International Journal of Obesity (1999) 23, 650±655 ß 1999 Stockton Press All rights reserved 0307±0565/99 $12.00 http://www.stockton-press.co.uk/ijo Stimulation of 1 expression in brown adipocytes by naturally occurring carotenoids

F Serra1, ML Bonet1, P Puigserver1, J Oliver1 and A Palou1*

1Laboratori de Biologia Molecular, Nutricio i Biotecnologia, Dept. Biologia Fonamental i CieÂncies de la Salut, Universitat de les Illes Balears, Valldemossa, Palma de Mallorca, Spain

OBJECTIVE: To assess the effect of naturally occurring carotenoids on brown adipocyte proliferation and differentia- tion. The rationale behind is that certain carotenoids have provitamin A activity in mammals, and that one of the active forms of vitamin A, (retinoic acid) is known to behave as a transcriptional activator of the key gene for brown fat , the one encoding the uncoupling protein thermogenin (UCP1). DESIGN: Con¯uent primary cultures of mice brown adipocytes were treated with various concentrations of carotenoids. Cell morphology, total culture protein content, the DNA synthesis rate, and the levels of UCP1, retinoic acid receptor a (RARa) and retinoid X receptor a (RXRa) were analysed. RESULTS: Treatment with b-carotene, a-carotene and lutein promoted UCP1 expression in a dose-dependent manner, with an effectiveness that was related to their potency as vitamin A precursors. Cell morphology, total culture protein content at con¯uence and DNA synthesis rate were unaffected after carotenoid treatment up to 10 mM. CONCLUSION: The results indicate that carotenoids can positively affect the expression of UCP1 without altering brown adipocyte proliferation.

Keywords: carotenoids; brown adipocytes; UCP1; b-cerotene

Introduction activator of the ucp1 gene of likely physiological signi®cance, since it induces UCP1 expression not only in cultured brown adipocytes7±9 but also in the The uncoupling protein thermogenin (UCP1) is an BAT of intact mice9 and, in addition, we have found inner-membrane mitochondrial protein that functions indications that the retinoid pathway of UCP1 induc- as a proton transporter, allowing the dissipation of the tion could be autoregulated10 and that the vitamin A proton gradient generated by the respiratory chain status of rodents can affect their BAT thermogenic upon nutrient oxidation, thereby uncoupling oxidative capacity.11 The effects of RA on brown adipocytes phosphorylation and producing heat.1;2 The UCP1 is appear to be mediated by heterodimers of retinoic acid characteristically expressed in mature brown adipo- receptors (RAR) and retinoid X receptors cytes, and its activity constitutes the molecular back- (RXR),8,10,12 as in other cell types,13 and the nucleo- ground for facultative thermogenesis in brown adipose tide sequence conferring RA responsiveness to the tissue (BAT). Related uncoupling proteins (UCP2, ucp1 gene has been identi®ed and characterized as a UCP3), with a distinct pattern of tissue expression, novel complex RA response element.12,14 have been recently identi®ed both in rodents and RA is the vitamin A form active in gene regulation. humans,3,4 reinforcing the potential importance of In mammals, the only route to vitamin A is via facultative thermogenesis in overall energy balance. carotenoid cleavage. b-carotene, in particular, is an The ucp1 gene is mainly regulated at the transcrip- important precursor of RA and its metabolism is the tional level, and it is physiologically induced by best known among carotenoids. Central cleavage of noradrenaline, acting through b3-adrenoceptors and the b-carotene molecule yields two molecules of subsequent cAMP production.5,6 The vitamin A deri- retinal which may either be reduced to retinol or vative retinoic acid (RA) is another transcriptional further oxidized to RA; as an alternative pathway, noncentral cleavage yields one molecule of b-apo- carotenal which can also be converted to RA, with or *Correspondence: Dr A Palou, Laboratori de Biologia Molecular, 15,16 Nutricio i Biotecnologia, Department de Biologia Fonamental i without the involvement of intermediate retinal. CieÁ ncies de la Salut, Universitat de les Illes Balears, Cra The main site of carotenoid metabolism is the intest- Valldemossa Km 7.5, Palma de Mallorca 07071, Spain. inal mucosa, but peripheral tissues such as the lung, E-mail: [email protected] Received 8 June 1998; revised 12 January 1999; accepted 1 kidney, liver and fat of several mammals, including February 1999 humans and rodents, can also convert b-carotene to Effects of carotenoids on UCP1 expression F Serra et al 651 RA.15,17 Liver and are the main sites of surface, therefore leading to very low protein recov- carotenoid deposition.15 eries per well. For harvesting, culture medium was With this background, we decided to address the removed and the cells were rinsed twice with phos- question as to whether carotenoids naturally occurring phate-buffered saline (PBS; 137 mM NaCl, 2.7 mM in food could display thermogenic properties, initially KCl and 10 mM phosphate buffer, pH 7.4) and then by investigating their effect on brown adipocyte cell scraped into 1.5 ml of PBS with a rubber policeman differentiation and proliferation in primary culture. and transferred to an Eppendorf tube. The cell suspen- sion were pelleted, resuspended in a minimal volume of PBS containing 0.2 mM phenylmethylsulfonyl ¯uoride and stored at 720C prior to sonication and Materials and methods analysis.

Materials Cell proliferation assay b-carotene, a-carotene, lutein, (Æ )-arterenol bitartrate Cell proliferation was measured using the 5-bromo-20- salt (noradrenaline) and all-trans retinoic acid (tRA) deoxy-uridine (BrdU) labeling and detection kit from were obtained from Sigma (Madrid, Spain). Cap- Boehringer Mannheim Biochemica, based on the cell santhin, capsorubin and canthaxanthin were puri®ed ELISA principle, following the manufacturer's from powder paprika by thin layer chromatography, instructions. Primary brown preadipocytes were cul- following the procedures in references 18 and 19. tured as described above but in a 96-well microtiter Newborn calf serum (NCS) and Dulbecco's modi®ed plate. The culture medium was changed on days 1, 3 Eagle's medium (DMEM) were from ICN (Barcelona, and 6 after inoculation. b-carotene or tRA at various Spain). Other cell culture reagents were supplied by concentrations was added on days 3 and 6. BrdU Sigma (Madrid, Spain), and routine chemicals used (10 mM) was added on day 6, and the cells were were from Merck (Barcelona, Spain) and Panreac further incubated for 24 h. Following this, the cells (Barcelona, Spain). Stock solutions of carotenoids were rinsed with PBS containing 10% fetal calf were prepared in ethanol daily; concentration was serum, ®xed in situ with ethanol/HCl and incubated calculated from the actual absorbance of each solution with 0.5 mM nucleases to partially digest the DNA so at an appropriate absorption maximum (450 nm for b- as to improve the accessibility of the antibodies to carotene, 444 nm for a-carotene, 445 nm for lutein) BrdU. Incorporated BrdU was detected with a mono- and the speci®c absorbance coef®cient (A%, the clonal anti-BrdU-POD antibody, Fab fragments, and absorbance of a solution of 1 g of solute in 100 ml the bound conjugate was visualized with the soluble of solution) at that wavelength of the corresponding chromogenic substrate 2,20-amino-bis-(3-ethyl- carotenoid dissolved in ethanol, according to refer- benzthiazoline-6-sulphonic acid) and measured using ence 19. an ELISA reader.

Primary cell culture Parameters determined BAT precursor cells were isolated from combined Protein was determined by the method of Bradford.22 cervical, axillar and interscapular brown fat depots UCP1, RARa and RXRa levels were analysed by of 4-week-old male NMRI mice (from Letica, Barce- immunoblotting. Fractionation of whole cell lysate lona, Spain), as earlier described.20,21 Precursor cells proteins by SDS-PAGE was according to Laemmli,23 were suspended in culture medium (cells from 1 in 10% polyacrylamide gels. Independent gels were mouse in 0.5 ml), pooled and inoculated in 6-well run for UCP1 detection (30 mg protein per lane) and multidish plates (0.2 ml of pooled ®nal cell suspension for RARa and RXRa detection (100 mg protein per per well containing 1.8 ml of culture medium). The lane); RARa was detected following stripping of the culture medium consisted of DMEM supplemented RXRa-stained blot. Electrotransfer and blocking of with 10% NCS, 4 nM insulin, 10 mM HEPES, 50 units the blots was as described.24 For UCP1 detection, of penicillin/ml, 50 mg of streptomycin/ml and 25 mg rabbit polyclonal antiserum raised against puri®ed rat of sodium ascorbate/ml. Cells were grown at 37Cin UCP1 obtained in our laboratory was used as primary an atmosphere of 8% CO2 in air. The culture medium antibody; its speci®city for BAT UCP1 (thermogenin) was changed on the ®rst and the third day after was con®rmed by immunoblotting of BAT homoge- inoculation. Cells were treated with carotenoids, nor- nate proteins, non-BAT homogenate proteins and adrenaline or tRA at the times and doses indicated in puri®ed thermogenin.24 For RARa and RXRa detec- the ®gure legends, after changing the culture medium; tion, rabbit polyclonal antisera raised against cDNA- control cells were processed in parallel and received deduced synthetic peptides contained in mouse the same volume of solvent. Carotenoid concentra- RARa1 (F region) and RXRa (A region), provided tions above 10 mM were not systematically included in by the IGBMC (Illkirch, France), were used; these the study because of their limited solubility in water; antibodies were previously shown to be speci®c for in addition, in preliminary trials, 100 mM b-carotene at least two species of RARa (of  58- and  54- was found to cause the cells to detach from the plastic kDa) and for a single species of RXRa (  54 kDa), Effects of carotenoids on UCP1 expression F Serra et al 652 respectively, in different cell systems and tissues, including primary brown fat cell cultures.10,25,26 The ®nal colour development was either with an alkaline phosphatase-conjugated substrate kit (Bio Rad, Madrid, Spain) or with an enhanced chemiluminis- cence kit (Amersham, Buckinghamshire, UK). For quantitative analysis, membranes or ®lms (Kodak X- Omat) were scanned with a BioImage computing densitometer (Millipore, Bedford, USA).

Results

Carotenoid treatment up to 10 mM ®nal concentration had no effect on primary brown adipocyte cell proliferation Treatment of brown adipocytes in primary culture with a-carotene, b-carotene or lutein up to 10 mM ®nal concentration for a period of 48 h before harvest- ing (from day 6 to day 8 of culture) had no apparent effect on cell morphology and did not affect the total culture protein content nor the DNA synthesis rate measured as BrdU incorporation into cellular DNA (not shown). This is in contrast with previous results of our group demonstrating an inhibitory effect of 10 mM tRA, when used under identical conditions, on brown adipocyte cell proliferation as indicated by a reduction of both the culture protein content and the DNA synthesis rate,9 which have been reproduced here (not shown).

Treatment with b-carotene, a-carotene and lutein induced the expression of UCP1 in con¯uent primary brown adipocytes The representative immunoblots in Figure 1 show the speci®c UCP1 levels of non-stimulated control cul- tures and of cultures stimulated with either 10 mM noradrenaline, 10 mM tRA or different concentrations of b-carotene, a-carotene and lutein. In these experi- ments, a single dose of the agents was added at con¯uence (day 6 after plating) and the cells were harvested 48 h later. The three carotenoids stimulated Figure 1 Representative immunobolts showing induction of UCP1 expression over the non-stimulated control UCP1 expression after treatment of primary cultures of brown level, and their effect was consistent and dose-depen- adipocytes with carotenoids. Cells were treated with 1075 M dent (Figure 2). b-carotene was the most effective of retionic acid (RA), 1075 M noradernaline (NA) or the indicated ®nal molar concentrations of b-carotene, a-carotene or lutein. A the carotenoids tested in inducing UCP1 expression; single dose of the agents was added on day 6 of culture, and the at 10 mM concentration, its effect was similar to that cells were harvested 48 h later. 30 mg of cellular lysate protein triggered by 10 mM noradrenaline and somewhat were run per lane in the SDS-PAGE. C, untreated control cells. lower than that triggered by 10 mM tRA. Other car- otenoids tested (capsanthin, capsorubin and canthax- anthin) had no signi®cant effect on the UCP1 levels we analyzed the steady-state RARa and RXRa protein when assayed under the same conditions. levels in brown adipocytes treated with different concentrations of b-carotene and tRA (from 1079 to 1075M, added at a single dose on day 5, 48 h before Comparison of the effects of b-carotene and retinoic harvesting), in comparison with those in control acid on the RARa and RXRa receptor levels non-treated cultures (Figure 3) tRA elicited a To go further into carotenoid regulation of UCP1 marked down-regulation of the steady-state levels expression and its relationship with retinoid regulation, of the RXRa and a slight down-regulation of the Effects of carotenoids on UCP1 expression F Serra et al 653

Figure 2 Dose-response curves showing the effect of carte- noids and tRA on the total UCP1 content of primary brown fat cell cultures. Cells were treated with the indicated agents as described in the legend of Figure 1. The UCPI content of cell lysates was quanti®ed by densitometric scanning of the immu- noblots. Data are expressed relative to the UCP1 content of the corresponding untreated control cells, set as the unit. Values represent the meansÆ S. E. M. of at least three independent experiments. A 6-fold induction of the UCP1 levels was attained after stimulation with 1075 M noradrenaline (6.01Æ 0.8).

54 kDa-RARa isoform levels, in accordance with previous results.10 b-carotene, on the other hand, triggered a decrease of the 54 kDa-RARa isoform levels only at the highest concentration tested (10 mM), and had no effect on the RXRa levels. The levels of the higher RARa isoform (58 kDa) were not affected by any of the two agents (not shown). Figure 3 Effect of b-carotene and tRA on the levels of RXRa and RARa (54 kDa isoform) of cultured primary brown adipocytes. A single dose of b-carotene or tRA (from 1075 to 1079 M) was added to the cultures on day 5; the cells were harvested on day 7 and the RXRa and RARa contents of cell lysates were quanti®ed by densitometric scanning of the corresponding immunoblots Discussion (100 mg cell lysate protein per lane). RXRa and RARa data are expressed relative to the corresponding levels in untreated control cultures. The results are the meansÆ S. E. M. of two In this work we show that certain carotenoids stimu- independent experiments. late UCP1 expression in primary cultures of brown adipocytes in a dose-dependent manner. The effect of limiting step in the b-carotene effect. The possibility carotenoids could be due to their conversion to vita- exists, however, that carotenoids have effects on min A and RA, which is known to behave as a brown adipocytes independent from RA. This is transcriptional activator of the ucp1 gene7±9. Some suggested by the fact that neither BrdU incorporation of our results and data from the literature are in into cellular DNA nor the total protein content of the agreement with this idea. Therefore, the effectiveness cultures were affected by 10 mM b-carotene but were of carotenoids as UCP1 induces (b-carotene > a-car- reduced by 10 mM RA, and especially by the lack of otene > lutein) qualitatively matched their potency as effect of b-carotene on the RXRa levels as compared vitamin A precursors,27,28 and it is known that several with RA. While the former result could simply re¯ect extra-intestinal tissues of mammals, including fat, are the fact that the actual intracellular concentration of able to convert b-carotene into RA.15,17 Indeed, it has the active species is higher in the cells treated with been suggested that the signi®cance of this local tissue RA than in those treated with b-carotene, the later conversion may be high considering the unstable result, when considered together with other results nature of RA, which cannot be stored and displays a presented here, strongly argues against the idea that very rapid metabolic clearance from body ¯uids.15 In all the effects of b-carotene are mediated by its addition, both b-carotene and RA triggered an upre- conversion to RA. If this were the case, since b- gulation of the UCP1 levels and a downregulation of carotene is 10-fold less active than RA in affecting the RARa levels, and b-carotene was about 10-fold UCP1 and RARa expression, it would also be 10-fold less active than RA in both effects; those are the kind less active in affecting RXRa expression, and a down- of results to be expected if conversion to RA is a rate regulation of the RXRa levels after treatment with Effects of carotenoids on UCP1 expression F Serra et al 654 10 mM b-carotene would have been found. We were 6 Palou A, Pico C, Bonet ML, Oliver P. Molecules in focus: the unable to detect any effect of 10 mM b-carotene on the uncoupling protein thermogenin. Int J Biochem Cell Biol 1998; 30: 7±11. RXRa levels. Therefore, RA-independent effects of 7 Cassard-Doulcier AM, Larose M, Matamala JC, Champigny carotenoids either as intact molecules or as metabolic O, Bouillaud F, Ricquier D. In vitro interactions between products different from RA on brown adipocytes nuclear proteins and uncoupling protein gene promoter should be considered. reveal several putative transactivating factors including Our results raise interest in the possibility that ets1, retinoid X receptor, thyroid hormone receptor, and a CACCC box binding protein. J Biol Chem 1994; 269: dietary carotenoids may affect BAT thermogenesis 24335±24342. and body weight of intact animals, an issue that 8 Alvarez R, De AndreÂs J, Yubero P, VinÄas O, Mampel T, deserves further investigation. Interestingly, the con- Iglesias R, Giralt M, Villarroya F. A novel regulatory pathway version of b-carotene to vitamin A seems to be self- of brown fat thermogenesis. Retinoic acid is a transcriptional limiting, so that high intakes of carotene usually do activator of the mitochondrial uncoupling protein gene. J Biol not lead to the toxic effects of high intakes of vitamin Chem 1995; 270: 5666±5673. 29 9 Puigserver P, VaÂzquez F, Bonet ML, Pico C, Palou A. In vitro A. It should be noted, however, that the extent of and in vivo induction of brown adipocyte uncoupling protein conversion of carotenoids to retinoids in vivo is highly (thermogenin) by retinoic acid. Biochem J 1996; 317: 827±833. species-dependent and that little is known about its 10 Bonet ML, Puigserver P, Serra F, Ribot J, VaÂzquez F, Pico C, regulation.15,16,29 Palou A. Retinoic acid modulates retinoid X receptor alpha and retinoic acid receptor alpha levels of cultured brown adipocytes. FEBS Lett 1997; 406: 196±200. 11 Bonet ML, Puigserver P, Pico C, Palou A. In vivo effects of retinoic acid on the thermogenic capacity of brown fat. Int J Obes 1997; 21: (Suppl 2) S19. Conclusion 12 Rabelo R, Reyes C, Schifman A, Silva E. A complex retinoic acid response element in the uncoupling protein gene de®nes a novel role for retinoids in thermogenesis. Endocrinology 1996; To our knowledge, this is the ®rst evidence of a 137: 3488±3496. 13 Chambon P. A decade of molecular biology of retinoic acid direct effect of carotenoids in promoting UCP1 receptors. FASEB J 1996; 10: 940±954. expression, and therefore in increasing the thermo- 14 Larose M, Cassard-Doulcier AM, Fleury C, Serra F, Cham- genic capacity of brown adipocytes. The present pigny O, Bouillaud F, Ricquier D. Essential cis-acting ele- results strengthen the idea that retinoid regulation ments in rat uncoupling protein gene are in an enhancer of the ucp1 gene may be of physiological relevance, containing a complex retinoic acid response domain. J Biol Chem 1996; 271: 31533±31542. by showing that brown adipocytes respond to natu- 15 Wang XD. Review: absorption and metabolism of beta-car- rally occurring carotenoids. otene. J Am Coll Nutr 1994; 134: 314±325. 16 Parker RS. Absorption, metabolism, and transport of carote- noids. FASEB J 1996; 10: 542±551. 17 Wang XD, Tang GW, Fox JG, Krinsky NI, Russell RM. Acknowledgements Enzymatic conversion of beta-carotene into beta-apo-carote- We thank Dr. P. Chambon, from IGBMC-LGME- nals and retinoids by human, monkey, ferret, and rat tissues. Arch Biochem Biophys 1991; 285: 8±16. U.184-ULP (Illkirch, France), for kindly providing 18 Gregory GK, Chent S, Philip T. Quantitative analysis of the anti-RARa and anti-RXRa antibodies. This work carotenoids and carotenoids esters in fruit by HPLC: red bell was supported by DGICYT of the Spanish Govern- peppers. J Food Sci 1987; 52: 1071±1073. ment (grants PB-92-0748-C04-01 and PB-94-1178) 19 Britton G. Carotenoids. In: Charlwood BV, Banthorpe DV and by the European Commission (Contract No. (eds). Methods in Plant Biochemistry. Academic Press: London, 1991, vol. 7, 473±518. ERBCHRX#CT940490). 20 Rehnmark S, Kopecky J, Jacobsson A, NeÂchad M, Herron D, Nelson BD, ObregoÂn MJ, Nedergaard J, Cannon B. Brown adipocytes differentiated in vitro con express the gene for the uncoupling protein thermogenin: effects of References hipothyroidism and . Exp Cell Res 1989; 1 Nicholls DG, Locke RM. Thermogenic mechanism in brown 182: 75±83. fat. Physiol Rev 1984; 64: 1±64. 21 Puigserver P, Herron D, Gianotti M, Palou A, Cannon B, 2 Cannon B, Nedergaard J. The biochemistry of an inef®cient Nedergaard J. Induction and degradation of the uncoupling tissue: . Essays Biochem 1985; 20: 110± protein thermogenin in brown adipocytes in vitro and in vivo. 164. Evidence for a rapidly degradable pool. Biochem J 1992; 284: 3 Fleury C, Neverova M, Collins S, Rainbault S, Champigny O, 393±398. Levi-Meyrueis C, Bouillard F, Seldin MF, Surwit RS, Ricquier 22 Bradford MM. A rapid and sensitive method for the quantita- D, Warden CH. Uncoupling protein-2: a novel gene linked tion of microgram quantities of protein utilizing the principle to obesity and hyperinsulinemia. Nature Genetics 1997; 15: of protein-dye binding. Anal Biochem 1976; 72: 248±254. 269±272. 23 Laemmli UK. Cleavage of structural proteins during the 4 Boss O, Samec S, Paolini-Giacobino A, Rossier C, Dulloo A, assembly of the head of bacteriophage T4. Nature 1970; Seydoux J, Muzzin P, Giacobino JP. Uncoupling protein-3: a 227: 680±685. new member of the family with tissue- 24 Puigserver P, Llado I, Palou A, Gianotti M. Evidence for speci®c expression. FEBS Lett 1997; 408: 39±42. masking of brown adipose tissue mitochondrial GDP-binding 5 Cannon B, Jacobsson A, Rehnmark S, Nedergaard J. Signal sites in response to fasting in rats made obese by dietary transduction in brown adipose tissue recruitment: noradrena- manipulation. Effects of reversion to standard diet. Biochem J line and beyond. Int J Obes 1996; 20 (Suppl): S36±S42. 1991; 279: 575±579. Effects of carotenoids on UCP1 expression F Serra et al 655 25 Gaub MP, Rochette-Egly C, Lutz Y, Ali S, Matthes H, 27 Furr HC. Carotenoids. Physiology. In: Macrae R, Robinson RK, Scheuer I, Chambon P. Immunodetection of multiple species Sadler MJ (eds) Encyclopaedia of Food Science, Food Technol- of retinoic acid receptor alpha: evidence for phosphorylation. ogy and Nutrition. Academic Press: London, 1993, 713±718. Exp Cell Res 1992; 201: 335±346. 28 Ganguly J. Biogenesis. In: Ganguly J (ed) Biochemistry of 26 Rochette-Egly C, Lutz Y, P®ster V, Heyberger S, vitamin A. CRC Press: Florida, 1989, 1±18. Scheuer I, Chambon P, Gaub MP. Detection of retinoid 29 Johnson EJ, Suter PM, Sahyoun N, Ribaya-Mercado JD, X receptors using speci®c monoclonal and polyclonal Russell RM. Relation between beta-carotene intake and antibodies. Biochem Biophys Res Comm 1994; 204: 525± plasma and adipose tissue concentrations of carotenoids and 536. retinoids. Am J Clin Nutr 1995; 62: 598±603.