Because de novo fatty acid synthesis in birds takes place mainly in the liver, most studies have been performed on hepatic tissues. The expression of some genes involved in lipid synthesis and secretion has been analyzed in the liver of lean and fat chickens. Bourneuf et al. [9] used a cDNA microarray to analyze the expression of genes in liver that are involved in pathways and mechanisms leading to adiposity, and found some genes differentially expressed between lean and fat chickens. Their research indicated that the mechanisms involved in the expression and regulation of lipogenic genes could play a key role in the ontogenesis of fatness in chickens from lean and fat lines. However, few studies have analyzed the functions of adipose tissue leading to adiposity in chickens.
Genome expression analysis aims to provide a broad and unbiased survey of the transcriptome, and requires true global coverage of a complex genome in a single microarray [10]. Significant progress has been made towards this goal and GeneChip microarrays have been improved significantly since they were invented [11].
The Chicken Genome Array, created by Affymetrix Inc. at the end of 2004 as part of the Ensembl annotation attempt at the complete chicken genome sequence (version 1, released May 2004), is a key research tool in chicken genomics. Our study provides, apparently for the first time, a comprehensive analysis of the chicken adipose tissue gene expression profile using this array. The gene expression profile results showed that 13,234–16,858 probe sets were detected in adipose tissue in 7-week-old chickens. Further study revealed that genes directly involved in lipid metabolism, including Spot14, LPL, adrenomedullin, pyruvate dehydrogenase kinase, isoenzyme 4 and the FABP family: FABP 3 (muscle and heart), FABP 4 (adipocyte) and FABP 5 (psoriasis-associated), were highly expressed in adipose tissue; FABP 4 is the most highly expressed gene in this tissue. These genes are mainly involved in fatty acid transport and degradation. Other genes that are closely related to lipid metabolism and obesity by signal transduction or regulation of transcription were also highly expressed: prosaposin (PSAP), retinol binding protein 7 (RBP7), retinoic acid receptor responder, IGFBP 7, thymosin beta 4, BMP7, superoxide dismutase 2 and annexin A5. In addition, many genes involved in the immune response and cytophylaxis, such as CD74 antigen, MHC class I glycoprotein, MHC class II beta 1 domain, MHC class II antigen alpha, lymphocyte antigen 6, beta-2 microglobulin, T-cell leukemia, Hsp70 and Hsp25, were also highly expressed in adipose tissue. The most highly expressed genes (A-FABP, LPL) were apparently not differentially expressed between fat and lean lines. This may not preclude a major involvement of these factors in regulating lipid metabolism in chick adipose tissue, and further work on protein expression is needed to understand the metabolism of this tissue better. Our results demonstrated primarily that the functions of adipose tissue in chicken are important in lipid metabolism by directly or indirectly regulating the synthesis, transport and degradation of lipids. Moreover, adipose tissue may play an important role in chicken immunity and cytophylaxis. These results will be help to clarify the adipose tissue gene expression profile and to choose appropriate methods for further study of these genes.
On the other hand, genes described as "not expressed", including obr,SREBP1, apoB,CCK,CCKR,IGF2 and lipoprotein (APOVLDLII), which have been studied extensively in other species (mainly Homo sapiens and Mus) and shown to affect obesity, were not detected in chicken adipose tissue in the present study. Possible reasons may be: (1) these genes are not expressed in chicken adipose tissue at all, or not at 7 weeks; (2) lipid metabolism in chicken is different from that in mammals. In birds, lipogenesis mainly occurs in the liver and is very limited in adipose tissue [12]. Although those genes were not detected in adipose tissue, they may regulate lipid metabolism indirectly. Another possible reason is gene interactions: the highly expressed genes may inhibit those with low expression levels and the "not expressed" ones when they participate the same biological responses.
Genetic variation in fatness was analyzed using two experimental lines (fat (FL) and lean (LL)) that were divergently selected for abdominal fat weight from a common genetic background. After selection for eight generations, the AFW and AFP differed significantly between the two lines (Figure 2). By consistency analysis within each line using hierarchical clustering of the total expressed genes, heterogeneity among individuals in each line was demonstrated (Figure 4). The hierarchical clustering results were barely satisfactory: the fat line showed higher consistency (about 87%) than the lean (about 77%). The consistency of each line is not sufficiently maintained. There may be two reasons for this. First, an insufficient number of generations have been used for selection, so the purity of each line will increase as selection continues. Second, our target trait is abdominal fat; after selection for eight generations the genes affecting abdominal fat may be better selected, but the genome array detects all genes expressed in adipose tissue, so genes affecting other traits, which have not been divergently selected, will reveal inter-individual differences within each chicken line. The higher consistency of the fat chicken line (about 87% compared to 77% for the lean line) shows that the fat line is purer than the lean line after being selected for eight generations. In addition, to make our study more reliable, the 7-3-16 individual was excluded from subsequent screening of differentially expressed genes because it deviated too much from the others.
Significance Analysis of Microarrays (SAM) has been considered a canonical algorithm for identifying differentially expressed genes in microarray data analysis. A total of 230 differentially expressed genes were screened by SAM. Fifteen genes that were differentially expressed between the two chicken lines were validated by real-time RT-PCR. To validate the results further, in addition to the samples used in the array experiments, four RNA pools (two from the fat line and two from the lean) were used for real-time RT-PCR analysis to reveal the differentially expressed genes more exactly. The results of this validation were encouraging: in all but one case (TNFAIP1), the real-time RT-PCR fold differences correlated well with the microarray data. Rajeevan et al. [13] evaluated the efficiency of real-time RT-PCR for validating the differentially expressed genes identified by microarrays. Their results indicated that genes showing less than a 2-fold difference in expression were not likely to be validated by real-time RT-PCR. Bourneuf et al. [9] considered that confirmation of genes with fold changes TNFAIP1 in the microarray was -1.6, which is difficult to confirm by real-time RT-PCR, so it is not correct to affirm that the two methods gave contradictory results for this gene.
Our analysis results revealed that the expression levels of some important genes implicated in lipid metabolism were up-regulated in fat chickens: propionyl-coenzyme A-carboxylase, acyl-CoA oxidase, pyruvate dehydrogenase complex, glycerol-3-phosphate dehydrogenase, phosphatidylinositol, lysophospholipase 3, low density lipoprotein-related protein 12, etc. Pyruvate dehydrogenase catalyzes oxidative decarboxylation of pyruvate to form acetyl-CoA, which is a central metabolite [14-16]. Acetyl-CoA enters the Krebs cycle, and is the donor of acetate for synthesis of fatty acids, ketone bodies and cholesterol. The expression level of glycerol-3-phosphate dehydrogenase increased in diet-induced obese animals [17]. Propionyl-coenzyme A-carboxylase is the key enzyme in the catabolism of odd-chain fatty acids, isoleucine, threonine, methionine and valine [18]. Phosphatidylinositol is an important lipid, both as a key membrane constituent and as a participant in essential signaling processes in all plants and animals. These genes regulate lipid metabolism mainly by enhancing the synthesis of fatty acids. In addition, the expression of genes involved in energy metabolism, such as thioredoxin-like 4B, 5-oxoprolyl-peptidase, etc., were markedly down-regulated in fat line chickens, and the expression of genes participating in gluconeogenesis or glycolysis, including glycerol kinase (GK), ATPase, beta 1,3-galactosyltransferase etc. were up-regulated in this line. In fact, it is well recognized that both energy metabolism [19] and glycometabolism [20,21] are highly related to obesity.
It is interesting that several genes in the protein tyrosine phosphatase pathway were more highly expressed in fat than in lean chickens: fms-related tyrosine kinase 1, receptor tyrosine kinase flk-1/VEGFR-2, a type III receptor tyrosine kinase, protein tyrosine phosphatase-like, member b, protein-tyrosine phosphatase MEG2, etc. This indicates that the protein tyrosine phosphatase pathway is very important in lipid metabolism, and further research is needed to elucidate the molecular mechanism.
Adipogenesis is promoted by the coordinated expression of many transcription factors [22,23]. De-differentiation, or loss of the adipocyte phenotype, has been observed in response to tumor necrosis factor and transforming growth factor β [24,25]. We also observed that several genes involved in tumorigenesis were differentially expressed between the two chicken lines. Expression of genes that inhibit tumorigenesis, such as tumor necrosis factor, alpha-induced protein 1, suppression of tumorigenicity 7 and thrombospondin 1, was down-regulated in fat line chickens, while the expression of genes promoting tumor formation, including ret proto-oncogene and turban tumor syndrome, was up-regulated. This indicates that tumorigenesis may be related to obesity. TNFAIP1 was first identified as a tumor necrosis factor a (TNFa) and interleukin-6(IL-6) inducible protein. It is induced rapidly and transiently by TNFa [26]. It is inferred that the in vivo effects of LPS on lipid metabolism are probably mediated by TNFa, which could be secreted by macrophages or by the adipose tissue itself [27,28]. Albalat et al. [29] suggested that TNFa could be a key modulator of lipid metabolism in fish. Taking these reports and our data into consideration, we presume that TNFa may play a significant role in chicken lipid metabolism.
Answers to all your Biology Questions
