Leptin regulates proinflammatory immune responses - CiteSeerX

Leptin regulates proinflammatory immune responses S. LOFFREDA,* S. Q. YANG,* H. Z. LIN,* C. L. KARP,*,† M. L. BRENGMAN,‡ D. J. WANG,‡ A. S. KLEIN,‡ G. B. BULKLEY,‡ C. BAO,* P. W. NOBLE,* M. D. LANE,§ AND A. M. DIEHL*,1 *Departments of Medicine, †Molecular Microbiology and Immunology, ‡Surgery, and §Biological Chemistry, Johns Hopkins University, Baltimore, Maryland 21205, USA

Obesity is associated with an increased incidence of infection, diabetes, and cardiovascular disease, which together account for most obesity-related morbidity and mortality. Decreased expression of leptin or of functional leptin receptors results in hyperphagia, decreased energy expenditure, and obesity. It is unclear, however, whether defective leptin-dependent signal transduction directly promotes any of the conditions that frequently complicate obesity. Abnormalities in tumor necrosis factor a expression have been noted in each of the above comorbid conditions, so leptin deficiency could promote these complications if leptin had immunoregulatory activity. Studies of rodents with genetic abnormalities in leptin or leptin receptors revealed obesity-related deficits in macrophage phagocytosis and the expression of proinflammatory cytokines both in vivo and in vitro. Exogenous leptin up-regulated both phagocytosis and the production of proinflammatory cytokines. These results identify an important and novel function for leptin: up-regulation of inflammatory immune responses, which may provide a common pathogenetic mechanism that contributes to several of the major complications of obesity.—Loffreda, S., Yang, S. Q., Lin, H. Z., Karp, C. L., Brengman, M. L., Wang, D. J., Klein, A. S., Bulkley, G. B., Bao, C., Noble, P. W., Lane, M. D., Diehl, A. M. Leptin regulates proinflammatory immune responses. FASEB J. 12, 57–65 (1998)

ABSTRACT

Key Words: obesity · macrophage · cytokine · phagocytic function · TNF · lipopolysaccharide

LEPTIN, THE PROTEIN ENCODED by the ob gene, is known to regulate appetite and energy expenditure. Obese/obese (ob/ob) mice, homozygous for a spontaneous mutation in the ob gene, fail to produce leptin and exhibit hyperphagia and obesity. Treatment of such mice with recombinant leptin results in decreased food intake and weight loss (1–3). It is not known whether leptin deficiency per se explains other aspects of the ob/ob phenotype, such as diabetes and hyperlipidemia. Recently, ectopic expression of tumor necrosis factor alpha (TNF-a)2 was documented in adipose tissues of obese rodents and humans (4, 5) and implicated in the pathogenesis of

both disorders (6). Diabetes and hyperlipidemia are thought to be primary factors in obesity-associated morbidity and mortality because they are associated with an increased susceptibility to infections and vascular disease (7–9). We recently reported that ob/ob mice exhibit increased sensitivity to endotoxin-induced liver injury and lethality, identifying a potential link between leptin deficiency and the dysregulated expression of endotoxin-inducible cytokines (10). Since these cytokines are known to regulate tissue sensitivity to insulin (11) as well as lipid metabolism (12, 13), it is conceivable that leptin deficiency promotes diabetes and hyperlipidemia by causing dysfunction of the immune system. Macrophages are important sources of cytokine generation in response to endotoxin stimulation (14, 15). To test the hypothesis that leptin itself might be an immune modulator, macrophages isolated from ob/ob mice and their lean, heterozygous (?/ob) littermates were evaluated with respect to macrophage phagocytic function and cytokine production before and after lipopolysaccharide stimulation in the presence and absence of recombinant leptin. To determine whether the effects of leptin were mediated directly by the leptin receptor, parallel studies were undertaken with macrophages from db/db mice (which lack functional leptin receptors) (16, 17) and their lean, heterozygous (?/db) littermates. The physiological relevance of the in vitro findings was evaluated by comparing the relative abilities of obese and lean rodents to phagocytose and kill circulating 51 Cr, 125I-labeled Escherichia coli and to produce proinflammatory cytokines when challenged with lipopolysaccharide (LPS). These studies demonstrate impaired phagocytosis and abnormal cytokine gene expression in these genetically obese animals, indicating that leptin regulates macrophage function, and suggest that defective leptin-dependent signaling could well contribute to obesity-re-

1 Correspondence: 912 Ross Bldg., Johns Hopkins University, 720 Rutland St., Baltimore, MD 21205, USA. 2 Abbreviations: TNF, tumor necrosis factor; LPS, lipopolysaccharide; IL, interleukin; RIA, radioimmunoassay; ANOVA, analysis of variance; ELISA, enzyme-linked immunosorbant assay; RE, reticuloendothelial.

0892-6638/98/0012-0057/$02.25 Q FASEB

/ 381c 0006 Mp 57 Tuesday Nov 25 11:05 AM LP–FASEB 0006

57

lated morbidity and mortality by altering the expression of TNF-a and related cytokines.

EXPERIMENTAL PROCEDURES Materials Human recombinant leptin was a generous gift from Amgen (Thousand Oaks, Calif.). To evaluate the purity of the recombinant protein, endotoxin levels were assessed by using a limulus lysate assay with a sensitivity of 12 pg/ml (Endogen, Cambridge, Mass.). No endotoxin contamination of the recombinant leptin was detected at experimental dilutions. Recombinant leptin was also subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis. The protein appeared as a single band with an apparent molecular mass of 16 kDa after silver staining. Lipopolysaccharide (LPS) from E. coli serotype 0111:B4 was obtained from Sigma (St. Louis, Mo.). Thioglycollate was purchased from DIFCO (Detroit, Mich.); plastic tissue culture dishes (Falcon, Franklin Lakes, N.J.) and medium came from GIBCO (Grand Island, N.Y.). The cDNA for murine TNF-a was generously provided by Dr. Bruce Beutler (18). TNF-a, interleukin 6 (IL-6), and IL-10 protein concentrations were measured by commercially available enzyme-linked immunosorbant assays (ELISAs) with a sensitivity of 50, 50, and 37 pg/ml, respectively (Endogen). IL12 p40 and p70 protein levels were measured by radioimmunoassays (RIAs) with a sensitivity of 30 and 10 pg/ml, respectively, as previously described (19). In vitro phagocytosis assays All mice and rats were purchased from the Jackson Laboratory (West Grove, Pa.). All animal experiments followed National Institutes of Health and Johns Hopkins Institutional guidelines, which ensure humane use of animal subjects. Obese (ob/ob, db/db) C57Bl/6 mice and their lean, heterozygous (?/ob, ?/db) littermates were injected i.p. with thioglycollate. Four days later, peritoneal macrophages or bone marrow were harvested as described previously (20). Macrophages were pooled separately from three ob/ob, three db/db, three ?/ ob, and three db/db mice. Cells were harvested from a second and third series of mice to perform replicate experiments. Freshly isolated peritoneal macrophages were cultured at a density of 0.5 1 106 cells/ml on plastic dishes in RPMI-1640 with 10% fetal bovine serum and 10 mM Hepes in the presence or absence of leptin (total volume 10 ml, final concentration 500 ng/ml) or vehicle (sterile water, total volume 10 ml) for 16 h. Half of the leptin-treated cultures and half of the control cultures were treated with LPS (1 mg/ml) 90 min before harvest. The remaining cultures were treated with an equal volume of vehicle (sterile water). Bone marrow (11106 cells/ml) was initially cultured for 5 days in 10% L cell-conditioned medium containing low glucose Dulbecco’s modified Eagle medium, 10% heat-inactivated fetal bovine serum, Lglutamine (2 mM), penicillin (100 U/ml), and streptomycin (100 mg/ml) to enrich the population for macrophages (20). Day 6 cultures of bone marrow-derived macrophages were treated with vehicle, LPS, and leptin as described above. To test the phagocytic function of the cultured peritoneal and bone marrow macrophages, duplicate cultures were incubated with Candida parasilopsis and stained with acridine orange/crystal violet (21). Phagocytic activity (i.e., the percentage of cells that had phagocytosed Candida) and indices of phagocytosis and intracellular killing were calculated as previously described (22). The phagocytic index has been defined 58

Vol. 12

January 1998

as the average number of intracellular Candida in each phagocytically active macrophage. The candicidal index was determined in the same macrophages by counting the number of dead (orange) intracellular Candida and expressing this number as a percentage of the total number of intracellular Candida (orange and green). Mean phagocytic activity, phagocytic index, and candicidal index were calculated for each of the three separate experiments by counting at least 500 macrophages per treatment group/experiment; results were averaged and evaluated by analyssis of variance (ANOVA). Cytokine assays in cultured macrophages To determine whether leptin promoted basal or LPS-stimulated cytokine production, ELISAs for TNF-a, IL-6, and IL-10 and RIAs for IL-12 p40 and p70 were performed on supernatants of stimulated macrophages. For these experiments, thioglycollate-elicited peritoneal macrophages were harvested from four separate groups (nÅ4 mice/group) of normal C57Bl/6 mice. Macrophages were cultured in the presence of medium alone, LPS (1 mg/ml), leptin (500 ng/ ml), or LPS / leptin. After 24 h, medium was harvested and concentrations of each cytokine were assayed in quadruplicate. Treatment-related differences were evaluated by the paired Student’s t test. To begin to evaluate the mechanism underlying the observed enhancement of LPS-induced cytokine production by leptin, total RNA was isolated from ob/ob, ?/ob, and normal C57BL-6 macrophages according to the method of Chomczynski and Sacchi (23). The RNA was separated by agarose gel electrophoresis under denaturing conditions and transferred to nylon membranes. Blots were incubated with radiolabeled murine TNF-a cDNAs, washed under stringent conditions, and exposed to film. To ensure that equivalent amounts of RNA (20 mg/lane) were present in each treatment group, variations in 18S RNA expression were evaulated on gels and membranes that had been stained with ethydium bromide. In addition, after hybridization with TNF-a cDNAS, all blots were striped and reprobed with cDNAs for GAPDH, a constitutively expressed gene (10). Evaluation of macrophage function in vivo In vivo phagocytic function The phagocytic function of macrophages within intact ob/ob, ?/ob mice, and normal C57BL-6 mice (nÅ6/group/experiment) was evaluated in two separate experiments by measuring the clearance and subsequent killing of intravenously administered 51Cr, I125-labeled E. coli as described (22). Each mouse received an average of 8 1 103 cpm in 108 bacteria suspended in 300 ml sterile saline via an indwelling internal jugular vein canula. Ninety minutes after injection of bacteria, the mice were killed and the liver, spleen, and blood were harvested. The tissue-specific radioactivity of each isotope was determined as described below and results were evaluated by ANOVA. Previous studies (22) have indicated that virtually all of the intravenously injected E. coli are cleared by the reticuloendothelial system (85{5% by the liver, lung, and spleen) within 6 min and that, thereafter, tissue 51Cr radioactivity remains constant for 6 h. The percent of E. coli that are cleared has been found by quantitative culture to be equal to the (51Cr activity in an organ/51Cr activity injected) 1 100. Quantitative culture of tissue homogenates has also shown that the magnitude of subsequent bacterial killing in an organ is directly proportional to the loss of 125I activity from that organ. Thus, at any given time, the difference between the accumulated

The FASEB Journal


LOFFREDA ET AL.

percent of 125I activity in an organ and the percent of 51Cr activity in that organ represents that fraction of E. coli that have been killed within that organ. The percent of cleared E. coli that have been killed is calculated by the following equation: {(51Cr activity in an organ/51Cr activity injected) 0 (125I activity in the organ/125I activity injected) divided by (51 Cr activity in the organ/51Cr activity injected)} 1 100. Cytokine production in vivo To further assess the effect of leptin-dependent signaling on proinflammatory cytokine production, intact obese ob/ob mice and fa/fa rats and their respective lean (?/fa rats and ?/ ob mice) littermates (nÅ20/group) were injected i.p. with LPS. Each obese mouse (weighing approximately 60 g each) and lean mouse (which weighedõ25 g each) received the same dose (10 mg) of LPS. To assure that differences in the cytokine response to LPS were not due to differences in body weight or to something other than leptin deficiency, additional experiments were performed with leptin-resistant fa/fa rats (which weighed about 600 g each) and their leptin-responsive ?/fa littermates (weighing approximately 250 g each), both of which received a dose of LPS that had been normalized for differences in body weight (0.5 mg/g body weight). In both experiments, animals were killed at several different time points (0, 0.5, 1, or 6 h) after LPS injection, and serum concentrations of TNF-a and IL-6 were measured as described above. Mean data from four animals per group/ time point were compared by ANOVA.

RESULTS Phagocytic function in ob/ob and db/db macrophages Macrophage phagocytic functions were evaluated in vitro using macrophages isolated from the peritoneal cavity and bone marrow of ob/ob mice, db/db mice, and their lean (?/ob, ?/db) littermates. Macrophages were incubated with C. parasilopsis to determine the proportion that were phagocytically active (Fig. 1A). As assessed by this protocol, peritoneal macrophages and bone marrow-derived macrophages exhibited similar phagocytic activities. As shown in Fig. 1B, 20– 25% of the macrophages that had been isolated from ?/ob and ?/db mice were phagocytically active compared to fewer than 15% of the macrophages from ob/ob and db/db mice (Põ0.03 for each lean group vs. its respective group of obese littermates). Treatment with human recombinant leptin at a dose of 500 ng/ml for 2 h increased the number of phagocytically active cells, so that 40–50% of the macrophages isolated from ?/ob, ob/ob, and ?/db mice contained Candida (Põ0.001 (0) vs. (/) leptin in both groups). As noted by others (24), we found that treatment with leptin also increased the phagocytic activity of macrophages harvested from normal C57Bl/6 mice (from 50% to almost 80%). Consistent with a direct, leptin-mediated effect on macrophage phagocytosis, leptin failed to improve phagocytosis in macrophages from db/db mice (which are known to

have defective leptin receptors). Thus, leptin modulated the proportion of macrophages that phagocytosed Candida in vitro via a mechanism that requires its interaction with functional leptin receptors. Phagocytically active macrophages from ?/ob, ?/ db, and ob/ob mice ingested, on average, a similar number of Candida/cell and exhibited similar baseline candicidal activities. Compared to candicidal activities of 30% in macrophages isolated from each of these groups, db/db phagocytic cells killed fewer than 10% of intracellular Candida (Põ0.04). Thus, although the number of Candida inside each phagocytically active db/db macrophage was similar to that observed in ob/ob, ?/ob, and ?/db macrophages, db/db macrophages appeared to be less capable of killing ingested Candida. Treatment with recombinant leptin did not consistently increase the phagocytic index or the candicidal activity in any group. The effect of leptin on cytokine release by macrophages Given evidence that leptin influences macrophage phagocytic function, we evaluated the effect of leptin on the expression of some monokines. Thioglycollate-elicited peritoneal macrophages from healthy C57Bl/6 mice released no detectable TNF-a, IL-6, IL-12, or IL-10 either spontaneously or after stimulation with leptin alone (data not shown). Leptin prestimulation led to a significant increase in the LPS-stimulated production of TNF-a, IL-6, and IL-12 from such macrophages (Fig. 2). The production of IL-1b and IL-10 were unaltered by leptin pretreatment (data not shown). The effect of leptin on cytokine gene expression in macrophages Northern blot analysis was performed to determine whether leptin could regulate cytokine gene expression. Thioglycollate-elicited peritoneal macrophages from ob/ob, ?/ob mice, and normal C57BL-6 mice were treated overnight with sterile saline or leptin (500 ng/ml) and harvested 90 min after treatment with LPS (1 mg/ml). Total RNA was isolated, separated by electrophoresis on denaturing gels, transferred to nylon membranes, and hybridized with a cDNA probe for murine TNF-a (Fig. 3). Under these conditions, neither the macrophages from obese mice nor those from lean mice expressed TNF-a until stimulated with LPS. Leptin alone failed to induce TNF-a mRNA expression. Expression of TNF-a mRNA was not consistently different in cells cultured with both LPS and leptin from that of cells cultured in the presence of LPS alone.

IMMUNE RESPONSE REGULATION BY LEPTIN


59

Figure 1. Phagocytic function of macrophages from lean (?/ob, ?/db) and obese (ob/ob, db/db) mice. Thioglycollate-elicited peritoneal macrophages were isolated, plated on plastic dishes overnight in medium with or without leptin (500 ng/ml), and then incubated for 2 h with Candida parasilopsis. Staining with acridine orange/crystal violet was performed to reveal live (green) and dead (orange) intracellular microorganisms so that differences in the proportion of phagocytically active cells (phagocytic activity), the number of intracellular Candida/phagocytically active cell (phagocytic index), and the ability of individual phagocytically active cells to kill ingested Candida (candicidal activity) could be evaluated (A). B) Phagocytic activity of duplicate cultures differed in the absence (open bars) and presence (hatched bars) of leptin (500 ng/ml). Results shown illustrate the means { standard errors of three separate experiments, each of which evaluated duplicate cultures of macrophages pooled from the three mice in each of the different obese and lean groups. *P õ 0.03 for each obese group vs. its respective lean control group. **P õ 0.001 for each group when compared to the same group treated with leptin.

Figure 2. Cytokine production by cultured peritoneal macrophages. Thioglycollate-elicited macrophages were isolated from wildtype C57Bl/6 mice. To evaluate production of proinflammatory cytokines, macrophages were cultured in the presence of medium alone, LPS (1 mg/ml), leptin (500 ng/ml), or LPS / leptin. After 24 h, medium was harvested and cytokine protein concentrations were measured in the medium by ELISA (for TNF-a and IL-6) or RIA (for IL-12 p40 and p70). Means and standard errors of quadruplicate assays from four different experiments are shown. Results were evaluated by the paired Student’s t test. *P õ 0.042, **P õ 0.007, / P õ 0.036, // P õ 0.001. None of these cytokines were detected in the medium of peritoneal macrophages cultured in the presence of vehicle or leptin alone (data not shown). 60

Vol. 12

January 1998

The FASEB Journal


LOFFREDA ET AL.

Figure 3. TNF-a mRNA expression in peritoneal macrophages. Thioglycollate-elicited peritoneal macrophages of ob/ ob, ?/ob and normal C57BL-6 mice were cultured for 16 h in the absence (0) or presence (/) of leptin (Lp) (final concentration 500 ng/ml). LPS (1 mg/ml) was added to half of the Lp-negative cultures (/LPS) and half of the Lp-positive cultures (/LPS/Lp) 90 min before harvest. Total RNA was isolated and TNF-a mRNA expression was evaluated by Northern blot analysis. To ensure that similar amounts (20 mg) of RNA were in each lane, the constitutively expressed genes, 18S RNA and GAPDH, were also evaluated on the same blots. Northern blot experiments were repeated twice using input RNA from different groups of mice. Macrophages from obese and lean mice always exhibited similar responses to LPS and Lp. A representative Northern blot demonstrating cytokine gene expression in ob/ob macrophages.

Phagocytic function of ob/ob and ?/ob mice in vivo We evaluated whether or not the changes in macrophage function in vitro were associated with changes in macrophage function in vivo. Clearance of intravenously administered 51Cr, 125I-labeled E. coli provides a good estimation of the phagocytic function of the reticuloendothelial (RE) system. Previous studies in normal mice and rats indicate that over 90% of intravenously injected bacteria are cleared from the circulation within 6 min by the liver, lung, and spleen, and that these bacteria then remain trapped in these RE organs for at least 6 h (21, 22). The liver clears approximately 80% of the injected bacteria, and quantitative culture of liver tissues has indicated that phagocytic killing can be tracked by measuring decline in liver 125I activity. These studies found that 80% of the cleared bacteria are phagocytosed and killed within 90 min in normal mice (22). This experimental protocol was used to evaluate RE system function in ob/ob and ?/ob mice. Results are summarized in Table 1. The net hepatic clearance of intravenously injected E. coli was similar in obese mice and their lean littermates 90 min postinjection. At this time point, almost 80% of the injected bacteria had been trapped within the liver in each group of mice. However, the livers of obese mice weighed about 2.5-fold more than those of their lean littermates and contained about twice as much total RNA (13.7{3 in ob/ob group vs. 7.4{0.4 mg RNA/ liver in ?/ob group). Previous Northern blot analyses indicated that expression of Pu-1/mg hepatic RNA is similar in obese and lean animals (10). Since Pu-1 is

specifically expressed in monocytes and macrophages, this suggests that the ob/ob mice have a larger number of hepatic macrophages than ?/ob mice. To take into account the obesity-associated differences in the total number of hepatic macrophages, 51Cr clearance results were normalized per gram of liver weight, and it became apparent that the hepatic clearance of E. coli bacteremia was significantly less efficient in the obese group than in the lean controls. More important, obese mice killed a significantly smaller fraction of the bacteria that had been cleared by the liver. Indeed, differences in this hepatic killing efficiency between obese and lean mice were apparent even without normalizing the data to control for intergroup differences in liver weight. Taken together, these observations suggest that RE system phagocytic function is impaired in ob/ob mice. LPS-induced cytokine production in rats with defective leptin receptors To determine whether other aspects of macrophage function may be abnormal in genetically obese animals, the in vivo cytokine response to LPS was compared in fa/fa rats, ob/ob mice, and their lean (?/fa and ?/ob) littermates. Obese and lean animals were injected i.p. with LPS and killed at various time points to obtain sera for cytokine measurements. As shown in Fig. 4, LPS induction of TNF-a and IL-6 was significantly attenuated in the obese animals compared to their lean littermates. Obesity was associated with a decreased production of cytokines regardless of whether or not the LPS dose was normalized for differences in body weight. Thus, fa/fa rats, which express abnormally few leptin receptors on plasma membranes (25), and ob/ob mice, which lack leptin (3), both exhibit a suboptimal cytokine response when challenged with LPS. Results of these in vivo studies complement our in vitro data, which indicate that leptin-dependent signals enhance macrophage synthesis of proinflammatory cytokines in response to LPS. DISCUSSION Obese patients have an increased incidence of infection, insulin resistance, hyperlipidemia, and atherosclerotic cardiovascular disease (7–9). Genetically obese rodent strains that fail to express either the adipocyte hormone leptin or its functional transmembrane receptor have been reported to have an increased incidence of similar problems (1–3, 16, 17, 25, 26). Although both short and long (transmembrane) forms of the leptin receptor are widely distributed in the body (24), the entire spectrum of physiological roles of leptin remains poorly under-



61

TABLE 1. In vivo phagocytic activity in ob/ob and ?/ob micea

Mouse weight (g) Liver weight (g) Hepatic clearance (E. coli) % injected bacteria % injected bacteria/g liver Hepatic killing efficiency (E. coli) % cleared bacteria killed % cleared bacteria killed/g liver Net hepatic killing (E. coli) % injected bacteria killed

?/ob

ob/ob

P value

30 { 2 1.3 { 0.1

52 { 4 3.4 { 0.3

õ0.001 õ0.001

77 58

{2 {2

79 23

{2 {2

n.s. õ0.01

75 57

{4 {2

65 19

{3 {2

õ0.02 õ0.001

58

{3

51

{2

õ0.05

a Mice were killed 90 min after i.v. injection of radiolabeled E. coli. The hepatic clearance of the bacteria (a measure of trapping / phagocytosis) and the phagocytic killing of cleared bacteria were determined as detailed in the Methods and Results. Results shown are means { SE of six mice/group. These findings were confirmed in a second experiment.

stood. Because leptin was initially identified as an important negative regulator of appetite (1), it is not surprising that leptin interactions with its receptors in the hypothalamus have been characterized most extensively. More recently, leptin has been implicated in female sexual maturation (27, 28), but no

evidence has yet been uncovered that explains why defective leptin-dependent signal transduction is so frequently associated with diabetes or heart disease. Because diabetes itself is an important cause of morbidity, and may also contribute to obesity-associated infections and vascular disease, considerable

Figure 4. Induction of cytokines in vivo. Obese ob/ob mice and fa/fa rats and their respective lean (?/ob and ?/fa) littermates were injected i.p. with LPS. In the mouse studies (A, B), both obese and lean mice received the same dose (10 mg/mouse) LPS. In the rat studies (C, D), the LPS dose was normalized for differences in body weight (0.5 mg/g body weight). Animals (nÅ5/ group) were killed at various time points (0, 0.5, 1, or 6 h) after LPS treatment. Serum concentrations of TNF-a and IL-6 were measured by commercially available ELISA kits, using either murine or rat recombinant cytokines as standards. Data are displayed as the mean { standard error of the cytokine concentration at each time point. These results were evaluated by ANOVA. *P õ 0.05 vs. the control group, **P õ 0.001 vs. the control group in a second experiment. 62

Vol. 12

January 1998

The FASEB Journal


LOFFREDA ET AL.

effort has been directed toward clarifying the pathogenesis of insulin resistance in obesity. Several groups have found increased expression of the proinflammatory cytokine TNF-a in the white adipose tissues of obese mice, rats (4), and humans (5). TNF-a has also been shown to inhibit insulin-dependent phosphorylation of IRS-1 (11). Thus, it has been suggested that obesity-related increases in TNF-a production by adipose tissue underlie the insulin-resistance that accompanies obesity (6, 11). However, the mechanisms driving the increased adipose expression of TNF-a in obesity remain obscure. In healthy, lean individuals, adipose tissue has not been identified as a major source of TNF-a. Tissue macrophages and circulating monocytes are the major producers of this cytokine; these cells are capable of synthesizing large amounts of TNF-a when exposed to gut-derived endotoxin or bacterial lipopolysaccharide (14, 29). Consistent with these observations, conditions (such as sepsis) that lead to endotoxemia are accompanied by increased circulating levels of TNF-a (12, 13). Moreover, insulin resistance and anorexia are also commonly associated with endotoxic states, and there is recent evidence that treatment with lipopolysaccharide increases adipose tissue production of leptin in rodents (30). However, whereas endotoxin-inducible cytokines are known to function as important modulators of the immune response to invading microorganisms (29, 31, 32), the possibility that leptin itself may have immunoregulatory actions has not heretofore been formally investigated. Some published information does support this concept. For example, leptin receptors have recently been identified on macrophages, and recombinant leptin has been found to induce production of M-CSF by cultured macrophages (24). Moreover, the long form of the leptin receptor resembles the gp130 family of cytokine receptors (33), and leptin is known to activate Janus kinases and certain signal transducers and activators of transcription, downstream components in the signaling pathways of several proinflammatory cytokines (34). By finding that leptin-deficient ob/ob mice cannot clear and kill circulating E. coli as efficiently as normal mice, our studies suggest that these genetically obese mice have an impaired immune system. The possibility that leptin-dependent signals regulate immune function is also supported by our observation that Zucker fa/fa rats, an obese and diabetic strain with deficient leptin receptors (25, 26), have suboptimal production of proinflammatory cytokines when challenged with LPS. Two other recent reports also suggest that fa/fa rats may have immune dysfunction. Plotkin and co-workers (35) noted that logarithmically fewer Candida albicans were required to establish tissue colonization in obese Zucker rats than in their lean littermates, and that the organs of obese rats had a 10-fold greater yeast/gram organ burden than did

lean rats. We found decreased uptake of intravenously administered fluorescent microspheres by hepatic macrophages in obese Zucker rats, documented abnormalities in their macrophage gene expression, and noted that these animals were more susceptible to the lethal and hepatotoxic actions of endotoxin (10). Thus, results obtained in different strains of genetically obese rodents with defects at different sites in the leptin-dependent signaling pathway support the hypothesis that leptin regulates the immune system. However, since both ob/ob mice and fa/fa rats are diabetic, these in vivo studies could not exclude the possibility that leptin deficiency may cause immune dysfunction indirectly by promoting obesity-related diabetes or some other systemic dysfunction. Our analysis of macrophage phagocytic functions in vitro addresses this issue directly. Results indicate that macrophages from ob/ob mice (which lack leptin) and db/db mice (which have dysfunctional leptin receptors) have an impaired ability to phagocytose and kill Candida. Recombinant leptin restores phagocytic function in cells from ob/ob mice, but not from db/db mice, indicating that leptin appears to regulate macrophage phagocytosis via mechanisms that require direct interaction with its receptor. The evidence that leptin can also increase phagocytic activity in macrophages isolated from lean, nondiabetic mice complements similar results reported by Gainsford et al. (24) and suggests that the phagocytic defect in ob/ob and db/db mice is a direct result of leptin deficiency rather than secondary to obesity-related diabetes. Immunodeficiency has also been suggested as a complication of obesity in normoglycemic humans based on evidence that mitogen-induced blastogenesis is decreased in peripheral blood lymphocytes isolated from obese, nondiabetic subjects (36). It is tempting to speculate that decreased macrophage phagocytic activity in obesity may result in increased exposure of peripheral adipocytes to gut-derived bacterial products that are normally cleared from the portal circulation by phagocytic cells that reside in the liver. The escape of bacterial endotoxin into the systemic circulation could trigger ectopic production of endotoxin-inducible proinflammatory cytokines in adipose tissue and explain previous observations that obesity increases TNF-a expression in peripheral fat (11). Variability in the basal exposure of hepatic macrophages to gut-derived products could also account for the relatively wide range of ‘‘baseline’’ serum TNF-a concentrations that have been reported in seemingly healthy obese rodents and humans. Because increased adipose tissue expression of TNF-a has been implicated in the pathogenesis of obesity-related diabetes, hyperlipidemia, and vascular disease, these results provide the first potentially direct pathogenetic link between leptin deficiency and the major complications of obesity.



63

These studies also indicate that leptin plays a role in regulating proinflammatory cytokine production in normal, nonobese animals. Increased macrophage production of TNF-a, IL-6, and IL-12 is a fundamental aspect of the primitive, innate immune response to invading microorganisms (37). Bacterial lipopolysaccharide is a potent inducer of macrophage cytokine production (14, 29, 31, 32). The present results indicate that leptin enhances the synthesis of these proinflammatory cytokines by cultured macrophages when these cells are treated with LPS. Comparison of the cytokine response to LPS injection in rodents with and without intact leptin-dependent signaling confirms the significance of these in vitro results by demonstrating that leptin is necessary for animals to mount an optimal cytokine response when challenged with endotoxin. In summary, these studies in intact genetically obese rodents and with macrophages isolated from these animals indicate that leptin activates macrophages and induces their expression of proinflammatory cytokines. These findings indicate an important novel function for this appetite-regulating hormone: up-regulation of inflammatory immune responses.

9.

10.

11.

12. 13. 14. 15. 16.

17.

18.

Thanks to M. Wysocka (Wistar Institute, Philadelphia, Pa.) for performing the IL-12 RIAs and to P. Cuomo (Johns Hopkins University, Baltimore, Md.) for his excellent technical assistance.

19.

REFERENCES

20.

1.

2.

3.

4.

5.

6.

7.

8.

64

Campfield, L. A., Smith, F. J., Buisez, Y., Devos, R., and Burn, P. (1995) Recombinant mouse OB protein: Evidence for a peripheral signal linking adiposity and central neural networks. Science 269, 546–549 Halaas, J. L., Gajiwala, K. S., Maffei, M., Cohen, S. I., Chait, B. T., Rabinowitz, D., Lallone, R. L., Burley, S. K., and Friedman, J. M. (1995) Weight-reducing effects of the plasma protein encoded by the obese gene. Science 269, 544–546 Pelleymounter, M. A., Cullen, M. J., Baker, M. B., Hecht, R., Winters, D., Boone, T., and Collins, F. (1995) Effects of the obese gene product on body weight regulation in ob/ob mice. Science 269, 540–543 Yamakawa, T., Tanaka, S., Yamakawa, Y., Kiuchi, Y., Isoda, F., Kawamoto, S., Okuda, K., and Sekihara, H. (1995) Augmented production of tumor necrosis factor-alpha in obese mice. Immunol. Immunopathol. 75, 51–56 Kern, P. A., Saghizadeh, M., Ong, J. M., Bosch, R. J., Deem, R., and Simsolo, R. B. (1995) The expression of tumor necrosis factor in human adipose tissue. Regulation by obesity, weight loss, and relationship to lipoprotein lipase. J. Clin. Invest. 95, 2111– 2119 Hotamisligil, G. S., Peraldi, P., Budavari, A., Ellis, R., White, M. F., and Spiegelman, B. M. (1996) Uncoupling of obesity from insulin resistance through a targeted mutation in aP2, the adipocyte fatty acid binding protein. Science 274, 1377–1379 Jouislahti, P., Tuomilehto, J., Vartiainen, E., Pekkanen, J., and Puska, P. (1996) Body weight, cardiovascular risk factors, and coronary mortality. 15-year follow-up of middle-aged men and women in eastern Finland. Circulation 93, 1372–1397 Gorsky, R. D., Pamuk, E., Williamson, D. F., Shaffer, P. A., and Koplan, J. P. (1996) The 25-year health care costs of women who remain overweight after 40 years of age. Am. J. Prev. Med. 12, 388–394

Vol. 12

January 1998

21.

22. 23. 24.

25. 26.

27. 28. 29.

Seidell, J. C., Verschuren, W. M., van Leer, E. M., and Kromhout, D. (1996) Overweight, underweight, and mortality. A prospective study of 48,287 men and women. Arch. Int. Med. 156, 958– 963 Yang, S. Q., Lin, H. Z., Lane, M. D., Clemens, M., and Diehl, A. M. (1997) Obesity increases sensitivity to endotoxin liver injury: Implications for the pathogenesis of steatohepatitis. Proc. Natl. Acad. Sci. USA 94, 2557–2562 Hotamisligil, G. S., Peraldi, P., Budavari, A., Ellis, R., White, M. F., and Spiegelman, B. M. (1996) IRS-1-mediated inhibition of insulin receptor tyrosine kinase activity in TNF-alpha- and obesity-induced insulin resistance. Science 271, 665–668 Tracey, K. J., and Cerami, A. (1993) Tumor necrosis factor, other cytokines and disease. Annu. Rev. Cell Biol. 9, 317–343 Vassalli, P. (1992) The pathophysiology of tumor necrosis factor. Annu. Rev. Immunol. 10, 411–452 Ulevitch, R. J., and Tobias, P. S. (1995) Receptor-dependent mechanisms of cell stimulation by bacterial endotoxin. Annu. Rev. Immunol. 13, 437–457 Nathan, C. F., Murray, H. W., and Cohen, Z. A. (1980) Current concepts. The macrophage as an effector cell. N. Engl. J. Med. 303, 622–626 Lee, G. H., Proenca, R., Montez, J. M., Carroll, K. M., Darvishzadeh, J. G., Lee, J. I., and Friedman, J. M. (1996) Abnormal splicing of the leptin receptor in diabetic mice. Nature (London) 379, 632–635 Chen, H., Charlat, O., Tartaglia, L. A., Woolf, E. A., Weng, X., Ellis, S. J., Lakey, N. D., Culpepper, J., Moore, K. J., Breitbart, R. E., Duyk, G. M., Tepper, R. I., and Morgenstern, J. P. (1996) Evidence that the diabetes gene encodes the leptin receptor: identification of a mutation in the leptin receptor gene in db/ db mice. Cell 84, 491–495 Caput, D., Beutler, B., Hartog, K., Thayer, R., Brown-Shimer, S., and Cerami, A. (1986) Identification of a common nucleotide sequence in the 3*-untranslated region of mRNA molecules specifying inflammatory mediators. Proc. Natl. Acad. Sci. USA 83, 1670–1674 Wysocka, M., Kubin, M., Vieira, L. Q., Ozmen, L., Garotta, G., Scott, P., and Trinchieri, G. (1995) Interleukin-12 is required for interferon-gamma production and lethality in lipopolysaccharide-induced shock in mice. Eur. J. Immunol. 25, 672–676 Noble, P. W., Lake, F. R., Henson, P. M., and Riches, D. W. H. (1993) Hyaluronate activation of CD44 induces insulin-like growth factor-1 expression by a TNF-alpha dependent mechanism in murine macrophages. J. Clin. Invest. 91, 2368–2377 Takao, S., Smith, E. H., Wang, D., Chan, C. K., Bulkley, G. B., and Klein, A. S. (1996) Role of reactive oxygen metabolites in murine peritoneal macrophage phagocytosis and phagocytic killing. Am. J. Physiol. 271, C1278–C1285 Klein, A., Zhadkewich, Margolick, J., Winkelstin, J., and Bulkley, G. (1994) Quantitative discrimination of hepatic reticuloendothelial clearance and phagocytic killing. J. Leukoc. Biol. 55, 248–252 Chomczynski, P., and Sacchi, N. (1987) Single step method for RNA isolation by acid guanidine thiocyanate-phenol-chloroform extraction. Anal. Biochem. 162, 156–161 Gainsford, T., Willson, T. A., Metcalf, D., Handman, E., McFarlane, C., Ng, A., Nicola, N. A., Alexander, W. S., and Hilton, D. J. (1996) Leptin can induce proliferation, differentiation, and functional activation of hemopoietic cells. Proc. Natl. Acad. Sci. USA 93, 14564–14568 Phillips, M. S., Liu, Q., Hammond, H. A., Dugan, V., Hey, P. J., Caskey, C. J ., and Hess, J. F. (1996) Leptin receptor missense mutation in the fatty Zucker rat. Nat. Genet. 13, 18–19 Chua, S. C., Jr., Chung, W. K., Wu-Peng, X. S., Zhang, Y., Liu, S. M., Tartaglia, L., and Leibel, R. L. (1996) Phenotypes of mouse diabetes and rat fatty due to mutations in the OB (leptin) receptor. Science 271, 994–996 Ahima, R. S., Dushay, J., Flier, S. N., Prabakaran, D., and Flier, J. S. (1997) Leptin accelerates the onset of puberty in normal female mice. J. Clin. Invest. 99, 391–395 Chehab, F. F., Mounzih, K., Lu, R., and Lim, M. E. (1997) Early onset of reproductive function in normal female mice treated with leptin. Science 275, 88–90 Ware, C. F., Van Arsdale, T. L., Crowe, P. D., and Browning, J. L. (1995) The ligands and receptors of the lymphotoxin system. Curr. Top. Immunol. 198, 175–218

The FASEB Journal


LOFFREDA ET AL.

30.

Sarraf, P., Frederich, R. C., Turner, E. M., Ma, G., Jaskowiak, N. T., Rivet, D. J., 3rd, Flier, J. S., Lowell, B. B., Fraker, D. L., and Alexander, H. R. (1997) Multiple cytokines and acute inflammation raise mouse leptin levels: potential role in inflammatory anorexia. J. Exp. Med. 185, 171–175 31. Trienchieri, G. (1995) Interleukin-12: A proinflammatory cytokine with immunoregulatory functions that bridge innate resistance and antigen-specific adaptive immunity. Annu. Rev. Immunol. 13, 251–276 32. Dalton, D. K., Pitts-Meek, S., Keshav, S., Figari, I. S., Bradley, A., and Stewart, T. A. (1993) Multiple defects of immune cell fucntion in mice with disrupted interferon-gamma genes. Science 259, 1739–1742 33. Tartaglia, L. A., Dembski, M., Weng, X., Deng, N., Culpepper, J., Devos, R., Richards, G. J., Campfield, L. A., Clark, F. T., and Deeds, J. (1995) Identification and expression cloning of a leptin receptor, OB-R. Cell 83, 1263–1271

34. Ghilardi, N., Ziegler, S., Wiestner, A., Stoffel, R., Heim, M. H., and Skoda, R. C. (1996) Defective STAT signaling by the leptin receptor in diabetic mice. Proc. Natl. Acad. Sci. USA 93, 6231–6235 35. Plotkin, B. J., Paulson, D., Chelich, A., Jurak, D., Cole, J., Kasimos, J., Burdick, J. R., and Casteel, N. (1996) Immune responsiveness in a rat model for type II diabetes (Zucker rat, fa/fa): susceptibility to Candida albicans infection and leucocyte function. J. Med. Microbiol. 44, 277–283 36. Tanaka, S., Inoue, S., Isoda, F., Waseda, M., Ishihara, M., Yamakawa, T., Sugiyama, A., Takamura, Y., and Okuda, K. (1993) Impaired immunity in obesity: suppressed but reversible lymphocyte responsiveness. Int. J. Obes. Relat. Metab. Disord. 11, 631– 636 37. Solbach, W. H., and Rollinghoff, M. (1991) Lymphocytes play the music but macrophages call the tune. Immunol. Today 12, 4–6



Received for publication May 29, 1997. Accepted for publication October 1, 1997.

65