1 Introduction

Carotenoids are liposoluble C-40-based isoprenoid pigments usually red, orange or yellow in color produced by plants and certain photosynthetic microorganisms, in which they function as structural and functional accessories of the photosynthetic apparatus and serve other specific functions. Carotenoids obtained through the diet (mainly fruits and vegetables) also serve specific functions in humans and other mammals. The one best known is to serve as precursors of vitamin A (retinol) and related retinoids such as retinal and retinoic acid that play important roles in the visual cycle and in gene regulation linked to many developmental and physiologic processes, respectively (Grune et al. 2010). The main provitamin A carotenoid in the human diet is β-carotene (BC), followed by α-carotene and β-cryptoxanthin. Non-provitamin A carotenoids – such as lycopene, lutein , and zeaxanthin, which are abundant in human diet and human serum − and the provitamin A carotenoids may serve other functions whose relevance for human health is still uncertain, such as acting as antioxidants and blue light filters for photoprotection (Fiedor and Burda 2014; Johnson 2014). The spectrum of processes impacted by the retinoids in mammals through genomic and non-genomic mechanisms is expanding (Brun et al. 2013). Additionally, novel biological activities of intact carotenoids and carotenoid-derived products other than the retinoids are emerging, which relate to their interaction with mammalian cell-signaling pathways and nuclear receptors transcription factors, sometimes similarly and sometimes quite differently to retinoids (Elliott 2005; Eroglu and Harrison 2013).

In recent years, a novel perspective of the function of carotenoids and carotenoid-derived products is emerging that connects these compounds to the control of adipocyte biology, lipid metabolism and body fat accumulation, with possible implications for the etiology and management of obesity and obesity-related metabolic diseases such as insulin resistance, diabetes and cardiovascular disease (Bonet et al. 2003, 2012; Tourniaire et al. 2009; Landrier et al. 2012; Bonet et al. 2015). This evidence is presented in this chapter.

2 Overview of Carotenoid Metabolism and Its Relationship with Vitamin A Metabolism

Dietary carotenoids are absorbed as part of mixed micelles consisting of lipids and bile components (Harrison 2012; von Lintig 2012; Reboul 2013; Shete and Quadro 2013). The intestinal absorption occurs via passive diffusion or through facilitated transport via scavenger receptor class B, member 1 (SR-BI) and perhaps other lipid transporters, such as CD36. In the enterocyte, BC undergoes central cleavage by β-carotene-15,15′-oxygenase (BCO1) to yield two molecules of retinal (also called retinaldehyde, Rald). BCO1 is a cytosolic enzyme specific for provitamin A carotenoids containing at least one nonsubstituted β -ionone ring (such as BC, α-carotene and β-cryptoxanthin) and is the key enzyme for retinoid production from BC, as evidenced from studies in knockout mouse models (Chapter 16, Hessel et al. 2007; Amengual et al. 2013). Enterocytes and many other cell types express a second carotenoid cleavage enzyme, β-carotene-9’,10′-oxygenase (BCO2), which locates to mitochondria and cleaves BC asymmetrically to generate diverse β-apocarotenals and β-apocarotenones. BCO2 has broad substrate specificity and it is also active on other carotenoids such as the acyclic carotene lycopene and oxygenated carotenoids (i.e. xanthophylls , such as lutein and zeaxanthin) (reviewed in von Lintig 2012). Some BCO2 cleavage products, in particular β-apo-10′-carotenol, can be converted into retinal, with the participation of BCO1 (Amengual et al. 2013).

Within the enterocyte, BC-derived retinal is converted to retinol which, together with preformed retinol in the diet, undergoes esterification to long-chain fatty acids and is packaged as retinyl ester in chylomicrons, for distribution to tissues. The efficiency of intestinal BC cleavage varies greatly across species. Adult rats and mice are efficient cleavers since they convert most absorbed BC to retinol; however, even following mild acute or chronic oral BC supplementation (20–35 mg/kg body weight (bw)/day), some BC is absorbed intact in mice, as revealed by accumulation in tissues including white adipose tissue (WAT) (Lobo et al. 2010a; Amengual et al. 2011a). In humans and other mammals, such as horses and ferrets, substantial amounts of absorbed BC (about 17–45 % of the ingested BC) escape intestinal cleavage and, together with other dietary carotenoids, is incorporated into chylomicrons and found associated with circulating lipoproteins. Intestinal BC absorption and conversion into retinoids are dependent on the animal’s vitamin A status through a feedback mechanism that involves the retinoic acid-dependent induction of an intestine-specific transcription factor, ISX, that negatively regulates the expression of both SR-BI and BCO1 (Lobo et al. 2010b).

Circulating carotenoids and retinyl esters in lipoproteins are taken up by tissues through the action of lipoprotein-specific receptors, such as SR-BI (which is the receptor for HDL), low-density lipoprotein receptor (LDLr), LDLr related protein-1 or lipoprotein lipase (LPL) (Shete and Quadro 2013), or the action of CD36, the latter implicated in particular in the cellular uptake of lycopene in adipocytes (Moussa et al. 2011). Liver, followed by adipose tissue, kidney, skin and lung are important sites of accumulation of BC and other carotenoids (Shete and Quadro 2013). The intracellular metabolism of carotenoids is not well known. Importantly, BCO1 and BCO2 are broadly expressed in peripheral tissues besides the intestinal mucosa; such widespread expression, together with the wide distribution of carotenoids within the body, has suggested that local tissue-specific conversion of carotenoids may contribute to the in situ generation of retinoids and other apocarotenoids capable of impacting tissue metabolism (von Lintig 2012). Additionally, BCO2 activity in mitochondria, the site of its subcellular localization, appears to be required to avoid mitochondrial dysfunction due to carotenoid accumulation in the organelle (Lindqvist and Andersson 2002; Amengual et al. 2011b; Lobo et al. 2012).

Vitamin A in the body is mainly stored as retinyl esters in liver stellate cells. Mobilization of vitamin A from hepatic stores depends on the 21-kDa retinol-binding protein (RBP, also known as RBP4) produced in hepatocytes and secreted into the circulation in a retinol-dependent manner. Extrahepatic tissues take retinol from the retinol-RBP (holo-RBP) complexes (see Fig. 15.1). The plasma membrane protein STRA6 (stimulated retinoic acid 6) is a high affinity RBP receptor and has been implicated in retinol uptake from holo-RBP complexes in tissues such as the eye (Kawaguchi et al. 2007; Berry et al. 2013), as well as in mediating cellular responses to holo-RBP related to the activation of the JAK-STAT pathway (Berry et al. 2011, 2013). A STRA6 paralog, RBPR2 (for RBP receptor 2) is characteristically expressed in liver and intestine and strongly induced in adipose tissue of obese mice (Alapatt et al. 2013). Retinol in cells can be esterified to fatty acids catalyzed by lecithin:retinol acyltransferase (LRAT) and possibly other acyltransferases or be reversibly metabolized to Rald via alcohol dehydrogenases (ADHs) or retinol dehydrogenases (RDHs). Rald can also be obtained from intracellular BC. Retinoic acid is produced from Rald, through irreversible oxidation catalyzed by aldehyde dehydrogenase-1 family of enzymes (Aldh1, also known as Raldh). Intracellular metabolism of retinoids takes place with them bound to specific binding proteins, which may represent an important mechanism to direct the various retinoids to specific metabolizing enzymes and targets. These intracellular binding proteins include cellular retinol-binding proteins (CRBPs), which chaperone retinyl ester biosynthesis catalyzed by LRAT, and cellular retinoic acid binding proteins (CRABPs) and fatty acid binding protein 5 (FABP5), which are involved in delivering retinoic acid to the nucleus and to specific nuclear receptors within it.

Fig. 15.1
figure 1

(continued) Overview of cellular carotenoid and retinoid metabolism. Circulating retinol (ROL) bound to retinol binding protein (RBP) is internalized in cells through the action of specific surface receptors (STRA6, RBPR2) or by diffusion across the plasma membrane. Efficient ROL uptake depends on the activity of lecithin:retinol acyltransferase (LRAT) in esterifying ROL to fatty acids to form retinyl esters (RE), which can be stored in lipid droplets. RE associated with circulating lipoproteins can be hydrolyzed to ROL by lipoprotein lipase (LPL) and taken up by the cells. Alternatively, circulating lipoproteins containing RE and carotenoids can be internalized in cells whole by endocytosis, through the action of lipoprotein receptors. Within the cell, RE is hydrolyzed by RE hydrolases (REH) to release ROL, which can be reversibly oxidized to retinaldehyde (Rald) by several dehydrogenases (of the RDH and ADH families). Rald is irreversibly oxidized to retinoic acid (RA) by the action of aldehyde dehydrogenases (ALDH). RA can also be taken up from the circulation where it is found bound to albumin, and is catabolized by the cytochrome enzymes (CYP) to oxidized products that are excreted from cells. Carotenoids within the cell can be cleaved by β-carotene oxygenases (BCO) to give rise to Rald (formed mainly upon cleavage of β-carotene and other provitamin A carotenoids by BCO1) and other metabolites (apocarotenals and apocarotenones). Retinoids (RA, Rald) and other carotenoid metabolites exert genomic effects by modulating the activity of distinct nuclear receptors (NR) and other transcription factors such as nuclear factor κB (NF-κB) and the nuclear factor erythroid 2-related factor 2 (Nrf2), through direct and indirect manners, and they exert as well non-genomic effects including reactive oxygen species (ROS) scavenging. Not shown in the Figure are specific cellular retinol and retinoic acid binding proteins which help solubilizing these retinoids and chaperoning their intracellular interactions

Full size image

3 Molecular Bases of the Biological Activity of Carotenoids and Carotenoid Derivatives in Adipose Tissue Biology and Obesity

Carotenoids and carotenoid conversion products impact gene expression and cell function through multiple mechanisms, notably by interacting with transcription factors of the nuclear receptor superfamily. Gene expression control by retinoic acid has been most studied (Bastien and Rochette-Egly 2004). Retinoic acid isomers are agonist ligands of two subfamilies of nuclear receptors, the retinoic acid receptors (RARs; 3 isoforms, α, β and γ) and the retinoid X receptors (RXRs; 3 isoforms: α, β and γ): all-trans retinoic acid (atRA) binds the RARs, whereas 9-cis retinoic acid binds both the RARs and the RXRs. Heterodimers of RAR with RXR control the expression of typical retinoid-target genes by binding to defined retinoic acid response elements in the gene promoter and modulating transcription in a manner that depends on atRA binding to the RAR moiety and subsequent recruitment of cofactor complexes (Bastien and Rochette-Egly 2004). Several genes encoding proteins in energy and lipid metabolism are up-regulated at the transcriptional level by RAR-dependent pathways, among them genes for uncoupling proteins one and three (UCP1 and UCP3, involved in inefficient substrate (fat) oxidation), medium-chain acyl-CoA dehydrogenase (involved in mitochondrial β-oxidation) and hormone sensitive lipase (involved in adipocyte lipolysis) (reviewed in (Bonet et al. 2012). RXRs are obligate dimerization partners for many other nuclear receptors besides the RARs, including the peroxisome proliferator activated receptors (PPARs), liver X receptor (LXR), farnesoid X receptor, pregnane X receptor, thyroid hormone receptor and vitamin D receptor. Some RXR heterodimers (so-called permissive) respond to ligands of either partner and are synergistically activated when both ligands are bound, providing a mechanism for widespread effects of retinoids on gene expression (Aranda and Pascual 2001). Notably, LXRs and PPARs, both of which are deeply involved in the control of different aspects of lipid metabolism, act on (at least some) target genes as permissive heterodimers with RXR (Willy et al. 1995; Mukherjee et al. 1997). Moreover, besides activating the canonical RARs, atRA (but not 9-cis retinoic acid) can behave as an activating agonist ligand of PPARβ/δ (but not the other PPAR isoforms) (Shaw et al. 2003), although there are results conflicting (Rieck et al. 2008). PPARβ/δ activation increases lipid catabolism in skeletal muscle and WAT, prevents the development of obesity and improves insulin sensitivity in obesity-prone mouse models (Luquet et al. 2005; Wang et al. 2003; Wang et al. 2004; Lee et al. 2006). The partitioning of atRA between RARs and PPARβ/δ in cells may depend on the relative expression levels of CRABP-II and FABP5, which appear to deliver atRA to RARs and PPARβ/δ, respectively (Schug et al. 2007).

Besides retinoic acid, other retinoids and apocarotenoids different from retinoids interact with nuclear receptor transcription factors to either antagonize or promote their action on target genes in cells. In particular studies have suggested that, besides atRA, Rald (the precursor of atRA and direct product of BCO1) (Kiefer et al. 2012), apo-10′-lycopenoic acid (a product of eccentric cleavage of lycopene) (Gouranton et al. 2011) and intact β-cryptoxanthin (Shirakura et al. 2011) can behave as RAR agonist ligands. On the other hand, some β-apocarotenoids resulting from eccentric cleavage of BC, such as β-apo-14′-carotenal, β-apo-14′-carotenoic acid and β-apo-13-carotenone appear to behave as RAR-antagonists, at least in cells of hepatic origin (Eroglu et al. 2012). Because some of these compounds are contained in foods and found in human plasma, it has been proposed they may act as naturally occurring retinoid antagonists (reviewed in (Eroglu and Harrison 2013). Studies have also suggested PPARγ antagonism by Rald (Ziouzenkova et al. 2007a), β-apo-14′-carotenal (a product of eccentric cleavage of BC) (Ziouzenkova et al. 2007b) and intact astaxanthin (Inoue et al. 2012).

Carotenoids and carotenoid derivatives impact mammalian biology by additional mechanisms, besides their physical interaction with nuclear receptors (Breitman and Takahashi 1996; Blomhoff and Blomhoff 2006; Bonet et al. 2012; Al Tanoury et al. 2013; Eroglu and Harrison 2013; Kaulmann and Bohn 2014). For instance, liganded RAR (i.e., atRA-bound RAR) can interfere with the activity of other transcription factors such as activator protein-1 or CCAAT-enhancer binding proteins (C/EBPs). In fact, interference with C/EBPs by liganded RAR is an important contributor to the inhibitory effect of atRA on adipocyte differentiation (adipogenesis) (Schwarz et al. 1997). Additionally, carotenoids modulate signaling pathways such as the nuclear factor κB (NF-κB) and the nuclear factor erythroid 2-related factor 2 (Nrf2) pathways, which are associated with inflammatory and oxidative stress responses, respectively. Finally, carotenoids and carotenoid derivatives may impact cell physiology and metabolism through extragenomic actions including scavenging of reactive oxygen species (ROS), retinoylation (acylation by retinoic acid) of proteins and activation of protein kinase cascades.

4 Aspects of Adipose Tissue Biology Affected by Carotenoids and Carotenoid Derivatives

Adipose tissue is an important site of carotenoid (Kaplan et al. 1990; Parker 1989) and retinol storage/accumulation (Tsutsumi et al. 1992). It has been estimated that 15–20 % of the total body retinol in rats is stored in WAT, in particular in the adipocytes (rather than in the stromal vascular cells) and mostly in the form of non-esterified retinol (as opposed to preferential storage of vitamin A as retinyl ester in the liver) (Tsutsumi et al. 1992). Carotenoids are found in adipocytes mainly with triacylglycerol in the lipid droplet, and also in association to cell membranes (Gouranton et al. 2008). In humans, carotenoid concentrations in abdominal fat depots show a strong association with both dietary carotenoid intake and plasma carotenoid concentrations (Kohlmeier and Kohlmeier 1995; Chung et al. 2009; El-Sohemy et al. 2002).

Pro-vitamin A carotenoids and retinol in adipocytes may serve to regulate systemic vitamin A homeostasis (in fact, WAT produces RBP and adipose retinol/retinyl esters stores are readily mobilized under conditions of dietary vitamin A deficiency (Frey and Vogel 2011)) and may also serve specific functions in adipose tissue and within the mature adipocytes. Different lines of evidence sustain a role for carotenoids and retinoids in adipose tissue biology. WAT expresses all intracellular binding proteins and enzymes involved in retinol and retinoic acid production and metabolism, including BCO1 and BCO2 (Hessel et al. 2007; Tourniaire et al. 2009; Landrier et al. 2012). Besides retinol, other retinoids including Rald and atRA have been detected in WAT (Kane et al. 2008a; Kane et al. 2008b). Studies suggest crosstalk of intracellular retinoid metabolism and lipid droplet dynamics, with an association of enzymes of retinoid metabolism with the lipid droplet coat which appears to be dependent on active acyl ester biosynthesis (Jiang and Napoli 2012, 2013). Animal (Sima et al. 2011) and human (Gerhard et al. 2014; Kiefer et al. 2012; Perez-Perez et al. 2009; Peinado et al. 2010) studies – including omic studies – have revealed a differential expression of genes for carotenoid/retinoid metabolizing enzymes in visceral and subcutaneous adipose tissue, which display important differences regarding developmental origin, metabolism, endocrinology, capacity for adipogenesis and the health risk they entail (Hamdy et al. 2006; Lee et al. 2013). Genetic ablation of different carotenoid/ retinoid-metabolising enzymes and transport proteins – such as BCO1, Aldh1a1, Rdh1, retinol saturase and CRBPs I and III – results in alterations of adiposity in mice (Hessel et al. 2007; Ziouzenkova et al. 2007a; Zhang et al. 2007; Schupp et al. 2009; Zizola et al. 2008, 2010; Kiefer et al. 2012) (reviewed in (Frey and Vogel 2011)). Finally, treatment/supplementation experimental studies indicate that specific carotenoids and carotenoid derivatives impact essential aspects of adipose tissue biology including the control of adipogenesis, adipocyte metabolism (relative capacities for fat storage and oxidation), oxidative stress and the production of regulatory signals and inflammatory mediators. These aspects are briefly introduced next.

4.1 Adipogenesis

Adipose tissue can greatly expand to facilitate energy storage through both increased lipid filling within existing mature adipocytes to increase adipocyte size (hypertrophic adipose tissue growth) and increased differentiation of adipocyte precursor cells to increase adipocyte number (hyperplastic adipose tissue growth or adipogenesis). The latter process is triggered by nutritional and hormonal signals that activate a cascade of transcription factors in preadipocytes including the C/EBPs and PPARγ (reviewed in (Farmer 2006)). PPARγ is considered the master regulator of adipogenesis and it is also required to maintain the adipocyte phenotype (Tamori et al. 2002) and to mediate high fat diet-induced adipocyte hypertrophy (Kubota et al. 1999), as it transactivates genes for proteins that facilitate uptake, cytosolic binding and activation of fatty acids for triacylglycerol synthesis and lipid droplet formation and maintenance (Dalen et al. 2004) and references therein)]. Even in (human) adult adipose tissue, about 10 % of adipocytes turn over every year (Spalding et al. 2008) and, both in rodents (Klyde and Hirsch 1979) and humans (Tchoukalova et al. 2010), adipogenesis can be induced by environmental cues such as consumption of a high fat diet. Proper adipogenesis is critical for maintaining health: defects in adipogenesis in the face of a positive energy balance (i.e. more energy ingested than expended per day) may result in ectopic fat deposition and lipotoxicity, leading to insulin resistance, diabetes and vascular complications (Virtue and Vidal-Puig 2010). However, many obese humans have more than the average number of adipocytes and weight loss decreases the volume of adipocytes but not adipocyte number, which may facilitate body weight regain (Arner and Spalding 2010). In fact, studies in rodents have suggested that increased adipocyte number might per se lead to obesity (Naaz et al. 2004). In this context, the control of adipogenesis emerges as a potential co-adjuvant therapeutic target in obesity when coupled to strategies enhancing a negative energy balance, and agents capable of tipping the adipocyte birth-death balance in favor of reducing the number of fat cells become of interest (Arner and Spalding 2010). Remarkably, accumulating evidence links carotenoids and carotenoid conversion products to the inhibition of adipogenesis and fat storage capacity of mature adipocytes through suppression of PPARγ, either by acting as direct PPARγ antagonists or by repressing PPARγ secondarily to RAR activation.

4.2 Adipocyte Energy Metabolism

Not all adipocytes in mammals have equal metabolic properties. Brown adipocytes in typical brown adipose tissue (BAT) depots, on the one hand, are rich in mitochondria, have a high oxidative capacity and are specialized in the regulated production of heat (nonshivering adaptive thermogenesis) through oxidation of fatty acids and other fuels linked to the subsequent dissipation of the proton electrochemical gradient generated by the respiratory chain via uncoupling protein 1 (UCP1), an inner mitochondrial membrane protein that behaves functionally as a proton transporter (Cannon and Nedergaard 2004). Typical white adipocytes in WAT depots, on the other hand, are poor in mitochondria, have a low oxidative capacity, do not express UCP1 and are specialized in the storage and release of energy, according to biological needs (besides their endocrine function). A third type are the so-called “brite” (from brown-in-white) or “beige” adipocytes: these are brown-like, UCP1-expressing adipose cells which can be induced in WAT depots in response to cold and a variety of nutritional and pharmacological factors, in a process known as browning or WAT-to-BAT remodeling (Bonet et al. 2013; Giralt and Villarroya 2013). Until recently, BAT was believed to play a negligible role in the adult human and thought to be only present in neonates. However, it has gained substantial interest since active BAT has been shown to be present in adults and BAT activity is negatively associated with increasing BMI in humans (Saito 2013). Increasing BAT activity and promoting the induction of beige adipocytes in WAT both represent attractive strategies to counteract obesity (Palou et al. 2013; Palou and Bonet 2013). Moreover, pharmacological and nutrient-dependent stimulation of mitochondrial oxidative capacity and fatty oxidation in WAT independent of UCP1 has been documented which may also contribute to a lean phenotype (Flachs et al. 2013). Carotenoids such as fucoxanthin and astaxanthin and retinoids such as atRA and Rald have been shown to impact on BAT thermogenic function and on WAT oxidative capacity and WAT browning to stimulate these processes/capacities, thus favoring energy consumption in adipose tissues (see below). Furthermore, feeding studies in rodents indicate that the thermogenic capacity of BAT is dependent on the animal’s vitamin A status, being reduced following a vitamin A-deficient diet and increased following a vitamin A (retinyl ester)-supplemented diet (Kumar et al. 1999; Bonet et al. 2000; Felipe et al. 2003).

4.3 Adipose Tissue Secretory Function and Inflammation

Besides storing energy, WAT has an important secretory function. Thus, WAT actively participates in the systemic control of energy balance, glucose homeostasis, insulin sensitivity, inflammation, vascular haemostasis and other processes through the secretion of signaling molecules, among them many signaling proteins (collectively named adipokines) produced by the adipocytes and/or cells of the stromal-vascular fraction, often in concert. The two roles of WAT are closely related to one another. Thus, the production of many of these signals, including endocrine signals and immunomodulatory factors, is altered in obesity, which is nowadays recognized as a state of chronic, low-grade local (WAT) and systemic inflammation, which links central obesity to metabolic disturbances characteristic of the metabolic syndrome (Wisse 2004; Wellen and Hotamisligil 2005). Both hypertrophied adipocytes and infiltrating macrophages in WAT are a source of inflammatory mediators in obesity. Anti-inflammatory properties of BC and non-provitamin A carotenoids (such as lycopene, astaxanthin or fucoxanthin, among other) have been demonstrated in different contexts and tissues/cell types and are thought to arise from the ability of these compounds to activate Nrf2, thus reducing oxidative stress, and to suppress NF-κB activation, thus inhibiting the downstream production of inflammatory cytokines (Kaulmann and Bohn 2014). Carotenoids and their conversion products can affect the inflammatory and secretory profile of adipose tissue by actively interacting with these pathways in adipocytes and adipose tissue macrophages, by impacting the expression of specific adipokines through effects on other specific transcription factors, and/or in a passive manner, i.e. secondary to their effects on adipocyte lipid content and body fat. Remarkably, cross-sectional and prospective studies in humans fairly consistently show that a higher intake and status of carotenoids is associated with lower levels of low-grade inflammation in relation to overweight, obesity and the metabolic syndrome (Calder et al. 2011).

4.4 Oxidative Stress in Adipocytes

Oxidative stress generated by adipose tissue in the context of obesity is an important pathogenic mechanism of obesity-associated metabolic syndrome (Furukawa et al. 2004; Le Lay et al. 2014) tightly linked to adipose tissue inflammation and secretion of systemic inflammatory mediators, since ROS activate inflammatory pathways. Cell and in vivo studies have suggested that carotenoids have antioxidant properties that stand both from their ability to scavenge several ROS and to interact with and potentiate the Nrf2 pathway, enhancing Nrf2 translocation to the nucleus and the subsequent activation of the expression of a collection of Nrf2 target genes encoding antioxidant and cytoprotective enzymes (Kaulmann and Bohn 2014). However, carotenoids may act as prooxidants depending on their concentration in cells, the cell oxidative environment and other factors (Palozza 1998; Rodriguez et al. 2005; Fuster et al. 2008; van Helden et al. 2009). In fact, studies suggest that, as in other cell types and tissues, carotenoids in adipocytes/adipose tissue may protect against oxidative stress or contribute to it, depending on the dose and cell/tissue factors.

5 Carotenoids as Modulators of Adiposity and Obesity: Cell an Animal Studies

5.1 Beta Carotene and Beta Carotene Derivatives

5.1.1 Cell Studies

BC (50 μM) inhibited the adipose conversion of murine 3 T3-L1 preadipocytes, the prototypical model for studies of adipogenesis (Kawada et al. 2000). This is most likely due to BC conversion into atRA, whose inhibitory effect on adipogenesis when applied at relatively high doses (0.1–10 μM) at defined, early stages of the process is well known (Murray and Russell 1980; Kuri-Harcuch 1982; Xue et al. 1996). Inhibition of adipogenesis by atRA is explained by several non-mutually exclusive mechanisms which translate into repression of PPARγ, the master transcription factor for adipogenesis. One is the interference of liganded RAR with the activity of the early transcription factor C/EBPβ on its downstream target genes in the adipogenic program (Schwarz et al. 1997; Marchildon et al. 2010). Additionally, atRA works upstream C/EBPβ in preadipocytes by inducing (at the transcriptional level, through an RAR pathway) specific proteins that inhibit adipogenesis (such as Pref-1, Sox9 and KLF2) (Berry et al. 2012). Retinoylation of regulatory proteins resulting in reduced PPARγ availability for transcription regulation also appears to play a role, this one independent of RAR and RXR (Dave et al. 2014). Promotion of apoptosis of primary rat preadipocytes and clonal preadipocytes by atRA has also been reported (Kim et al. 2000; Chawla and Lazar 1994).

Remarkably, atRA effects on adipogenesis are concentration-dependent. Thus, whereas relatively high doses of atRA suppress adipogenesis, atRA induces the opposite effect, i.e. it promotes adipogenesis, when applied at low doses (1 pM to 10 nM range) (Safonova et al. 1994). Moreover, stem cell commitment into the adipocyte lineage requires a time-defined treatment with atRA (Dani et al. 1997; Bost et al. 2002), and endogenous RA production appears to be required for efficient adipogenesis of 3 T3-L1 fibroblast cell lines (Reichert et al. 2011).

BC metabolites other than atRA have been shown to repress adipogenesis of preadipocyte cell lines. This is the case of Rald (Kawada et al. 2000; Ziouzenkova et al. 2007a) and certain β-apo-carotenals resulting from asymmetric cleavage of BC, such as β-apo-8′-carotenal (Kawada et al. 2000) and, especially, β-apo-14′-carotenal (Ziouzenkova et al. 2007b). Rald and β-apo-14′-carotenal behave as weak RAR agonists, but their inhibitory effect on adipogenesis has been traced to their ability to suppress PPARγ- and RXR-mediated responses trough RAR- independent mechanisms and possibly following their direct physical binding to PPARγ and RXR (Ziouzenkova et al. 2007a, b).

Besides impacting adipogenesis, carotenoids and carotenoid derivatives impact metabolism of mature adipocytes. In particular, exposure to BC (2 μM) boosted the production of atRA and reduced lipid content and the expression of PPARγ and PPARγ target genes in mature 3 T3-L1 adipocytes (Lobo et al. 2010a). These effects were BCO1- and RAR-dependent, as they were abrogated by both a BCO1 inhibitor and a pan-RAR antagonist (Lobo et al. 2010a). Importantly, BC effects in 3 T3-L1 adipocytes were not reproduced upon incubation of the cells with vitamin A (retinol), suggesting that BC is a key precursor for retinoid production in mature adipocytes (Lobo et al. 2010a). Remarkably, while BCO1-derived products repress PPARγ, BCO1 itself is encoded in a PPARγ target gene (Boulanger et al. 2003) that is induced during adipocyte differentiation (Lobo et al. 2010a), which is suggestive of a negative feed-back loop to keep control of PPARγ.

Early work showed that BC and other provitamin A carotenoids can induce UCP1 expression in primary brown adipocytes (Serra et al. 1999), possibly reflecting local conversion to atRA, which is a potent signal for UCP1 gene transcription (Alvarez et al. 1995; Puigserver et al. 1996; del Mar Gonzalez-Barroso et al. 2000). The retinoid responsiveness of the UCP1 gene is explained by the presence of both a noncanonical retinoic acid response element and PPAR response element in its promoter (Larose et al. 1996; Rabelo et al. 1996; Sears et al. 1996), and is mediated by RARs and RXRs (Alvarez et al. 2000) and also by p38 mitogen-activated protein kinase (p38 MAPK) (Teruel et al. 2003).

Besides inducing UCP1 in brown adipocytes, retinoids stimulate the acquisition of brown adipocyte features in white adipocytes. Thus, exposure to atRA increased basal lipolysis and fatty acid oxidation rate, triggered changes in gene expression consistent with increased fatty acid mobilization, oxidation and turnover and, upon prolonged treatment, reduced intracellular lipid content in mature 3 T3-L1 adipocytes (Mercader et al. 2007). More recently, atRA has been shown to induce oxidative phosphorylation and mitochondria biogenesis in adipocytes (Tourniaire et al. 2015). Even if atRA treatment failed to induce UCP1 in 3 T3-L1 adipocytes (Mercader et al. 2010; Murholm et al. 2013), it potently induced UCP1 in other murine models of white adipocytes, such as adipocytes derived from primary mouse embryo fibroblasts (MEFs) and mature C3H10T1/2 adipocytes (Mercader et al. 2010; Murholm et al. 2013), in a p38 MAPK-dependent (Mercader et al. 2010), RAR-dependent (Murholm et al. 2013) and PPARγ coactivator 1α -independent manner (Mercader et al. 2010; Murholm et al. 2013). However, atRA failed to induce UCP1 expression in human adipocyte cell lines and primary human white adipocytes (Murholm et al. 2013). Rald was also capable of inducing UCP1 expression in murine MEFs-derived adipocytes, though at higher effective doses than atRA (10 μM vs 1 μM for atRA), as it could be expected for a Rald effect following its intracellular conversion to atRA (Mercader et al. 2010). Subsequently, evidence was presented that Rald can induce UCP1 expression in cultured white adipocytes independently of its conversion to atRA, following binding to and activation of RAR (Kiefer et al. 2012).

Cell-autonomous effects of carotenoids and carotenoid derivatives on the secretory function of mature adipocytes have been demonstrated. Elevated levels of the adipokines resistin (Steppan et al. 2001; Stofkova 2010), leptin (Huang et al. 2004; Stofkova 2009) and RBP4 (Yang et al. 2005; Esteve et al. 2009) have been associated with insulin resistance and inflammation in rodents and humans. We and others have shown that exposure to atRA suppresses adipocyte production of leptin (Bonet et al. 2000; Menendez et al. 2001; Hollung et al. 2004), resistin (Felipe et al. 2004) and RBP4 (Mercader et al. 2008) through transcriptional mechanisms that involve both RAR- and RXR-dependent pathways (Felipe et al. 2004; Felipe et al. 2005; Mercader et al. 2008). Additionally, exposure to atRA limited pro-inflammatory cytokine (interleukins 6 and 1β) expression in TNFα-treated 3 T3-L1 adipocytes (Gouranton et al. 2011). Adiponectin is an adipokine quite specific of adipocytes that is down-regulated in obesity and has well-established insulin-sensitizing, anti-inflammatory and anti-atherogenic effects (Turer and Scherer 2012). In our hands, atRA treatment did not affect adiponectin expression in 3 T3-L1 adipocytes (or atRA-treated mice) (Bonet, Ribot and Palou, unpublished results); however, exposure to BC (20 μM, from days 4 to 8 of differentiation) induced adiponectin expression in differentiating 3 T3-L1 adipocytes (Kameji et al. 2010). All in all, the scenario is suggestive of a provitamin A-independent effect of BC on adiponectin production in adipocytes.

Reports dealing specifically with carotenoids activity on oxidative stress in adipocytes are scarce and suggest that, as in other cell types, carotenoids might be antioxidant or prooxidant in fat cells. For instance, whereas pre-treatment with BC (10 μM) suppressed TNFα-induced ROS production in 3 T3-L1 adipocytes, suggesting that intracellular BC accumulation enables elimination of ROS in these cells (Kameji et al. 2010), exposure to BC (10–30 μM) led to mitochondrial dysfunction in immortalized Chub-S7 human preadipocytes (Sliwa et al. 2012), likely linked to increased oxidative stress.

5.1.2 Animal Studies

Animal studies linking dietary BC supplementation, vitamin A status and retinoid treatment to changes in body adiposity and responses to a high fat diet are addressed in this section.

Dietary BC supplementation (∼35 mg/kg bw/day, 14 weeks) was shown to reduce body fat content without affecting body weight in wild-type mice (Amengual et al. 2011a). The anti-adiposity effect of BC is linked to the suppression of PPARγ and PPARγ target genes in WAT, as indicated by both transcriptomic and targeted gene and protein expression analyses (Amengual et al. 2011a; van Helden et al. 2011), and is dependent on BC conversion to retinoids via BCO1, as it was not seen in BCO1-deficient (BCO1-/-) mice, despite these mutants accumulated in WAT depots large amounts of BC and of the BCO2 product β-10´-apocarotenol (derived from BC) upon supplementation (Amengual et al. 2011a). Importantly, following dietary BC supplementation, BC accumulation was detected in serum and WAT of wild-type mice, supporting the notion that a BCO1-dependent, local production of retinoids from BC may modulate fat storage capacity in adipocytes (Amengual et al. 2011a). In ferrets, which absorb intact BC efficiently but are not as efficient as humans and rodents in converting BC to vitamin A (Lederman et al. 1998; Lee et al. 1999), 6-month BC supplementation (3.2 mg/kg bw/day) did not reduce body fat (Murano et al. 2005; Sanchez et al. 2009), whereas treatment with atRA did (Sanchez et al. 2009). Whether dietary BC can effectively counteract diet-induced obesity in animal models has not been studied, to our knowledge.

Chronic feeding of a diet restricted in retinol (vitamin A-deficient diet) led to increased adiposity and PPARγ expression in WAT depots in rodents (Ribot et al. 2001; Esteban-Pretel et al. 2010), and diets low in carotenoids and vitamin A are used traditionally to favor the development of adipose tissue and the formation of intramuscular fat in the cattle (Kawada et al. 1996; Gorocica-Buenfil et al. 2007), the so-called bovine marbling. Reciprocally, chronic dietary vitamin A supplementation (as retinol or retinyl palmitate, at 40- to 50-fold the control dose) led to reduced adiposity in lean rats (Kumar et al. 1999) and genetically obese WNIN/Ob rats (Jeyakumar et al. 2006, 2008; Sakamuri et al. 2011). However, vitamin A supplementation had only a modest effect counterbalancing the development of diet-induced obesity in obesity-prone mice (Felipe et al. 2003), and did not affect body weight and adiposity gain in response to a cafeteria diet in rats (Bairras et al. 2005). Moreover, there are studies pointing to a pro-obesogenic effect of excess intake of preformed vitamin A (retinol) in critical periods in early life, through mechanisms that may relate to changes in WAT development (Redonnet et al. 2008; Granados et al. 2013) (see also section 7).

In keeping with an anti-adiposity action of retinoids, atRA treatment reduces adiposity and enhances glucose tolerance and insulin sensitivity in both lean and obese rodents (Puigserver et al. 1996; Bonet et al. 2000; Ribot et al. 2001; Felipe et al. 2004, 2005; Mercader et al. 2006, 2008; Strom et al. 2009; Berry and Noy 2009; Amengual et al. 2010; Manolescu et al. 2010). These effects have been evidenced with different atRA dosages (from 0.25 to 100 μg/g bw/day), routes of administration and treatment durations (reviewed in (Bonet et al. 2012)). The slimming action of atRA is not due to reduced food intake, and it has been traced to increased oxidative metabolism and energy expenditure in tissues including BAT, WAT, skeletal muscle and the liver, and to reduced PPARγ expression and activity in WAT depots. atRA treatment in rodents activates BAT with induction of UCP1 (Puigserver et al. 1996; Kumar and Scarpace 1998; Bonet et al. 2000) and in WAT it increases the expression of genes linked to mitochondria biogenesis and function, thermogenis and fatty acid oxidation – including UCP1 protein in the subcutaneous (inguinal) WAT depot – and the appearance of adipocytes with a multilocular distribution of intracellular fat, which are paramount features of WAT browning (Mercader et al. 2006; Tourniaire et al. 2015). Enhancement of lipid catabolism in skeletal muscle (Felipe et al. 2003; Amengual et al. 2008; Berry and Noy 2009) and liver cells (Amengual et al. 2010, 2012) also contribute to the slimming effect of atRA treatment in mice. Most effects of atRA on adiposity in rodents appear to be dependent on RAR and PPARβ/δ activation (Amengual et al. 2008; Berry and Noy 2009; Bonet et al. 2012). Acute atRA treatment also reduced adiposity and increased multilocularity and UCP1 content in the retroperitoneal adipose depot of ferrets (Sanchez et al. 2009). Importantly, simultaneous chronic treatment with atRA has been shown to suppress high fat diet-induced adipogenesis and adipocyte hypertrophy, thus counteracting the development of diet -induced obesity in obesity-prone (C57BL/6) mice (Berry et al. 2012; Noy 2013).

The immediate precursor of atRA, Rald, also appears to have an anti-adiposity action per se in vivo. Mice lacking Aldh1a1 − which is the main aldehyde dehydrogenase involved in atRA production from Rald in adipocytes/ adipose tissue (Reichert et al. 2011; Sima et al. 2011) − had increased Rald levels in WAT and resisted diet-induced obesity due to hypermetabolism (Ziouzenkova et al. 2007a). Moreover, the in vivo knockdown of Aldh1a1 in WAT (through antisense oligonucleotide) limited the development of diet-induced obesity in obesity-prone mice (Kiefer et al. 2012) and administrating Rald or a Aldh1a1 inhibitor reduced subcutaneous fat mass in obese ob/ob mice (Ziouzenkova et al. 2007a). Similar to atRA treatment, the anti-adiposity effect of Aldh1a1 knockout/ knockdown has been related to WAT browning, since lack of Aldh1a1 enhanced cold-induced thermogenesis and induced a BAT-like transcriptional program in visceral WAT of the knockout mice, without changes in BAT (Kiefer et al. 2012), and WAT-selective Aldh1a1 knockdown in adult obese mice triggered WAT browning (Kiefer et al. 2012). It appears, therefore, that atRA and Rald have somewhat redundant effects on adipose tissue metabolism, likely owing to their shared ability to activate RARs.

In good concordance with results in adipocyte cell models, treatment studies in rodents have revealed a down-regulatory effect of atRA treatment on the adipose production of leptin (Kumar and Scarpace 1998; Bonet et al. 2000; Menendez et al. 2001; Hollung et al. 2004), resistin (Felipe et al. 2004) and RBP4 (Mercader et al. 2008). Elevated levels of these three adipokines associate with inflammation and insulin resistance in humans and rodents (Steppan et al. 2001; Stofkova 2010; Huang et al. 2004; Stofkova 2009; Yang et al. 2005; Esteve et al. 2009), and their down-regulation was paralleled by improved insulin sensitivity in the atRA-treated mice (Felipe et al. 2004; Mercader et al. 2008). atRA-induced down-regulation has been demonstrated in humans besides rodents for leptin (Menendez et al. 2001; Hollung et al. 2004). Remarkably, down-regulation of RBP4 by atRA is adipocyte-specific, as hepatic RBP4 expression was unaffected by atRA treatment (Mercader et al. 2008); this is of interest, since it is specifically RBP4 of adipose origin which has been related to inflammation and insulin resistance (although any physical difference between hepatic and adipose RBP4 remains to be established, to our knowledge). Similar to atRA treatment, dietary vitamin A supplementation to rodents resulted in reduced adipose expression and circulating levels of resistin and leptin to an extent that largely exceeded the reduction of adipose mass (Felipe et al. 2004; Kumar et al. 1999; Felipe et al. 2005), and opposite changes of leptin (i.e., up-regulation) were demonstrated in mice fed a vitamin A deficient-diet (Bonet et al. 2000). Moreover, transcriptome analysis revealed leptin, RBP4 and resistin among the top fifty down-regulated genes in inguinal WAT of BC-supplemented wild-type mice, and these effects were provitamin A-dependent, as they were absent in BC-supplemented BCO1-/- mice (Amengual et al. 2011a).

5.2 Cryptoxanthin

β-cryptoxanthin is a provitamin A carotenoid which displays both structural and functional similarities to BC. Oral supplementation with β-cryptoxanthin (0.8 mg/kg bw/day contained in 400 mg of a powder derived from Satsuma mandarins, Citrus unshiu Marc., 8 weeks) reduced body weight, visceral adipose tissue mass, adipocyte hyperthorphy and serum lipid concentrations in a genetic obese mouse model (Tsumura Suzuki obese diabetic), independent of changes in food intake (Takayanagi et al. 2011). Anti-obesity effects have also been reported for mango (Mangifera indica L.) pulp (10 % w/w) – which is a rich source of β-cryptoxanthin, violaxanthin and BC (Cano and de Ancos 1994); supplementation with mango reduced body fat gain and improved glucose tolerance and insulin sensitivity in mice on a high fat diet (Lucas et al. 2011).

The anti-adiposity effect of β-cryptoxanthin in rodents might be in keeping with studies in differentiating 3 T3-L1 adipocytes showing reduction of lipid accumulation following exposure of the cells to β-cryptoxanthin (1–10 μM) (Shirakura et al. 2011; Goto et al. 2013), although there are results conflicting (Okada et al. 2008). β-cryptoxanthin effects on adipogenesis entail RAR activation and subsequent PPARγ down-regulation (Shirakura et al. 2011). Results from in vitro nuclear receptor binding assays indicated that β-cryptoxanthin can efficiently bind RARs (but not PPARγ), which raises the possibility that β-cryptoxanthin acts per se as a RAR agonist to down-regulate PPARγ in adipocytes (Shirakura et al. 2011).

5.3 Astaxanthin

Astaxanthin is a natural antioxidant , non-provitamin A carotenoid, abundant in marine animals. Astaxanthin (6–30 mg/kg bw/day) prevented visceral fat accumulation, metabolic syndrome, and insulin resistance induced by high fat diet feeding in mice and rats (Ikeuchi et al. 2007; Arunkumar et al. 2012), and reduced oxidative stress markers in adipose tissue and skeletal muscle of the high fat diet-fed mice (Arunkumar et al. 2012). The anti-adiposity action of astaxanthin was not due to changes in food intake and was traced to enhanced systemic fatty acid utilization, as indicated by reduced respiratory quotient in indirect calorimetry tests (Ikeuchi et al. 2007). Whether this enhanced fat catabolism occurred in adipose tissue depots was not addressed. Astaxanthin has also been shown to inhibit rosiglitazone (a PPARγ ligand)-induced adipogenesis of 3 T3-L1 cells by antagonizing PPARγ transcriptional activity, possibly upon direct binding, since results of in vitro assays (CoA-BAP) indicated that astaxanthin is able to selecetively bind PPARγ (but not PPARα or PPARβ/δ) (Inoue et al. 2012).

5.4 Fucoxanthin

Fucoxanthin is an orange colored carotenoid present in edible brown seaweeds, such as Undaria pinnatifida (Wakame), Hijikia fusiformis (Hijiki), Laminaria japonica (Ma-Kombu) and Sargassum fulvellum. It is a non-provitamin A xanthophyll whose distinct structure includes an unusual allenic bond (reviewed in (Miyashita et al. 2011)). Fucoxanthin or fucoxanthin-rich seaweed extract, alone or as part of mixtures with other selected agents, counteracts the development of dietary obesity in susceptible mice (when added to the obesogenic diet at 0.05–2 %, w/w) and reduces abdominal WAT in genetically obese KK-Ay mice (when added to the control diet at 0.2 % but not in lean mice. Fucoxanthin reduces WAT mass in obese animals by favoring fatty acid oxidation, UCP1 induction and heat production in abdominal WAT (Maeda et al. 2005, 2007; Jeon et al. 2010; Okada et al. 2011; Hu et al. 2012). Notably, fucoxanthin intake promotes WAT browning at doses at which it does not affect UCP1 expression in BAT, suggesting a WAT selective effect (Maeda et al. 2005). The actual mammalian targets for interaction with fucoxanthin metabolites in mature white fat cells remain to be identified, but enhancement of the sensitivity of adipocytes to sympathetic (SNS) nerve stimulation (which favors thermogenic activation) (Maeda et al. 2009) and activation of AMP-dependent protein kinase (AMPK) (Kang et al. 2012) appear to contribute to the fucoxanthin effects in WAT. Obese (KK-Ay) mice (but not lean mice) fed a fucoxanthin-supplemented (0.2 %) diet had reduced expression levels of pro-inflammatory factors such as monocyte chemoattractant protein-1 and tumor necrosis factor α (TNFα), and a reduced infiltration of macrophages in visceral WAT compared with control animals (Hosokawa et al. 2010). Likewise, treatment with fucoxanthin reduced the expression of pro-inflammatory factors, oxidative stress markers (maleic dialdehyde) and macrophage infiltration in the mammary gland of mice on a high fat diet, although analysis of classical WAT depots was not included in this report (Tan and Hou 2014). Overall, animal studies sustain anti-adiposity and anti-inflammatory actions of fucoxanthin in WAT of obese animals and animals under obesogenic diets.

Interaction of fucoxanthin with adipogenesis of preadipose cells has been described that could contribute to its anti-adiposity action. Fucoxanthin and its metabolite, fucoxanthinol (which is found in WAT of treated animals) suppressed adipogenesis of 3 T3-L1 preadipocytes by down-regulating PPARγ when applied at intermediate and late stages of the adipogenic process (Maeda et al. 2006), although it enhanced adipogenesis when applied at an early stage (coincident with preadipocyte clonal expansion) (Kang et al. 2011). More recently, inhibition of 3 T3-L1 preadipocyte differentiation by Xanthigen™ (a source of fucoxanthin plus pomegranate seed oil extract rich in the conjugated linolenic acid punicic acid) has been traced to down-regulation of PPARγ and C/EBPs, up-regulation of Sirtuin 1, activation of AMPK and modulation of FoxO pathways (Lai et al. 2012). Fucoxanthinol attenuated the TNFα-induced expression of pro-inflammatory cytokines and chemokines in differentiating 3 T3-F442A adipocytes (Hosokawa et al. 2010) and the response of RAW 264.7 macrophages (a macrophage cell line) to pro-inflammatory agents such as lipopolysaccharide or palmitic acid (Kim et al. 2010; Hosokawa et al. 2010). This is of interest because in the obese WAT there is a vicious cycle between adipocytes and macrophages contributing to inflammation. Inhibition of the NF-κB pathway and of mitogen-activated protein kinase pathways (such as the JNK pathway) underlies the anti-inflammatory action of fucoxanthin in adipocytes and macropohages (Kim et al. 2010). In mature adipocytes, exposure to fucoxanthin (up to 10 μM) did not affect basal lipolysis but reduced glucose uptake (Kang et al. 2011).

5.5 Other Carotenoids

Animal studies indicate an anti-adiposity action for some other carotenoids. For instance, crocetin (50 mg/ kg bw/day) – which is a natural antioxidant carotenoid abundant in saffron, Crocus sativus Linn – prevented visceral fat accumulation, metabolic syndrome, and insulin resistance induced by high fat diet feeding in mice and rats, without affecting food intake (Sheng et al. 2008). The effects were traced to enhanced hepatic fatty acid oxidation for crocetin (Sheng et al. 2008). In another report, supplementation with violaxanthin-rich crude thylakoids prepared from spinach leaves reduced body weight and body fat gain in mice on a high fat diet without affecting energy intake, suggestive of the involvement of a metabolic mechanism (Emek et al. 2010).

The acyclic carotenoid lycopene commonly found in tomatoes is well known for its antioxidant properties and evidence is increasing that lycopene or tomato preparations can decrease inflammatory markers, and may improve diseases with chronic inflammatory backgrounds such as obesity (Ghavipour et al. 2013). Lycopene and its metabolite, apo-10´-lycopenoic acid, have been reported to display anti-inflammatory effects in adipocytes (Gouranton et al. 2010; Gouranton et al. 2011) and macrophage-like cells (RAW 264.7) (Marcotorchino et al. 2012) through inhibition of the NF-κB pathway (Gouranton et al. 2010). A supplementation study in animals revealed an anti-inflammatory effect of dietary lycopene on the inflamed obese WAT independent of effects on adiposity (Luvizotto Rde et al. 2013). Thus, in obese Wistar rats lycopene supplementation (10 mg/kg bw/day during 6 weeks) did not affect body weight or adiposity, but decreased leptin, resistin, interleukin 6 and monocyte chemoattractant protein-1 gene expression in gonadal adipose tissue and plasma concentrations of the former three proteins (Luvizotto Rde et al. 2013). Cell studies showed no effect of lycopene and apo-10´-lycopenoic acid on adipogenesis, even if apo-10´-lycopenoic acid was able to activate the RAR and impact the transcription of certain RAR target genes in adipocytes (Gouranton et al. 2011). Overall, results indicate that lycopene does not exert anti-adiposity action but, as in other cell types, has anti-inflammatory properties in adipocytes and obese WAT tissue.

6 Carotenoids as Modulators of Adiposity and Obesity: Human Studies

6.1 Human Epidemiological Studies

Serum levels of carotenoids including BC are reduced in overweight and obese individuals, both adults and children/adolescents (e.g. (Moor de Burgos et al. 1992; Decsi et al. 1997; Yeum et al. 1998; Sarni et al. 2005; Burrows et al. 2009)). Furthermore, several large population-based, cross-sectional epidemiological studies indicate an inverse association between carotenoid concentrations in blood and BMI and other measures of obesity including adiposity, in some cases even when adjusted for other factors associated with carotenoid concentration, such as intake of fruit and vegetables, fat, fibre, alcohol, supplement use, smoking, gender and lipid concentrations (Brady et al. 1996; Strauss 1999; Neuhouser et al. 2001; Wallstrom et al. 2001; Ford et al. 2002; Kimmons et al. 2006; Suzuki et al. 2006; Andersen et al. 2006; de Souza Valente da Silva et al. 2007; Wang et al. 2008; Gunanti et al. 2014). Cross-sectional studies have also reported lower serum carotenoid concentrations in adults and children with the metabolic syndrome (Beydoun et al. 2012; Beydoun et al. 2011). In fact, independent associations of low serum carotenoids with risk factors/biomarkers of the metabolic syndrome including increased insulin resistance index (HOMA-IR), fasting insulinemia, oxidized LDL, glycosylated hemoglobin, and circulating levels of inflammatory markers such as C-reactive protein have been reported, that persisted after adjusting for confounders including BMI or other obesity-related measures (Ford et al. 1999; Kritchevsky et al. 2000; Erlinger et al. 2001; Ford et al. 2003; van Herpen-Broekmans et al. 2004; Sugiura et al. 2006; Hozawa et al. 2007; Beck et al. 2008; Wang et al. 2008). These latter associations are generally considered to be related to antioxidant and anti-inflammatory activities of carotenoids.

Several arguments are generally provided to explain the inverse association between carotenoid concentrations in blood and BMI or adiposity. First, BMI and serum carotenoids may be correlated because of dietary and other lifestyle factors that affect them both. Second, since serum carotenoids are partially fat soluble, adipose tissue may act as a sink for them, so that relatively fewer are located in the blood (Brady et al. 1996; Wallstrom et al. 2001). However, the concentration of carotenoids in adipose tissue and isolated adipocytes is also lower in obese people (Chung et al. 2009; Virtanen et al. 1996; Kabagambe et al. 2005; Osth et al. 2014). A recent report showed that isolated adipocytes from obese subjects contain 50 % lower concentrations of BC than cells from lean or non-obese subjects (Osth et al. 2014). Third, adipose tissue in obesity generates oxidative stress (Higdon and Frei 2003), and the carotenoids may be reduced because of defending against this stress and being consumed upon their action as antioxidants (see (Andersen et al. 2006) and references therein), leading to reduced carotenoids levels in both blood and adipose tissue in obesity.

Notwithstanding the aforementioned explanations, we believe a mechanistic link cannot be ruled out in view of data indicating that, in adipocytes, BC rather than retinol may function as the precursor for the local synthesis of retinoids capable of exerting an anti-adiposity action (Lobo et al. 2010a). Reduced BC per adipocyte in obesity could reflect increased BC consumption in antioxidant reactions needed to neutralize increased reactive species, and perhaps also an attempt to counteract adipocyte hypertrophy by increasing retinoid production from BC. Whatever its origin, if not corrected through dietary consumption, once established reduced BC per adipose cell could contribute to the maintenance and further development of pathological obesity by limiting BC-derived anti-adiposity retinoid signaling (Fig. 15.2). Moreover, obese subjects may have a lowered capability to convert BC to retinoids, since an inverse association of BMI with BC conversion efficiency as assessed by a stable-isotope reference method was demonstrated in humans (Tang et al. 2003) and another study found a slower rate of BC decline in serum after cessation of dietary supplementation in individuals with the highest BMI (Wise et al. 2009). Reduced efficiency of BC conversion to retinoids in the obese subjects could further contribute to their obese state.

Fig. 15.2
figure 2

Proposed involvement of β-carotene adipose stores in counteracting adiposopathy. β-Carotene (BC) and possibly other provitamin A carotenoids in adipocytes may serve to scavenge reactive oxygen species (ROS) and to produce retinoids such as retinoic acid (RA) capable of repressing adipocyte hyperthrophy (by suppressing PPARγ and enhancing energy utilization, see text). These activities may help counteracting adipose tissue pathological expansion and keeping the individual lean. However, BC as such is consumed in the course of these activities. An inadequate intake of dietary carotenoids on the top of an obesogenic diet and lifestyle would led to depletion of adipose BC stores and suppression of local RA production, this contributing to increased ROS, adipocyte hyperthrophy, and hence the development of pathological obesity

Full size image

An association of vitamin A status (as serum retinol levels) with human obesity is less clear. Some of the studies reporting an inverse association of carotenoids in blood and obesity found instead serum retinol to be constant across BMI values or between obese and non-obese groups (Decsi et al. 1997; Neuhouser et al. 2001; Sarni et al. 2005; de Souza Valente da Silva et al. 2007), likely reflecting homeostatic regulation of circulating retinol through controlled storage in and release from the liver. A recent study reported higher concentrations of serum retinol (and lower of serum carotenoids) to be associated with increased probability of overweight and obesity in children (Gunanti et al. 2014). Other studies, in overweight and obese (adult) subjects, found the opposite, i.e. lower serum retinol to be associated with increased BMI (Viroonudomphol et al. 2003; Botella-Carretero et al. 2010).

Studies have also investigated possible associations of dietary intake of vitamin A or carotenoids as evaluated through food frequency questionnaires with measurements of adiposity. Early studies highlighted vitamin A as one of few micronutrients with a frequently inadequate intake among population samples with a high prevalence of obesity (Wolfe and Sanjur 1988; Vaughan et al. 1997). Accordingly, an inverse association between preformed vitamin A intake and several measurements of adiposity − such as body weight, BMI, waist circumference and waist-to-hip ratio – after adjusting for total energy intake has been reported in healthy young adults (Zulet et al. 2008). Higher total carotenoid intakes, mainly those of BC and lycopene, were found to be associated with lower waist circumferences and visceral and subcutaneous fat mass and lower prevalence of metabolic syndrome in a cross-sectional study involving middle-aged and elderly men (n = 374) (Sluijs et al. 2009). A high intake of carotenoids derived from a high consumption of vegetables and fruit, as in the Mediterranean diet, was shown to associate with lower development of metabolic syndrome traits including increased waist circumference in a prospective study involving 3232 subjects (Kesse-Guyot et al. 2013). Altogether, the results of these studies point to an association between higher dietary intakes of vitamin A and carotenoids and reduced adiposity.

6.2 Human Intervention Studies

Randomized controlled intervention studies specifically designed to test the impact of dietary carotenoid supplementation on adiposity are scarce, and have been conducted mainly for BC, β-cryptoxanthin and fucoxanthin. Two small pilot double blind placebo-controlled studies in overweight and obese children using similar doses of BC (3–4 mg/day as part of supplements) reported significant changes in the 6 month rate of accrual in abdominal adiposity (Canas et al. 2012, 2014). In the first of these two studies, lean and overweight children underwent daily supplementation with an encapsulated supplement of fruit and vegetable juice concentrate (providing approximately 3.75 mg of BC, 117 mg of vitamin C, 22.5 IU of vitamin E, 210 mg of folate and 30 mg of calcium per day) or placebo in the presence of nutritional counseling; the supplement led to increased serum BC levels and resulted in a reduction in abdominal fat mass in conjunction with an improvement in insulin resistance in the overweight children (Canas et al. 2012). In the second study, obese children completed a 2-week intense lifestyle intervention program followed by 6 months of supplementation with Jarrow Formulas Carotene All® complex (providing daily 5000 IU of β- and α-carotene , 20 mg of lutein, 4 mg of zeaxanthin, 20 mg of lycopene, 1 mg of astaxanthin and 20 mg of vitamin E) or placebo; in the treatment group, reductions in both subcutaneous and visceral adiposity relative to baseline values were reported together with concomitant increases in serum adiponectin, while these parameters changed in the opposite, unwanted, direction in the placebo group (Canas et al. 2014).

The administration of β-cryptoxanthin extracted from Satsuma mandarin (0.5 mg/day as part of test drink) to moderately obese Japanese males resulted in increased levels of β-cryptoxanthin in serum and led to reductions in body weight, visceral fat and waist circumference (Tsuchida et al. 2008; Takayanagi and Mukai 2014). Another 3-week long study in 17 postmenopausal obese women supplemented with a beverage containing β-cryptoxanthin (4.7 mg/day) reported no differences in body weight or BMI but a fourfold increase in adiponectin serum levels after the treatment, suggestive of a reduction in adiposity (data on body fat mass were not include in this report) (Iwamoto et al. 2012).

Regarding fucoxanthin, in a small randomized double blinded placebo-controlled trial involving adult non-seaweed consuming subjects in Ecuador with at least one symptom of metabolic syndrome, researchers showed that 6 grams per day of dietary brown seaweed containing fucoxanthin consumed for 2 months resulted in decreased waist circumference in women and improved systolic blood pressure (Teas et al. 2009). Another 16 week study investigating the effects of the fucoxanthin-containing product XanthigenTM reported statistically significant reductions in body weight, waist circumference, body fat content and serum triglycerides in obese, non-diabetic female volunteers with non-alcoholic fatty liver disease and normal liver fat content compared to baseline (Abidov et al. 2010).

7 Programming Effects of Vitamin A on Adipose Tissue Expandability

Nutritional factors at critical stages in early life can condition the susceptibility to obesity and metabolic alterations later in life (reviewed in (Pico and Palou 2013)). In this sense, the net impact of vitamin A supplements on body adiposity may be developmental stage-dependent. We have shown that treatment with a moderate, threefold excess vitamin A (as retinyl ester) during the suckling period – which is a critical period in the development of adipose tissue in the rat (Cryer and Jones 1979) – favors the accumulation of small, immature adipocytes, with a reduced expression of PPARγ and an increased expression of proliferating cell nuclear antigen, a classical marker of proliferative status (Granados et al. 2013), which might be in line with known anti-adipogenic action of retinoids . Nevertheless, these changes appeared to favor the hyperplasic component of fat expansion upon an obesogenic stimulus, since vitamin A-treated animals gained more adiposity – but not body weight – than their controls on a high fat diet owing mainly to a higher increase in WAT cellularity (DNA content) (Granados et al. 2013). A previous report already evidenced a synergic effect of exposure to a fourfold excess vitamin A (retinol) and high fat diet on WAT expansion in young (3-week-old) rats that paralleled a higher proliferation competence of precursor cells isolated from the fat depots of excess vitamin A fed animals’ (Redonnet et al. 2008). These studies point, therefore, to a pro-obesogenic effect of excess vitamin A intake (even if moderate) in early life, likely by influencing the proliferative status of adipocytes.

Interestingly, different from preformed vitamin A intake, a threefold excess vitamin A intake as BC during the suckling period did not elicit changes in the developing WAT of rats at weaning, even though BC was readily absorbed and partially metabolized by the suckling rats, as indicated by BC accumulation in serum and liver and enhanced atRA-dependent transcriptional responses in intestine and liver (Musinovic et al. 2014). This latter work establishes a new potential model in studies of BC action, and suggests that BC supplementation may serve to replenish liver retinol stores in infants/children while avoiding eventual unwanted effects of early-life supplementation with pre-formed vitamin A on the susceptibility to obesity later in life (Musinovic et al. 2014).

8 Summary and Concluding RemarkS

Carotenoids and carotenoids conversion products seem to play a substantial role in the control of key aspects of adipose tissue biology including the production of novel adipocytes from precursor cells (adipogenesis) and the metabolic and secretory capacities of mature adipocytes, such as those for hyperthrophy, browning and endocrine and pro-inflammatory signal production. Mechanisms of action are emerging and notably include physical and/or functional interaction with transcription factors of the nuclear receptor superfamily and with pro-inflammatory and antioxidant signaling pathways in adipocytes and cells of the stromal-vascular fraction. Studies support an anti-adiposity and anti-inflammatory action of specific carotenoids and carotenoid conversion products in obesity. Interestingly, some of these compounds – such as fucoxanthin, astaxanthin and the BC-derived retinoids atRA and Rald– exert both suppressive effects on PPARγ activity and adipogenesis and activating effects on lipid oxidation and thermogenesis in mature brown and white adipocytes and other cell types. These compounds might, therefore, help moderating the formation of new adipocytes under obesogenic conditions and resetting adipocyte cell number in obese subjects while simultaneously favoring the dissipation of excess energy. Moreover, adipose tissue is an important site of carotenoid accumulation and results suggest that local production of retinoids from BC may modulate fat storage capacity in adipocytes. Human epidemiological studies reveal a low intake and status of carotenoids among obese subjects, including obese children/adolescents, which needs to be addressed. So far, there have been limited human intervention studies with carotenoid-rich supplements or extracts to reduce adiposity or obesity-related co-morbidities; even if these studies are insufficient to prove a cause-effect relationship, the scenario they depict is encouraging, since both in children and adults the interventions had beneficial effects on the accrual of body fat, abdominal fat and related risk parameters. Importantly, beneficial effects of BC supplements on adiposity in humans have been achieved under mild BC supplementation, at doses quite lower than the ones that caused concern and controversy in large intervention trials in the past (The Alpha-Tocopherol BCCPSG 1994; Omenn et al. 1996; Gallicchio et al. 2008). In summary, the literature reviewed herein supports a role of specific carotenoids and carotenoid derivatives in the prevention of excess adiposity , and suggests that carotenoids requirements may be dependent on body composition, among other factors.