Search

In the present study, we investigated the role of the phytoestrogen genistein and 17beta-estradiol in human bone marrow stromal cells, undergoing induced osteogenic or adipogenic differentiation. Profiling of estrogen receptors (ERs)-alpha, -beta1, -beta2, -beta3, -beta4, -beta5, and aromatase mRNAs revealed lineage-dependent expression patterns. During osteogenic differentiation, the osteoblast-determining core binding factor-alpha1 showed a progressive increase, whereas the adipogenic regulator peroxisome proliferator-activated receptor gamma (PPARgamma) was sequentially decreased. This temporal regulation of lineage-determining marker genes was strongly enhanced by genistein during the early osteogenic phase. Moreover, genistein increased alkaline phosphatase mRNA levels and activity, the osteoprotegerin:receptor activator of nuclear factor-kappaB ligand gene expression ratio, and the expression of TGFbeta1. During adipogenic differentiation, down-regulation in the mRNA levels of PPARgamma and CCAAT/enhancer-binding protein-alpha at d 3 and decreased lipoprotein lipase and adipsin mRNA levels at d 21 were observed after genistein treatment. This led to a lower number of adipocytes and a reduction in the size of their lipid droplets. At d 3 of adipogenesis, TGFbeta1 was strongly up-regulated by genistein in an ER-dependent manner. Blocking the TGFbeta1 pathway abolished the effects of genistein on PPARgamma protein levels and led to a reduction in the proliferation rate of precursor cells. Overall, genistein enhanced the commitment and differentiation of bone marrow stromal cells to the osteoblast lineage but did not influence the late osteogenic maturation markers. Adipogenic differentiation and maturation, on the other hand, were reduced by genistein (and 17beta-estradiol) via an ER-dependent mechanism involving autocrine or paracrine TGFbeta1 signaling.

Peroxisome proliferator-activated receptor gamma (PPAR-gamma) activation in adipose tissue is known to regulate genes involved in adipocyte differentiation and lipid metabolism. However, the role of PPAR-gamma in muscle remains unclear. To examine the potential regulation of genes by PPAR-gamma in human skeletal muscle, we used semiquantitative RT-PCR to determine the expression of PPAR-gamma, lipoprotein lipase (LPL), muscle carnitine palmitoyl transferase-1 (mCPT1), fatty acid-binding protein (FABP), carnitine acylcarnitine transferase (CACT), and glucose transporter-4 (GLUT4) in freeze-dried muscle samples from 14 male subjects. These samples were dissected free of adipose and other tissue contamination, as confirmed by minimal or absent adipsin expression. Between individuals, the messenger ribonucleic acid concentration of PPAR-gamma varied up to 3-fold, whereas LPL varied up to 6.5-fold, mCPT1 13-fold, FABP 4-fold, CACT 4-fold, and GLUT4 up to 3-fold. The expression of LPL (r2 = 0.54; P = 0.003), mCPT1 (r2 = 0.42; P = 0.012), and FABP (r2 = 0.324; P = 0.034) all correlated significantly with PPAR-gamma expression in the same samples. No significant correlation was observed between the expression of CACT and PPAR-gamma or between GLUT4 and PPAR-gamma. These findings demonstrate a relationship between PPAR-gamma expression and the expression of other genes of lipid metabolism in muscle and support the hypothesis that PPAR-gamma activators such as the antidiabetic thiazolidinediones may regulate fatty acid metabolism in skeletal muscle as well as in adipose tissue.

Both bone mass and serum leptin levels are increased in obesity. Because osteoblasts and adipocytes arise from a common precursor in bone marrow, we assessed the effects of human recombinant leptin on a conditionally immortalized human marrow stromal cell line, hMS2-12, with the potential to differentiate to either the osteoblast or adipocyte phenotypes. By RT-PCR and Western immunoblot analysis, the hMS2-12 cells expressed messenger RNA (mRNA) and protein for the leptin receptor. Leptin did not affect hMS2-12 cell proliferation, but resulted in dose- and time-dependent increases in mRNA and protein levels of alkaline phosphatase, type I collagen, and osteocalcin, and in a 59% increase in mineralized matrix. Leptin increased mRNA levels of lipoprotein lipase at 3 days, but decreased mRNA levels of adipsin and leptin at 9 days and decreased lipid droplet formation by 50%. Leptin did not affect the expression of Cbfa1 or peroxisome proliferator-activated receptor-gamma2, transcription factors involved in commitment to the osteoblast and adipocyte pathways, respectively. Thus, leptin acts on human marrow stromal cells to enhance osteoblast differentiation and to inhibit adipocyte differentiation. Our data support the hypothesis that leptin is a previously unrecognized, physiological regulator of these two differentiation pathways, acting primarily on maturation of stromal cells into both lineages.

Complement plays important roles in host immune defences, and recent studies suggest that adipose tissue is an important site of production for some complement proteins. Starvation has been associated with low complement levels, but studied populations have usually had concomitant opportunistic infections or other conditions which might affect complement levels. To determine the impact of body weight and changes in body weight on serum complement, we investigated levels of complement proteins in otherwise healthy patients with a wide range of body weights, including patients with anorexia nervosa before and after treatment, obese dieters before and after weight loss, and normal weight controls. We found that complement proteins of the alternative pathway (C3, B, and D), alternative pathway haemolytic activity (AP50) and the inhibitors H and I were low in starving anorectics and normalized with weight gain. C3a levels were comparable in anorectics at low weight and after weight gain, indicating that low serum complement levels were attributable to hypoproduction and not complement cascade activation with consumption. Further, levels of C3, B, AP50, H and I, but not D, were higher than controls in obese patients and decreased toward normal after weight loss. Overall, percentage of ideal body weight, changes in body weight, and serum transferrin were each highly correlated with serum levels of complement proteins. We conclude that levels of alternative pathway complement components are determined in part by factors that influence body weight and by weight changes, possibly due to changes in production in adipose tissue or at other sites.

Preadipocyte factor-1 (Pref-1), a novel gene product isolated from murine preadipocyte 3T3-L1 cells, is thought to function as a negative regulator of adipocyte differentiation. We investigated the regulation of Pref-1 expression in 3T3-L1 preadipocytes during proliferation, growth arrest, and early differentiation in the presence and absence of three well described differentiation antagonists: interleukin-11 (IL-11), transforming growth factor-beta, and tumor necrosis factor-alpha. Northern blot analysis was used to determine messenger RNA (mRNA) steady state expression of Pref-1 and two differentiation-specific genes, adipsin and glycerol-3-phosphate dehydrogenase. We confirmed that Pref-1 mRNA is abundant in proliferating preadipocytes and that its expression is dramatically reduced early in differentiation. However, proliferating and growth-arrested cells treated with the differentiation inhibitor IL-11 demonstrated a modest decrease in Pref-1 mRNA abundance. Transforming growth factor-beta and tumor necrosis factor-alpha had little effect. The reduction of Pref-1 mRNA was most dramatic in differentiating preadipocytes treated with IL-11, occurring despite inhibition of adipogenesis, as judged by cell morphology and adipocyte-specific gene expression (adipsin and glycerol-3-phosphate dehydrogenase). This effect of IL-11 on Pref-1 suggests that different mechanisms are responsible for the IL-11-induced and the differentiation- associated down-regulation of Pref-1, thus dissociating Pref-1 regulation from differentiation. We conclude that Pref-1 expression is not a reliable marker of preadipocytes, and that decreased Pref-1 abundance does not function as a trigger for adipocyte differentiation.

Adipsin is a serine protease with complement factor D activity that is synthesized by adipocytes and secreted into the blood stream. Expression of adipsin is deficient in models of genetic (ob/ob, db/db) and acquired (monosodium glutamate-lesioned) obesity, but the cellular mechanisms responsible for this deficiency are unknown. Because hyperinsulinemia is frequently associated with obesity, we evaluated the effects of this hormone and insulin-like growth factor 1 (IGF-1) on adipsin secretion and adipsin messenger RNA (mRNA) levels in 3T3-F442A adipocytes. In the present study, we report that in fully differentiated adipocytes (after 11 days post confluence), insulin exposure progressively decreases adipsin secretion by 40%, 67%, and 78% after 2, 4, and 6 days of treatment. The inhibition of adipsin secretion by insulin is the result of a corresponding decrease in adipsin mRNA and is specific since two other differentiation-dependent fat cell mRNAs encoding aP2 (a fatty acid binding protein) and glycerophosphate dehydrogenase (GPD), are unaffected. Insulin suppresses adipsin gene expression via high affinity insulin receptors, because physiological levels of insulin produce this effect, and dose-response curves for insulin stimulation of 2-deoxyglucose uptake and glucose utilization are similar to insulin's effect on adipsin. In contrast, insulin when present during days 1-8 post confluence (during differentiation) markedly increases adipsin secretion and adipsin mRNA levels. This stimulation is due to the ability of insulin to accelerate differentiation as evidenced by corresponding increases in aP2 and GPD mRNAs as well. Insulin and IGF-1 are equipotent in this effect, suggesting that both insulin and IGF-1 receptors can mediate this response. In summary, during the differentiation of 3T3-F442A adipocytes, insulin stimulates adipsin gene expression by accelerating differentiation. As the cells become mature adipocytes, they acquire some differentiation-dependent factor, which couples insulin receptor stimulation to inhibition of adipsin gene expression. This model should aid our search for the molecular links between insulin receptor stimulation and altered gene expression.

Adipsin gene expression is greatly diminished in certain forms of genetic and acquired obesity. In the present study we evaluate the time course for the development of adipsin deficiency in obesity and its regulation by the sympathomimetic-thermogenic drug mixture ephedrine and caffeine. Previously, it was unknown whether adipsin deficiency occurred before or after the development of massive obesity. In the first series of experiments in which mice were treated with monosodium glutamate (MSG) for the first week of life, we demonstrate that adipsin deficiency occurs early in the development of MSG-induced obesity as evidenced by decreased circulating adipsin concentrations by 1 week of age and deficient adipsin mRNA levels in white adipose tissue (WAT) by 2 weeks. In db/db mice, diminished circulating adipsin was noted at 2 weeks of age. In both models, decreased adipsin gene expression precedes the development of marked obesity. Little is known about the factors which regulate adipsin gene expression in obesity. Common to the ob/ob, db/db and MSG models is diminished thermogenesis and sympathetic nervous system activity. In a second series of experiments we sought to determine whether adipsin deficiency in obesity could be corrected by treatment with ephedrine and caffeine (E+C), a sympathomimetic-thermogenic mixture previously shown to increase thermogenesis and reverse obesity in some models. In the present study, E+C treatment of MSG obese mice reversed obesity and markedly increased serum adipsin and adipsin mRNA levels in WAT and brown adipose tissue (BAT). In ob/ob mice, however, E+C treatment produced a negligible increase in adipsin mRNA levels in WAT and BAT as well as serum adipsin concentrations and this correlated with only a very small decrease in obesity. Thus, the ability of E+C to increase adipsin gene expression correlated with its ability to reverse obesity in these two models. Finally, the effect of E+C on adipsin gene expression may not be exerted directly on the fat cell since treatment of cultured 3T3-F442A adipocytes and isolated rat adipocytes in primary culture produced no effect on adipsin mRNA or secreted protein despite a lipolytic effect as measured by increased glycerol release. In summary, decreased adipsin gene expression occurs early in the development of MSG and db/db obesity and is markedly increased in the MSG model by the sympathomimetic-thermogenic drug mixture, E+C, which also reverses obesity. Elucidation of the factors responsible for these effects may enhance our understanding of fat cell gene regulation and obesity.

The C3 convertase of the C3b feedback cycle was generated in the fluid phase from C3b and factors B and D. It was shown to be an effective decomplementing agent when reacted with human serum. Many of the effects observed were dependent on the operation of the feedback during the ongoing reaction. Oxidization of factor B with iodine did not significantly increase the potency or the stability of the enzyme. Treatment of C3b with antrypol inhibited the formation of the enzyme. In preliminary experiments in the rat, the convertase was successfully formed in vivo, accompanied by cleavage of C3, consumption of C5, C6 and C7, and a biphasic change in the circulating neutrophils.

The factors influencing the ultracentrifugation characteristics of immune complexes generated in the presence of fresh normal human serum have been analysed. In the absence of alternative pathway factors B, D or Properdin, the size of complexes was increased. When classical pathway function was blocked, in C1q deficient serum or in the presence of Mg EGTA, although the proportion of complexes remaining in solution were reduced their size was similar to those formed in normal human serum. In C2 deficient serum, a heterogeneous population of complexes was generated. In all instances repletion with the appropriate missing complement component reversed the abnormality. We conclude that there is normally a rapid sequential process of classical followed by alternative pathway activation leading to stable soluble complexes. In the absence of C1 activation the alternative pathway process requires precipitation of the antigen-antibody aggregates whereas in normal serum these events occur in the fluid phase. We suggest that in C2 deficient serum the C1 and/or C4 reacted complexes fail to activate the alternative pathway efficiently.

We have isolated, mapped and sequenced adipsin, the adipocyte differentiation-dependent serine protease gene. This gene, which is present in a single form in the mouse, spans 1.7 kilobases and contains five exons. While the basic exon structure characteristic of serine protease genes is conserved in adipsin, there is also a fusion of two exons that are separate in other serine proteases. The sequence data also suggests a mechanism of alternative splicing which appears to account for the generation of two adipsin mRNA species differing by only three nucleotides and encoding two different signal peptides. To investigate the control of adipsin expression we have examined the effects of tumor necrosis factor (TNF) on adipocytes. The level of adipsin RNA is dramatically decreased by hormone treatment, but the change occurs more slowly than for other fat cell mRNAs, such as glycerophosphate dehydrogenase. These results show that adipsin is a novel serine protease gene whose expression is regulated by a macrophage-derived factor which modulates expression of other adipocyte-specific RNAs.