Female-Predominant Expression of Fatty Acid Translocase/CD36 in Rat and Human Liver

NINA STÅHLBERG, ELIZABETH RICO-BAUTISTA, RACHEL M. FISHER, XUXIA WU, LOUISA CHEUNG, AMILCAR FLORES-MORALES, GUNNEL TYBRING, GUNNAR NORSTEDT, AND PETRA TOLLET-EGNELL Department of Molecular Medicine (N.S., E.R.-B., X.W., L.C., A.F.-M., G.N., P.T.-E.), and Atherosclerosis Research Unit, Downloaded from https://academic.oup.com/endo/article/145/4/1972/2878908 by guest on 01 October 2021 King Gustaf V Research Institute (R.M.F.), Karolinska Hospital, Karolinska Institute, SE-171 76 Stockholm, Sweden; and Department of Laboratory Medicine, Division of Clinical Pharmacology, Huddinge University Hospital, Karolinska Institute (G.T.), SE-141 86 Stockholm, Sweden

The aim of this study was to identify genes for hepatic fuel predominant transcripts was fatty acid translocase/CD36, metabolism with a gender-differentiated expression and to with 18-fold higher mRNA levels in the female liver and 4-fold determine which of these that might be regulated by the fe- higher mRNA levels in males treated with GH, compared with male-specific secretion of GH. Effects of gender and continu- untreated males. This gender-differentiated expression was ous infusion of GH to male rats were studied in the liver using confirmed at mRNA and protein levels in the rat and at the cDNA microarrays representing 3200 genes. Sixty-nine tran- mRNA level in human livers. Although purely speculative, it scripts displayed higher expression levels in females, and 177 is possible that higher levels of fatty acid translocase/CD36 in displayed higher expression in males. The portion of GH- human female liver might contribute to the sexually dimor- regulated genes was the same (30%) within the two groups of phic development of diseases resulting from or characterized gender-specific genes. The male liver had a higher expression by disturbances in lipid metabolism, such as arteriosclerosis, of genes involved in fuel metabolism, indicating that male rats hyperlipidemia, and insulin resistance. (Endocrinology 145: might have a greater capacity for high metabolic turnover, 1972–1979, 2004) compared with females. Most notable among the female-

T IS WELL known that the body composition differs be- pattern) infusion of GH (8). Interestingly, these different I tween the sexes, and there is also increasing evidence that types of GH treatment induced opposite changes in the same fuel metabolism differs between genders (1–3). It is therefore set of genes. This implies that previously reported gender- likely that tissues from male and female individuals have related effects of GH, such as body growth, might include a different metabolic functions and that genes from different sexually dimorphic metabolism of lipids. metabolic pathways are expressed in a sex-specific manner. Gender differences are also seen in the manifestations of Hepatic sex differences at the level of gene expression have different disorders, some of which may have links to sex- previously been demonstrated and shown to be regulated by specific liver functions. The male predominance of the met- different hormones (4), including sex steroids and peptide abolic syndrome is well established, and men have an earlier hormones. One hormone in the latter category is GH. The onset and a higher incidence of coronary heart disease (9). existence of a sexually dimorphic pattern of GH secretion The sexually dimorphic development of lipid-related disor- was first described in the rat (5), but has also been demon- ders motivates investigations that seek to clarify gender dif- strated in other species, including humans (6). In adult male ferences in the physiology and pathophysiology of the liver. rats, GH is secreted in episodic bursts at 3- to 4-h intervals, Generating sex-specific transcript profiles should help to elu- with low or undetectable levels between peaks, whereas cidate molecular mechanisms behind sex-specific functions females have a more continuous pattern of secretion. In ro- of the liver. The aim of the present investigation was to dents, biological effects of this sexually differentiated pattern identify genes involved in the metabolism of carbohydrates, of GH secretion includes sex differences in body weight gain and longitudinal bone growth (7), but also hepatic steroid lipids, and protein with a gender-differentiated expression metabolism (4). We have recently demonstrated that episodic and to determine which of these might be regulated by the (male pattern) GH treatment of old male rats repress the sex-specific secretion of GH. We have collected transcript expression of hepatic gene products involved in lipogenesis profiles from male and female rat livers using cDNA mi- and induce genes for fatty acid oxidation (7a). In a related croarrays representing 3200 genes. Gender differences in study, old male rats were treated by continuous (female gene expression have been compared with the effect of continuous infusion of GH to male rats. Most notable among the female-predominant gene products was fatty acid translo- Abbreviations: FAT/CD36, Fatty acid translocase/CD36; LCFA, long-chain fatty acid; RNase, ribonuclease. case/CD36 (FAT/CD36), which was shown to have 18-fold higher mRNA levels in the female liver. This gender-differ- Endocrinology is published monthly by The Endocrine Society (http:// www.endo-society.org), the foremost professional society serving the entiated expression was confirmed at mRNA and protein endocrine community. levels in the rat and at the mRNA level in humans.

1972 Ståhlberg et al. • Sex-Differentiated Expression in the Liver Endocrinology, April 2004, 145(4):1972–1979 1973

Materials and Methods to the method of Melton et al. (15). Samples were analyzed in triplicate, Animals and the results were determined as counts per minute of mRNA per microgram of total RNA. The results are expressed as the mean Ϯ sd, Ten-week-old male and female Sprague Dawley rats (B&K Universal and a t test was used for the determination of statistical significance. AB, Hull, UK) were maintained under standardized conditions. Five male rats were treated with bovine GH (gift from Pharmacia & Upjohn AB, Stockholm, Sweden) by continuous infusion from Immunoblotting osmotic minipumps (model 2001, B&K Universal AB) at a dose of 5 Tissue homogenates were resolved by SDS-PAGE on 7.5% polyacryl- ␮g/h. After 1 wk of treatment, the rats were killed, and tissues removed amide gels and transferred to polyvinylidene difluoride membranes and frozen in liquid nitrogen. The animal experiments were approved with a Trans-Blot SD semidry transfer cell (Bio-Rad Laboratories, Her- by the institutional animal care and use committee. cules, CA). The membranes were blocked for2hinTris-buffered saline Human liver samples (10 mm Tris, pH 8.0, and 150 mm NaCl) containing 0.05% (vol/vol) Tween 20 and 10% (wt/vol) milk powder, incubated for 2 h with mono- Human liver tissue was obtained from a donor liver bank established clonal anti-FAT/CD36 antibodies (antimurine CD36, Cascade Bio- at the Department of Clinical Pharmacology at Huddinge University sciences, Winchester, MA) diluted 1:2500 in Tris-buffered saline, Downloaded from https://academic.oup.com/endo/article/145/4/1972/2878908 by guest on 01 October 2021 Hospital. The liver specimens which for various reasons could not be washed, and incubated with the secondary antibody (1:10,000; donkey used for organ transplantation, included seven women and five men antimouse IgG, Jackson ImmunoResearch Laboratories, West Grove, between 22 and 50 yr of age. None of the subjects had been diagnosed PA). After additional washing steps antibody binding was visualized with metabolic disorders. The donors were checked for hepatitis B and using an ECL detection system (Pierce Chemical Co., Rockford, IL). C serology, and all were HIV negative. All livers were processed to be used for transplantation and were transfused with 5–10 liters Belzer UW Immunohistochemistry Cold Storage Solution (ViaSpan, DuPont Pharmaceuticals, Wilmington, DE). Before transfusion, benzyl penicillin (0.12 mg), actaprid (1.4 mg), Cryostat sections (5 ␮m) from rat livers were used in a standard and dexamethasone (16 mg) were added (all per liter of UW solution). immunohistochemical protocol (avidin-biotin-peroxidase). After rins- The livers were handled within the ischemic time of about 16 h, cut into ing the sections in PBS, nonspecific endogenous peroxidase activity was small pieces, snap-frozen in liquid nitrogen, and stored at Ϫ80 C. The blocked by 3% hydrogen peroxide in methanol for 10 min. After 10-min study was approved by the ethics committee of Huddinge University washing, sections were exposed to a 30-min nonimmunoblock using Hospital, Karolinska Institute. Consent was obtained from the donor or diluted normal horse serum (Vectastain, Vector Laboratories, Inc., Bur- their relatives to use the liver tissue for scientific purposes. lingame, CA) in PBS. The sections were incubated with monoclonal anti-FAT/CD36 antibodies (antimurine CD36, Cascade Biosciences), di- cDNA microarray analysis luted 1:500 in PBS containing 2% BSA, at 4 C overnight. Negative con- trols were obtained by replacing the primary antibody with normal The generation (10), use, and analysis (11) of microarrays represent- horse serum. As secondary antibody, a biotinylated horse antimouse ing 3200 cDNAs have been described previously. Total RNA was iso- antibody was used (Vector Laboratories, Inc.), and the slides were in- lated from livers using TRIzol reagent (Invitrogen Life Technologies, cubated with horseradish peroxidase-avidin-biotin complex (Vectastain Gaithersburg, MD). The aim of this study was to investigate not indi- ABC Elite, Vector Laboratories, Inc.). The site of bound enzyme was vidual variations, but, rather, the common changes representative of the visualized by 3,3-diaminobenzidine (DAB-kit, Vector Laboratories, Inc.). whole group. Therefore, equal amounts of total RNA from five animals The sections were counterstained with hematoxylin and dehydrated in the same experimental group were pooled before cDNA labeling. before mounting with Pertex (Histolab, Gothenburg, Sweden). Although pooling of samples might lead to the signals being confounded by mixed individuals, it also minimizes the biological noise. In the first set of experiments, each hybridization compared Cy3-labeled cDNA re- Determination of hepatic lipid content verse transcribed from RNA isolated from male rats with Cy5-labeled Cellular lipids were extracted from rat liver homogenates using chlo- cDNA isolated from females. In another set of experiments, Cy3-labeled roform and methanol (2:1, vol/vol). The extracts were dried and dis- cDNA derived from livers of nontreated males were compared with Cy5- solved in 200 ␮l isopropanol. The lipid contents were determined using labeled cDNA from male rats that had been treated with bovine GH for 1 kits for total cholesterol (ABX Diagnostics, Montpellier, France), free wk. Each experiment was analyzed in triplicate on the same cDNA pool. cholesterol (Wako Chemical GmbH, Neuss, Germany), and triglycerides The significance of the expression ratios of both sex and GH replace- (Roche, Indianapolis, IN). Samples were analyzed in duplicate, and the ment studies were estimated using the significance analysis of microar- results were determined as micromoles of lipid per gram of liver tissue. ray technique (12). A q value was assigned for each of the detectable The amount of cholesteryl esters was calculated as the difference be- genes in the array. This value is similar to the familiar P value, measuring tween total and free cholesterol. The results are expressed as the mean Ϯ the lowest false discovery rate at which the differential expression (the se.At test was used for the determination of statistical significance. ratio between control and experimental cDNA) of the gene is called significant. In this study, genes with a q value less than 5% were considered significantly differentially expressed. Genes with higher q values Quantification of human FAT/CD36 mRNA expression were excluded from further analysis. To this statistically based criterion a further requirement was added for differential expression based on the Total RNA was isolated from human livers using TRIzol reagent, as absolute changes in gene expression ratios. A value of 1.5 was chosen described above for rat tissues. The RNA concentration was carefully to denote differences (increased or decreased expression) in the level of determined spectrophotometrically, and the quality of the RNA samples was examined on a denaturing agarose gel. A total of 1.0 ␮g RNA was hybridization between control and experimental cDNA. This cut-off was Ϫ - empirically chosen based on previous validation studies using different transcribed into cDNA using 200 U SuperScript II RNase H reverse tran scriptase (Life Technologies, Inc.) and oligo(deoxythymidine) independent techniques. Previous results from our laboratory have shown 12–18 primer. Quantification of CD36 and ␤ that Cy5/Cy3 ratios around 1.5 correspond to higher ratios when calculated -actin mRNAs was performed by Taq- using results obtained from ribonuclease (RNase) protection analysis (8, 10, Man real-time semiquantitative PCR according to the manufacturer’s 11, 13, 14). In terms of fold regulation, DNA chip analysis tended to un- protocol, using an ABI PRISM 7000 Sequence Detection System instru- derestimate differences compared with RNase protection assay. ment and software (PE Applied Biosystems, Foster City, CA). CD36 was measured using a predeveloped TaqMan Assay-on-Demand (assay ID, Solution hybridization analysis Hs00169627_m1, PE Applied Biosystems). Primers and probe sequences for ␤-actin were designed using the Primer Express software (PE Ap- Total RNA was isolated from rat tissues as described above, and plied Biosystems): forward primer, 5Ј-CTGGCTGCTGACCGAGG-3Ј; mRNA levels were measured using a solution hybridization/RNase reverse primer, 5Ј-GAAGGTCTCAAACATGATCTGGGT-3Ј; and probe, protection assay. A specific 35S-labeled cRNA probe was transcribed in 5Ј-(6-FAM)CCTGAACCCCAAGGCCAACCG(TAMRA)-3Ј. vitro from the cDNA vector construct corresponding to nucleotides Standard curves were constructed using dilutions of purified PCR- 1886–2058 within the FAT/CD36 mRNA sequence (L19658), according amplified target sequences. Results are expressed in arbitrary units. 1974 Endocrinology, April 2004, 145(4):1972–1979 Ståhlberg et al. • Sex-Differentiated Expression in the Liver

Differences in loading of RNA were accounted for by expressing CD36 To identify gene products that might be involved in the expression relative to ␤-actin. Data analysis was performed using t test. sexually dimorphic development of metabolic disorders, the differentially expressed gene products were next grouped Results into functional categories to enable an overview of the effects Expression profiles were collected from rat liver to identify of gender and GH treatment on specific cellular pathways. genes with a gender-differentiated expression. Of 3200 genes The male liver was found to have a higher expression of printed on the arrays, approximately 1800 were detected in the several genes involved in glucose, lipid, and amino acid liver. Among these, 246 transcripts were expressed in a sexually metabolism compared with the situation in females (Table 2). differentiated manner; 69 transcripts (4%) displayed a statisti- Only four metabolic gene products displayed a higher ex- cally significant increase (at least 1.5-fold) in females, whereas pression in female livers. Most notable among these was 177 (10%) had a higher level in males. To identify sex-specific FAT/CD36, which was shown to have 18-fold higher mRNA genes dependent on the sexually differentiated pattern of GH levels in the female liver. This finding might be of special secretion, we compared mRNA expression in livers from un- interest, because FAT/CD36 has been implicated in both Downloaded from https://academic.oup.com/endo/article/145/4/1972/2878908 by guest on 01 October 2021 treated male rats with livers exposed to 1 wk of continuous GH reverse cholesterol transport and long-chain fatty acid infusion. GH infusion of male rats changed the expression lev- (LCFA) uptake in various tissues. As a consequence, plasma els in 16% of the total number of hepatic genes. By comparing cholesterol is elevated in FAT/CD36 gene knockout mice the list of gender-differentiated genes with the list of GH- (21). Furthermore, hepatic clearance of LCFA is substantially regulated genes, various GH-regulated genes with gender- faster in females than in males (22), a fact that has been differentiated expression could be identified. Twenty-one suggested to underlie known gender-related differences in female-predominant genes were induced by continuous infu- lipoprotein metabolism and associated disease states. sion of GH to male rats, indicating that these genes might be The finding that females had higher levels of hepatic FAT/ female-predominant due to their ability to become induced by CD36 indicates that females, in situations of increased avail- the continuous presence of GH. Similarly, 51 male-predomi- ability of circulating fatty acids, will be better equipped to nant genes were reduced by GH and might thus be under import these lipids. As previous reports have concluded that negative influence of female-specific GH secretion. FAT/CD36 is only expressed in the liver at low levels (23), Continuous treatment of male rats with GH through osmotic we wanted to confirm our observation by further studies. minipumps is well known to feminize the liver at the level of mRNA expression in liver was compared with that in adi- gene expression (4). This feminizing effect of GH could be pose tissue, where the presence of FAT/CD36 is well doc- observed in the present study through the induction of umented (23). When liver and fat RNA samples were pre- CYP2C12 and CYP2C7 (Table 1), which represent GH-regu- pared from individual rats and hybridized with a FAT/ lated female-specific (16) and female-predominant (17) gene CD36-specific cRNA probe in an RNase protection/solution products, respectively. Similarly, the male-specific CYP2C13 hybridization assay, FAT/CD36 mRNA levels were approx- (18), carbonic anhydrase III (19), and ␣2u-globulin (20) were imately 6 times higher in fat compared with liver, but only expressed at higher levels in male liver and were reduced by if the RNA was obtained from male animals (Fig. 1). If female continuous GH treatment. The expression of these genes is tissues were compared, the liver had higher levels of this dependent on the male-specific intermittent pattern of GH se- transcript. FAT/CD36 mRNA levels were approximately 8 cretion in the rat and is repressed by continuous infusion of GH. times higher in female livers compared with male livers, and Taken together, these results indicate that the animals had been GH was shown to induce this transcript in male liver. In able to respond sufficiently to the infusion of GH. contrast to the situation in liver, adipose FAT/CD36 displayed approximately 2 times lower mRNA levels in female TABLE 1. Effects of gender and continuous GH treatment on the expression of hepatic genes previously characterized as sex specific rats and was not affected by GH treatment, indicating that and GH regulated this gene is differently regulated in fat compared with liver. One likely explanation for the opposite effect of gender on Female GH mp male Accession no. Gene product the FAT/CD36 mRNA levels in liver and fat might be that vs. male vs. male tissue-specific signaling molecules are involved in regulating J03786 CYP2C12 7.18 6.24 the expression of this gene. Furthermore, as the gender effect AW914620 CYP2C7 4.58 2.88 J00737 ␣2u-Globulin 0.02 0.06 was greater than the GH effect (in both tissues), other hor- AF037072 Carbonic anhydrase III 0.03 0.02 mones, such as sex steroids, might regulate this gene. U09742 CYP3A2 0.07 0.02 To gain more information about the hormonal regulation J02861 CYP2C13 0.09 0.15 of FAT/CD36 expression in rat liver, we used our recently Comparison of changes in gene expression in female vs. male rats developed database for microarray data (EndoGED) to (female vs. male) and GH-treated vs. untreated male rats (GH mp male search for expression data associated with this gene. The vs. male); n ϭ five rats/group. The data are presented as ratios between the levels of test to reference cDNA that is hybridized to spotted DNA. database can be found at www.cmm.ki.se/endoged and con- The significance of the expression ratios was estimated using the sig- tains detailed information about the animal experiments nificance analysis of microarray technique. Genes with a q value higher behind the data collected therein. The data extracted from than 5% were not considered significantly differentially expressed and EndoGED related to FAT/CD36 are summarized in Table 3. were therefore excluded from further analysis. The magnitude of the Based on data from previously described experiments (24), changes was calculated as the mean of three or four independent assays on the same cDNA pool. Values highlighted in bold are those represent- estrogen treatment did not induce hepatic FAT/CD36 ex- ing mRNAs that were increased at least 1.5 times, and values in plain pression in castrated female rats. However, continuous treat- text are those that were decreased at least 1.5 times. ment of GH (female pattern) induced FAT/CD36 mRNA Ståhlberg et al. • Sex-Differentiated Expression in the Liver Endocrinology, April 2004, 145(4):1972–1979 1975

TABLE 2. Effects of gender and continuous GH treatment on the expression of genes involved in fuel metabolism in rat liver

GH mp Accession no. Gene product Female vs. male male vs. male Glucose oxidation AW914470 Glycerol-3-phosphate dehydrogenase 0.35 GP:U68535 Aldo-keto reductase 0.67 S63233 Phosphoglycerate mutase type B 0.59 AF034577 Pyruvate dehydrogenase kinase 4 1.89 NM_012624 Pyruvate kinase L-type 0.59 AW140918 Succinyl-CoA synthetase ␤ 1.59 BM986362 NADP-dependent isocitrate dehydrogenase 0.67 1.68 Glucose synthesis AW144708 Glucose-6-phosphatase catalytic subunit 0.39 3.17 K03248 Phosphoenolpyruvate carboxykinase 0.17 Downloaded from https://academic.oup.com/endo/article/145/4/1972/2878908 by guest on 01 October 2021 BF281428 UDP-glucose pyrophosphorylase 0.22 0.36 Pentose phosphate pathway BF281439 Glucose-6-phosphate dehydrogenase 1.70 6.77 Lipid mobilization NM_012598 Lipoprotein lipase 1.83 L19658.1 FAT/CD36 18.13 3.75 M13508 Apolipoprotein A-IV 0.43 NM_012824 Apolipoprotein C1 0.66 M14201 ACBP/DBI 1.69 Lipid synthesis K01934 Hepatic product spot 14 0.49 1.67 BF281429 Fatty-acid synthase 0.56 3.69 AW141104 Acyl-CoA synthetase 5 2.83 AW140731 ␣-Propionyl-CoA carboxylase 0.52 J02585 Stearyl-CoA desaturase-1 0.40 1.67 BF281436 Farnesyl pyrophosphate synthetase 2.11 2.47 Z49204 NADP transhydrogenase 0.57 Lipid oxidation BF281420 Carnitine palmitoyltransferase 1 0.60 0.31 AW141072 Carnitine octanoyltransferase 0.36 BF281486 Short chain L-3-hydroxyacyl-CoA dehydrog. 0.46 AF095449 3-Hydroxyacyl-CoA dehydrogenase 0.42 AF121345 Peroxisomal phytanoyl-CoA hydroxylase 0.26 0.18 X05341 3-Oxoacyl-CoA thiolase 0.62 AW140999 CYP4A3 0.65 Amino acid turnover BF281435 Ornithine aminotransferase 1.88 J04791 Ornithine decarboxylase 1.52 BF281456 Ornithine decarboxylase antizyme 0.48 NM_013177 Aspartate aminotransferase 0.60 M35266 Cystein dioxygenase 0.44 M29599 Glutamine synthetase 0.57 Comparison of changes in gene expression in female vs. male rats (female vs. male) and GH-treated vs. untreated male rats (GH mp male vs. male); n ϭ five rats/group. The data are presented as ratios between the levels of test to reference cDNA that is hybridized to spotted DNA. The significance of the expression ratios was estimated using the significance analysis of microarray technique. Genes with a q value higher than 5% were not considered significantly differentially expressed and were therefore excluded from further analysis. The magnitude of the changes was calculated as the mean of three or four independent assays on the same cDNA pool. Values highlighted in bold are those representing mRNAs that were increased at least 1.5 times and values in plain text are those that were decreased at least 1.5 times. CoA, Coenzyme A. levels in hypophysectomized (10) or 2-yr-old (8) male rats, final gender difference observed in this study. Work is now confirming the positive effect of GH on this gene. Further- in progress to validate this hypothesis. more, GH treatment of primary cultures of rat hepatocytes To examine whether the sex-differentiated mRNA expres- was also shown to increase the level of FAT/CD36 mRNA in sion in rat liver is translated into protein, FAT/CD36 expression the absence of any other hormone except insulin, indicating was studied by immunoblotting and immunohistochemistry. a direct effect of GH on the hepatocyte. Interestingly, epi- As shown in Fig. 2, immunoblotting with FAT/CD36 mono- sodic treatment (male pattern) with GH led to reduced ex- clonal antibodies revealed higher levels of a protein with the pression in 18-month-old male rats. Bearing in mind the apparent molecular mass of 85 kDa in female livers, indicating screening nature of DNA microarray analysis, these data a female-predominant expression of hepatic FAT/CD36 pro- indicate that GH is an important regulator of this gene, and tein. The reason why multiple bands appear on the blot is depending on the mode of treatment (continuous or epi- unknown, but might be related to different degrees of protein sodic), the gene might be either induced or repressed. It may glycosylation (25). The same antibody recognized a protein in therefore be speculated that the added effects of continuous hepatic cryostat sections (Fig. 3). Immunostaining of FAT/ GH secretion in females (which is stimulatory) and episodic CD36 was mainly confined to hepatocytes, although a weak GH secretion in males (which is inhibitory) will produce the presence could be observed in some Kupffer cells (arrowhead). 1976 Endocrinology, April 2004, 145(4):1972–1979 Ståhlberg et al. • Sex-Differentiated Expression in the Liver

FIG. 1. Effects of gender and GH treatment on FAT/CD36 mRNA expression in liver and adipose tissue from rats. Comparison of changes in FAT/CD36 mRNA expression in female vs. male rats and GH-treated (continuous infusion) vs. untreated male rats (n ϭ five rats/group) determined by RNase protection/solution hybridization analysis. Individ- ual RNA samples were analyzed in triplicate, and the results were expressed as the mean Ϯ SD. The effect of gender was significant in both tissues, but the GH effect was only significant Downloaded from https://academic.oup.com/endo/article/145/4/1972/2878908 by guest on 01 October 2021 in liver. **, P Ͻ 0.001; *, P Ͻ 0.005 (vs. untreated male rats).

TABLE 3. Effects of different hormonal treatments on hepatic FAT/CD36 mRNA expression in rats

Exp. Control (reference) Treatment (test) Treatment vs. control 1 Hypophysectomized males hGH mp 2.27 Ϯ 0.06 2 24-month-old males hGH mp 6.64 Ϯ 2.69 3 18-month-old males hGH injections 0.19 Ϯ 0.0 4 Primary cultures of hepatocytes bGH 24 h (100 ng/ml) 2.33 Ϯ 0.57 5 Sham-operated females Castration 0.88 Ϯ 0.13 6 Castrated females 17␤-estradiol (2.5 mg/d), 1 d 1.20 Ϯ 0.03 7 Castrated females 17␤-estradiol (2.5 mg/d), 4 d 0.86 Ϯ 0.08 8 Castrated females 17␤-estradiol (2.5 mg/d), 7 d 1.09 Ϯ 0.27 The table was generated from results obtained in different microarray experiments. Detailed information about the listed experiments, as well as the associated expression data, can be found at www.cmm.ki.se/endoged. The data are presented as ratios between the levels of test to reference cDNA that is hybridized to spotted DNA (FAT/CD36). The magnitude of the changes was calculated as the mean Ϯ SD of three or four independent assays. All rats were 3 months old except for the animals in experiments 2 and 3, which were older. The rats in experiments 1 and 2 were infused with GH at a dose of 0.34 ␮g/g body weight⅐d for 3 wk. The animals in experiment 3 were injected twice daily with 1 ␮g/g body weight⅐d for 1 wk.

der difference was detected for total cholesterol. The hepatic content of cholesteryl ester was significantly lower in the females, but was not affected by GH treatment. As the total amount of cholesterol was the same in the three groups of animals, the lower amount of cholesteryl ester in females might be due to a lower capacity to convert free cholesterol

FIG. 2. Effects of gender on FAT/CD36 protein expression in rat tis- into cholesteryl esters. Although further studies on lipid sues. FAT/CD36 protein levels were analyzed by immunoblotting in metabolism, including assays for fatty acid oxidation and male adipose tissue (A), male liver (B–E), or female liver (F–I) from secretion of lipoprotein particles, are required to better ad- eight different individuals. Different amounts of protein were loaded for the different types of tissue (100 ␮g from fat and 250 ␮g from liver). dress what type of metabolic changes would be expected from a high expression of FAT/CD36, the sex differences Arrows indicate strongly stained hepatocytes in female rats, might not appear unless the rats are exposed to increased whereas the arrows with broken lines indicate the weak staining levels of circulating fatty acids. This question might therefore in male rats. The intensity of FAT/CD36 staining in hepatocytes be better addressed if the experiments described in this re- from female livers (Fig. 3A) seems to be stronger than that in port were repeated on rats fed a high fat diet. male livers (Fig. 3B). Finally, FAT/CD36 mRNA expression was investigated in In attempts to relate gender differences of FAT/CD36 human liver samples using TaqMan real-time semiquanti- expression to a specific phenotype, the effect of gender and tative PCR. Human liver tissue was obtained from postmor- GH treatment was determined on liver triglyceride and cho- tem subjects, including seven women and five men between lesteryl ester contents. As shown in Fig. 4, the hepatic tri- 22–50 yr of age. When FAT/CD36 mRNA levels were quan- glyceride content was not different between the sexes, but tified and related to levels of ␤-actin, it could be concluded increased after GH treatment. This effect of GH has been that hepatic FAT/CD36 expression is also female-predomi- shown previously (26) and fits with the data presented in nant in humans (Table 4), although the differences are not as Table 2, showing increased expression of gene products in- marked as in rodents. The mean Ϯ sd FAT/CD36 expression volved in lipogenesis in GH-treated rats. Similarly, no gen- in male and female livers were 1.56 Ϯ 0.33 and 3.28 Ϯ 1.31, Ståhlberg et al. • Sex-Differentiated Expression in the Liver Endocrinology, April 2004, 145(4):1972–1979 1977

In the above-described experiments we have used cDNA microarrays as a screening instrument to obtain an overview of differentially expressed genes. The finding that FAT/CD36 is a gene with a sex-specific liver expression caught our attention, especially as this seems to also be the case in humans. To pave the way for further studies, we found it relevant to conduct an in silico analysis of the FAT/CD36 promoter. It is reasonable to assume that regulatory regions in gene promoters are conserved between species if they regulate the gene in a similar manner. Rodent compared with human FAT/CD36 promoters share regions of conserved sequences. These regions contain more than 50 different conserved putative DNA-binding elements (see the supplemental data published on The Endocrine Downloaded from https://academic.oup.com/endo/article/145/4/1972/2878908 by guest on 01 October 2021 Society’s Journals Online web site at http://endo.endojournals. org) that motivate further investigations. Some of the conserved DNA-binding elements that could be of interest are pointed out in Fig. 5, where the organization of the mouse gene promoter is shown. Hepatocyte nuclear factor-3␤ has previously been shown to be involved in gender-differentiated and GH pattern- dependent gene regulation (27, 28). Also, c-Fos (29) and nuclear factor-␬B (30) have shown gender-related differences in expression under certain circumstances. Further investigations are needed to explore whether any of these transcription factors is involved in the gender-specific regulation of FAT/CD36 expression.

Discussion This study describes gender-dimorphic expression profiles from rat liver with particular emphasis on fuel metabolism, which may set the basis for gender-related differences in this metabolically important tissue. Our results indicate that male rats have a higher expression of genes encoding important proteins for glucose oxidation, glycogen production, lipid synthesis, fatty acid oxidation, and amino acid turnover. This observation might explain why male rats have a higher metabolic rate than females and is in line with the fact that males have a bigger muscle mass and higher growth rate. We have recently observed that genes involved in hepatic fatty acid oxidation and glycogen production are induced by intermittent GH treatment in old male rats. As this type of treatment is believed to mimic the male-specific episodic secretion of GH, the results suggested that this set of genes might be male-predominant. Indeed, data obtained in the present study show that the same genes had a higher hepatic expression in male rats. As the higher growth rate in males is also dependent on episodic GH secretion, this might be a way to ensure that increased body growth is always coupled to increased utilization of fatty acids for energy ex- penditure. A high rate of hepatic ␤-oxidation in male rats may increase the ability of the liver to produce glucose by glucone- FIG. 3. Effects of gender on FAT/CD36 expression in rat liver. Cry- ogenesis. In addition, a higher level of UDP pyrophosphorylase ostat sections show anti-FAT/CD36 staining in normal female (A) and might lead to a relatively higher deposition of hepatic glycogen male liver (B). C, Negative control. Representative sections are in males compared with females. Both UDP pyrophosphory- shown. Scale bars,50␮m. lase and carnitine palmitoyltransferase 1 were male-predominant, repressed by continuous GH in the present study, and respectively. The smaller gender effect in human samples could shown to be induced by intermittent GH treatment in a pre- indicate species-specific mechanisms behind the expression of vious study (7a). Therefore, these genes seem to be under the this gene, or it might be due to the fact that gender-unrelated same type of GH regulation as has previously been described factors affected the FAT/CD36 expression in the human livers. for other male-predominant gene products, such as CYP2C13 1978 Endocrinology, April 2004, 145(4):1972–1979 Ståhlberg et al. • Sex-Differentiated Expression in the Liver

FIG. 4. Effects of gender and GH treatment on hepatic lipid content. Total cholesterol, free cholesterol, cholesterol ester, and triglycerides were determined in hepatic lipid

extracts from male, female, and GH-treated Downloaded from https://academic.oup.com/endo/article/145/4/1972/2878908 by guest on 01 October 2021 male rats. Lipid extraction and analyses were performed as described in Materials and Methods. Data are given as the mean Ϯ SE. **, P Ͻ 0.01; *, P Ͻ 0.05 (vs. untreated male rats).

TABLE 4. Effect of gender on human hepatic FAT/CD36 mRNA expression

Gender Age CD36 ␤-Actin CD36/␤-actin F 44 12.30 (29.21) 4.30 (23.28) 2.86 F 44 49.80 (27.27) 10.22 (22.06) 4.87 F 40 5.62 (30.35) 1.45 (24.82) 3.88 F 35 6.20 (30.21) 2.24 (24.20) 2.77 F 28 2.81 (31.33) 0.84 (25.59) 3.34 F 34 16.80 (28.81) 19.28 (21.16) 0.87 F 44 5.04 (30.51) 1.16 (25.12) 4.34 FIG. 5. Organization of the mouse FAT/CD36 promoter. The figure M 39 7.95 (29.86) 5.26 (23.02) 1.38 shows the 5Ј-untranslated region of the mouse FAT/CD36 gene with M 22 6.31 (30.19) 4.51 (23.21) 1.40 the alternative promoters preceding exons 1a and 1b. Translation M 28 15.20 (28.95) 13.50 (21.66) 1.13 starts in exon 3. Putative binding sites for hepatocyte nuclear fac- M 32 3.18 (31.16) 1.65 (24.63) 1.93 tor-3␤ (HNF-3␤), c-Fos, and nuclear factor-␬B (NF-kappaB) are in- M 50 4.64 (30.62) 2.51 (24.04) 1.85 dicated. Other binding sites that were conserved between species, and Quantification of FAT/CD36 or ␤-actin mRNAs was performed by their exact location, can be found in the supplemental data. TaqMan real-time semiquantitative PCR. Results are given in arbitrary units. The CT values for the expression of FAT/CD36 or ␤-actin are given in parenthesis. Student’s t test showed that FAT/CD36 was not believed to be a major site for FAT/CD36 expression. expression in females (F) was significantly different from males (M). However, during the preparation of this manuscript, Zhang P Ͻ 0.05. et al. (32) described the production of a monoclonal Ab (UA009) that recognizes an antigen characterized as FAT/ (18), CYP3A2 (31), carbonic anhydrase III (19), and ␣2u-globulin CD36. Immunostaining using UA009 revealed high FAT/ (20). CD36 expression in female hepatocytes. Moreover, a clear Among the gender-differentiated genes classified as being gender difference was reported, with higher expression in important in intermediary metabolism, FAT/CD36 was the liver of female rats compared with males, confirming the found to display the biggest difference in mRNA levels be- results obtained in the present study. tween the sexes, with a very high level in female livers. This If FAT/CD36 is involved in the uptake of LCFAs by the gene was further shown to have at least 2 times higher liver, the female-predominant hepatic expression of this gene mRNA expression in human female livers compared with might explain the long-recognized difference in LCFA up- male livers. FAT/CD36 is a cell surface glycoprotein that take and utilization in livers of male and female rats (33). functions as a receptor/transporter for LCFAs. FAT/CD36 Moreover, as fatty acids also act as signals involved in reg- has been shown to be expressed by various cell types, in- ulating gene expression, a greater capacity to import these cluding platelets, monocytes/macrophages, and endothelial molecules might lead to a sexually differentiated pattern of cells, and tissues with active LCFA metabolism, such as gene expression during situations of increased availability of adipose, small intestine, and heart. Until recently, the liver fatty acids. Several genes encoding proteins involved in the Ståhlberg et al. • Sex-Differentiated Expression in the Liver Endocrinology, April 2004, 145(4):1972–1979 1979 metabolism of glucose and lipids are known to be regulated 10. Flores-Morales A, Ståhlberg N, Tollet-Egnell P, Lundeberg J, Malek R, Quackenbush J, Lee N, Norstedt G 2001 cDNA microarray analysis of the in by fatty acids. Down-regulated genes include pyruvate ki- vivo effects of hypophysectomy and growth hormone treatment on gene ex- nase, glucose-6-phosphatase, fatty acid synthase, spot 14, pression in the rat. Endocrinology 142:3163–3176 and stearyl CoA desaturase 1. Interestingly, all of these gene 11. Pang S, Dillner K, Wu X, Pousette A, Norstedt G, Flores-Morales A 2002 Gene expression profiling of androgen deficiency predicts a pathway of prostate products were expressed at a lower level in the livers from apoptosis that involves genes related to oxidative stress. Endocrinology 143: female rats in this study. Whether this observation is related 4897–4906 to the higher level of FAT/CD36 in female livers must await 12. Tusher V, Tibshirani R, Chu G 2001 Significance analysis of microarrays applied to the ionizing radiation response. Proc Natl Acad Sci USA 98:5116–5121 further investigations. It is tempting, however, to speculate 13. Flores-Morales A, Gullberg H, Fernandez L, Stahlberg N, Lee N, Vennstrom that the lower level of FAT/CD36 observed in male rats and B, Norstedt G 2002 Patterns of liver gene expression governed by TR␤. Mol Endocrinol 16:1257–1268 male humans might lead to a reduced capacity to repress 14. Iyer VR, Eisen MB, Ross DT, Schuler G, Moore T, Lee JC, Trent JM, Staudt hepatic lipogenesis when circulating levels of LCFAs in- LM, Hudson Jr J, Boguski MS, Lashkari D, Shalon D, Botstein D, Brown PO 1999 The transcriptional program in the response of human fibroblasts to crease. This observation is in line with the idea that sex serum. Science 283:83–87 differences in arteriosclerosis might be determined not only 15. Melton DA, Krieg PA, Rebagliati MR, Maniatis T, Zinn K, Green MR 1984 Downloaded from https://academic.oup.com/endo/article/145/4/1972/2878908 by guest on 01 October 2021 Efficient in vitro synthesis of biologically active RNA and RNA hybridization by sex-related differences in the extracellular hormonal en- probes from plasmids containing a bacteriophage SP6 promotor. Nucleic Acids vironment, but also by sex-specific cellular characteristics Res 12:7035–7056 mediating responses to a variety of stimuli. Previous studies 16. Mode A, Wiersma-Larsson E, Gustafsson J-Å 1989 Transcriptional and post- transcriptional regulation of sexually differentiated rat liver cytochrome P-450 have shown that there is a pathophysiological connection by growth hormone. Mol Endocrinol 3:1142–1147 between FAT/CD36 and cardiovascular disorders (34). Our 17. Bandiera S, Ryan D, Levin W, Thomas P 1986 Age- and sex-related expression of cytochromes p450f and P450g in rat liver. Arch Biochem Biophys 248:658–676 finding that FAT/CD36 expression is sex different leads us 18. Ryan D, Iida S, Wood A, Thomas P, Lieber C, Levin W 1984 Characterization to speculate that this gene might be involved in the sexually of three highly purified cytochromes P-450 from hepatic microsomes of adult male rats. J Biol Chem 259:1239–1250 dimorphic development of diseases resulting from or char- 19. Jeffery S, Wilson C, Mode A, Gustafsson J-Å, Carter N 1986 Effects of acterized by disturbances in lipid metabolism, such as arte- hypophysectomy and growth hormone infusion on rat hepatic carbonic an- hydrases. J Endocrinol 110:123–126 riosclerosis, hyperlipidemia, and insulin resistance. 20. Chatterjee B, Demyan W, Gustafsson J, Harris M, Ho¨kfelt T, Norstedt G, Roy A 1986 Effect of anterior hypothalamic deafferentation and continuous growth ␣ Acknowledgments hormone infusion on the hepatic synthesis of 2u-globulin in the male rat. J Endocrinol 108:351–355 21. Febbraio M, Abumrad N, Hajjar D, Sharma K, Cheng W, Pearce S, Silver- Received July 14, 2003. Accepted December 9, 2003. stein R 1999 A null mutation in murine CD36 reveals an important role in fatty acid and lipoprotein metabolism. J Biol Chem 274:19055–19062 Address all correspondence and requests for reprints to: Dr. Nina 22. Luxon BA, Holly DC, Milliano MT, Weisiger RA 1998 Sex differences in Ståhlberg, Karolinska Institute, Department of Molecular Medicine, multiple steps in hepatic transport of palmitate support a balanced uptake CMM L8:01, Karolinska Hospital, 171 76 Stockholm, Sweden. E-mail: mechanism. Am J Physiol 274:G52–G61 [email protected]. 23. Abumrad N, el-Maghrabi M, Amri E, Lopez E, Grimaldi P 1993 Cloning of This work was supported by grants from the Swedish Medical Research a rat adipocyte membrane protein implicated in binding or transport of long- Council, 13X-08556, 32X-14053-01, 14GX-13571-01A, 3902, the Swedish So- chain fatty acids that is induced during preadipocyte differentiation. Homol- ogy with human CD36. J Biol Chem 268:17665–17668 ciety of Medical Research, the Fredrik and Ingrid Thuring Foundation, the 24. Wu X, Pang S, Sahlin L, Blanck A, Norstedt G, Flores-Morales A 2003 Gene Tore Nilsson Foundation for Medical Research, the Magnus Bergvall Foun- expression profiling of the effects of castration and estrogen treatment in the dation, the Swedish Medical Association, the Swedish Heart Lung Foun- rat uterus. Biol Reprod 2003 69:1308–1317 dations, and the Loo and Hans Osterman Foundation. 25. Greenwalt D, Lipsky R, Ockenhouse C, Ikeda H, Tandon N, Jamieson G 1992 N.S. and E.R.-B. contributed equally to this study. Membrane glycoprotein CD36: a review of its roles in adherence, signal trans- duction, and transfusion medicine. Blood 80:1105–1115 26. Elam M, Simkevich C, Solomon S, Wilcox H, Heimberg M 1988 Stimulation References of in vitro triglyceride synthesis in the rat hepatocyte by growth hormone treatment in vivo. Endocrinology 122:1397–1402 1. Blaak E 2001 Gender differences in fat metabolism. Curr Opin Clin Nutr Metab 27. Legraverend C, Eguchi H, Stro¨m A, Lahuna O, Mode A, Tollet P, Westin S, Care 4:499–502 Gustafsson J-Å 1994 Transactivation of the rat CYP2C13 gene promoter in- 2. Tarnopolsky M, Ruby B 2001 Sex differences in carbohydrate metabolism. volves HNF-1, HNF-3, and members of the orphan receptor subfamily. Bio- Curr Opin Clin Nutr Metab Care 4:521–526 chemistry 33:9889–9897 3. Tipton K 2001 Gender differences in protein metabolism. Curr Opin Clin Nutr 28. Delesque-Touchard N, Park S, Waxman D 2000 Synergistic action of hepa- Metab Care 4:493–498 tocyte nuclear factors 3 and 6 on CYP2C12 gene expression and suppression 4. Zaphiropoulos PG, Mode A, Norstedt G, Gustafsson J-Å 1989 Regulation of by growth hormone-activated STAT5b. Proposed model for female specific sexual differentiation in drug and steroid metabolism. Trends Pharmacol Sci expression of CYP2C12 in adult rat liver. J Biol Chem 275:34173–34182 10:149–153 29. Porsh-Ha¨llstro¨m I, Gustafsson J-Å, Blanck A 1989 Effects of growth hormone 5. Isaksson OG, Ede´n S, Jansson J-O 1985 Mode of action of pituitary growth on the expression of c-myc and c-fos during early stages of sex-differentiated hormone on target cells. Annu Rev Physiol 47:483–499 rat liver carcinogenesis in the resistant hepatocyte model. Carcinogenesis 6. Rose S, Municchi G, Barnes K, Kamp G, Uriarte M, Ross J, Cassorla F, Cutler 10:2339–2343 GJ 1991 Spontaneous growth hormone secretion increases during puberty in 30. Kono H, Wheeler M, Rusyn I, Lin M, Seabra V, Rivera C, Bradford B, Forman D, Thurman R 2000 Gender differences in early alcohol-induced liver injury: normal girls and boys. J Clin Endocrinol Metab 73:428–435 ␬ ␣ 7. Clark RG, Jansson JO, Isaksson O, Robinson IC 1985 Intravenous growth role of CD14, NF- B, and TNF- . Am J Physiol 278:G652–G661 hormone: growth responses to patterned infusions in hypophysectomized rats. 31. Pampori N, Shapiro B 1996 Feminization of hepatic cytochrome P450s by nominal levels of growth hormone in the feminine plasma profile. Mol Phar- J Endocrinol 104:53–61 macol 50:1148–1156 7a.Tollet-Egnell P, Parini P, Stahlberg N, Lonnstedt I, Lee NH, Rudling M, 32. Zhang X, Fitzsimmons R, Cleland L, Ey P, Zannettino A, Farmer E, Sincock Flores-Morales A, Norstedt G 2004 Growth hormone-mediated alteration of P, Mayrhofer G 2003 CD36/fatty acid translocase in rats: distribution, isolation fuel metabolism in the aged rat as determined from transcript profiles. Physiol from hepatocytes, and comparison with the scavenger receptor SR-B1. Lab Genom 16:261–267 Invest 83:317–332 8. Tollet-Egnell P, Flores-Morales A, Ståhlberg N, Malek R, Lee N, Norstedt G 33. Ockner R, Burnett D, Lysenko N, Manning J 1979 Sex differences in long chain 2001 Gene expression profile of the aging process in rat liver: normalizing fatty acid utilization and fatty acid binding protein concentration in rat liver. effects of growth hormone replacement. Mol Endocrinol 15:308–318 J Clin Invest 64:172–181 9. Wingard D, Suarez L, Barrett-Connor E 1983 The sex differential in mortality 34. Tanaka T, Sohmiya K, Kawamura K 1997 Is CD36 deficiency an etiology of from all causes and ischaemic heart disease. Am J Epidemiol 117:165–172 hereditary hypertrophic cardiomyopathy? J Mol Cell Cardiol 29:121–127

Endocrinology is published monthly by The Endocrine Society (http://www.endo-society.org), the foremost professional society serving the endocrine community.