, oxidative stress, and atherosclerosis [40]. FABP4 is strongly involved in glucose and lipid metabolism, inflammation, and insulin resistance [76]. In one of the most recent systematic reviews by Bellos et al., targeting the potential association of 10 novel adipokines (i.e., apelin, chemerin, FABP4, fibroblast

growth factor-1, monocyte chemoattractant protein-1, nesfatin-1, omentin-1, resistin, vaspin, and visfatin) with GDM, only FABP4 was found to be the most promising predictor of this metabolic pregnancy complication [99].

In fact, a large number of studies have shown that plasma levels of FABP4 are increased in obesity and T2DM, and that circulating FABP4 concentration correlated with clinical outcomes, such as body mass index, insulin resistance, and dyslipidemia [33]. Since the plasma-circulating FABP4 can modulate the function of several types of cells, it becomes extremely interesting to try to develop effective therapeutic strategies against the pathogenesis of metabolic and vascular diseases in this respect [33].

Furthermore, several studies have shown that serum FABP4 levels of GDM patients were higher than those found in women with normal pregnancies [30,65,73,76]. According to the current state of scientific knowledge, females with a previous history of GDM are much more prone to suffering from T2DM, obesity, and metabolic syndrome in the future. There is a huge need to use current research results regarding improved GDM management strategies, including primary prevention for mothers who are at risk of developing subsequent complications. Considering previous findings, it seems that FABP4 may be used as a predictive marker for mothers with a history of GDM. Further studies should focus on the evaluation of FABP4 in the pathogenesis of T2DM following GDM.

**Author Contributions:** Conceptualization, M.T.; Data curation, M.T., J.P.-M., Z.K.-T.; Writing-review and editing, ˙ M.T., Z.K.-T.; Visualization, J.P.-M.; Supervision, B.L.-G., J.M. ˙

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

**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/).

*Review*

## **Obesity-Induced Adipose Tissue Inflammation as a Strong Promotional Factor for Pancreatic Ductal Adenocarcinoma**

## **Hui-Hua Chang and Guido Eibl \***

Department of Surgery, David Geffen School of Medicine at UCLA, Los Angeles, CA 90095, USA **\*** Correspondence: GEibl@mednet.ucla.edu; Tel.: +1-310-825-5743

Received: 14 June 2019; Accepted: 2 July 2019; Published: 3 July 2019

**Abstract:** Pancreatic ductal adenocarcinoma (PDAC) is expected to soon become the second leading cause of cancer related deaths in the United States. This may be due to the rising obesity prevalence, which is a recognized risk factor for PDAC. There is great interest in deciphering the underlying driving mechanisms of the obesity–PDAC link. Visceral adiposity has a strong correlation to certain metabolic diseases and gastrointestinal cancers, including PDAC. In fact, our own data strongly suggest that visceral adipose tissue inflammation is a strong promoter for PDAC growth and progression in a genetically engineered mouse model of PDAC and diet-induced obesity. In this review, we will discuss the relationship between obesity-associated adipose tissue inflammation and PDAC development, with a focus on the key molecular and cellular components in the dysfunctional visceral adipose tissue, which provides a tumor permissive environment.

**Keywords:** obesity; adipose tissue inflammation; pancreatic ductal adenocarcinoma

## **1. Obesity and Pancreatic Ductal Adenocarcinoma**

Pancreatic cancer, of which pancreatic ductal adenocarcinoma (PDAC) accounts for the vast majority, is an aggressive disease with an overall 5-year survival rate of ~8% [1]. Currently, as the third leading cause of cancer mortality in both men and women, pancreatic cancer is expected to soon become the second leading cause of cancer-related deaths in the United States [2]. One of the recognized risk factors for PDAC and several other gastrointestinal cancers is obesity [3–8]. A recent analysis by the National Institutes of Health reported that 16.9% of all PDAC cases in the Unites States are estimated to be attributable to excess body weight (in comparison, the proportion of PDAC cases attributable to cigarette smoking was 10.2%) [9]. Besides promoting tumorigenesis, obesity might also complicate the treatment of established cancers [10], not only by altering the pharmacokinetics and pharmacodynamics of administered anti-cancer drugs [11], but also by modulating the tumor microenvironment (TME) [12]. Notably, worse prognosis is reported in patients who are either overweight or obese at the time of PDAC diagnosis [13,14]. Considering the current obesity pandemic, there is great interest in deciphering the underlying biological basis of this compelling epidemiological link.

The mechanisms underlying the obesity–cancer link are likely multifactorial, with metabolic perturbations and inflammation being the current research focus. Other mechanisms, such as alterations in the gut microbiome, changes in gastrointestinal peptides, and sex hormones, certainly play an important role as well. For example, obesity-associated insulin resistance and hormonal changes (e.g., compensatory hyperinsulinemia and elevated insulin-like growth factors (IGFs)) are among the various driving mechanisms postulated to explain the obesity–cancer link [15]. In accordance with this notion, type 2 diabetes, hyperinsulinemia, and increased circulating IGF-1 are established risk factors for pancreatic and other types of cancers [16–20]. In addition to the insulin/IGF axis, increasing

attention has focused on adipose tissue (AT) dysfunction characterized by a pro-inflammatory state, which can contribute not only to the pathogenesis of insulin resistance [21] but also to the development and promotion of cancer. Utilizing a genetically engineered mouse model of PDAC and diet-induced obesity, our own data suggest that inflammatory responses in the AT surrounding the pancreas is a strong promotional factor for PDAC growth and progression [22]. AT inflammation associated with increased abdominal adiposity, which involves various cellular and extracellular components (e.g., infiltrated macrophages and cytokines), can shape the peri-tumoral micro-environment favorable for PDAC development, both locally and systemically. More detailed cellular and molecular characteristics of dysfunctional and inflamed AT associated with the obese state, and their roles in tumor development, will be discussed in the later sections.

## **2. The Role of Adipose Tissue Inflammation in Obesity-Promoted Tumor Development**

## *2.1. Classification and Physiological Roles of AT*

White adipose tissue (WAT), the most abundant type of AT in adult humans, stores excess energy in the form of triglycerides, which can be mobilized via lipolysis to release fatty acids for usage by other tissues. Based on anatomical locations, AT can also be organized into different types of depots: Subcutaneous and visceral. These depots are heterogeneous with regard to their morphological, molecular, and metabolic profiles [23], and exhibit major gender differences. Visceral adipose tissue (VAT), which encompasses omental, mesenteric, and other intra-abdominal fat pads, is considered of great relevance in obesity-related comorbidities, as increased VAT area is more closely associated with metabolic dysfunction and cancer than subcutaneous fat [7,23–25]. Surgical removal of VAT prior to high fat diet feeding is shown to prevent obesity-induced insulin resistance in mice [26]. Since VAT is in close proximity to digestive organs, including the pancreas, and is the primary source of systemic and local chronic inflammation during obesity [27], it is particularly pertinent to the topics covered in this review. For example, VAT (but not subcutaneous fat) and pancreatic fatty infiltration are significantly correlated with pancreatic intraepithelial neoplasia (PanIN), the precursor lesions of PDAC [28]. Interestingly, the increased PDAC incidence in obese conditional KrasG12D (KC) mice was largely observed in male animals [29], which correlated with a greater expansion of VAT in obese male mice, as compared to obese female mice that preferentially gained subcutaneous fat [22].

Besides being a key regulator of energy homeostasis, AT is regarded as an important endocrine organ secreting various factors collectively known as "adipokines", and growingly recognized as a prominent immune site harboring macrophages and other types of immune cells. Overall, this dynamic organ is composed of mature adipocytes and the "stromal-vascular fraction" comprising a mixture of mesenchymal, hematopoietic, and endothelial cell types. Under homeostatic conditions, the cells with diverse functional roles act cooperatively to support the metabolic and physiological functions of the organ. Specifically, healthy AT maintains an anti-inflammatory phenotype characterized by the secretion of anti-inflammatory adipocytokines (e.g., adiponectin) and enrichment of regulatory immune cells, such as M2 macrophages, eosinophils, Th2 leukocytes, and myeloid-derived suppressor cells that limit AT inflammation. In obesity, on the other hand, the regulatory anti-inflammatory network in AT is disrupted and skewed to a pro-inflammatory state, with substantial changes in adipokines and the number and function of AT immune cells (see next section), leading to sustained chronic inflammation.

## *2.2. Obesity-Associated AT Inflammation*

Chronic over-nutrition is manifested as adipose tissue expansion via hyperplasia (increase in adipocyte numbers) and/or hypertrophy (enlargement of existing adipocytes). Adipocyte hyperplasia, which occurs more readily in subcutaneous depots than in VAT, is normally associated with no/weak inflammation and improved insulin sensitivity and therefore regarded as a "metabolically healthy" expansion of AT. Conversely, visceral adipocyte hypertrophy, which is often associated with cellular

stress, pro-inflammatory responses, and impaired insulin sensitivity, is recognized as unhealthy pathological expansion and an important contributor to the obesity-related metabolic disorders [30]. When hypertrophic adipocytes are overloaded with lipids, they become hypoxic and undergo stress responses and cell death, which are correlated with a pro-inflammatory secretome, increased pro-inflammatory immune cell infiltration, and ectopic lipid deposition in other metabolically active organs [31], each of which may contribute to progressive insulin resistance and the pathogenesis of type 2 diabetes.

Adipocyte hypertrophy is associated with a shift in the adipokine secretion profile [32]. Specifically, there is a significant increase in the production of pro-inflammatory factors, such as leptin, tumor necrosis factor-α (TNF-α), interleukin 6 (IL-6), monocyte chemoattractive protein-1 (MCP-1), and plasminogen activator inhibitor-1 (PAI-1). On the contrary, the release of adiponectin, a major anti-inflammatory adipokine, is reduced with increased adiposity in response to AT inflammation [33]. These adipocyte-derived mediators, which link adipocytes with adipose-resident immune cells, lead to an altered composition of the adipose stromal-vascular fraction characterized by reduced numbers of anti-inflammatory cells (e.g., eosinophils, T-regs, Th2 cells) and an increased recruitment of pro-inflammatory Th1 cells, neutrophils, IFNγ-producing natural killer (NK) cells, and monocytes/macrophages [34,35].

The detailed landscape of immune cell changes in AT during obesity is extremely complex and virtually every immune cell type plays a role in obesity-associated metabolic diseases [36]. However, the most prominent feature is monocyte infiltration and differentiation into pro-inflammatory macrophages, which surround necrotic adipocytes and form the histological hallmark of AT inflammation called "crown-like structures" [37]. AT macrophages (ATM), which can account for up to 50% of the cellular content in the obese AT, are the primary source of pro-inflammatory cytokines (e.g., IL-6, TNF-α, IL1-β), and have been shown to be centrally important in obesity-related metabolic dysfunction [38–41]. The increase in ATM abundance during obesity is crucial for exacerbating the inflammatory response and is mainly caused by the recruitment of circulating bone marrow-derived monocytes through an MCP-1 (CCL2)/CCR2- dependent mechanism [42–44]. While resident ATMs in the lean state display an anti-inflammatory M2-polarized phenotype, obesity-associated ATM polarize towards the pro-inflammatory M1 state in response to the disturbed immune and cytokine profile and other environmental cues [35]. Interestingly, the oversimplified M1/M2 dichotomy is revised by the recent discovery of the "metabolically-activated" type ATM, which is metabolically activated by excessive saturated free fatty acids (FFAs), hyperinsulinemia, and hyperglycemia and is functionally distinct from the classical M1 macrophage activated by bacterial challenge [45]. As mentioned above, it is well established that AT inflammation and ATMs play a key role in the pathogenesis of metabolic syndromes, mainly via secreted mediators. However, much less is known about their role in cancer, especially PDAC.

Besides intrinsic danger signals, exogenous factors, such as gut microbiota-derived pathogen-associated molecular patterns (PAMPs), are also implicated in the reprogramming of immune cells toward pro-inflammatory subtypes in obese AT. Diet-induced obesity is associated with a robust change in gut microbiota composition [46,47] and impaired intestinal barrier integrity, correlating with an increased expression of pro-inflammatory cytokines in the AT [48,49]. Induced by gut dysbiosis and leaky intestinal tight junctions, circulating levels of lipopolysaccharide (LPS; a cell wall component of gram-negative bacteria) are found to be 2 to 3-fold higher in obese humans and rodents as compared to their lean counterparts, a phenomenon known as "metabolic endotoxemia" [50,51]. Elevated LPS levels in tissues, e.g., AT, can stimulate the innate immune response by activating Toll-like receptor 4 (TLR4) on macrophages, leading to the induction of NF-κB and pro-inflammatory cytokine production. Using germ-free mouse models, microbiota-derived LPS has been shown to promote macrophage infiltration and M1 polarization in WAT, independent of its role in regulating fat mass and glucose metabolism [52]. Additionally, mice fed a diet rich in saturated lipids (lard) exhibited reduced insulin sensitivity, enhanced WAT inflammation, and

increased activation of TLRs compared with mice fed a fish oil diet, and such phenotypic differences were in part attributable to differences in gut microbiota composition [53]. These authors further demonstrated that lard-induced ATM accumulation was mediated by adipocyte-produced MCP-1, which was induced by microbial-derived factors through activation of TLR4 and the downstream adaptor molecules, MyD88 and TIR-domain-containing adapter-inducing interferon-β (TRIF) [53]. Together, these data strongly suggest that intestinal microbiota contribute to changes in immune cell function and AT inflammation during diet-induced obesity.

## *2.3. AT Inflammation in PDAC Development: The Molecular Links*

Inflammation has long been associated with the development of cancer, including PDAC [54,55]. In a mouse model of obesity-promoted KrasG12D-driven pancreatic neoplasia, with obesity induced by a high-fat high-calorie diet (HFCD), we observed a higher percentage of advanced PanIN lesions [56] and increased PDAC incidence [29], accompanied by prominent signs of inflammatory immune cell infiltration and elevation of pro-inflammatory cytokines in the pancreas [56] as well as the peri-pancreatic AT [22]. In addition, deficiency in hormone-sensitive lipase (HSL) was associated with AT inflammation and an acceleration of pancreatic cancer development in conditional KrasG12D mice [57]. Cumulating evidence suggest a model in which diet-induced obesity leads to a robust inflammatory response in visceral peri-pancreatic (or intra-pancreatic [28]) adipose tissue, which may in turn promote inflammation and neoplastic progression in the neighboring pancreas, via soluble factors secreted from adipocytes and/or adipose-infiltrating immune cells.

## 2.3.1. Leptin, Adiponectin, and Other Adipokines

Obese individuals are characterized by dramatic changes in adipokine production, and several of these adipokine changes are associated with cancer development. The most prominent examples are the increased levels of the pro-inflammatory leptin [58,59] and the reduced levels of the anti-inflammatory adiponectin [60,61]. In several cancer models, leptin has been shown to promote cell proliferation, migration, and invasion through activation of the Janus Kinase/Signal Transducer and Activator of Transcription (JAK/STAT) pathway and the subsequent oncogenic phosphatidylinositol 3-kinase (PI3K)/Akt and extracellular signal-regulated kinase (ERK) signaling, leading to increased expression of inflammatory cytokines, angiogenic factors, and apoptotic proteins [61–64]. Adiponectin, on the other hand, seems to exert antagonistic effects on tumor growth and progression through activation of AMPK and protein tyrosine phosphatase 1B (PTP1B) [62,65]. However, the roles of circulating leptin and adiponectin in PDAC development remain debatable [66], as some studies found higher adiponectin/leptin ratios associated with pancreatic cancer [67,68]. Also, adiponectin transcription is activated by nuclear receptor 5A2 (NR5A2), an important risk factor identified by a genome-wide association study (GWAS) for pancreatic cancer [69]. Further, in our recent study using HSL deficient conditional KrasG12D mice, decreased plasma leptin levels were observed in animals with advanced PDAC [57], suggesting that leptin may not play a central mechanistic role in PDAC promoted by AT inflammation. Besides systemic effects, adipokines generated by adjacent adipose tissue might act locally directly on tumor cells [70,71] or modulate inflammation and immunity in the pancreatic TME [72,73]. In that regard, it is noteworthy that lipocalin 2, a novel adipokine elevated in obesity subjects [74], has been shown to promote obesity-associated PDAC by stimulating the pro-inflammatory response in the TME [75].

## 2.3.2. Pro-Inflammatory Cytokines and Chemokines

Although the events initiating obesity-associated AT inflammation are not completely understood, increased pro-inflammatory factors produced by adipocytes and infiltrating ATMs or other immune cells in the dysfunctional AT create a systemic and local inflammatory environment. Some of these factors elevated in obese subjects, such as tumor-promoting cytokines (e.g., TNF-α and IL-6) and monocyte chemo-attractants (e.g., MCP-1 [76,77]), are implicated in PDAC development and progression.

Increased expression of TNF-α is observed in obese AT [40,78], where this key pro-inflammatory cytokine is thought to act locally as an autocrine or paracrine cytokine, since systemic TNF-α levels do not necessarily reflect local changes in TNF-α concentrations [78,79]. A central role for TNF-α in obesity-associated early pancreatic tumor promotion has been demonstrated by genetic deletion of tumor necrosis factor receptor 1A (*TNFR1*) [80]. Functionally, TNF-α activates JNK and NF-κB pathways, which are known to promote cell proliferation, invasion, and metastasis in various cancers, including human PDAC, where NF-κB is constitutively activated [81–83]. NF-κB activation by TNF-α can crosstalk with Notch signaling, which cooperates with oncogenic KRAS to promote PDAC progression [84]. In addition, activated NF-κB can stimulate the production of chemokines (such as MCP-1), by adipocytes and pre-adipocytes, leading to increased infiltration of pro-inflammatory macrophages [85,86]. Further, NF-κB is a critical transcription factor of M1 macrophages, regulating the expression of a variety of inflammatory genes, including those encoding IL-6, IL-1β, MCP-1, and cyclooxygenase-2 (COX-2), shown to play important roles in obesity-related PDAC promotion [55]. Adipose expression of IL-6, another key modulator in inflammation-associated tumorigenesis [87,88], is also upregulated in obese individuals [40,79]. IL-6 exerts its pro-tumorigenic and pro-invasive activities through the activation of the STAT3 signaling pathway, which is essential for PDAC development in the Kras-driven mouse model [89–91] and observed in human PDAC [92,93]. The effects of IL-1 are similar to those of TNF-α, and IL1β gene promoter single nucleotide polymorphisms (SNPs) are linked with pancreatic cancer risk [94]. These key cytokines, whose production is increased with obesity, also play critical roles in insulin resistance [78,95], adipokine regulation [96], and tumor stroma modulation [12,97,98], indirectly influencing tumor cell growth. Importantly, these inflammatory cytokines are associated with a poor prognosis in PDAC [99–101]. Overall, PDAC progression is characterized by convergent activation of inflammatory transcriptional factors (i.e., NF-κB and STAT3) [102], which can be stimulated by the pro-inflammatory cytokines whose secretion is markedly enhanced in dysfunctional AT with increased adiposity (especially visceral adiposity).

## 2.3.3. Microbiota and Obesity-Promoted PDAC

The gut microbiome could be closely related to obesity-associated chronic inflammation though TLR activation in the immune cell compartment, amplifying inflammatory responses and cytokine production that can fuel tumor growth. Although the effects of gut microbiota on tumorigenesis are mostly highlighted in other inflammation-driven, obesity-related gastrointestinal malignancies (e.g., colon and liver cancer) [103], it is well-documented that dysbiosis of oral bacteria is associated with increased risk for pancreatic cancer [104–106]. In experimental mouse models, LPS and TLR4 activation could enhance the severity of acute pancreatitis [107,108]. Also, ligation of TLRs, possibly through NF-κB and MAP kinase pathways, are shown to exacerbate pancreatic fibro-inflammation and accelerate Kras-driven pancreatic tumorigenesis [109–111]. However, less is known regarding how obesity-induced gut dysbiosis and AT inflammation contribute to PDAC development. Based on evidence linking the microbiome and inflammation-associated PDAC, and the notion that metabolic endotoxemia can be a strong promoter of ATM infiltration and M1 polarization, gut bacteria are likely to play a role in the interface of obesity, AT inflammation, and PDAC. Regardless, these interactions need to be further tested in the context of obesity.

## 2.3.4. Other Factors Associated with AT Dysfunction in Obesity

Obesity and the inflammatory environment are typically associated with oxidative stress, another important feature of dysfunctional AT, conferring genetic instability that can promote the acquisition of oncogenic mutations in neighboring pancreatic cells [112,113]. The elevated oxidative stress can result from enriched reactive oxygen species (ROS) and reactive nitrogen intermediates (RNIs) generated by inflammatory cells, as well as mitochondrial dysfunction and lipid oxidation related to increased release of surplus free fatty acids (FFA) and ectopic fat deposition. Interestingly, in our previous study involving a mouse model of obesity-promoted PDAC, exome sequencing of advanced pancreatic

intraepithelial neoplasia (PanIN) lesions identified numerous genetic variants unique to the HFCD (and diet-induced obesity) group [29]. These genetic alterations are found in genes involved in oncogenic pathways that are commonly implied in PDAC, including the insulin and PI3K/Akt pathway.

## **3. Interventional Perspectives**

Based on the increasingly recognized link between obesity-induced AT inflammation and PDAC development, AT inflammation has become an intriguing target for PDAC interception. There are several possibilities to disrupt the promoting effects of (obesity-induced) AT inflammation on PDAC growth and progression. Besides general health-promoting strategies aimed at reducing obesity (e.g., weight reduction), which certainly will positively impact AT inflammation, a detailed understanding of the mechanistic link between AT inflammation and PDAC will lead to targeted approaches. Inhibiting major pro-inflammatory mediators secreted by inflamed AT or blocking cytokine receptors in the pancreas might be a promising approach. For example, anti-TNF-α antibodies and antagonists of IL-1 and IL6 receptors exhibit anti-tumor effects in pre-clinical models of PDAC [101,114–117]. However, the blockade of these cytokines, although proven to be effective in treating other inflammatory diseases, has had limited success in patients of metabolic diseases [118,119] or PDAC [120,121]. Given the complexity of the immune cell changes and inflammatory responses in the AT during obesity, with the multitude of inflammatory mediators, targeting single or even few cytokines simultaneously might not be the best strategy. Blocking major pathways driving the inflammatory response in obese AT seems to be preferable. An intriguing approach with clear and rapid translational potential is the use of Food and Drug Administration (FDA)-approved drugs, e.g., statins.

Statins are lipid-lowering drugs that inhibit HMG-CoA reductase, which plays a central role in the production of cholesterol. High cholesterol levels have been associated with cardiovascular disease (CVD). Statins have been found to reduce CVD and mortality in those who are at high risk [122]. The evidence is strong that statins are effective for treating CVD in the early stages of a disease (secondary prevention) and in those at elevated risk but without CVD (primary prevention). There is intense interest in repurposing statins in cancer [123]. A recent meta-analysis of 95 cohorts including 1,111,407 individuals concluded that statin therapy has potential survival benefit for patients with malignancy [124]. Several large studies support the preventive effect of statins in selected cancers. Although some early epidemiologic studies reported no beneficial effects of statins on PDAC risk [125–128], several recent studies associated statins with a significantly lower risk of cancer, including PDAC [129–134]. In a large case-control study, statin use was associated with a 34% reduced PDAC risk, with a stronger association in male subjects [133]. In a case control study of 408 patients with PDAC and 816 matched controls, statin use was associated with reduced PDAC risk (odds ratio: 0.61; 95% CI: 0.43–0.88) [135]. In this study, the protective effect of statin was dose-dependent and stronger in obese patients. Very recently, a meta-analysis of 26 studies containing more than 3 million participants and 170,000 PDAC patients found a significant decrease in PDAC risk with statin use (RR: 0.84; 95% CI: 0.73–0.97) [136]. In patients with chronic pancreatitis, statins lowered the risk of progression and pancreatic cancer [137]. In addition, statins have been shown to improve survival after resection of early PDAC, indicating a potential benefit for secondary prevention [138–140].

Besides their cholesterol lowering effects, statins are known to have anti-inflammatory properties [141]. Statins attenuate AT inflammation in various human conditions and experimental models [142–145]. The improvement of AT inflammation by statins correlated with a reduction of inflammatory cell infiltration. In addition, statins have been found to decrease circulating MCP-1 levels, which are elevated in obese patients [76,146]. Importantly, statins elicited an anti-inflammatory M2-polarization of macrophages [145,147]. In our own study using a genetically engineered mouse model of pancreatic cancer promoted by diet-induced obesity, we found that oral administration of simvastatin attenuated early pancreatic neoplastic progression [148], which was accompanied by a marked reduction of inflammation in the VAT (unpublished; Figure 1).

**Figure 1.** Representative histology of mesenteric adipose tissue of conditional KrasG12D mice fed an obesogenic high fat high calorie diet (HFCD) for 3 months supplemented without (upper panel) or with simvastatin (sim; lower panel). Quantification of crown-like structures demonstrates significant elevation of adipose tissue inflammation in obese HFCD-fed mice, which was abrogated by **Figure 1.** Representative histology of mesenteric adipose tissue of conditional KrasG12D mice fed an obesogenic high fat high calorie diet (HFCD) for 3 months supplemented without (upper panel) or with simvastatin (sim; lower panel). Quantification of crown-like structures demonstrates significant elevation of adipose tissue inflammation in obese HFCD-fed mice, which was abrogated by simvastatin (sim). \*: *p* < 0.01 vs. CD; #: *p* < 0.01 vs. HFCD

simvastatin (sim). \*: *p* < 0.01 vs. CD; #: *p* < 0.01 vs. HFCD

prevention/interception, in particular in the setting of obesity.

Mechanistically, statins inhibit 3-hydroxy-3-methlyglutaryl coenzyme (HMG-CoA) reductase, the rate-limiting enzyme that controls the conversion of HMG-CoA to mevalonic acid. Mevalonic acid is a precursor of isoprenoids, which are essential for the post-translational modification of small G-proteins, e.g., Ras homolog gene family, member A (RhoA), cell division control protein 42 homolog (Cdc42), and Ras-related C3 botulinum toxin substrate 1 (Rac1), to attach to lipid membranes. Activation of Rho family members GTPases (guanosine triphosphate hydrolases) modulate the actin cytoskeleton, which has been shown to be an important regulator for macrophage polarization [149]. Our own unpublished data showed that lipophilic statins significantly attenuate pro-inflammatory signaling in macrophage cell lines through inhibition of the mevalonic acid/protein prenylation/actin cytoskeleton pathway. Besides their beneficial effects on AT inflammation, statins also have significant effects on PDAC cells [150–152]. Overall, statins show great promise for PDAC Mechanistically, statins inhibit 3-hydroxy-3-methlyglutaryl coenzyme (HMG-CoA) reductase, the rate-limiting enzyme that controls the conversion of HMG-CoA to mevalonic acid. Mevalonic acid is a precursor of isoprenoids, which are essential for the post-translational modification of small G-proteins, e.g., Ras homolog gene family, member A (RhoA), cell division control protein 42 homolog (Cdc42), and Ras-related C3 botulinum toxin substrate 1 (Rac1), to attach to lipid membranes. Activation of Rho family members GTPases (guanosine triphosphate hydrolases) modulate the actin cytoskeleton, which has been shown to be an important regulator for macrophage polarization [149]. Our own unpublished data showed that lipophilic statins significantly attenuate pro-inflammatory signaling in macrophage cell lines through inhibition of the mevalonic acid/protein prenylation/actin cytoskeleton pathway. Besides their beneficial effects on AT inflammation, statins also have significant effects on PDAC cells [150–152]. Overall, statins show great promise for PDAC prevention/interception, in particular in the setting of obesity.

## **4. Conclusions**

**4. Conclusion**  There is clear evidence that obesity increases the risk of pancreatic cancer. Obesity-induced AT inflammation is increasingly recognized as an important driver of the pathophysiologic process underlying the obesity–PDAC link. Dramatic changes in immune cell number and function in obese AT, especially VAT, with accompanying increased production of various pro-inflammatory cytokines may thereby promote the proliferation and survival of transformed cells in the adjacent pancreas. Although numerous immune cells have been shown to be involved in maintaining sustained AT inflammation in obesity, AT macrophages clearly play a central and critical role in orchestrating the inflammatory response. Given the importance of AT inflammation as a strong promotional driver of obesity-associated PDAC, strategies to inhibit obesity-induced AT inflammation are an intriguing approach. In this regard, the re-purposing of FDA-approved drugs aimed at reducing obesity-induced AT inflammation holds great translational significance. Recent epidemiologic data and preclinical evidence strongly suggest a role of statins, especially lipophilic statins, as promising drugs for PDAC prevention/interception. Besides their direct effects on PDAC cells, current data suggest that the anti-cancer properties of statins might at least partially include There is clear evidence that obesity increases the risk of pancreatic cancer. Obesity-induced AT inflammation is increasingly recognized as an important driver of the pathophysiologic process underlying the obesity–PDAC link. Dramatic changes in immune cell number and function in obese AT, especially VAT, with accompanying increased production of various pro-inflammatory cytokines may thereby promote the proliferation and survival of transformed cells in the adjacent pancreas.Although numerous immune cells have been shown to be involved in maintaining sustained AT inflammation in obesity, AT macrophages clearly play a central and critical role in orchestrating the inflammatory response. Given the importance of AT inflammation as a strong promotional driver of obesity-associated PDAC, strategies to inhibit obesity-induced AT inflammation are an intriguing approach. In this regard, the re-purposing of FDA-approved drugs aimed at reducing obesity-induced AT inflammation holds great translational significance. Recent epidemiologic data and preclinical evidence strongly suggest a role of statins, especially lipophilic statins, as promising drugs for PDAC prevention/interception. Besides their direct effects on PDAC cells, current data suggest that the anti-cancer properties of statins might at least partially include their anti-inflammatory effects on obese AT.

their anti-inflammatory effects on obese AT. **Author Contributions:** Writing—original draft preparation, H.-H.C; writing—review and editing, G.E. **Author Contributions:** Writing—original draft preparation, H.-H.C.; writing—review and editing, G.E.

**Funding:** This research was funded by the National Institutes of Health, National Cancer Institute, P01 CA163200, and the Hirshberg Foundation for Pancreatic Cancer Research.

**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/).

*Review*

## **New Insights into the Liver–Visceral Adipose Axis During Hepatic Resection and Liver Transplantation**

**María Eugenia Cornide-Petronio 1,**† **, Mónica B. Jiménez-Castro 1,**† **, Jordi Gracia-Sancho 2,3 and Carmen Peralta 1,3,\***


Received: 11 July 2019; Accepted: 17 September 2019; Published: 18 September 2019

**Abstract:** In the last decade, adipose tissue has emerged as an endocrine organ with a key role in energy homeostasis. In addition, there is close crosstalk between the adipose tissue and the liver, since pro- and anti-inflammatory substances produced at the visceral adipose tissue level directly target the liver through the portal vein. During surgical procedures, including hepatic resection and liver transplantation, ischemia–reperfusion injury induces damage and regenerative failure. It has been suggested that adipose tissue is associated with both pathological or, on the contrary, with protective effects on damage and regenerative response after liver surgery. The present review aims to summarize the current knowledge on the crosstalk between the adipose tissue and the liver during liver surgery. Therapeutic strategies as well as the clinical and scientific controversies in this field are discussed. The different experimental models, such as lipectomy, to evaluate the role of adipose tissue in both steatotic and nonsteatotic livers undergoing surgery, are described. Such information may be useful for the establishment of protective strategies aimed at regulating the liver–visceral adipose tissue axis and improving the postoperative outcomes in clinical liver surgery.

**Keywords:** adipose tissue; liver; inflammation; steatosis; liver resection; liver transplantation; lipectomy

## **1. Introduction**

In the last decade, adipose tissue has emerged as an essential and highly active metabolic and endocrine organ [1–3]. The basic function of adipocytes is to take up free fatty acids (FFA) from circulating lipoprotein complexes and esterify them into triacylglycerides [4]. During times of metabolic demand, hydrolysis of triacylglyceride releases FFA to generate adenosine triphosphate (ATP) [5]. These adipocyte processes, termed lipogenesis and lipolysis, respectively, are primarily governed through hormonal pathways [6]. However, one of the most important characteristics of adipose tissue is its function in whole-body energy homeostasis, mediated principally through the endocrine system [4]. Adipose tissue expresses and secretes a variety of bioactive molecules, known as adipokines, which may exert their effects in adipose tissue and in other organs [7]. Adipokines include leptin, interleukin (IL)-6, other cytokines, adiponectin, complement components, adipsin, plasminogen activator inhibitor-1 (PAI-1), and proteins of the renin–angiotensin system, among others [7]. Collectively, adipokines modulate the crosstalk between adipose tissue and other metabolic organs, including the liver [8]. Thus, adipokines directly target the liver through the portal vein [9] and have significant effects on liver diseases [4].

The hypoxia and subsequent oxygen delivery restoration to the liver, namely, hepatic ischemia–reperfusion (I/R), is one of the major pathophysiological events and causes of morbidity and mortality in liver resections and transplantation, being more evident in the presence of hepatic steatosis [10–14]. Despite the attempts to solve this issue, hepatic I/R is an unresolved problem in clinical practice [15]. The cellular mechanisms involved in liver I/R injury are numerous and complicated [14], which led to discrepancies in our understanding of this pathology [16]. For instance, the mechanisms underlying I/R injury in conditions of cold ischemia associated with liver transplantation (LT) are different from those that occur in conditions of warm ischemia associated with liver resections. In addition, hepatic steatosis is associated with an increased postoperative complication index and mortality after liver resection and transplantation, and the mechanisms responsible for hepatic damage and regenerative failure are different in steatotic versus nonsteatotic livers [15]. The investigations focused on the role of adipose tissue are of clinical and scientific relevance since the prevalence of obesity ranges from 24–45% of the population and consequently is expected to increase the number of steatotic livers submitted to surgery, which poorly tolerate I/R damage, resulting in liver dysfunction and regenerative failure [17–23]. In addition, it has been reported that adipose tissue exerts both pathological or, on the contrary, protective effects on damage and regenerative response [24]. It should be noted that functional differences between lean and obese adipose tissue have been extensively described [25–27] and summarized, as seen in Figure 1. Briefly, adipose tissue from lean individuals is a connective tissue of low density with small insulin-sensitive adipocytes that secrete adipokines involved in energy homeostasis, angiogenesis, and antioxidant processes. However, the rigidity of adipose tissue from obese individuals is caused by the increment of connective fiber content. Hypertrophic insulin-resistant adipocytes secrete different inflammatory mediators, resulting in adipose tissue dysfunction, impaired angiogenesis, and cell death [25–27]. Moreover, obesity induces changes in the secretion of adipokines from adipose tissue to the circulation [28–30] and increases the inflammatory response and oxidative stress in adipose tissue [31–35]. Therefore, investigations focused on evaluating the liver–adipose tissue axis in steatotic and nonsteatotic livers subjected to hepatic resections or transplants are highly useful in the establishment of specific therapies to prevent both hepatic I/R injury and regenerative failure in liver surgery.

In the first part of this review, we highlight the actual knowledge of the crosstalk between the adipose tissue and the liver during liver surgery. In addition, the different experimental models and pharmacological strategies aimed at regulating potential dysfunctions in the adipose tissue–liver axis in liver surgery are presented, focusing on the strengths and limitations. Clinical results on the role of adipose tissue in the postoperative outcomes after liver surgery are also discussed.

*Cells* **2019**, *8*, x FOR PEER REVIEW 3 of 15

**Figure 1.** Schematic illustration of functional differences between lean and obese adipose tissue. Abbreviations: IL, interleukin; NO, nitric oxide; TNFα, tumor necrosis factor α. **Figure 1.** Schematic illustration of functional differences between lean and obese adipose tissue. Abbreviations: IL, interleukin; NO, nitric oxide; TNFα, tumor necrosis factor α.

#### In the first part of this review, we highlight the actual knowledge of the crosstalk between the **2. Relevance of Adipose Tissue in Experimental Models of Liver Resection**

adipose tissue and the liver during liver surgery. In addition, the different experimental models and pharmacological strategies aimed at regulating potential dysfunctions in the adipose tissue–liver axis in liver surgery are presented, focusing on the strengths and limitations. Clinical results on the role of adipose tissue in the postoperative outcomes after liver surgery are also discussed. **2. Relevance of Adipose Tissue in Experimental Models of Liver Resection**  Crosstalk between the adipose tissue and the liver is an important event both in the physiological function of the liver and in the development of liver diseases [4]. Indeed, adipose Crosstalk between the adipose tissue and the liver is an important event both in the physiological function of the liver and in the development of liver diseases [4]. Indeed, adipose tissue inflammation is a well-recognized sign of obesity [36,37], one of its major consequences being the alteration of the secretion of adipokines that drain to the liver via the portal vein, a notion known as the "portal theory" [37–39]. This dysfunctional adipose tissue–liver axis is supported by the specific disruption in adipocytes of inflammatory mediators (apoptosis antigen 1 or cluster of differentiation 95; Fas/CD95) and/or inflammatory signals (c-Jun N-terminal kinase-1, JNK1) in different mouse models, resulting in protection against hepatic steatosis [40,41].

tissue inflammation is a well-recognized sign of obesity [36,37], one of its major consequences being the alteration of the secretion of adipokines that drain to the liver via the portal vein, a notion known as the "portal theory" [37–39]. This dysfunctional adipose tissue–liver axis is supported by the specific disruption in adipocytes of inflammatory mediators (apoptosis antigen 1 or cluster of differentiation 95; Fas/CD95) and/or inflammatory signals (c-Jun N-terminal kinase-1, JNK1) in different mouse models, resulting in protection against hepatic steatosis [40,41]. It has been suggested that lipids are the preferred energy substrate for nonsteatotic livers in conditions of partial hepatectomy (PH) without I/R [42–44]. Briefly, hypoglycemia that follows PH induces catabolism of peripheral adipose stores followed by hepatic accumulation of systemically derived fat and subsequent liver regeneration [45–47]. This is supported by the fact that glucose administration could block the mobilization of fatty acids from adipose tissue by the liver to obtain energy [48]. Parameters of lipid metabolism have been reported during hepatic regeneration: esterification rate of fatty acids from adipose tissue is higher and lipogenesis is raised [49,50]. In fact, the remaining liver after PH expresses the lipoprotein lipase, which could take up fatty acids from circulating triacylglycerides [51]. Moreover, different studies have described that mice lacking lipid metabolism-associated genes have reduced hepatic adipogenesis and regeneration liver failure [52,53]. In line with this, the surgical relevance of the lipid lowering effect of omega-3 fatty acids has been studied in steatotic livers in the setting of hepatic I/R injury without PH [54] or in PH without I/R [55] (Figure 2). It should be noted that the process of liver regeneration requires careful It has been suggested that lipids are the preferred energy substrate for nonsteatotic livers in conditions of partial hepatectomy (PH) without I/R [42–44]. Briefly, hypoglycemia that follows PH induces catabolism of peripheral adipose stores followed by hepatic accumulation of systemically derived fat and subsequent liver regeneration [45–47]. This is supported by the fact that glucose administration could block the mobilization of fatty acids from adipose tissue by the liver to obtain energy [48]. Parameters of lipid metabolism have been reported during hepatic regeneration: esterification rate of fatty acids from adipose tissue is higher and lipogenesis is raised [49,50]. In fact, the remaining liver after PH expresses the lipoprotein lipase, which could take up fatty acids from circulating triacylglycerides [51]. Moreover, different studies have described that mice lacking lipid metabolism-associated genes have reduced hepatic adipogenesis and regeneration liver failure [52,53]. In line with this, the surgical relevance of the lipid lowering effect of omega-3 fatty acids has been studied in steatotic livers in the setting of hepatic I/R injury without PH [54] or in PH without I/R [55] (Figure 2). It should be noted that the process of liver regeneration requires careful regulation of lipid accumulation. In fact, Yang et al. suggested that Smad interacting protein-1 (SIP1) (a key factor linked to the transforming growth factor-β (TGF-β), bone morphogenetic protein (BMP), and Wnt signaling pathways) is one of the mechanisms involved in the hepatic lipid accumulation and, consequently, in the process of liver regeneration under PH without I/R [56] (Figure 2). It is well known that in order for the liver to regenerate, the provision of fatty acids, phospholipids, and cholesterol to the liver is essential for the maintenance of the rate of formation of the membranes of dividing liver

regulation of lipid accumulation. In fact, Yang et al. suggested that Smad interacting protein-1 (SIP1) (a key factor linked to the transforming growth factor-β (TGF-β), bone morphogenetic protein cells [50,51]. On the other hand, important questions for steatotic livers arise from these data since further studies will be required to elucidate whether steatosis may be reduced to avoid the vulnerability of steatotic livers to I/R and the regenerative failure, or instead if drugs aimed at increasing the levels of hepatic triglycerides should be used during surgery and thus conserve the energy required for liver regeneration.

The studies mentioned above in liver surgery have been reported in settings of I/R without PH or in PH without I/R. However, it should be noted that in the clinical setting, PH is usually performed under vascular occlusion. Thus, if our aim is the establishment of new protective strategies in the clinical setting of hepatic resection, the experimental conditions used at the bench-side should simulate as close as possible the clinical reality. Fortunately, some studies have evaluated the contribution of adipose tissue to liver injury and regeneration in PH with I/R conditions. Firstly, Mendes-Braz et al. demonstrated that the relevance of adipose tissue in hepatic damage and regeneration depends on the type of liver [57]. In this sense, adipose tissue is not required for the regeneration of nonsteatotic livers subjected to PH with I/R. In contrast, it is necessary to promote regeneration and reduce injury in steatotic livers. Taking these data into account, glucose or lipid emulsion was administered in obese and lean animals undergoing PH + I/R. Glucose or lipid treatment in nonsteatotic livers protected against hepatic damage and regenerative failure. In obese animals, glucose treatment did not protect steatotic livers against damage but improved their regeneration. However, lipid treatment conferred protection against damage and regenerative failure [57]. Mendes-Braz et al. suggest that in addition to the function of adipose tissue as a lipid precursor for new membrane synthesis, the requirement of systemic adipose stores during regeneration of steatotic livers might be based on the endocrine role of adipose tissue as a source of different adipokines, which are essential signals for liver regeneration [57].

In addition to the studies related to the role of adipose tissue as a source of energy substrates or inductors of hepatic lipid accumulation, other studies have investigated the potential contribution of adipose tissue as a source of bioactive molecules such as visfatin, cortisol, and soluble forms of the VEGF receptor 1 (sFlt1).

Elias-Miró et al. found the injurious effects of visfatin in PH with I/R and that steatotic livers were more vulnerable to upregulated visfatin than nonsteatotic livers. The administration of visfatin exacerbated damage and regenerative failure in steatotic livers following PH with I/R. Treatment with resistin maintained low levels of visfatin in steatotic livers by blocking its hepatic reuptake from adipose tissue and consequently prevented the injurious effects of visfatin on hepatic damage and regenerative failure [58] (Figure 2).

In pathologic states, adipose tissue may also secret a range of hormones including cortisol, which may be taken up from the circulation by the liver [1,59]. Cornide-Petronio et al. reported that in instances of PH with I/R, the contributory potential of adipose tissue (as a cortisol source) is dependent on baseline liver status (steatotic versus nonsteatotic livers). In such surgical conditions, the authors found that cortisol levels in adipose tissue and liver were elevated only in obese animals [59]. In addition, cortisol administration under PH with I/R conditions exacerbated hepatic damage and regenerative failure only in obese animals. Indeed, in obese animals, alterations in enzymatic regulation of cortisol metabolism caused cortisol accumulation in steatotic livers, whereas in lean animals, compensatory mechanisms mainly based on the clearance of hepatic cortisol were shown to prevent intrahepatic cortisol and its deleterious effects [59].

Interestingly, Bujaldon et al. recently examined the effects of vascular endothelial growth factor type A (VEGFA) on damage and regeneration in steatotic and nonsteatotic livers submitted to PH with I/R. The authors reported that VEGFA levels were decreased in both steatotic and nonsteatotic livers after surgery, but the exogenous VEGFA administrated was only able to reach nonsteatotic livers, reducing the incidence of postoperative complications following surgery. Unexpectedly, the authors found that circulating VEGFA was sequestered by the high circulating levels of the sFlt1 released from adipose tissue, so VEGFA could not reach the steatotic liver to exert its effects, ultimately exacerbating damage and regenerative failure [60]. Thus, the concomitant administration of VEGFA and an antibody

against sFlt1 was required to avoid binding of sFlt1 to VEGFA. This was associated with high VEGFA levels in steatotic livers and protection against damage and regenerative failure [60] (Figure 2). of VEGFA and an antibody against sFlt1 was required to avoid binding of sFlt1 to VEGFA. This was associated with high VEGFA levels in steatotic livers and protection against damage and regenerative failure [60] (Figure 2).

*Cells* **2019**, *8*, x FOR PEER REVIEW 5 of 15

authors found that circulating VEGFA was sequestered by the high circulating levels of the sFlt1

ultimately exacerbating damage and regenerative failure [60]. Thus, the concomitant administration


**Figure 2.** Strategies aimed at regulating hepatic damage and regenerative failure considering the adipose tissue–liver crosstalk during partial hepatectomy (PH) with or without I/R [54–60]. Abbreviations: I/R, ischemia–reperfusion; PH, partial hepatectomy; sFlt1, soluble form of the VEGF receptor 1; SIP1, smad interacting protein 1; VEGFA, vascular endothelial growth factor type A. **Figure 2.** Strategies aimed at regulating hepatic damage and regenerative failure considering the adipose tissue–liver crosstalk during partial hepatectomy (PH) with or without I/R [54–60]. Abbreviations: I/R, ischemia–reperfusion; PH, partial hepatectomy; sFlt1, soluble form of the VEGF receptor 1; SIP1, smad interacting protein 1; VEGFA, vascular endothelial growth factor type A.

#### **3. Relevance of Adipose Tissue in Experimental Models of Liver Transplantation 3. Relevance of Adipose Tissue in Experimental Models of Liver Transplantation**

To our knowledge, the relevance of adipose tissue (as a source of fatty acids and related lipid substrates as well as bioactive molecules) on lipid metabolism, hepatic damage, and regeneration associated with transplantation remains to be elucidated (Figure 3). In addition, the few experimental studies on LT [61–63] have described the levels of adipokines in the liver but not in adipose tissue. In this vein, it has been demonstrated in experimental studies that adiponectin, resistin, and visfatin levels were not modified in recipients when nonsteatotic livers were subjected to transplantation, whereas in recipients of steatotic liver grafts, the presence of hepatic steatosis down-regulated both adiponectin and resistin levels under such surgical conditions, whereas no changes in visfatin levels were observed [61,62]. The role of adipose tissue as a potential source of adiponectin, resistin, or visfatin was unexplored. Nevertheless, the effects of such bioactive To our knowledge, the relevance of adipose tissue (as a source of fatty acids and related lipid substrates as well as bioactive molecules) on lipid metabolism, hepatic damage, and regeneration associated with transplantation remains to be elucidated (Figure 3). In addition, the few experimental studies on LT [61–63] have described the levels of adipokines in the liver but not in adipose tissue. In this vein, it has been demonstrated in experimental studies that adiponectin, resistin, and visfatin levels were not modified in recipients when nonsteatotic livers were subjected to transplantation, whereas in recipients of steatotic liver grafts, the presence of hepatic steatosis down-regulated both adiponectin and resistin levels under such surgical conditions, whereas no changes in visfatin levels were observed [61,62]. The role of adipose tissue as a potential source of adiponectin, resistin, or visfatin was unexplored. Nevertheless, the effects of such bioactive molecules on hepatic damage and regenerative failure were investigated in steatotic and nonsteatotic livers under PH with I/R conditions. As expected, hepatic damage in recipients of steatotic liver grafts was unaltered under

pharmacological regulation of visfatin. However, the treatment with either exogenous adiponectin or resistin in steatotic liver grafts improved the postoperative outcomes after transplantation. In addition, the activation of adenosine monophosphate-activated protein kinase (AMPK) by pharmacological drugs such as AICAR (cell-permeable adenosine analog that is a selective activator of AMPK) or ischemic preconditioning (PC)—which increased both adiponectin and resistin in steatotic liver grafts of recipients submitted to transplantation—resulted in protection against hepatic damage [62]. In experimental models of ex vivo LT, it has been reported that the addition of leptin to preservation solutions was able to increase the signal transducer and activator of transcription-3 levels and to reduce damage in nonsteatotic grafts submitted to transplantation [63]. steatotic liver grafts was unaltered under pharmacological regulation of visfatin. However, the treatment with either exogenous adiponectin or resistin in steatotic liver grafts improved the postoperative outcomes after transplantation. In addition, the activation of adenosine monophosphate-activated protein kinase (AMPK) by pharmacological drugs such as AICAR (cell-permeable adenosine analog that is a selective activator of AMPK) or ischemic preconditioning (PC)—which increased both adiponectin and resistin in steatotic liver grafts of recipients submitted to transplantation—resulted in protection against hepatic damage [62]. In experimental models of ex vivo LT, it has been reported that the addition of leptin to preservation solutions was able to increase the signal transducer and activator of transcription-3 levels and to reduce damage in nonsteatotic grafts submitted to transplantation [63].

*Cells* **2019**, *8*, x FOR PEER REVIEW 6 of 15

nonsteatotic livers under PH with I/R conditions. As expected, hepatic damage in recipients of


All the studies mentioned above [61–63] reported the levels as well as the role of adipokines in liver grafts in experimental models of LT. However, the levels of adipokines in adipose tissue as well as the potential involvement of adipose tissue in the hepatic levels of adipokines following transplantation were not evaluated in such studies. In our view, further investigations to address this issue are of clinical and scientific relevance. In fact, in clinical practice, most of the liver grafts are All the studies mentioned above [61–63] reported the levels as well as the role of adipokines in liver grafts in experimental models of LT. However, the levels of adipokines in adipose tissue as well as the potential involvement of adipose tissue in the hepatic levels of adipokines following transplantation were not evaluated in such studies. In our view, further investigations to address this issue are of clinical and scientific relevance. In fact, in clinical practice, most of the liver grafts are obtained from brain-dead donors [64]. It is well known that brain death is associated with cerebral

trauma and usually caused by hypoxia [65]. In addition, an experimental study aimed at evaluating the role of adipose tissue in the metabolism of rats with brain injury showed that resistin levels were increased in subcutaneous fat of rats with traumatic brain injury [66]. Moreover, it should be taken into account that adipose tissue is considered to be an important source of adipokines, such as leptin, adiponectin, and resistin, and that adipokines directly access the liver from adipose tissue through the portal vein [9,67]. Therefore, the relevance of adipose tissue in the hepatic damage associated with transplantation should be considered in further experimental research of LT to elucidate whether the regulation of adipose tissue functions could improve the quality of donor organs and postoperative outcomes after transplantation.

Altogether, the current knowledge emphasizes the relevance of further characterizing the role of mediators released from peripheral adipose tissue on damage and regenerative failure in both steatotic and nonsteatotic livers undergoing surgery. All of this is required to provide novel therapeutic approaches that can be transferred to clinical liver surgery and consequently increase the number of available donors for transplantation and improve recovery for patients subjected to liver resections.

#### **4. Relevance of Adipose Tissue in Patients Undergoing Liver Surgery**

Interestingly, the extent of visceral adipose tissue as well as serum levels of adipokines have been evaluated in patients undergoing general surgery. Nevertheless, from our knowledge, pharmacological modulation of adipokine actions has not been reported in the clinical practice of liver surgery [68–71]. Indeed, the significance of visceral adipose tissue remains controversial in the surgical setting [71].

In liver resections associated with hepatocellular carcinoma (HCC), preoperative visceral adiposity, as well as low muscularity (since obesity might be associated with a decrease in muscle mass), was closely related to postoperative death and HCC recurrence [72–74]. In addition, it has been reported that greater fat accumulation in skeletal muscle has been associated with a worse prognosis and survival after PH in patients with HCC, even with adjustment for other known predictors [75]. Moreover, prospective studies and meta-analyses have suggested that obese patients have increased risk and a poorer prognosis for many types of cancer [72–74]. All of these results in PH are in line with those observed in living donor liver transplantation (LDLT), since patients with a high degree of muscle steatosis and visceral adiposity show worse survival rates after transplantation compared with patients without obesity or with normal musculature [76]. Nevertheless, these findings are challenged by opposite observations. Indeed, preoperative abdominal computed tomographic (CT) scans in patients undergoing major hepatic resection associated with cancer suggest that obesity does not correlate with poor outcomes after major surgery [77]. Interestingly, neither preoperative visceral adiposity nor low muscularity were poor prognostic factors in patients undergoing liver resection for colorectal liver metastases [78]. In addition, some studies focused on liver resections of different cancer types showed that patients with a higher body mass index (BMI) survive longer than normal-weight patients after surgery [79–82]. It should be noted that CT measurement enables specific quantification of visceral adipose tissue, which is not reflected by BMI.

The contradictory results in clinical practice, the so-called obesity paradox, might occur due to the different methodologies used to evaluate and measure adipose tissue [71], but different types of surgery (resection vs. transplantation) as well as liver pathologies should also be noted.

Changes in adipokine levels in patients subjected to PH have been reported, suggesting that early-phase elevation of serum levels of hepatocyte growth factor (HGF), leptin, and macrophage colony-stimulating factor (M-CSF) could be associated with the acceleration of liver regeneration [83]. In line with this, plasmatic adipokines after LDLT have been mainly reported as biochemical markers to evaluate the risk of fibrosis progression in patients transplanted due to hepatitis C [84]. However, in these studies, the role of adipose tissue as a source of adipokines was not evaluated.

#### **5. Experimental Strategies to Evaluate Adipose Tissue in Liver Surgery 5. Experimental Strategies to Evaluate Adipose Tissue in Liver Surgery**

#### *5.1. Lipectomy 5.1. Lipectomy*

The literature describes the surgical excision of adipose tissue (lipectomy) to evaluate the function of adipose tissue in physiological conditions and different pathologies [37,40,41]. However, few studies have attempted to discern the role of adipose tissue on adipokine levels and hepatic damage and regenerative failure in liver surgery of PH with I/R [56–60], and no studies have evaluated the effects of a lipectomy in livers submitted for transplantation (Figure 4). The literature describes the surgical excision of adipose tissue (lipectomy) to evaluate the function of adipose tissue in physiological conditions and different pathologies [37,40,41]. However, few studies have attempted to discern the role of adipose tissue on adipokine levels and hepatic damage and regenerative failure in liver surgery of PH with I/R [56–60], and no studies have evaluated the effects of a lipectomy in livers submitted for transplantation (Figure 4).

*Cells* **2019**, *8*, x FOR PEER REVIEW 8 of 15

**Figure 4.** Schematic illustration of lipectomy effects in liver surgery. Abbreviations: ATP, adenosine triphosphate; I/R, ischemia–reperfusion; LT, liver transplantation; PH, partial hepatectomy; sFlt1, soluble form of the VEGF receptor 1; VEGFA, vascular endothelial growth factor type A. **Figure 4.** Schematic illustration of lipectomy effects in liver surgery. Abbreviations: ATP, adenosine triphosphate; I/R, ischemia–reperfusion; LT, liver transplantation; PH, partial hepatectomy; sFlt1, soluble form of the VEGF receptor 1; VEGFA, vascular endothelial growth factor type A.

The results obtained using lipectomy in PH with I/R indicate that, in contrast with nonsteatotic livers, adipose tissue is required for liver regeneration and to reduce damage in the presence of The results obtained using lipectomy in PH with I/R indicate that, in contrast with nonsteatotic livers, adipose tissue is required for liver regeneration and to reduce damage in the presence of

steatosis [57] (Figure 4). Interestingly, adipose tissue does not seem to be an energy source for the nonsteatotic liver since ATP levels were unchanged after lipectomy [57]. Regarding adipokines, steatosis [57] (Figure 4). Interestingly, adipose tissue does not seem to be an energy source for the nonsteatotic liver since ATP levels were unchanged after lipectomy [57]. Regarding adipokines, removal of adipose tissue using lipectomy in lean animals undergoing PH with I/R resulted in plasmatic and hepatic levels of adiponectin, resistin, visfatin, and cortisol, similar to those observed under PH with I/R conditions [57–59]. The levels of sFlt1 were reduced in plasma in lean animals lipectomized and undergoing PH + I/R, indicating that adipose tissue might be a potential source of sFlt1 [60]. As it has been suggested that circulating VEGFA is sequestered by the sFlt1 released by adipose tissue in lean animals in conditions of PH + I/R [60], this was associated with increases in hepatic VEGFA (Figure 4). In obese animals, the reduced hepatic ATP levels were more evident than in lean animals. However, similar to that occurring in lean animals, adipose tissue does not seem to be an energy source for the steatotic liver, since ATP levels were unchanged after lipectomy. Adipose tissue seems to be an adiponectin source for the steatotic liver, since the induction of lipectomy in obese animals reduced both plasmatic and hepatic levels of adiponectin compared with the results obtained under PH with I/R conditions [57]. The contribution of adipose tissue as a source of resistin was irrelevant since obese animals undergoing PH with I/R showed high levels of plasmatic and hepatic resistin levels, whereas these resistin levels were unaltered under lipectomy conditions [58] (Figure 4). In obese animals undergoing PH with I/R, the visfatin levels were increased and reduced in liver and plasma, respectively. When adipose tissue was removed in obese animals undergoing PH with I/R, circulating visfatin levels were reduced, whereas hepatic visfatin accumulation was unaltered [58]. Thus, PH with I/R induced the release of visfatin from adipose tissue to circulation and reduced the generation of visfatin by the liver [58]. In obese animals undergoing PH with I/R, plasmatic and hepatic cortisol levels were increased [61]. In contrast, lipectomy in obese animals reduced cortisol levels in plasma but not in the steatotic liver [59], although the potential contribution of adipose tissue in the hepatic levels of cortisol of obese animals cannot be discounted. Indeed, changes in the enzymes engaged in cortisol generation and clearance were detected in adipose tissue of obese animals undergoing PH with I/R [59]. The reduced plasmatic levels of sFlt1 in obese animals lipectomized and undergoing PH with I/R were more evident than in lean animals [60]. This increased the circulating VEGFA bioavailability and, consequently, increased the opportunity of VEGFA to be taken up by the steatotic liver [60] (Figure 4).

Reduced adiponectin and resistin levels were observed only in steatotic livers when obese animals were subjected to LT [62] (Figure 4). The role of adipose tissue on hepatic adipokine levels, hepatic damage, and regeneration following LT remains to be elucidated. In our view, given the key role of adipose tissue in steatotic and nonsteatotic livers undergoing PH with I/R, strategies based in the adipose tissue removal should been used to study the crosstalk liver–adipose tissue in LT, mainly in the presence of steatosis.

## *5.2. Transgenic Animal Models*

The use of transgenic animal models has improved our understanding of the pathophysiology of adipose tissue. The main focus of transgenic animal models has been the expression or knockout of selected genes, specifically in adipose tissue, identifying and characterizing promoter regions that confer adipose–tissue specific expression [7]. For instance, to target both white and brown adipose tissue, the promoters for adipocyte lipid binding protein aP215 and for phosphoenolpyruvate carboxykinase are usually used, whereas to target only brown adipose tissue, the mitochondrial uncoupling protein-1 (UCP-1) promoter is used [85–87]. In our view, the potential applications of transgenic animal models with overexpression or knockout of adipose tissue-selected genes might be of scientific and clinical interest to evaluate the adipose tissue–liver axis in hepatic resections and transplantation, since in different surgical conditions, hepatic diseases might be improved by directly targeting adipose tissue, rather than liver tissue per se.

## **6. Conclusions**

The role of adipose tissue on damage and regenerative failure in experimental liver surgery depends on the type of surgical procedure (PH with or without I/R) as well as the type of liver (steatotic versus nonsteatotic) submitted to liver surgery. This should be taken into account for the establishment of protective strategies modulating the liver–adipose tissue axis, which would be specific for each surgical procedure and type of liver, as it has been reported in the present review. Further clinical studies and appropriate methods for adipose tissue measurement will be required to elucidate the significance of visceral adipose tissue in the clinical scenario of surgical hepatic resections. The use of experimental models of lipectomy as well as transgenic animal models with expression or knockout of adipose tissue-selected genes might be of scientific and clinical relevance to elucidate the contribution of the adipose tissue–liver axis, as well as the role of adipose tissue as an energy substrate and/or a source of different adipokines and hormones in livers subjected to transplantation. This would provide novel therapeutic approaches to be transferred to clinical conditions to improve the post-transplantation outcomes and consequently increase the number of available donors for transplantation.

**Funding:** This research was supported by the Ministerio de Ciencia, Innovación y Universidades (RTI2018-095114-B-I00) Madrid, Spain; the European Union (Fondos Feder, 'una manera dehacer Europa'); CERCA Program/Generalitat de Catalunya; and Secretaria d'Universitats I Recerca del Departament d'Economia i Coneixement (2017 SGR-551) Barcelona, Spain.

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

## **Abbreviations**


## **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/).

## *Review* **Obesity as a Risk Factor for Severe COVID-19 and Complications: A Review**

**Fien Demeulemeester \*, Karin de Punder , Marloes van Heijningen and Femke van Doesburg**

Natura Foundation, 3281 NC Numansdorp, The Netherlands; k.depunder@naturafoundation.com (K.d.P.); m.vanheijningen@naturafoundation.com (M.v.H.); f.vandoesburg@naturafoundation.com (F.v.D.) **\*** Correspondence: f.demeulemeester@naturafoundation.com

**Abstract:** Emerging data suggest that obesity is a major risk factor for the progression of major complications such as acute respiratory distress syndrome (ARDS), cytokine storm and coagulopathy in COVID-19. Understanding the mechanisms underlying the link between obesity and disease severity as a result of SARS-CoV-2 infection is crucial for the development of new therapeutic interventions and preventive measures in this high-risk group. We propose that multiple features of obesity contribute to the prevalence of severe COVID-19 and complications. First, viral entry can be facilitated by the upregulation of viral entry receptors, like angiotensin-converting enzyme 2 (ACE2), among others. Second, obesity-induced chronic inflammation and disruptions of insulin and leptin signaling can result in impaired viral clearance and a disproportionate or hyper-inflammatory response, which together with elevated ferritin levels can be a direct cause for ARDS and cytokine storm. Third, the negative consequences of obesity on blood coagulation can contribute to the progression of thrombus formation and hemorrhage. In this review we first summarize clinical findings on the relationship between obesity and COVID-19 disease severity and then further discuss potential mechanisms that could explain the risk for major complications in patients suffering from obesity.

**Keywords:** coagulopathy; COVID-19; cytokine storm; inflammation; leptin; obesity; SARS-CoV-2

## **1. Introduction**

The novel coronavirus SARS-CoV-2 that first appeared in the Chinese city of Wuhan is still causing a global pandemic with a death toll that already exceeds 2,430,000 people and continues to grow every day (number from John Hopkins University). SARS-CoV-2 is the virus that causes COVID-19, a clinical picture characterized by fever, coughing, muscle pain and fatigue and can evolve into hyperinflammation, cytokine storm, ARDS and COVID-associated-coagulopathy (CAC) [1,2]. A large number of patients severely ill with COVID-19 arriving at the ICU are overweight or suffer from obesity [3]. These conditions, as well as smoking, age, type II diabetes and cardiovascular diseases, appear to be major risk factors for serious complications and increased mortality in COVID-19 patients [4–6].

Understanding the contributing factors to COVID-19 disease severity and complications in the context of obesity is of key importance for the development of therapeutic interventions, as well as for advancing preventative strategies in this high-risk group. Therefore, in this review we first summarize clinical findings on obesity and COVID-19 disease outcomes. Next, we discuss possible underlying mechanisms linking obesity to major disease complications as a result of SARS-CoV-2 infection Here we focus on the metabolic- and immune-related consequences of obesity on COVID-19 disease course.

**Citation:** Demeulemeester, F.; de Punder, K.; van Heijningen, M.; van Doesburg, F. Obesity as a Risk Factor for Severe COVID-19 and Complications: A Review. *Cells* **2021**, *10*, 933. https://doi.org/10.3390/ cells10040933

Academic Editor: Javier Gómez-Ambrosi

Received: 17 March 2021 Accepted: 14 April 2021 Published: 17 April 2021

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

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

## **2. Obesity**

## *2.1. Obesity Is a Common Disease Associated with Chronic Inflammation and Insulin and Leptin Resistance*

The prevalence of obesity has increased worldwide over the last 50 years. In 2015 the mean prevalence of obesity in adults of selected countries was 19.5% and ranged from 3.7% in Japan to 38.2% in the United States [7]. Obesity (BMI <sup>≥</sup> 30 kg/m<sup>2</sup> ) is a major risk factor for the development of non-communicable diseases and is defined by the WHO as abnormal or excessive fat accumulation that might impair health [8].

Fat or adipose tissue, originally regarded as a simple organ for storing energy, is currently viewed as one of the most important endocrine organs [9–11]. Fat cells, or adipocytes, produce cytokine-like hormones, called adipokines, which play a major role in metabolism and inflammation [11]. Adipocytes in different regions fulfill different functions. Distinctions should be made between brown and white adipocytes, between ectopic and non-ectopic and between visceral and subcutaneous adipose tissue to accurately estimate the impact of the adipose tissue on the patient's health. For example, visceral fat, but not so much subcutaneous fat accumulation, promotes systemic inflammation and is associated with impairments of glucose and lipid metabolism accompanied by insulin resistance [12–14].

Obesity is associated with chronic inflammation, resulting from immune cell activity in dysfunctional (visceral) adipose tissue. In obesity, the excessive presence and hypertrophy of adipocytes result in hypoxia, cell stress and apoptosis. The hypoxic environment induces the infiltration of immune cells into the adipose tissue as a result of the expression of chemo attractive molecules [15]. Also, hypertrophic adipocytes produce multiple proinflammatory adipokines, such as such as tumor necrosis factor-α (TNF-α), interleukin (IL)-6 and leptin [16].

Leptin informs the brain about the amount of energy stored in the adipose tissue [17,18]. The percentage of total body fat is the most important determinant of leptin levels [19]. Given their high percentage of total body fat, individuals suffering from obesity show elevated leptin concentrations [20], a condition also referred to as hyperleptinemia. Persistent hyperleptinemia is often accompanied with central leptin resistance, due to which appetite and satiety regulation is disrupted [21]. In addition to the regulation of hunger and satiety, leptin also has pro-inflammatory properties [22], further contributing to a chronic inflammatory state in individuals suffering from obesity [23].

## *2.2. Obesity and COVID-19 Disease Severity*

In the following section, we summarize clinical findings supporting an association between obesity and COVID-19 disease severity (see also Table 1). We searched the Pubmed database up to March 15, 2021. In our search strategy, the combination of the following keywords was used: obesity OR BMI OR overweight OR adiposity OR adipose tissue AND COVID-19 OR SARS-CoV-2. Our search was limited by published full-text article in English language. Most studies investigating this association used BMI categories as the predictor variable [24–37]. Four cross-sectional studies did not specify the used classification of obesity [38–41] presumably using the WHO guidelines defining obesity as BMI of 30 or higher. Outcome variables included: hospitalization [36,38,42], ICU admission [31,35,37,38,43–45], intubation [24,25,29,37,38], invasive mechanical ventilation [26,31,34], disease severity [27,28,30,33] and death [24–29,32,38–41]. Of the eleven studies investigating the association between BMI and mortality in hospitalized COVID-19 patients, ten studies observed an increased mortality rate in patients that were overweight (BMI ≥ 25 to <30) [24,32], or suffering from obesity (BMI ≥ 30) [25,29,38–41], or severe obesity (BMI ≥ 35) [26,27]. One study observed no difference in in-hospital deaths between normal and overweight (BMI > 28) patients, but overweight patients did show more severe disease symptoms [28]. Because BMI does not discriminate between fat and lean body mass and poorly reflects fat distribution, four studies used measures of visceral adipose tissue (VAT) obtained by high-resolution computed tomography [42–45]. In COVID-19

patients reporting at the emergency department, subsequent admission to the ICU was associated with 30% higher VAT and 30% lower subcutaneous adipose tissue (SAT) [43] with VAT scores being the best ICU admission predictor. Another study observed no differences in BMI but did see higher VAT and a higher VAT/total adipose tissue (TAT) ratio in hospitalized patients compared to outpatients that did not require hospitalization [42]. Finally, in a different study it was observed that both VAT and TAT were associated with ICU admission [45]. To summarize, numerous studies indicate that obesity is associated with increased disease severity and mortality in COVID-19.

**Table 1.** Clinical Studies Investigating the Association Between Obesity and Disease Severity in COVID-19 Patients.



**Table 1.** *Cont*.

Abbreviations: VAT = visceral adipose tissue; SAT = subcutaneous adipose tissue; TAT = total adipose tissue; BMI = body mass index (kg/m<sup>2</sup> ); IMV = invasive mechanical ventilation; ICU = intensive care unit.

## **3. Underlying Mechanisms Linking Obesity to Major Complications as a Result of SARS-CoV-2 Infection**

To gain a better understanding of the pathophysiology of COVID-19 in patients suffering from obesity, in the next section and in Figure 1 we provide an overview of the mechanistic pathways linking obesity with COVID-19 disease severity, with a focus on the metabolic and immunological consequences of obesity on COVID-19 disease course. Other features of obesity, like impaired respiratory mechanics and pulmonary function and the co-existence of metabolic disorders like diabetes and cardiovascular disease within a single individual also increase the risk for severe COVID-19 and complications [46], but are beyond the scope of this review. In this overview we discuss, *in vitro*, animal and human *in vivo* studies, including clinical trials. Studies were excluded when not indexed and when methodology did not reach minimal criteria. *Cells* **2021**, *10*, x FOR PEER REVIEW 5 of 19 and human *in vivo* studies, including clinical trials. Studies were excluded when not indexed and when methodology did not reach minimal criteria.

**Figure 1.** Schematic representation of different mechanisms through which obesity can promote COVID-19 disease severity and risk for complications. Obesity is often accompanied by insulin and leptin resistance which impairs viral clearance. Next, obesity is characterized by large hypoxic adipocytes infiltrated with immune cells and M1 macrophages leading to a chronic inflammatory state, hypercoagulability and hyperferritinemia. ACE2 produced by adipocytes could provide viral entry into the adipose tissue. In this way the adipose tissue could possibly function as a reservoir for the virus. The constant tissue expansion and tissue remodeling accompanying obesity in concert with high cell stress can upregulate the expression of other potential SARS-CoV-2 receptors, such as csGRP78, HSPG and NRP-1 in adipose tissue and other organs. Obesity-associated endothelial dysfunction, enhanced production of PAI-1 and vitamin K deficiency all increase the risk of developing COVID-19 associated coagulopathy. Abbreviations: IL-6=Interleukin-6; TNFα=Tumor necrosis factor α; RAS=Renin angiotensin system; ACE=Angiotensin converting enzyme; HSPG=Heparan sulfate proteoglycan; NRP-1=Neuropilin-1; UPR=Unfolded protein response; csGRP78= Cell surface glucose related protein 78; PAI-1=Plasminogen activator inhibitor 1. **Figure 1.** Schematic representation of different mechanisms through which obesity can promote COVID-19 disease severity and risk for complications. Obesity is often accompanied by insulin and leptin resistance which impairs viral clearance. Next, obesity is characterized by large hypoxic adipocytes infiltrated with immune cells and M1 macrophages leading to a chronic inflammatory state, hypercoagulability and hyperferritinemia. ACE2 produced by adipocytes could provide viral entry into the adipose tissue. In this way the adipose tissue could possibly function as a reservoir for the virus. The constant tissue expansion and tissue remodeling accompanying obesity in concert with high cell stress can upregulate the expression of other potential SARS-CoV-2 receptors, such as csGRP78, HSPG and NRP-1 in adipose tissue and other organs. Obesity-associated endothelial dysfunction, enhanced production of PAI-1 and vitamin K deficiency all increase the risk of developing COVID-19 associated coagulopathy. Abbreviations: IL-6 = Interleukin-6; TNFα = Tumor necrosis factor α; RAS = Renin angiotensin system; ACE = Angiotensin converting enzyme; HSPG = Heparan sulfate proteoglycan; NRP-1 = Neuropilin-1; UPR = Unfolded protein response; csGRP78 = Cell surface glucose related protein 78; PAI-1 = Plasminogen activator inhibitor 1.

viral entry by the upregulation of some of these host cell receptors.

cellular receptors and on S protein priming by host cell proteases. Obesity can promote

Viral entry into target cells is facilitated by binding of the SARS-CoV-2 spike (S) protein to the membrane protein angiotensin-converting enzyme 2 (ACE2) and the priming

*3.1. Obesity Facilitates Viral Entry of SARS-CoV-2* 

3.1.1. Obesity Increases ACE2 Expression

## *3.1. Obesity Facilitates Viral Entry of SARS-CoV-2*

Cell entry of SARS-CoV-2 virus depends on binding of the viral spike (S) proteins to cellular receptors and on S protein priming by host cell proteases. Obesity can promote viral entry by the upregulation of some of these host cell receptors.

## 3.1.1. Obesity Increases ACE2 Expression

Viral entry into target cells is facilitated by binding of the SARS-CoV-2 spike (S) protein to the membrane protein angiotensin-converting enzyme 2 (ACE2) and the priming of the S protein by serine protease TMPRSS2 and the endosomal cysteine proteases cathepsin B and L expressed by host cells [47,48]. ACE2 is part of the renin-angiotensin system (RAS) and plays an important role in regulating systemic blood pressure. Renin induces proteolytic cleavage of angiotensinogen into angiotensin (Ang)-I, which is then converted into Ang-II by ACE1. The action of ACE1 induces vasoconstriction and inflammation while ACE2 acts as a counter-regulator of the RAS. ACE2 induces vasodilatation and has an anti-inflammatory effect, by converting the pro-inflammatory Ang-II into Ang-1-7 which opposes the actions of Ang-II. ACE2 is expressed in a wide variety of human tissues such as the small intestine, testis, kidneys, heart, thyroid, lungs and brain [49]. Adipose tissue also expresses ACE2, together with all other components of the RAS. The RAS has been shown to play an important role in lipid metabolism and vice versa [50–53]. For example, the expression of ACE2 in adipocytes seems to be promoted by a high-fat diet [54]. Also, analysis of available transcriptome data showed that diet-induced obese mice displayed increased expression of ACE2 in the lung epithelium [55]. The expression of ACE2 was correlated with the expression of genes that code for sterol response element binding proteins 1 and 2, transcription factors controlling lipid synthesis and adipogenesis [55].

In addition, a number of studies indicate that ACE2 expression can be upregulated by a wide range of pro-inflammatory cytokines [56,57]. The chronic inflammatory state that characterizes obesity could therefore potentially facilitate viral entry [56]. Moreover, it has been suggested that viral pools can lodge in adipose tissue and promote shedding, immune activation, and chronic excessive cytokine release [58]. Earlier studies demonstrated that other viruses such as influenza A, HIV and cytomegalovirus use the adipose tissue as a reservoir [59].

Binding of SARS-CoV-2 to ACE2, which prevents ACE2 from exerting its enzymatic activities [60], has further consequences for immune system functioning. Decreased enzymatic ACE2 activity results in a reduction of anti-inflammatory Ang-1-7 and an accumulation of pro-inflammatory Ang-II. This process further increases the risk for COVID-19 related immune system complications such as ARDS.

In conclusion, obesity and its related pro-inflammatory state can promote viral entry, viral shedding and excessive immune activation through the upregulation of ACE2.

### 3.1.2. Activation of the Unfolded Protein Response by GRP78

Another receptor that provides viral entry for SARS-CoV-2 is glucose-regulated protein 78 (GRP78) [61,62]. GRP78, also called immunoglobulin heavy chain binding protein (BiP), is part of the heat shock protein 70 family and is a molecular chaperone found in the lumen of the endoplasmic reticulum (ER). GRP78 functions as an anti-apoptotic regulator that protects cells against cell death induced by ER stress [62]. GRP78 protects against cell death by ensuring the correct folding and assembly of proteins in the ER and by initiating the degradation of misfolded proteins. This evolutionarily preserved cellular homeostatic response to ER stress, also named the ER-stress or unfolded protein response (UPR) [63] can be disrupted by viruses in the establishment of acute, chronic and latent infections [64].

Normally, new foreign viral proteins produced inside host cells would be immediately degraded by the UPR. Therefore, viruses that acquired the ability to disrupt the process designed to degrade unrecognized protein have a clear advantage. Binding GRP78 is a mechanism commonly used by viruses like the Ebola virus, Zika virus, Dengue virus, Japanese encephalitis virus, Coxsackie A9, MERS-CoV and the Borna disease virus to

ensure safe entry into host cells and to facilitate viral replication [65–71]. Also, the S proteins of the SARS-CoV-2 virus binds GRP78 with high affinity [61,72].

Obesity and various factors associated with obesity induce ER stress [73] and consequently UPR activation. These factors, including hypoxia, reactive oxygen species, insulin resistance, nutritional imbalance and excessive fat storage [74–76], stimulate GRP78 expression as a mechanism of the UPR to restore normal cell functions [77,78]. For example, evidence from *in vitro* and *in vivo* studies shows that dyslipidemia is associated with GRP78 overexpression in adipocytes (especially in white adipose tissue) [79], pneumocytes [80], neurons of the hypothalamus [81] and hepatocytes [82].

GRP78 overexpression induces its translocation to the cell membrane or cell surface. When GRP78 is located on the cell surface, it is referred to as cell surface (cs)-GRP78. This acts as a multi-functional receptor that can bind various ligands and plays a crucial role in apoptosis and inflammation, among other things [62]. By binding to cs-GRP78, SARS-CoV-2 could ensure safe entry into host cells [61,72].

#### 3.1.3. Heparan Sulfate Proteoglycans and Neuropilin-1

Heparan sulfate proteoglycans (HSPGs) are located in the extracellular matrix and at the surface of the cell, where they act as co-receptors for various ligands. HSPGs are widely expressed and mediate many biological activities, including angiogenesis, blood coagulation, developmental processes, and cell homeostasis. The binding of cytokines, chemokines and growth factors to HSPGs at the cell surface prevents their degradation [83]. A significant amount of research has established that many different types of viruses can interact with HSPGs. For some of these viruses, these interactions are essential for internalization into host cells [84]. This is also the case for the internalization of SARS-CoV-2 [85].

Syndecans are the major family of transmembrane HSPGs and are present on virtually all nucleated human cells. Syndecans have been shown to facilitate the cellular entry of SARS-CoV-2 *in vitro*. Among syndecans, syndecan-4 was most efficient in mediating SARS-CoV-2 uptake, yet overexpression of other isoforms, including syndecan-1 and the neuronal syndecan-3, also increased SARS-CoV-2 internalization [85,86]. Syndecan 4 is widely expressed in most adult tissues [87] and, among other functions, is involved in lipid metabolism, by clearing pro-atherogenic remnant lipoproteins from the circulation [88]. In addition, syndecan-4 is crucial for adipocyte differentiation and proliferation [89]. Syndecan 4 is upregulated upon activation of the pro-inflammatory transcription factor nuclear factor-κB (NF-κB) [90,91] and its expression is therefore markedly increased under inflammatory conditions [92].

Neuropilin-1 (NRP-1), a membrane-bound co-receptor for growth factors such as vascular endothelial growth factor (VEGF), also facilitates viral entry of SARS-CoV-2 *in vitro* [93]. NRP-1 is strongly involved in adipogenesis and is highly upregulated during adipose-derived stem cell differentiation [94].

In obesity, chronic inflammation, higher circulating pro-atherogenic lipoprotein levels [95] and the constant expansion of adipose tissue accompanied with the differentiation and proliferation of adipose-derived stem cells, could increase syndecan (especially syndecan-4) and NRP-1 expression in various tissues. The up-regulation of syndecans and NRP-1 could facilitate cellular entry of SARS-CoV-2. However, further studies need to confirm this theory.

## *3.2. Obesity Related Insulin Resistance Contributes to an Impaired Immune Response to SARS-CoV-2 Infection*

## 3.2.1. Obesity Induces Insulin Resistance

Insulin resistance is a consequence of the impairment of insulin signaling in insulinresponsive cells, like hepatocytes, myocytes and adipocytes. In the adipose tissue, insulin resistance is promoted by excess lipid accumulation, causing hypoxia and inflammation. The subsequent invasion of macrophages in the adipose tissue that release pro-inflammatory cytokines further impair insulin signaling (see also Section 3.2.2). As a consequence of

adipocyte insulin resistance higher levels of free fatty acids (FFA) leave the fat tissue and enter into the circulation. The increase in circulating FFA and pro-inflammatory mediators further impairs insulin action in other metabolically active organs and tissues, including skeletal muscle and the liver, leading to systemic insulin resistance, which is associated with impaired glucose transport [96,97]. It has been shown that the size of the visceral adipose tissue and adipocyte size in humans is directly associated with systemic insulin resistance [14].

#### 3.2.2. Insulin Resistance Is Induced by Adipocytes and Related Immune Cells

Adipose tissue macrophages can be divided into the classical M1 and alternatively activated M2 macrophages. Adipose tissue from obese individuals contains elevated numbers of M1-like macrophages, which produce pro-inflammatory cytokines, such as TNF-α and IL-6 [98,99]. These pro-inflammatory cytokines inhibit insulin signaling pathways in multiple tissues [100].

The production of TNF-α by M1 macrophages is positively related to the size of the adipose tissue mass. In the adipose tissue TNF-α induces phosphorylation and inactivation of insulin receptors and activates lipolysis, which increases the FFA load. The production of IL-6 by adipocytes and related immune cells is also associated with the amount of body fat. IL-6 induces the production of the pro-inflammatory acute-phase protein C-reactive protein (CRP) and increases fibrinogen levels, resulting in a prothrombotic state. It also promotes adhesion molecule expression by endothelial cells and activates local RAS pathways [101].

Chronic elevations of pro-inflammatory cytokines, such as TNF-α and IL-6, also directly influence COVID-19 disease course. Because TNF-α plays an important role in promoting ARDS, obese individuals with chronically elevated serum TNF-α are at greater risk of developing this life-threatening complication [102]. In patients with COVID-19, IL-6 levels are significantly elevated and associated with adverse clinical outcomes. Meta-analysis of mean IL-6 concentrations demonstrated 2.9-fold higher levels of IL-6 in hospitalized COVID-19 patients with complications compared to patients without complications [103].

## 3.2.3. Systemic Insulin Resistance Impairs the Immune Response

Insulin acts upon immune cells and therefore, systemic insulin resistance can have a substantial impact on the functioning of the adaptive and innate immune system [100]. Animal studies have shown that insulin signaling is essential for optimal T cell effector function [104]. In humans, insulin-resistant individuals displayed delayed innate immunerelated pathway activation after respiratory viral infection compared to insulin-sensitive individuals, suggesting an impairment of the early acting innate immune response [105].

It is likely to assume that chronic inflammation and impairments of the immune response as a consequence of obesity-induced insulin resistance may reduce efficient viral clearance and drive organ injury in the development of severe COVID-19 [106].

## *3.3. Disrupted Leptin Signaling in Obesity can Induce Hyperinflammation during SARS-CoV-2 Infection*

#### 3.3.1. Leptin Is an Important Regulator of Energy Metabolism

Leptin is a pleiotropic protein secreted primarily from white adipose tissue into the bloodstream and can be transported across the blood-brain barrier. Through its effects on the central nervous system (CNS) and peripheral tissues, leptin is an important regulator of energy homeostasis, metabolism, neuroendocrine and immune system function [107].

Obesity proceeds from a chronic energy imbalance and is characterized by persistent hyperleptinemia and central leptin resistance. Under physiological conditions, leptin informs the brain about the energy status in the periphery, but in obesity, signaling to regulatory centers in the brain that normally inhibit food intake and regulate body weight and energy homeostasis is disrupted. Mechanisms underlying leptin resistance, include disruption of leptin signaling in hypothalamic and other CNS neurons, impaired leptin transport across blood-brain barrier, hypothalamic inflammation, ER stress, and autophagy [107,108].

### 3.3.2. Leptin Modulates the Immune System

Leptin has pro-inflammatory properties and upregulates the secretion of proinflammatory cytokines [18,22]. Leptin signaling results in the modulation of both the innate and adaptive immune system on multiple levels. Leptin signaling occurs primarily through the binding of leptin to the long isoform of the leptin receptor, followed by activation of the JAK/STAT pathway [2]. In the innate immune system, increased leptin production stimulates chemotaxis and neutrophil survival, induces pro-inflammatory cytokine production [109,110], and higher expression of adhesion molecules by eosinophils and basophils [111,112]. Monocyte activation and proliferation as well as the production of pro-inflammatory cytokines and chemotaxis are also stimulated by leptin [22]. Certain immune cells, more specifically those that contain the long isoform of the leptin receptor, may become unresponsive to leptin, or leptin resistant, when exposed to high leptin levels for an extended period of time [113,114]. Therefore, chronic hyperleptinemia, as seen in obesity, can have detrimental effects on the immune response due to both chronic pro-inflammatory effects and immune cell dysfunction [21].

Additionally, the adaptive immune system responds to leptin. Leptin induces a shift towards the more pro-inflammatory Th1 response [21], activates T and B lymphocytes, and inhibits regulatory T cells [18,115,116]. Regulatory T cells are involved in suppressing an excessive immune response [117,118]. It was demonstrated in patients suffering from obesity that leptin levels and BMI were inversely correlated with the number of regulatory T cells [119].

In obesity, leptin levels can be further increased due to infection or sepsis [120]. Elevated leptin levels, in combination with the obesity-induced pro-inflammatory state, further increases the risk for the development of a disproportionate or hyper-inflammatory response upon SARS-CoV-2 infection. Furthermore, SARS-CoV-2 infection has been shown to increase the expression of the gene that encodes for suppressor of cytokine signaling 3 (SOCS3) in lung epithelium [55]. This gene is a key regulator of inflammation and an inhibitor of leptin signaling. Increased SOCS3 expression as a result of SARS-CoV-2 infection could therefore further impair leptin signaling and negatively influence the immune response in patients suffering from obesity [121].

#### *3.4. Hyperferritinemia as a Result of Hyperinflammation can Induce a Cytokine Storm*

Ferritin is an iron-binding molecule that stores iron in a biologically available form for vital cellular processes and protects proteins, lipids and DNA from the potential toxicity of this metal element. Ferritin is a marker of the acute-phase response, and its secretion is regulated by pro-inflammatory cytokines. However, the origin of circulating serum ferritin during inflammatory conditions is still debated [122]. While some describe serum ferritin as a leakage product of damaged cells [123], increasing evidence shows that circulating serum ferritin levels may play a critical role in the inflammatory process [91]. Serum ferritin may protect the host during active infection by limiting the availability of iron to pathogens, a phenomenon called 'nutritional immunity' [124–126].

COVID-19 is also accompanied by a rise in circulating ferritin levels. Serum ferritin levels can be used as a diagnostic marker and even a predictor of COVID-19 severity [127–131]. The elevated ferritin levels observed in COVID-19 patients are probably a consequence of the inflammatory process induced by SARS-CoV-2 infection and actively act as enhancer of the inflammatory process in more severe COVID-19 [132]. Alternatively, it has been suggested that ferritin levels increase due to the break-down of red blood cells by the SARS-CoV-2 virus. By breaking down red blood cells and then attacking the hemoglobin 1-beta chain, the virus could separate iron from the porphyrin ring to eventually hijack the porphyrin ring, producing a rise in free iron and subsequently increasing ferritin levels [133]. Although hemoglobin-related biomarkers such as serum ferritin, progressively increase as

the severity of COVID-19 increases [134–136], there is very limited evidence supporting this hypothesis [133]. This hypothesis has subsequently been refuted by DeMartino and colleagues, confirming that disease markers such as hemoglobin, iron and ferritin did not differ between critically ill COVID-19 patients and non-COVID ARDS patients [137].

Several studies have shown a relationship between obesity and elevated ferritin serum levels [138]. Chronic inflammation, as a result of the increased release of leptin and pro-inflammatory cytokines by adipocytes and related immune cells, [11,21,23,120] likely contributes to the manifestation of elevated ferritin levels in individuals with obesity.

Infection with SARS-CoV-2 in individuals with obesity as pre-existing condition can progress into a hyperferritinaemic state as both SARS-CoV-2 and obesity initiate the release of ferritin. Ferritin further stimulates macrophages to produce pro-inflammatory cytokines, mainly IL-1, IL-6 and IL-17. At overly elevated serum ferritin levels, macrophages produce so many cytokines that the situation can evolve into a cytokine storm [122]. The cytokine storm occurring with COVID-19 could be considered a hyperferritinemia syndrome [139] and is a significant adverse development in the course of the disease, translating into a marked increased risk of death [140].

#### *3.5. Obesity-Related Risk Factors for Developing Coagulopathy in COVID-19 Patients*

About a third of ICU patients with COVID-19 develop thrombotic complications [141,142]. The characteristics of CAC seem to be more complex than the development of thromboinflammation triggered by systemic inflammation in response to infection [143]. For example, the presence of systemic microthrombi and hemorrhage in SARS-CoV-2 affected organs indicates a malfunction in the coordination of coagulation and fibrinolysis [144]. Moreover, CAC is commonly associated with increased D-dimer and fibrinogen levels, indicating there is initially no suppression of fibrinolysis [145]. Abnormalities in other coagulation biomarker such as prothrombin time and platelet count are less frequent and seem less affected by SARS-CoV-2 infection [146].

In the following section we provide an overview of parameters that are dysregulated in obesity and could contribute to increase the risk of developing CAC (See also Figure 1).

### 3.5.1. Hyperleptinemia

Hyperleptinemia is a risk factor for developing thrombi. Leptin affects blood clotting by enhancing prothrombotic and antifibrinolytic protein expression in vascular and inflammatory cells and thereby producing a more hypofibrinolytic state [147]. A large population-based cohort study from The Netherlands, examining associations between serum leptin concentrations and coagulation factor concentrations and parameters of platelet activation, showed that serum leptin concentrations were positively associated with concentrations of coagulation factor VIII and IX [148]. Hyperleptinemia and leptin resistance have also been described as risk factors for the development of cardiovascular diseases [149,150]. As elevated leptin levels are a key feature of obesity [149,151,152], the prothrombotic state produced by hyperleptinemia, increases the risk for developing thrombotic complications in COVID-19.

## 3.5.2. PAI-1 Production by Adipocytes

Plasminogen activator inhibitor 1 (PAI-1) is the primary physiological inhibitor of plasminogen activation. Elevations in plasma PAI-1 compromise normal fibrin clearance mechanisms and promote thrombosis. PAI-1 is also produced by adipocytes and its production is dramatically upregulated in obesity [153,154]. PAI-1 mRNA expression was demonstrated in the visceral and subcutaneous fat of obese rats [155] and in adipose tissue from human subjects [156]. In both rats and humans, VAT produced significantly more PAI-1 compared to SAT. These results are consistent with the observation that cardiovascular risk is most closely correlated with central obesity [157]. PAI-1 produced by adipocytes and vascular endothelium is involved in tissue expansion and angiogenesis necessary during adipose tissue development [158].

Several molecular mechanisms are involved in the upregulation of PAI-1 mRNA expression in obesity. For example, different cytokines including TNF-α and transforming growth factor (TGF)-ß, triglycerides, free fatty acids and insulin all stimulate PAI-1 expression in adipose tissue [153]. Leptin, additionally, contributes to a prothrombotic state by increasing the expression of PAI-1 in vascular endothelium [159]. Thus, obesity-induced upregulation of PAI-1 further increases the risk for coagulopathy, following SARS-CoV-2 infection.

#### 3.5.3. Endothelial Dysfunction

The endothelium serves as a dynamic barrier that separates blood from interstitia. Endothelial cells respond rapidly to changes in the circulation and become activated according to environmental needs [160]. The multitude of physiological functions of the endothelium are still a topic of extended research. Endothelium is recently described as an active regulator of lipid and glucose homeostasis [161].

Endothelial cells play a major role in vascular homeostasis and blood coagulation. Endothelial dysfunction is considered a hallmark of metabolic diseases and is characterized by a loss of molecular cell functions and inevitably causing coagulopathy. Endothelial dysfunction contributes to various pathological states such as atherothrombosis, arterial thrombosis (stroke, visceral and peripheral artery occlusive diseases), venous thrombosis, intravascular coagulation and thrombotic microangiopathies [160].

Oxidative stress is considered the major cause of endothelial dysfunction. Obesity generates oxidative stress through different pathways, such as hyperglycemia, known to trigger vascular damage by inducing the accumulation of reactive oxygen species (ROS) [162]. Hyperglycemia also activates NF-κB, a transcription factor that mediates vascular inflammation [162]. The exacerbated production of pro-inflammatory cytokines by adipose tissue further increases oxidative stress levels and promotes the upregulation of procoagulant factors and adhesion molecules in the endothelium, the downregulation of anticoagulant regulatory proteins, increases thrombin generation, and enhances platelet activation [163]. Not coincidentally, endothelial dysfunction is a common feature of comorbidities that increase the risk for severe COVID-19, including hypertension, obesity, diabetes mellitus, coronary artery disease and heart failure.

Furthermore, because ACE2 is expressed abundantly on vascular endothelial cells of both small and large arteries and veins [164], SARS-CoV-2 infection of endothelial cells can further aggravate endothelial dysfunction. Endothelial damage and dysfunction may thus be the result of cellular infection by SARS-CoV-2, as well as a consequence of obesityassociated excessive systemic inflammation [165]. Evidently, patients with preexisting endothelial dysfunction are more vulnerable to develop severe complications, including coagulopathy, following SARS-CoV-2 infection (See also Figure 1).

## 3.5.4. Vitamin K

Vitamin K is a fat-soluble vitamin, required for the carboxylation of vitamin Kdependent proteins (VKDP). Intrahepatic VKDP include coagulation factors II, VII, IX, X, and different anticoagulant proteins. Extra-hepatic VKDP include diverse gammacarboxyglutamate (Gla) proteins involved in maintaining bone homeostasis, as well as inhibiting ectopic calcification [166]. Vitamin K is not a single entity but a family of structurally related molecules such as phylloquinone, also referred to as vitamin K1, and menaquinone or vitamin K2.

Obesity is linked with vitamin K deficiency [167–169]. Vitamin K deficiency results in decreased vitamin K-dependent carboxylation and phosphorylation of Gla-proteins and, as a consequence, into elevated levels of circulating desphosphorylated-uncarboxylated matrix Gla protein (dp-ucMGP). Recent research studied the relationship between obesity and serum dp-ucMGP in a cohort of 278 Chinese Han people. The results demonstrated that serum dp-ucMGP level was positively associated with visceral fat index, waist height ratio, but not BMI [167]. In addition, lower vitamin K2 levels were observed in obese hemodialysis patients compared to non-obese patients [168]. Also, a lower vitamin K1

status was observed in patients with obesity compared to healthy individuals [169]. Yet, differences in dietary K1 intake could not fully explain this observation. It has been suggested that vitamin K accumulates in adipose tissue, thereby reducing the bioavailability of this vitamin in individuals with obesity [169]. This theory could be plausible since a similar mechanism has been established for other fat-soluble vitamins [170,171]. However, understanding the mechanisms underlying the association between obesity and vitamin K deficiency remains a topic for further investigation.

Vitamin K status was also assessed in COVID-19 patients by measuring dp-ucMGP. Levels of dp-ucMGP were significantly elevated in COVID-19 patients compared to controls and were higher in COVID-19 patients with unfavorable disease outcome. Carboxylated matrix Gla-protein protect against degradation and calcification of vasculature and elastic fibers in the extracellular matrix of the lungs. In COVID-19, elastic fiber degradation combined with calcification of these fibers due to low vitamin K status could aggravate lung injury and lung fibrosis [172].

Thus, to summarize, we can argue that several factors associated with obesity increase the risk for CAC, including elevated leptin levels, increased PAI-1 levels, endothelial dysfunction and low vitamin K levels. The latter is possibly also involved in the development of more severe lung injury in COVID-19.

#### **4. Conclusions and Implications for Further Research**

Obesity cannot simply be defined as an excess of fat cells. Adipose tissue releases many active substances, such as adipokines and components of the RAS, all influencing the brain and metabolic- and immune system. Being obese increases the risk of SARS-CoV-2 infection and complications via several mechanisms. First, viral entry is enhanced due to increased ACE2, csGRP78 and presumably HSPG and NRP-1 expression levels in various cell types, like pneumocytes and adipocytes. Second, the immune system is unable to provide an adequate immune response leading to impaired viral clearance. Eventually, the immune system can overreact as a result of pro-inflammatory 'priming' due to excessive cytokine production by adipose tissue and its related immune cells and high ferritin levels, eventually triggering a cytokine storm [122,139]. Finally, hyperleptinemia, PAI-1 production by adipocytes and endothelial cells, endothelial dysfunction and low levels, or low bioavailability, of vitamin K, all increase the risk for the development of thrombus formation and hemorrhage.

An important lesson learned from the coronavirus pandemic is the importance of a healthy lifestyle to positively influence the course of COVID-19 disease. A non-processed nutrient-rich diet [173–175], limited excessive or overly energy-rich food, sufficient and intensive exercise [176], sufficient sleep [177] and avoiding chronic psycho-emotional stress [178] are all efficient health-promoting measures in the prevention of obesity [179]. We also advocate an integrated multidisciplinary approach in the fight against COVID-19. Future research must identify causes of severity and complications to develop efficient preventive measures and curative interventions.

**Author Contributions:** F.D. developed the concepts, reviewed the literature and drafted the manuscript. M.v.H. provided editorial assistance. K.d.P. and F.v.D. developed the concepts and provided final editorial oversight. All authors have read and agreed to the published version of the manuscript.

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

**Data Availability Statement:** Not applicable.

**Acknowledgments:** We are grateful for the input of all the members of the Natura Foundation research team including Michaëla Van Langeveld, Margo Peinemann, Aldert Hoogland, Lotte Bos, Iris Brunsmann and André Frankhuizen.

**Conflicts of Interest:** Authors F.D., K.d.P., M.v.H. and F.v.D. were sponsored by Bonusan B.V. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.

## **References**


MDPI St. Alban-Anlage 66 4052 Basel Switzerland Tel. +41 61 683 77 34 Fax +41 61 302 89 18 www.mdpi.com

*Cells* Editorial Office E-mail: cells@mdpi.com www.mdpi.com/journal/cells

Academic Open Access Publishing

mdpi.com ISBN 978-3-0365-8674-8