HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Citation Map Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Valet, P. Articles by Langin, D. Search for Related Content PubMed PubMed Citation Articles by Valet, P. Articles by Langin, D. Social Bookmarking                      What's this?

Journal of Lipid Research, Vol. 43, 835-860, June 2002
Copyright © 2002 by Lipid Research, Inc.

Review

Understanding adipose tissue development from transgenic animal models

Philippe Valet¹, Geneviève Tavernier, Isabelle Castan-Laurell, Jean Sébastien Saulnier-Blache and Dominique Langin¹

INSERM Unit 317, Louis Bugnard Institute, Paul Sabatier University, Bldg. L3, Rangueil University Hospital, 31403 Toulouse Cedex 4, France

¹ To whom correspondence should be addressed. e-mail: valet{at}toulouse.inserm.fr, langin{at}toulouse.inserm.fr

	ABSTRACT

TOP ABSTRACT INTRODUCTION TRANSGENIC TECHNIQUES APPLIED TO... TRANSCRIPTION FACTORS AND... LIPID AND GLUCOSE METABOLISM ENDOCRINE RESPONSES AND SIGNAL... ADIPOCYTE SECRETIONS GENETIC ABLATION OF ADIPOSE... CONCLUSIONS REFERENCES

 
The World Health Organization has recognized obesity as a health problem of pandemic proportions. Recent work led to major breakthroughs in the understanding of the molecular basis of adipose tissue development with the cloning and characterization of numerous genes involved in fat cell differentiation and metabolism. Transgenesis has proved very useful in establishing the physiological roles of these genes. Here we review transgenic models made to study adipose tissue's metabolic and trophic responses. Genetic modifications unexpectedly associated to alterations of adipose tissue development are also examined because of their potential involvement in obesity and energy balance regulation. After a description of the methodologies commonly used, we review the data obtained on transcription factors, metabolism, signal transduction, secreted products, and models of lipodystrophy. An overview of such integrative studies leads to a better understanding of the physiology of adipose tissue development.

Alterations in expression levels of proteins involved at different steps of a regulatory pathway highlight the complementary roles of genes in the regulation of adipose tissue development. However, lack of phenotypes also illustrates the capacity of animals to set up adaptive mechanisms.

Abbreviations: ACC, acetylCoA carboxylase; ADD1, adipocyte determination and differentiation factor 1; AGT, angiotensinogen; aP2/ALBP, adipocyte lipid binding protein; AR, adrenergic receptor; ASP, acylation-stimulating protein; AT, adipose tissue; BAT, brown adipose tissue; 11ß HSD-1, 11ß hydroxysteroid deshydrogenase type 1; C/EBP, CCAAT-enhancer binding protein; CMV, cytomegalovirus; DGAT, diacylglycerotransferase; Cre, cyclization recombination; DT-A, diphtheria toxin A; EDG, endothelial differentiation gene; ES, embryonic stem; FAS, fatty acid synthase; FAT, fatty acid transporter; FOXC2, Forkhead box C2; GH, growth hormone; GLUT 4, glucose transporter 4; HMG, high mobility group protein; HSL, hormone sensitive lipase; IGF, insulin growth factor; IR, insulin receptor; IRES, internal ribosome entry site; IRS, insulin receptor substrate; iNOS, inducible NO synthase; PEPCK, phosphoenolpyruvate carboxykinase; PPAR, peroxisome-proliferator activated receptors; PTP1B, protein tyrosine phosphatase 1B; RXR, retinoid X receptor; SREBP, sterol responsive element binding protein; TNF, tumor necrosis factor ; TZD, thiazolidinediones; UCP, uncoupling protein; WAT, white adipose tissue

Supplementary key words fat cell • obesity • gene invalidation • additive or random insertional transgenesis • targeted transgenesis

	INTRODUCTION

TOP ABSTRACT INTRODUCTION TRANSGENIC TECHNIQUES APPLIED TO... TRANSCRIPTION FACTORS AND... LIPID AND GLUCOSE METABOLISM ENDOCRINE RESPONSES AND SIGNAL... ADIPOCYTE SECRETIONS GENETIC ABLATION OF ADIPOSE... CONCLUSIONS REFERENCES

	TRANSGENIC TECHNIQUES APPLIED TO ADIPOSE TISSUE STUDIES

TOP ABSTRACT INTRODUCTION TRANSGENIC TECHNIQUES APPLIED TO... TRANSCRIPTION FACTORS AND... LIPID AND GLUCOSE METABOLISM ENDOCRINE RESPONSES AND SIGNAL... ADIPOCYTE SECRETIONS GENETIC ABLATION OF ADIPOSE... CONCLUSIONS REFERENCES

Animal models
The mouse remains the species of choice to produce transgenic animals and to analyze the consequences of the modifications introduced in its genome. The laboratory mouse is the premier animal model for the study of human diseases. The major and sometimes obvious reasons are briefly that i) its small size means that it is easy to manipulate and, it lives in colony so keeping mouse lines requires less room than growing other transgenic mammals; ii) mating can occur at only 6 weeks and the generation time is around 12 weeks; iii) numerous (6–12) animals per litter are obtained; and iv) prices, depending on the strain used, are relatively low. The mouse model is made even more attractive because of the extensive and varied genetic tools available. These tools include a high-resolution genetic map, near complete sequencing of the genome, and technologies that allow direct and specific manipulation of its genome. From 1980 to 1985, the mouse represented the only species in which developmental physiologists and molecular biologists were able to introduce foreign DNA by microinjection into the fertilized egg or zygote. Furthermore, isolation of pluripotent embryo-derived stem cells, which are essential for homologous recombination, is only possible from the mouse.

Before starting a transgenesis protocol, it remains very important to spend time on the choice of the appropriate mouse strain. It must possess characteristics that allow easy preparation and injection of embryo, and the capacity to develop the expected phenotype. The inbred strains C57BL/6, BALB/cBy, 129/Sv, and the F1 hybrids C57BL/6 x CBA/Ca or C57BL/6 x DBA/2 have proved very useful. However, mice with various genetic backgrounds could behave differently and present various metabolic responses when submitted to the same pharmacological or physiological stimuli. As early as in the 1970s, Hummel and collaborators (5) studied the influence of the genetic background on the db/db mutation that is now known to be associated to a lack of the long-form of the leptin receptor. The db/db C57BL/6 mice develop obesity whereas the db/db C57BL/Ks mice are obese and diabetic. Chronic intraperitoneal infusion of leptin significantly reduced body fat in lean db/+ C57BL/6 but not in C57BL/Ks mice (6). Comparison of the effects of varying dietary macronutrient content on the body composition in AKR or SWR mice showed that the AKR strain had a greater percentage of carcass fat and was more responsive to the effects of dietary fat composition compared with the SWR strain (7). Differences in cold-sensitivity between uncoupling protein (UCP) 1-deficient congenic and F1 hybrid mice have also been reported (see below) (8).

The rat is the other rodent used as transgenic animal, e.g., as cardiovascular model of hypertension (9). It also represents a tool in the pharmacological studies related to diabetes (10), to lipoprotein metabolism (11) or to obesity (see below). However, practical and technical disadvantages of the rat model include difficulties in seeing the vaginal plug of the female and the low percentage of transgenic founders varying from 5% to 10% according to the strain (vs. 25% in the mouse). Moreover, although embryonic stem (ES) cells were isolated in vitro, cultured, and used to produce chimeras by injection into rat blastocysts, these chimeras have never demonstrated germ line transmission (12).

Genetic mutations of genes responsible for human disease do not always result in appropriate pathological models in rodents. Moreover, a number of invasive techniques, e.g., to study cardiovascular function, are available, but are limited to single experiments because the mouse does not survive the analysis. For these reasons, some investigators have chosen as transgenic models larger mammals like the rabbit. This animal, more relevant in cardiovascular research than the rodents as a model for human hypertrophic cardiomyopathy (13) and in lipoprotein metabolism and/or atherosclerosis (14), also presents some drawbacks however. Sexual maturity occurs between 20 and 24 weeks (6 to 8 in the mouse), the cost of generating and maintaining a rabbit colony is high, only 50% of the zygotes recovered are fertilized (vs. 80% to 90% in mouse), mosaicism is frequently seen in the founder, and the success rate for generating transgenic animals is very low (1%). Nevertherless, the rabbit represents the smallest domestic animal that can be used to produce recombinant protein with therapeutic and commercial interest in its blood or milk (15).

Additive or random insertional transgenesis
The main application of additive transgenesis has been to express genes specifically in AT. This became possible because of the in vivo characterization of promoter regions conferring AT-specific expression (Table 1). The most widely used promoters are the adipocyte lipid binding protein (aP2/ALBP) and phosphoenolpyruvate carboxykinase (PEPCK) promoters (16, 17) that target both white adipose tissue (WAT) and brown adipose tissue (BAT), and the UCP1 promoter that targets only BAT (18, 19). Brinster and Palmiter established the characteristics of a transgene, which are essential to achieve proper expression (20). Linearized constructs integrate more efficiently than circular DNA (21), plasmid sequences are undesirable (22), and introns increase the expression of the transgene and the frequency of transgenic mice (23). A poly(A) tail is also an essential structure in the design of the transgene, because the activation of mRNA degradation is tightly coupled to deadenylation (24). Transcription terminators localized in the 3' untranslated region, and particularly the stem-loop structures also belong to the elements involved in the enhancement of the transcription. Upstream of the open reading frame, a minimum length of the 5' untranslated region (20 to 100 nucleotides) is required to obtain efficient translation. Generally, the mRNAs translated at a high rate possess a short 5' untranslated region with a Kozak consensus sequence around the AUG (25). During the initiation of translation, the 40S ribosomal subunit binds to the cap structure (m⁷GpppN) and scans the 5' untranslated region in order to reach a functional initiation codon. Some mRNAs without a cap are, however, efficiently translated. In that case, the ribosomes bind to an internal ribosome entry site (IRES). The IRES can therefore be used to favor translation of a second cistron in a bicistronic mRNA. Several gene products could then be expressed under the control of the same promoter (26). It is often of great interest to target the nuclear compartment (e.g., for expression of transcription factors); this is done by adding to the transgene a short sequence named nuclear localization signals (27). Other elements controlling gene expression are located either upstream of the transcription initiation site and/or downstream of the poly(A) addition site. They include matrix and scaffold attachment regions, AT-rich sequences joining the nuclear matrix in the vicinity of genes, and locus control regions (28). Matrix and scaffold attachment regions generally enhance gene expression in cooperation with resident regulatory elements, thereby blocking chromosomal position effects. Intact locus control regions are required for complete insulation, i.e., to establish position-independent, copy number-dependent expression of transgenes with homologous or heterologous promoters. To avoid the engineering of complicated DNA constructs and to ensure a pattern of expression identical to that of the endogenous mouse gene, a large genomic fragment containing the most important regulatory sequences can be microinjected (29).

Besides targeted expression, other applications can be achieved using additive transgenesis. Tissue-specific ablation is possible through linkage of the sequence coding for the poisonous element of a toxin [such as diphtheria toxin subunit A (DT-A)] to WAT or BAT promoters (30, 31). The DT-A protein inactivates elongation factor-2 resulting in inhibition of protein synthesis and cell death. In this approach, cell killing is irreversible provided that threshold levels of intracellular toxin are attained. A conditional or inducible system exists that is based on the expression of the thymidine kinase gene of the herpes simplex virus (32). In the presence of nucleoside analogs, only the cells expressing thymidine kinase are killed. Suppression of gene expression in vivo can be achieved through targeted ablation (see below) but also through the expression of antisense RNA (33, 34). This method for gene disruption is technically easier to undertake than homologous recombination. Furthermore, it may represent a means to overcome in utero lethality by using promoter regions only active at birth. However, the decrease in the targeted protein expression is variable. Targeted oncogenesis has been used to induce hibernomas using the large tumor antigen of the simian virus 40 and derive immortalized brown fat cell lines (35, 36).

The DNA solution is generally injected into the pronucleus of a one-cell fertilized embryo which is reimplanted in the uterus of a pseudopregnant foster mother at this stage of development or after the first cellular division. This is the method of choice in the mouse. Pronuclear injection of DNA often results in multiple copies of the transgene arranged as head to tail concatamers inserted randomly into the host genome generally at a single integration site. Sometimes many integration sites co-exist that could represent an advantage in generating two or more lines from the same founder. Thus, screening of the F1 generation must rely on Southern blot analysis and not only on PCR amplification of the transgene. Upon breeding, transgenic DNA is inherited as a simple Mendelian trait. If this is not observed, it is assumed that integration of the transgene occurred after the first round of replication in the single fertilized egg cell and that the resulting transgenic mouse is mosaic for the transgene.

Targeted transgenesis
Targeted transgenesis using homologous recombination is the technique of choice to invalidate (knockout) or to introduce a subtle and precise mutation into a gene (Table 2). Homologous recombination relies on the naturally occurring but rare event of recombination between the targeting vector and its endogenous counterpart. The prerequisite is to know the exon-intron structure and the sequence of at least part of the targeted gene. The gene targeting construct is made up with genomic parts of sufficient lengths that surround a positive selectable gene encoding an enzyme conferring antibiotic resistance (neomycin, puromycin, or hygromycin). This vector is then electroporated into ES cells. Evans and Kaufman have shown that it is possible to establish culture cells derived from the inner cell mass of the early embryo at the blastocyst stage (37). Mouse ES cells remain diploid even after being cultured for several weeks and conserve the ability to proliferate vigorously in an undifferentiated state.

In order to distinguish random insertion, also named illegitimate recombination, from homologous recombination, the thymidine kinase cassette is generally attached to the 3' end of the targeting vector (38). Hence, the number of ES clones to analyze is considerably diminished. Cells from a positive clone are introduced into recipient blastocysts and then these embryos are reintroduced into a pseudopregnant female. To determine the positive animals, the easiest technique is coat color selection. Indeed, ES cells usually come from 129/Sv mice with agouti color and the blastocysts are from C57BL/6 black mice. The more the coat of the offspring is agouti, the more the ES cells have been involved in the development of all cell types of the embryo. The chimeric animals are mated with C57BL/6 mice in order to obtain heterozygous animals for the mutation. It is usually necessary to get the mutation on the two alleles to obtain a phenotype, thus hemizygous offspring must be intercrossed in order to obtain homozygous mice.

Systems for conditional expression of transgenes
The importance of being able to control the expression and disruption of a transgene became clear when it was impossible to establish transgenic lines because the transgene product, or lack of it, was deleterious for the animal's development or impaired the fertility of the adult. Moreover, an alteration of gene expression at early embryonic stages is more likely to activate compensatory pathways that mask phenotypes. New approaches have been developed that place the genetic manipulations under stringent temporal and spatial control. Classical systems for inducible mammalian gene expression have typically encountered limitations such as low level of expression, basal leakiness, and toxic or nonspecific effects of inducing agents (39). Inducible systems that combine functional domains from prokaryotic, eukaryotic, and viral proteins to create chimeric transactivators capable of modulating gene expression in a drug-dependent manner were then developed. The other technology was based on the use of an enzymatic system where the enzyme is able to recombine DNA fragments at small specific sites.

The first system developed by Gossen and Bujard is an allosteric off switch where the transcription of the gene of interest is suppressed by the addition of tetracycline or its structural analog doxycycline (40, 41). In this system, the repressor of Escherichia coli Tn10 tetracycline resistance operon (tetR) is fused to the activation domain of the herpes simplex virus VP16 to produce the transactivator tTA, which binds to tet operator (tetO) DNA sequences upstream of target genes and activates transcription. In the allosteric off switch, the permanent addition of tetracycline or its analog to the drinking water to avoid transcriptional activity may lead to deleterious effects. Therefore, an alternative system, the "allosteric on switch" system, has been developed. Using chemical mutagenesis and genetic selection, a mutant TetR was identified that only binds tetO in the presence of the antibiotic. Fusing this to VP16 produces a new transactivator named rtTA (42, 43). In mice, different lines with transgenic transactivator, usually under the control of the cytomegalovirus (CMV) promoter and responder with TetO minimal promoter driving the gene of interest, have to be generated and crossed to obtain the double-transgenic offspring. Both transgenes segregate independently, making further breeding experiments tedious. Moreover, the copy number of the individual components in the genome of double-transgenic offspring may vary in different combinations and could be unbalanced. Steroid hormones and their receptors constitute the molecular basis of alternative allosteric on switch systems (44).

Using the recombinase activity of the cyclization recombination (Cre) gene from the P1 bacteriophage, conditional transgenesis and knockout became available. Cre recombinase catalyzes the recombination of two 34 bp-long consensus sites, the loxP (locus of X-over of P1) sites, without additional cofactors. With the Cre recombinase system, it is possible to excise loxP-flanked DNA segments (45). The structure of the conditional transgene is simple. The promoter and the gene of interest are separated by a loxP-flanked Stop region, which does not allow transcription initiated from the promoter to go through. When the Stop region is removed by Cre-mediated excision, the gene is expressed. Many variations have been added to this system. The conditional transgene can be expressed under the control of lineage/cell-specific promoter and the Cre recombinase under the control of either lineage/cell type-specific or ubiquitous promoters. The conditional transgene may also be driven by a ubiquitous promoter while the Cre expression is lineage/cell type-specific. This approach has been used to study the relationship between brown and white adipocyte lineages (46). It is possible to control Cre recombinase expression in a temporal manner using fusion protein between Cre recombinase and a mutated form of the ligand-binding domain of steroid nuclear receptor. These chimeric proteins do not bind the corresponding endogenous hormone, progesterone or estradiol, but keep the capacity to bind antagonists like RU486 or Tamoxifen (47, 48). Finally, some attempts were made to combine both the tetracycline Tet-off system and Cre technology in transgenic mice (49).

Another major application of the Cre/LoxP system is spatial and/or temporal gene invalidation (50–53). Through homologous recombination, the wild-type allele is replaced by a functional allele containing LoxP sites encompassing coding exons. Mice with the floxed allele are crossed with transgenic mice expressing the Cre recombinase to produce tissue-specific gene knockout (54). The system is also used to remove the selectable marker from the targeted allele. Indeed there is convincing evidence that the presence of a selectable marker expression cassette can influence the expression of the neighboring genes (55). Such interference can create a phenotype that does not reflect the role of the targeted gene. Furthermore, a neomycin gene resistance construct has a cryptic acceptor site between the phosphoglycerokinase promoter and the neo^r coding region. This creates abnormal splicing events when the phosphoglycerokinase promoter-neo^r cassette is sitting in an intron with the same transcriptional orientation as the gene (56, 57).

	TRANSCRIPTION FACTORS AND REGULATION OF ADIPOGENESIS

TOP ABSTRACT INTRODUCTION TRANSGENIC TECHNIQUES APPLIED TO... TRANSCRIPTION FACTORS AND... LIPID AND GLUCOSE METABOLISM ENDOCRINE RESPONSES AND SIGNAL... ADIPOCYTE SECRETIONS GENETIC ABLATION OF ADIPOSE... CONCLUSIONS REFERENCES

 
A tremendous amount of information has been collected on the molecular regulation of adipocyte differentiation (58). The most comprehensive set of data comes from studies on preadipocyte cell lines. When treated with appropriate media, the fibroblast-like preadipocytes undergo differentiation and acquire the morphology and characteristics of lipid-laden adipocytes. Despite the wealth of information obtained on these models, it is important to keep in mind the inherent differences with in vivo AT development. The immortalized preadipocyte cells are aneuploid. This property may induce differences in gene expression compared with adipocytes. Moreover, the cells are cultured out of the normal environment for AT, e.g., normal extracellular matrix in the presence of several cell types that can interact with each other. In that respect, the recent production of animal models with altered expression of key transcription factors for fat cell development has provided essential support for the current model of adipocyte differentiation (Fig. 1) . Several classes of transcription factors and nuclear factors have been implicated in the control of adipocyte differentiation (58). Two groups of factors appear to be essential, CCAAT-enhancer binding proteins (C/EBPs) and peroxisome-proliferator activated receptors (PPARs). In vivo experiments, however, suggest the involvement of other proteins, but their exact contribution to the transcriptional network of adipogenesis awaits further studies.

View larger version (28K): [in this window] [in a new window]   Fig. 1. The transcriptional control of adipogenesis. After proliferation of preadipocytes, the differentiation is promoted by several families of transcription factors. CCAAT-enhancer binding proteins (C/EBP)ß and C/EBP are first expressed, followed by peroxisome-proliferator activated receptors (PPAR), which in turn activate C/EBP. C/EBP exerts positive feedback on PPAR to maintain differentiation. Sterol responsive element binding protein 1c (also named adipocyte determination and differentiation factor 1; ADD1/SREBP1) can increase the transcriptional activity of PPAR. These factors induce the expression of genes that characterize the differentiated adipocyte phenotype. Forkhead box C2 (FOXC2) also activates genes that stimulate adipocyte differentiation (C/EBP, PPAR, and ADD1/SREBP1). References in parentheses correspond to transgenic overexpression or gene knockout of the proteins.

PPAR and retinoid X receptors
PPARs are members of the nuclear receptor superfamily. PPARs heterodimerize with retinoid X receptors (RXR) to bind DNA and activate transcription. PPAR has been shown to play a critical role in adipocyte differentiation. Two protein isoforms that differ in their amino terminus region have been characterized. PPAR1 is expressed in several cell types including fat. PPAR2 is almost exclusively expressed in AT. Targeted disruption of the PPAR gene provokes cardiac malformation around embryonic day 10 due to a placental defect and in utero lethality (64–66). To bypass this developmental stage that precedes the appearance of AT, three different approaches have been used which showed that PPAR was essential for fat development. First, to rescue the placental defect, chimeric embryos were produced with diploid PPAR^2/- cells and wild-type tetraploid cells that develop into extraembryonic lineages, such as placenta, but cannot contribute to the embryo formation (64). One homozygous animal developed to term. Although it had several defects and died shortly after birth, the pup lacked BAT. Second, a study of PPAR^2/- chimeric mice showed that adipocytes in WAT came exclusively from wild-type cells whereas other organs contained a mix of wild-type and negative cells (66). Third, ES cells or embryonic fibroblasts from PPAR null mice did not differentiate into adipocytes (66). Cells from heterozygous animals had impaired lipid accumulation with decreased expression of C/EBP, indicating that the cross talk between the two factors works in both directions (65). Furthermore, cell lines null for PPAR were generated from mouse embryonic fibroblasts containing a floxed allele and a null allele (67). After immortalization, cells were infected with adenovirus expressing Cre recombinase to inactivate the floxed allele. In these cells, PPAR but not C/EBP restored adipogenesis. The data strongly suggest that PPAR is the direct modulator of adipogenesis while the primary role of C/EBP is maintenance of PPAR level. The phenotype of heterozygous mice proved to be very informative (65, 68). Fed a standard diet, the animals had similar weight gain and fat mass as wild-type mice. However, they were resistant to high fat diet-induced obesity. Besides a direct role of PPAR on the development of adipocyte hypertrophy, the phenotype may result from an increased expression of leptin accompanied by a decrease of food intake and an increase in energy expenditure. This is somewhat paradoxical because leptin production is usually proportional to adipocyte size. However, it has been shown that PPAR agonists repress the leptin promoter activity. The ubiquitously expressed PPARß (also named ) has also been proposed to play a role in adipocyte differentiation. PPARß null mice develop normally except that they are smaller than wild-type littermates (69, 70). Gonadal fat stores are reduced because of a decrease in cell number rather than cell size. To determine whether the reduction in fat pad mass was due to a loss of PPARß function in adipocytes, mice with a selective depletion of PPARß in AT were produced (70). No difference was observed between wild-type and transgenic animals, indicating that the decrease of fat mass in PPARß^2/- mice was a consequence of its expression in other tissues than AT.

As stated above, RXRs are the indispensable partners of PPARs. WAT expresses high levels of RXR; however, its role cannot be investigated in RXR^2/- mice because the fetuses die in utero (71, 72). To alleviate this problem, specific ablation of RXR was performed in AT using the Cre-Lox system (53). The authors produced a transgenic mouse line [aP2-Cre-ER^T2(tg/0)] that expresses a tamoxifen-inducible fusion protein between the Cre recombinase and a mutated ligand-binding domain of the human estrogen receptor under the control of the aP2/ALBP promoter. Several rounds of crosses generated aP2-Cre-ER^T2(tg/0)/RXR^L2/- mice in which the RXR DNA binding domain is floxed on one allele and the RXR gene is disrupted on the other. The technological tour de force yielded mice with spatiotemporal control of RXR expression. Treatment of 4-week-old mice with Tamoxifen induced Cre-mediated excision of the floxed allele selectively in adipocytes. Mice with adipocyte ablation of RXR did not develop obesity under a high fat diet or administration of monosodium glutamate, which provokes lesions in the hypothalamus. Adipocyte RXR null mice had an impaired increase in plasma free fatty acid levels during fasting. The phenotype of the mice fed high fat diet is reminiscent of that of PPAR^1/- mice suggesting that PPAR/RXR heterodimers are indeed essential for the formation of hypertrophic adipocytes. Data in fasted mice reveal that RXR is not only important for fat accretion but also for fat mobilization.

PPAR has also recently been proposed as the target for agouti in adipocytes and thus involved in insulin sensitivity (73). Dominant mutations of the agouti gene cause the development of obesity and insulin resistance. In terms of tissue distribution, species differences have been described since, unlike mouse agouti, human agouti is expressed in AT. Transgenic mice expressing agouti in adipocyte do not become obese or diabetic but exhibit high plasma levels of glucose and leptin. Signal transducers and activators of transcription such as STAT-1 and STAT-3 and PPAR protein levels were elevated in transgenic mice. Moreover, these mice increased weight gain when a daily injection of insulin was performed. The authors suggest that insulin triggers the onset of obesity and that adipocyte agouti potentiates this effect through a direct action on PPAR expression (74, 75).

Sterol responsive element binding proteins
Sterol responsive element binding protein 1c (also named adipocyte determination and differentiation factor 1; SREBP1c/ADD1) is a member of the basic helix-loop-helix leucine zipper family of transcription factors expressed in BAT, WAT, liver, and kidney. Studies in cellular models have shown that SREBP1c/ADD1 positively influences adipogenesis and stimulates the expression of genes involved in lipogenesis. However, in vivo, the role of SREBP1c/ADD1 in adipogenesis is more elusive. Eighty five percent of SREBP1 null mice die in utero at embryonic day 11 (76). The surviving animals appear normal with no effect on fat mass. The effect on AT gene expression was moderate with no changes in lipogenic genes such as lipoprotein lipase (LPL), acetylCoA carboxylase (ACC), and fatty acid synthase (FAS). Overexpression of a constitutively active form of SREBP1c/ADD1 in AT unexpectedly results in lipodystrophy (see below). SREBP2 is a member of the SREBP family expressed in AT in lower amounts than SREBP1c. Expression of a dominant positive form of SREBP2 was followed in liver and AT using the PEPCK promoter (77). No change in lipogenic enzyme gene expression was observed. However, transgene expression resulted in increased expression of LDL receptor and cholesterol biosynthetic enzyme mRNAs. This suggests that SREBP2 is a relatively specific activator of cholesterol synthesis in WAT.

Forkhead box C2
Study of the LPL promoter showed that members of the winged helix/forkhead gene family were expressed in differentiated adipocytes (78). Forkhead box C2 (FOXC2) is exclusively expressed in WAT and BAT of adult mice and humans. However, mice lacking FOXC2 die embryonically or perinatally with aortic arch, craniofacial, and skeletal defects (79, 80). To investigate its role in AT, overexpression was achieved with the aP2/ALBP promoter fused to the FOXC2 cDNA. Transgenic mice were leaner with a reduced amount of intraabdominal WAT (81). White fat cells from these depots looked like brown fat cells with a reduced size, multilocular lipid droplets, and numerous mitochondria. The increased thermogenic capacity of WAT might contribute to the lean phenotype. Profound changes in gene expression underlie the new function of WAT. There was an increase in genes important for energy dissipation, such as PPAR coactivator 1, cytochrome oxidase II, and UCP1. Moreover, the sensitivity of the ß-adrenergic signal transduction pathway was enhanced. The increased mRNA levels for several components of the insulin-signaling cascade may explain the increased insulin sensitivity of the transgenic animals fed a high-fat diet. Concomitantly, expression of genes involved in adipogenesis such as C/EBP, PPAR, and SREBP1c/ADD1 was also increased. Altogether, FOXC2 appears as a master gene. Its induction protects against obesity and diet-induced insulin resistance but also participates in the maintenance of the white fat cell phenotype.

Other transcription factors
Other factors expressed in WAT and BAT have been shown to influence adipogenesis. The AP-1 family consists of dimeric complexes of Fos- and Jun-related proteins. Overexpression of fosB, a natural and functional truncated form of FosB, led to increased bone formation and reduced WAT (82). Bone marrow stromal cell culture from fosB mice treated with either osteogenic or adipogenic agents showed a significant decrease in adipose cell number and maturation and increased expression of genes associated with the osteoblast lineage. Adipocytes and osteoblasts are cells of mesenchymal origin that are believed to originate from a common pluripotent precursor. It is therefore conceivable that fosB positively regulates osteoblastogenesis at the expense of adipogenesis.

A role for high mobility group protein Ic (HMGIc) in AT development was recently found in vivo. HMG proteins bind to an adenine/thymine-rich region on the minor groove of DNA and have been shown to play a role in chromatin structure as well as gene-specific transcriptional activation. HMGIc is only expressed during embryonic and fetal stages in line with a role in development. Rearrangements of the HMGIc gene have been frequently detected in benign tumors of mesenchymal origin such as lipomas. Truncation of HMGIc seems to be responsible for cell transformation. Indeed, ubiquitous expression of a truncated HMGIc protein in transgenic mice led to an increased development of WAT early in life and a high incidence of lipomas (83, 84). Conversely, mice with disruption of one or two alleles of the HMGIc gene are resistant to diet-induced obesity (85). Lack of HMGIc expression in leptin-deficient Lep^ob/Lep^ob mice resulted in a decrease in fat pad weights due to a decrease in fat cell number. No differences in gene expression were observed in adipocytes. Together, the data suggest that HMGIc has an important role in fat cell precursor proliferation and/or commitment.

	LIPID AND GLUCOSE METABOLISM

TOP ABSTRACT INTRODUCTION TRANSGENIC TECHNIQUES APPLIED TO... TRANSCRIPTION FACTORS AND... LIPID AND GLUCOSE METABOLISM ENDOCRINE RESPONSES AND SIGNAL... ADIPOCYTE SECRETIONS GENETIC ABLATION OF ADIPOSE... CONCLUSIONS REFERENCES

 
Lipoprotein metabolism
Transgenic models have contributed to clarify the respective roles of many proteins involved in lipid metabolism (Fig. 2) . Links have been established between fatty acid transport in lipoprotein particles and AT development. Transgenic mice expressing human apolipoprotein C-I (apoC-I) in liver show reduced amounts of visceral fat depots and a lack of subcutaneous AT (86). Similarly, VLDL receptor-deficient mice are resistant to genetic and diet-induced obesity (87). These data suggest that fatty acid delivery to AT is impaired. They are also consistent with the role of VLDL receptor as a docking protein for triglyceride-rich lipolysis and for the defective binding of apoC-I enriched lipoprotein particles to the receptor. Overexpression of apoA-II, the second most abundant component of HDL, leads to increased fat mass and promotes insulin resistance with reduced skeletal muscle uptake of glucose (88). This finding shows that a primary disturbance in lipoprotein metabolism can result in traits associated with insulin resistance. LPL, located on the capillary endothelium of extrahepatic tissues, catalyses the rate-limiting step in the hydrolysis of triglycerides from circulating chylomicrons and VLDL. Most LPL is found in AT and skeletal muscle, where the released free fatty acids are stored or oxidized, respectively. Hence, LPL plays a key role in fat partitioning and, through delivery of fatty acids to skeletal muscle, may play an important role in the genesis of insulin resistance as a result of competition between fatty acid and glucose. LPL deficient mice are normal at birth, but develop lethal hypertriglyceridemia within the first day of life (89). To directly assess the role of LPL in AT, LPL heterozygous knockout mice have been crossed with transgenic mice expressing human LPL in skeletal muscle and heart (90). Through backcross, mice expressing LPL exclusively in muscle were obtained. Growth and body composition were not altered by the lack of LPL in AT on a standard genetic background. However, when AT LPL deficiency was obtained on an Lep^ob/Lep^ob background, the rate of weight gain was decreased due to an impaired accumulation of lipid in AT. Triglyceride content was increased in skeletal muscle suggesting partial reallocation of dietary fat storage from AT to skeletal muscle. The chemical nature of the lipid stored in AT was markedly modified in LPL deficient mice, though data suggests that the development of fat stores in AT LPL deficient mice relies on endogenous fat synthesis. On a genetically obese background, this compensatory mechanism does not keep up with the massive weight gain programmed in Lep^ob/Lep^ob mice due to leptin deficiency.

View larger version (37K): [in this window] [in a new window]   Fig. 2. Elements of signal transduction pathways involved in white adipocyte metabolism. Lipolysis: The ß1-, ß2-, ß3-, and the 2-adrenergic receptors (AR) are, respectively, positively and negatively coupled to adenylyl cyclase (AC) and to cAMP production by heterotrimeric G proteins (Gs and Gi). cAMP produced by activation of adenylyl cyclase activates protein kinase A (PKA), which stimulates phosphorylation of hormone-sensitive lipase (HSL). Perilipins and adipocyte lipid binding protein (ALBP) are other members of the lipase complex influencing lipolytic capacity. Activation of HSL catalyzes the hydrolysis of triglycerides. Insulin antilipolytic effect is mediated by the activation of insulin receptor (IR), insulin receptor substrates (IRS), phosphatidylinositol 3-kinase (PI3-K), protein kinase B (PKB), and the phosphodiesterase 3B (PDE3B), which hydrolyzes cAMP into 5'AMP. Glucose transport: The glucose transporter 4 (GLUT 4) is translocated to the plasma membrane in response to insulin by a PI3-K-dependent pathway that involves PKB. Uptake of triglycerides: triglycerides contained in lipoprotein particles like chylomicrons (CL) and VLDL are hydrolyzed by LPL. The released fatty acids are re-esterified for storage as triglycerides by the diacylglyceroltransferase (DGAT). Fatty acids can also be formed from glucose (lipogenesis). Acylation-stimulating protein (ASP) and CD36/FAT represent proteins that are also in fatty acid uptake. References in parentheses correspond to transgenic overexpression or gene knockout of the proteins.

Adipose tissue lipolysis
Hormone-sensitive lipase (HSL) is classically considered as the key enzyme catalyzing the rate-limiting step of AT lipolysis (98, 99). This view is supported by numerous biochemical, physiological, and clinical studies. However, recent data from HSL deficient mice led to a reassessment of the role of HSL in WAT and BAT fat mobilization (100, 101). Catecholamine-induced lipolysis is markedly blunted as expected, but basal (or unstimulated) lipolysis is unaltered in isolated adipocytes suggesting the existence of a lipase different from HSL. Although no major change in the weight of fat pads was observed, lipid metabolism was altered in the knockout mice. WAT from HSL deficient mice accumulated diglycerides demonstrating that the enzyme catalyzed the rate-limiting step in diglyceride catabolism (102). During fasting, i.e., when the enzyme activity is maximal in wild-type animals, HSL^-/- mice showed decreased plasma free fatty acid and triglyceride levels (103). Alteration of triglyceride-rich lipoprotein metabolism was due to a downregulation of VLDL synthesis in liver and an upregulation of LPL activity in skeletal muscle and WAT.

Proper activation of lipolysis also relies upon proteins that are not directly involved in the catalytic process. ALBP/aP2 is an intracellular fatty acid-binding protein highly expressed in adipocytes. Its interaction with HSL N-terminal region may avoid local accumulation of fatty acid during lipolysis and prevent their deleterious effects. It would also allow the fatty acids to be shuttled out of AT. Consistent with such a role for ALBP is the observation that ALBP-null mice exhibit a decreased lipolytic capacity (104, 105). The ALBP-deficient mice show minor alterations of lipid metabolism under a standard chow diet as a consequence of functional compensation by the keratinocyte fatty acid binding protein (106, 107). However, the lack of ALBP protects against hyperinsulinemia and insulin resistance in high-fat-diet-induced or genetically obese mice (106, 108). Access to the lipid droplet constitutes another potential mechanism for the control of lipolysis. Perilipins are proteins covering the large lipid droplets in adipocytes. They shield stored triglycerides from cytosolic lipases. It has been hypothesized that, upon phosphorylation, perilipins allow access to the lipid droplet and thereby allow lipases to interact with their substrates. In two independent studies, ablation of perilipin resulted in mice with decreased fat mass and increased lean body mass (109, 110). The mice are resistant to diet-induced obesity. Moreover, double mutant Lepr^db/db/Plin^-/- mice are protected against the obesity phenotype due to mutation in the leptin receptor. No hepatic steatosis or alteration of the lipid profile was observed, which might be due to the increased metabolic rate of the mutant animals. Basal lipolysis is increased in perilipin-deficient adipocytes, which is in line with a role of perilipin as a suppressor of lipolysis in quiescent cells. However, the results of ß-adrenergic stimulated lipolysis differ between the two studies. Martinez-Botas et al. observed that the basal lipolysis in Plin^-/- mice was similar to the maximal lipolytic capacity of wild-type fat cells and that there was no further stimulation by a ß-adrenergic agonist (109). The data suggest that, without perilipin, adipocytes have a permanent lipolytic drive. In contrast, Tansey et al. report that the increase of glycerol and free fatty acid release induced by a ß-adrenergic agonist was markedly blunted in Plin^-/- adipocytes, which would indicate that perilipin is a necessary cofactor for full lipolytic stimulation (110). The reasons for the discrepancy are unclear. However, the issue needs to be solved as it implies different functions for perilipin. If perilipin does suppress basal and stimulated lipolysis, one can envisage a futile cycle of lipogenesis and lipolysis in Plin^-/- animals that could contribute to the increased metabolic rate. Alternatively, the increased metabolic rate may derive from the increased lean body mass, which is an intriguing aspect of the phenotype, as perilipin is not expressed in skeletal muscle.

Glucose metabolism
The glucose transporter 4 (GLUT4) is the major transporter in tissues in which glucose uptake is stimulated by insulin such as skeletal muscle and WAT. A decrease in GLUT4 level might therefore be responsible for the insulin resistance observed in type 2 diabetes. Skeletal muscle accounts for most of the mass of insulin-responsive tissues. Hence, it has been postulated that in vivo alterations in glucose disposal is due to skeletal muscle. Indeed, GLUT4 heterozygous knockout mice develop muscle insulin resistance and diabetes, which is prevented by transgenic complementation of GLUT4 in skeletal muscle (111, 112). However, the amount of GLUT4 is not decreased in muscle cells in diabetic people whereas it is in their fat cells. Transgenic techniques have therefore been used to modify the level of GLUT4 expression in WAT. GLUT4 overexpression in WAT and BAT was achieved using the aP2/ALBP promoter linked to a genomic fragment encompassing all coding exons of the GLUT4 gene (113). In vivo glucose tolerance is enhanced in transgenic mice. The 6- to 9-fold increased GLUT4 expression in WAT resulted in increased basal and insulin-stimulated glucose transport. Interestingly, young transgenic mice showed increased fat mass resulting from an increase in fat cell number without a change in fat cell size. However, in old female mice, adipocyte size increases. To gain further insight into the role of adipose GLUT4, inactivation of the GLUT4 gene was performed selectively in WAT and BAT by crossing mice with a floxed GLUT4 allele with transgenic mice expressing the Cre recombinase under the control of the ap2/ALBP promoter (52). GLUT4 levels were reduced by more than 70% in BAT and WAT without change in GLUT1 expression. GLUT4 expression was preserved in skeletal muscle and heart. No apparent growth retardation or cardiac abnormalities were observed in contrast to mice totally deficient in GLUT4 (114). Unlike overexpression, GLUT4 targeting in AT does not affect body weight or fat mass of mice eating a standard chow diet. Reduced basal and markedly blunted insulin-stimulated glucose uptake was observed in isolated adipocytes but not in skeletal muscle ex vivo. In vivo, the animals were intolerant to glucose and resistant to insulin. Insulin-stimulated whole-body glucose uptake was decreased. As expected, in vivo insulin-stimulated glucose transport was reduced in WAT and BAT but, surprisingly, the impairment was also observed in skeletal muscle despite normal GLUT4 expression. Moreover, the ability of insulin to suppress hepatic glucose production was blunted. The data clearly suggest that impaired expression of GLUT4 in AT may lead to insulin resistance in WAT but also in skeletal muscle and liver leading to glucose intolerance and hyperinsulinemia. This provocative discovery renews the interest in the role of WAT GLUT4 in the development of type II diabetes but also questions the transmission of impaired insulin action from AT to skeletal muscle and liver. Fat cell-secreted products such as leptin, tumor necrosis factor (TNF) and fatty acids were excluded as potential culprits. Other adipocyte-derived molecules such as adiponectin could be involved.

Glucose metabolism through the hexosamine pathway has been hypothesized to mediate some of the toxic effects of hyperglycemia. The first and rate-limiting enzyme of the pathway is glutamine:fructose-6-phosphate amidotransferase which catalyses the formation of glucosamine-6-phosphate. Overexpression in skeletal muscle and AT using the GLUT4 promoter leads to whole body insulin resistance with a defect in GLUT4 translocation that impairs glucose uptake in skeletal muscle (115, 116). The animals are hyperleptinemic due to increased expression of the leptin gene in WAT (117). This in vivo result is consistent with in vitro data, which suggest that the end product of the hexosamine pathway, UDP-N-acetyl glucosamine, modulates leptin gene transcription through the O-linked glycosylation of the transcription factor Sp1 (118). The hexosamine pathway may serve as a glucose-sensing device linked to leptin expression and glucose disposal.

Thermogenesis
The thermogenic function of BAT results from the expression in this tissue of the uncoupling protein UCP1. UCP1 is expressed in the inner membrane of the mitochondrion where it promotes the dissipation of the proton electrochemical gradient across the inner mitochondrial membrane. The activation of this pathway results in energy dissipation as heat instead of ATP synthesis and is involved in the regulation of body temperature and body weight. Mice lacking UCP1 show no difference in resting metabolic rate but have blunted ß-adrenergic stimulated oxygen consumption (119). In the initial study on mice with a mixed 129/SvPas and C57BL/6J background, 85% of the animals were sensitive to cold, showing the pivotal function of UCP1 for thermogenesis but also suggesting that variations in the genetic background could alter cold sensitivity. Through backcross, UCP1-deficient mice congenic on 129/SvImJ and C57BL/6J backgrounds were obtained (8). The animals showed a profound cold-sensitive phenotype. However, mice on a (129/SvImJ x C57BL/6J) F1 hybrid background were cold resistant despite the absence of UCP1. The cold resistance on the hybrid background is a dramatic example of heterosis or hybrid vigour. The compensation does not come from the UCP1 homologs, UCP2 and UCP3, which are expressed in BAT. Determination of the unknown alternate mechanisms leading to protection against cold in mice without UCP1 may be of great interest for adult humans that express very little BAT. Unlike partial ablation of BAT (see below), UCP1 deficiency does not cause hyperphagia or obesity, suggesting that other factors in BAT are responsible for diet-induced thermogenesis.

Targeted disruptions of the ubiquitously expressed UCP2 and BAT and skeletal muscle-specific UCP3 do not result in alterations of WAT development and response to cold (120–123). However, ectopic expressions of UCPs in WAT and skeletal muscle have provided models of resistance to obesity. Expression of UCP1 in WAT of C57BL/6J mice was achieved through the use of the aP2/ALBP promoter (124). UCP1 level represented 2–10% of the level normally expressed in BAT. In BAT, the overall level of UCP1 was nearly identical in transgenic and wild-type mice. On a standard chow diet, differences in body weight appeared after 6 months of age. The first effect of the transgene was a reduction of the subcutaneous fat pad. At 1 year of age, most fat pads were diminished with the noticeable exception of gonadal fat pads. This regional difference was not due to differences in the level of transgene expression. To determine whether the transgene could alter the development of obesity, the aP2-UCP1 transgenic mice were crossed with the genetically obese C57BL/6J-A^vy/+ mice. Expression of the transgene markedly reduced fat masses in A^vy mice. As in C57BL/6J mice, the effect was more pronounced in subcutaneous AT. Similarly, transgenic C57BL/6J mice were resistant to diet-induced obesity despite food intake comparable to that of non-transgenic animals (125). These data show that expression of UCP1 in WAT protects against genetic and dietary obesity with a more pronounced effect in subcutaneous fat depots compared with intraabdominal fat pads. The mechanisms underlying the phenotype of the aP2-UCP1 mice have been intensively studied. Energy dissipation was depressed in BAT whereas WAT expressing UCP1 showed increased oxygen consumption (126). Despite the inherently low respiratory capacity of WAT, the increase in thermogenesis conferred by UCP1 in all depots may contribute to a slight shift in energy balance leading to a decreased weight gain. However, alternative mechanisms may be involved. Fatty acid synthesis was decreased in aP2-UCP1 mouse WAT as a consequence of UCP1-mediated uncoupling of oxidative phosphorylation (127). Targeting UCP1 into skeletal muscle also resulted in a mouse model resistant to obesity (128). UCP1 levels in muscle were 1% of BAT levels. Fed a chow diet, the transgenic mice weighed less. Furthermore, they were resistant to diet-induced obesity as a result of increased resting metabolic rate. With both types of diet, the transgenic mice had better glucose tolerance and insulin sensitivity than control mice. Decreased competition between fatty acid and glucose in transgenic mice fed a high-fat diet does not seem to contribute to the improved glucose tolerance since, as observed in endurance-trained athletes, there is enhanced expression of LPL and increased triglyceride content in skeletal muscle. The improved skeletal muscle glucose transport may therefore result from increased uncoupling of the respiratory chain as shown in cellular models. Strong overexpression of human UCP3 in skeletal muscle results in a similar phenotype, except that the animals, despite their body weights, which are lower than their non-transgenic littermates, are hyperphagic (129). Hence, whereas neither UCP1 nor UCP3 knockout mice have an increased susceptibility to obesity, overexpression of UCP1 and UCP3 can prevent obesity. Compensatory mechanisms may explain the lack of response in knockout models. It remains to be seen whether pharmacological strategies targeting UCPs, either through increased expression or modulation of their activities, influence energy balance in humans.

Increased thermogenesis may also come from an unexpected route. Inactivation of a translational inhibitor, the eukaryotic translation initiation factor 4E-binding protein 4E-BP1, leads to high metabolic rate and small white fat pads. WAT of the knockout animals expresses UCP1 (130). The upregulation of the thermogenic protein may, in part, result from the increased translation of PPAR coactivator 1, a transcriptional coactivator involved in mitochondriogenesis and adaptive thermogenesis. Besides providing a new model of leanness and obesity resistance, this study reveals the importance of translation in the regulation of energy expenditure.

	ENDOCRINE RESPONSES AND SIGNAL TRANSDUCTION

TOP ABSTRACT INTRODUCTION TRANSGENIC TECHNIQUES APPLIED TO... TRANSCRIPTION FACTORS AND... LIPID AND GLUCOSE METABOLISM ENDOCRINE RESPONSES AND SIGNAL... ADIPOCYTE SECRETIONS GENETIC ABLATION OF ADIPOSE... CONCLUSIONS REFERENCES

 
Adrenergic receptors
Among the nine pharmacologically and genetically distinct adrenergic receptors (AR), four are expressed in adipocytes (Fig. 2). The ₂- and ß-ARs have opposite signal transduction pathways and are known to participate in the regulation of BAT and WAT development and metabolism (131). In addition to variations in receptor number, lipolytic rates in AT are thought to be affected by the expression of particular ß-AR subtypes and by the ratio of ₂/ß-AR. Transgenic mice overexpressing the human ß₁-AR have been generated (132). The aP2/ALBP promoter was used to specifically target the transgene to differentiated white and brown adipocytes (16). Transgenic mice gained weight more slowly and had reduced adipose stores compared with transgenic littermates, especially in response to a high fat diet. Moreover, brown adipocytes appeared in the subcutaneous white fat pads. The in vivo phenotypic effects are in agreement with the in vitro responses to ß₁-AR stimulation in isolated fat cells, i.e., increased lipolytic activity of the adipocytes and greater energy expenditure through heat production by the additional population of brown adipocytes.

In rodents, the ß₃-AR is expressed in fat at a much higher level than ß₂ and ß₁-AR (133) and has therefore been proposed to be the major regulator of adrenergic responses in spite of its lower affinity for endogenous catecholamines. Moreover, ß₃-AR selective agonists have been proposed as potential anti-obesity drugs based on their impressive effects on energy expenditure, insulin levels, and food intake in rodents. To study the physiological relevance of ß₃-AR, mice lacking the receptor were generated (134, 135). Surprisingly, ß₃-AR^-/- mice show only a modest tendency to become overweight even when fed a high fat diet. However, a rise in total body fat was observed. Decreased action of the ß₃-AR was once thought to be responsible for the development of obesity. It is clear now that the absence of the receptor is not sufficient. Specific expression of ß₃-AR in BAT or BAT and WAT confirmed that the expression of ß₃-AR was indispensable in white and brown adipocytes to rescue the effects of ß₃-AR agonists on oxygen consumption, insulin secretion, and food intake and that ß₃-AR in other tissues were not required (136). White adipocytes appeared to be predominantly involved in the ß₃-AR agonist effects on insulin levels and food intake while effect on oxygen consumption was associated to brown adipocytes.

Although ß₃-AR agonists have been described as potent anti-obesity drugs in rodents, their relevance remains questionable in humans. Rodents have large amounts of thermogenic brown adipocytes and high levels of ß₃-AR in WAT compared with humans. To understand the differences in ß₃-AR sites of expression, fragments of human genomic DNA encompassing the ß₃-AR gene were microinjected into ß₃-AR knockout mouse oocytes (137). The transgenic mice were used to identify tissues where the human ß₃-AR promoter/enhancer is active and to delineate the cis acting regions involved in tissue-specific expression. Human ß₃-AR mRNA was expressed only in brown adipocytes with little or no expression in WAT, liver, stomach, small intestine, skeletal muscle, or heart. This strategy can prove very useful for the precise mapping of cis-acting elements conferring tissue-specific expression and for the study of physiological or pharmacological mechanisms controlling human ß₃-AR gene expression.

In human white adipocytes, the ß-adrenergic response to catecholamines can be totally counteracted by the ₂-adrenergic pathway. A large body of evidence indicates that the ratio of ₂/ß-AR in different fat pad depots affects the lipolytic rate and is closely associated with the enlargement of AT in obese subjects (131). Because of the high levels of ß₃-ARs and the very low expression of ₂-ARs in WAT, rodents do not mimic human adrenergic receptivity. To assess the importance of the ₂/ß-AR balance in vivo, gene targeting and transgenic approaches were combined to create mice with increased ₂/ß-AR balance in AT (138). Using the aP2/ALBP promoter, expression of human ₂A-AR was targeted in the ß₃-AR null mouse AT. Such "human-like" mice developed high-fat-diet-induced obesity associated to adipocyte hyperplasia rather than hypertrophy. None of the plasma parameters such as glucose or insulin levels were modified in obese mice while a slight increase in leptin levels was observed. In these conditions, the lack of insulin resistance could be associated to adipocyte hyperplasia. No apparent phenotype was noticed in mice expressing both ß₃- and ₂-ARs, clearly demonstrating that the obese phenotype required the interactions between two genes and diet, i.e., the presence of ₂-ARs, the absence of ß₃-ARs, and a high fat diet.

GTP binding proteins
When considering the control of lipolysis in adipocytes, the activity of adenylyl cyclase and consequently the levels of intracellular cAMP are crucial for HSL activation. Key elements involved in the transmembrane control of adenylyl cyclase activity are heterotrimeric GTP-binding proteins because of their capacity to be coupled with multiple receptors. The Gs and Gi subunits are able to stimulate or inhibit adenylyl cyclase activity, respectively. Gs deficiency provokes embryonic lethality (139). In heterozygous animals, a different phenotype is observed depending on the sex of the transgenic genitor. Tissue-specific imprinting of the Gs gene has been reported in numerous tissues, including WAT and BAT. A decreased fat mass is observed in +/p- mice from heterozygous male genitors, while m-/+ mice from heterozygous female genitors become obese in the early adulthood (140). The effects on energy metabolism also depend on parental inheritance, i.e., +/p- mice are hypermetabolic and +/m- mice are hypometabolic. It is likely that decreased adipocyte Gs expression in +/m- mice leads to obesity because cAMP stimulates lipolysis and thermogenesis in mouse WAT and BAT. The explanation of the phenotype described in +/p- mice remains unclear although a rise in sympathetic nervous system activity could contribute to the phenotype. A better insulin sensitivity is observed in lean males and, surprisingly, in obese females (141).

The inhibitory pathway of adenylyl cyclase has also been genetically modified by an inducible RNA antisense strategy directed against the Gi2 subunit (33, 34, 142). The PEPCK promoter was used for its capacity to drive the transgene in liver and fat and to be inducible by cAMP. The expression of the antisense RNA induced an increase in basal cAMP levels, no modifications of the ß-AR stimulatory pathway, and severe blunting in the response to the inhibitory agonists such as adenosine A1. Transgenic mice in which the transgene becomes active at birth by a rise in basal levels of cAMP displayed a smaller liver and less AT. The animals developed type 2 diabetes with hyperinsulinemia, impaired glucose tolerance, and resistance to insulin. This data suggesting that Gi mimics insulin action was confirmed by targeted expression of the GTPase-deficient constitutively active Gi2 in fat and muscle (143). The same animals were used to define protein-tyrosine phosphatase 1B (PTP1B) as a step in the cellular pathway leading to Gi2-mediated enhancement of insulin signaling (142, 144).

Although the Gq-protein kinase C signaling pathway has only been poorly documented in AT, a similar antisense strategy has been used to impair Gq expression (145). Gq deficiency in liver and AT at birth was associated to increased body and fat masses and reduced lipolytic response. Since ₁-ARs are coupled to phosphoinositide hydrolysis and protein kinase C activation, the authors suggested a potential role for ₁-ARs and intracellular Ca²⁺ in the inhibition of lipolysis in agreement with other studies (146).

Protein kinase A
Surprisingly, adipocyte adenylyl cyclase, one of the major components of transmembrane signaling associated with the control of lipolysis has not been investigated using transgenic techniques. Stimulation of the lipolytic cascade involves the phosphorylation of HSL by the cAMP-activated protein kinase, which is composed of two regulatory and two catalytic subunits. Among the four regulatory subunit genes, the RIIß isoform is abundant in BAT, WAT, and brain. Targeted disruption of the RIIß subunit produces lean mice resistant to obesity when fed a high fat diet (147–149). In both brown and white adipocytes, a compensatory rise in the RI subunit has been described. This isoform switch is associated with an increased UCP1 expression in BAT and enhanced basal lipolysis in WAT due to the higher binding capacity of RI to cAMP. The disruption of both RIß and RIIß genes leads to the same preservation of cAMP-dependent regulation by the compensatory rise in RI protein half-life without a change in gene transcription. However, the ability of ß-AR agonists to stimulate lipolysis is strongly compromised in WAT. Finally, RI null mice show early embryonic lethality with severe developmental abnormalities confirming RI modulation as an essential mechanism in the safeguard of pleiotropic cAMP cellular responses (150).

Insulin and insulin signaling
Insulin is the primary hormone to inhibit lipolysis and is responsible for signaling the storage and utilization of glucose. The application of transgenic mouse technology to the study of diabetes has been extensively reviewed (151). Here, we will focus on the