INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

LEPTIN, THE PRODUCT of the obese gene (ob), is an adipose tissue-derived factor originally described for its ability to cause body weight loss and food intake reduction in genetically obese (ob/ob) and normal mice (18, 31, 45). A high-affinity receptor for leptin (OB-R) has been described (37). OB-R is structurally related to the gp130 signal-transducing molecule associated with the receptors of the cytokines of the interleukin (IL)-6 family (20, 37), which include IL-6, IL-11, leukemia inhibitory factor (LIF), oncostatin M (OM), ciliary neurotrophic factor (CNTF), and cardiotrophin-1 (CT-1) (23). OB-R is also structurally related to the -subunit of the receptor for IF (LIF-R; Ref. 37), which has also been reported to be a subunit of the receptors for OM, CNTF, and CT-1 (11, 14, 32). OB-R exists in five alternatively spliced isoforms, OB-Ra-e (9, 16, 20, 21, 24, 37). All OB-R isoforms share the same extracellular domain and, except for OB-Re, which is soluble, a single transmembrane sequence (9, 16, 20, 21, 24, 37). OB-Rb has a long cytoplasmic tail that contains two JAK binding boxes and a STAT binding consensus sequence, activates JAK and STAT proteins, and is responsible for leptin signaling (5, 9, 16, 20, 21, 24, 37). The other isoforms have no (OB-Re) or short (OB-Ra, c, and d) cytoplasmic tails that still contain a JAK binding box, and their function remains unclear (9, 16, 20, 21, 24, 37). OB-Ra, however, has recently been reported to be capable of mediating leptin signaling, although less effectively than OB-Rb (30). OB-R signaling capabilities, therefore, by involving JAK and STAT proteins, appear to be similar to those of gp130 and LIF-R, which are known to activate the JAK-STAT pathway (23). Interestingly, it has been noted that the leptin three-dimensional structure resembles one of the cytokines that activate the JAK-STAT pathway (26). OB-R isoforms, including the signal transducing OB-Ra and OB-Rb, are found not only in the brain, especially in regions such as the choroid plexus and the hypothalamus, where they may mediate the body weight control activity expressed by leptin, but also in several other organs (9, 13, 16, 25, 37), where they may mediate possible functions of leptin in addition to body weight control.

gp130-user cytokines share a number of in vivo activities. We have reported that all gp130-user cytokines are able to induce the acute-phase protein serum amyloid A (SAA) and potentiate the elevation of serum levels of corticosterone and IL-6 induced by low doses of IL-1 (6). Furthermore, IL-6, IL-11, LIF, CNTF, and CT-1 have been found to inhibit tumor necrosis factor (TNF) production on endotoxin or super antigen administration (1, 3, 4, 7, 8), and, in intriguing similarity with leptin, IL-6, IL- 11, LIF, and CNTF have been observed to be able to cause body weight loss and/or food intake reduction (33, 38). To reinforce the concept of a potential functional relationship between gp130-user cytokines and leptin, two findings have recently been published: one, that leptin mRNA is induced by lipopolysaccharide (LPS) in the adipose tissue (17), suggesting that leptin takes part in the cytokine cascade triggered by LPS, of which IL-6 family members are important components (23, 27, 39, 40); and the other, that leptin stimulates hematopoiesis (10, 13, 29), in resemblance to IL-6, IL-11, LIF, and OM, which have been attributed hematopoietic properties, especially the one of increasing platelet count (19, 22, 28, 38, 42).

The aim of this study is to investigate the possibility that leptin is functionally related to the IL-6 family and shares in vivo activities with gp130-user cytokines. We tested leptin in mice for the ability to induce SAA, to potentiate the induction by IL-1 of serum corticosterone and IL-6, to inhibit the induction by LPS of serum TNF, to cause body weight loss, and to change peripheral blood cell counts. LIF was used as the reference cytokine, because its receptor involves both gp130 and LIF-R (15, 23).

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Reagents and mice. Recombinant murine leptin [<3.8 endotoxin units (EU)/mg] and recombinant human (rh) LIF (<7.6 EU/mg) were from Amgen (Thousand Oaks, CA). rhIL-1 was a kind gift from Sclavo, Siena, Italy. LPS (phenol-extracted preparation from Escherichia coli O55:B5) was purchased from Sigma (St. Louis, MO). Male CD-1 mice of ~25 g body weight (Charles River Laboratories, Como, Italy) were used in the experiments of SAA induction, IL-1 potentiation, and inhibition of LPS-induced serum TNF. Female BALB/c mice of ~19 g body wt (Charles River Laboratories, Wilmington, MA) were used in the experiments of platelet count increase and body weight loss. Mice were housed in rooms maintained at constant temperature and humidity and subjected to a 12:12-h light-dark cycle. Mice received standard laboratory diet and water ad libitum. Procedures involving animals and their care were conducted in conformity with institutional guidelines that are in compliance with national and international laws and policies (Italian Legislative Decree 116/92, GU suppl. 40, 1992; EEC Council Directive 86/609, OJL 358, 1987; US National Research Council, Guide for the Care and Use of Laboratory Animals, 1996).

SAA induction, potentiation of corticosterone, IL-6 induction by IL-1, and inhibition of LPS-induced TNF. Leptin was given intravenously at a dose of 0.5 or 5 mg/kg alone or in association with IL-1 (100 ng/mouse) or LPS (100 ng/mouse). LIF was similarly given at a dose of 0.5 mg/kg. Control mice received saline. Blood was taken from the retroorbital plexus 8 h after the administration of leptin or LIF for SAA determination, 2 h after for corticosterone and IL-6, and 1.5 h after for TNF. Experiments were conducted on groups of five mice.

SAA was measured in serum by ELISA with the use of a commercially available kit (Hemagen Diagnostics, Waltham, MA); results were expressed in micrograms per milliliter. Corticosterone was measured by radioimmunoassay as previously described (12) with the use of an antiserum purchased from Sigma Chemical and [³H]corticosterone purchased from Amersham (Little Chalfont, UK); results were expressed in nanograms per milliliter. IL-6 was measured by bioassay as hybridoma growth factor, as previously described (36), with the use of 7TD1 cells and rhIL-6 (a kind gift from Dr. J. Van Snick, Bruxelles, Belgium) as a standard; results were expressed in units per milliliter. TNF was also measured by bioassay as cytotoxic factor as previously described (2), with the use of L929 cells and rhTNF (a kind gift from Dr. W. Fiers, Ghent, Belgium) as a standard; results were expressed in nanograms per milliliter.

We also investigated in vitro if leptin has hybridoma growth factor activity on 7TD1 cells, which IL-6 is able to support (41). To this end leptin was added to the 7TD1 cells in culture under the same experimental conditions used to test IL-6 hybridoma growth factor activity in the bioassay referenced above.

Change in body weight and peripheral blood cell counts. For 7 consecutive days mice received a daily intraperitoneal injection of 0.5 or 5 mg/kg murine leptin or 0.5 mg/kg LIF or saline at 9:00 AM. This time of treatment was used because it has been reported that LIF-induced body weight loss and platelet increase peak after seven daily injections (38). Body weight was recorded at 9:00 AM before treatment at the beginning of the study and after 2, 5, and 7 days. Twenty-four hours after the seventh injection, mice were bled by cardiopuncture to count blood cells. In one experiment, to assist in the analysis of a possible effect of leptin on the peripheral blood cell counts, mice were treated daily with 5 mg/kg of leptin or saline for up to 7 consecutive days and bled 24 h after the last injection as above. We also investigated if leptin has the ability to mitigate the loss of peripheral blood cells induced by the intraperitoneal administration of 150 mg/kg of 5-fluorouracil (5-FU) (34). Mice were treated daily with 5 mg/kg of leptin or saline starting 1 h after the injection of 5-FU and for the 6 following days and bled 24 h after the seventh injection as above. Blood cells were counted with an H1E cell counter (Technicon, Tarrytown, NY).

Statistical analysis. Results are expressed as means ± SD. Differences between groups were analyzed by the Student's t-test. Because body weight was repeatedly measured in each individual, differences in body weight change within and between groups were tested by ANOVA for repeated measures. When peripheral blood cell counts were serially measured in different groups, differences in blood cell counts within and between groups were tested by two-way ANOVA.

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

SAA induction, potentiation of corticosterone and IL-6 induction by IL-1, and inhibition of LPS-induced TNF. At variance with LIF, leptin did not induce circulating SAA at either the 0.5 or the 5 mg/kg dose. SAA was not detectable, i.e., it was <1 µg/ml, given the sensitivity limit of the assay, in the circulation after the injections of leptin at either dose or after the injection of saline, whereas it reached the concentration of 77 ± 21 µg/ml after the injection of LIF (n = 5).

Also at variance with LIF at the dose of 0.5 mg/kg, leptin did not potentiate the induction by a low dose of IL-1 of either serum corticosterone or IL-6 (Figs. 1 and 2). However, at the dose of 5 mg/kg, leptin showed potentiation effect, resulting in higher serum levels of corticosterone and IL-6 than saline but in much lower levels than LIF (Figs. 1 and 2). The potentiation effect of 5 mg/kg of leptin was only 25% of 0.5 mg/kg LIF for corticosterone and 10% for IL-6. Whereas LIF also had some ability to increase the circulating levels of corticosterone and to induce detectable serum levels of IL-6 when it was injected alone, leptin did not show this ability at either dose (Figs. 1 and 2). At variance with LIF, leptin did not inhibit the induction by LPS of serum TNF at either the 0.5 or the 5 mg/kg dose (Fig. 3).

View larger version (16K): [in this window] [in a new window]   Fig. 1.   Serum corticosterone levels in mice that received leptin (0.5 or 5 mg/kg) or leukemia inhibitory factor (LIF; 0.5 mg/kg) alone or in association with interleukin (IL)-1 (100 ng/mouse). There were 5 mice/group.

View larger version (13K): [in this window] [in a new window]   Fig. 2.   Serum IL-6 levels in mice that received leptin (0.5 or 5 mg/kg) or LIF (0.5 mg/kg) alone or in association with IL-1 (100 ng/mouse). ND, not detectable. Sensitivity limit of the assay is 50 U/ml. There were 5 mice/group.

View larger version (13K): [in this window] [in a new window]   Fig. 3.   Serum tumor necrosis factor (TNF) levels in mice that received leptin (0.5 or 5 mg/kg) or LIF (0.5 mg/kg) alone or in association with LPS (100 ng/mouse). There were 5 mice/group.

We also investigated in vitro whether leptin shared with IL-6 hybridoma growth factor activity and found that leptin does not support the growth of the 7TD1 cells, even when it is used at a concentration as high as 60 µg/ml (data not shown).

Change in body weight and in peripheral blood cell counts. Leptin caused body weight loss at either dose and so did LIF, whereas saline-treated mice showed an increase in body weight (Fig. 4). Already after a single injection, leptin at either dose or LIF caused a body weight change different from saline (Fig. 4; P < 0.05). Mice treated with LIF developed a much higher body weight loss than mice treated with leptin at either dose, and mice treated with leptin at 5 mg/kg developed higher body weight loss than mice treated with leptin at 0.5 mg/kg (Fig. 4). The body weight loss induced by LIF was dramatic. The body weight loss at the end of the study brought about by the 5 mg/kg dose of leptin was 15% of the one caused by LIF.

View larger version (24K): [in this window] [in a new window]   Fig. 4.   Body weight (BW) change during 7 days of follow-up in mice treated daily with leptin (0.5 or 5 mg/kg) or LIF (0.5 mg/kg) or saline. There were 12 mice in the LIF-treated group and 17 mice in the other groups. a P < 0.001 and aa P < 0.04 for differences in body weight change within groups; b P < 0.001 for differences in body weight change with the saline-treated mice; c P < 0.001 and cc P < 0.01 for differences in body weight change with the mice treated with leptin at the 0.5 mg/kg dose; and d P < 0.001 for differences in body weight change with leptin at the 5 mg/kg dose.

Leptin did not affect any of the peripheral blood cell counts after 7 days of treatment at either dose (Table 1). On the contrary, LIF increased the platelet and the red blood cell counts and decreased the total white cell count with decreases in the individual counts of neutrophils, eosinophils, basophils, monocytes, and lymphocytes (Table 1). The peripheral blood cell counts were monitored during the first 7 days of treatment with leptin (5 mg/kg) or saline and found to be similar between the two groups at all time points (Fig. 5). Leptin also did not affect the peripheral blood cell counts after the administration of 5-FU (data not shown).

View larger version (33K): [in this window] [in a new window]   Fig. 5.   Peripheral blood cell counts during 7 days of follow-up in mice treated daily with 5 mg/kg of leptin or saline. There were 18 mice/group. RBC, red blood cells; WBC, white blood cells; PLT, platelets; PMN, neutrophil granulocytes.

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

This study shows that leptin, in addition to causing body weight loss, displays only some of the in vivo activities typical of cytokines of the IL-6 family and only if used at a dose that exceeds the one sufficient to affect body weight. Leptin does not show any ability to change peripheral blood cell counts, although it has been reported to have properties of hematopoietic growth factor in vitro.

We tested leptin for a number of activities commonly displayed in mice by cytokines of the IL-6 family, which all depend on gp130 for signal transduction in target cells (23). This was done with the rationale that structural and functional homology exist between OB-R on the one hand and gp130 and also LIF-R on the other (5, 9, 16, 20, 21, 24, 37). Leptin at a dose of 0.5 mg/kg, similar to gp130-user cytokines, caused body weight loss. However, it did not induce the acute-phase protein SAA, it did not potentiate the elevation of serum levels of corticosterone and IL-6 induced by a low dose of IL-1, it did not inhibit the induction by LPS of serum TNF, and it did not affect peripheral blood cell counts. Leptin at the dose of 5 mg/kg also caused body weight loss and, similar to gp130-user cytokines, potentiated the induction by IL-1 of serum corticosterone and IL-6. However, unlike gp130-user cytokines, it did not induce SAA and did not inhibit the induction by LPS of TNF in the circulation. Moreover, the IL-6-like effects shown by 5 mg/kg of leptin were minimal compared with those observed with 0.5 mg/kg of LIF, which was chosen as the reference cytokine, because its receptor involves both gp130 and LIF-R (15, 23). LIF was also able to increase the circulating levels of corticosterone and to induce detectable serum levels of IL-6 when it was injected alone, whereas leptin did not show these abilities. The body weight loss caused by leptin at either the 0.5 or 5 mg/kg dose was also minimal compared with the one brought about by LIF. Whereas leptin evidently acted more to prevent body weight gain rather than to induce body weight loss (after 7 days of treatment, 0.5 mg/kg of leptin caused a body weight loss of 1% and 5 mg/kg of 3%, whereas saline-treated controls showed 5% body weight gain), by contrast LIF caused a dramatic loss of body weight (22%).

Despite the structural and functional homology at the receptor level, leptin seems to be partly related to the cytokines of the IL-6 family as to in vivo activities. In addition to regulating body weight, the activity that resembled those of gp130-user cytokines the most was the potentiation of the induction by IL-1 of circulating corticosterone (5 mg/kg of leptin resulted in 25% of the effect of 0.5 mg/kg of LIF on IL-1-induced corticosterone). Interestingly, signal-transducing OB-R isoforms are expressed in the hypothalamic region (16, 25, 37), where they have been connected to the body weight control function of leptin and where they could act to trigger the hypothalamus-pituitary-adrenal axis together with IL-1 (44). However, signal transducing OB-R isoforms are also expressed in the liver, although OB-Rb much less than in the hypothalamus (9, 16, 25), and yet leptin failed to induce SAA, which is produced by hepatocytes (35) and whose appearance in the circulation promptly follows stimulation with IL-6, IL-11, LIF, OM, CNTF, and CT-1, i.e., all the gp130-user cytokines (6). Furthermore, it has been suggested that macrophages express OB-R and are directly regulated by leptin (13). We found that leptin at the dose of 5 mg/kg has no ability to inhibit the induction by LPS of serum TNF, whereas it has a demonstrable, although minimal, ability to potentiate the induction by IL-1 of serum IL-6, two effects that depend on the production of TNF and IL-6 by cells of the monocyte/macrophage lineage. It seems therefore that the modest relationship between leptin and IL-6 family members as to in vivo activities cannot be accounted for simply by the lack of signal-transducing OB-R isoforms on effector cells. It is conceivable that differences exist between the signal-transducing isoforms of OB-R on the one hand and gp130 and LIF-R on the other in the recruitment of third messengers to an extent sufficient to prevent "spilling-over" from the intracellular signal-transduction channels involved in the action of leptin toward those involved in the action of the gp130-user cytokines. This might explain why only a high dose of leptin has activities typical of gp130-user cytokines, yet only some of them and in minimal amount, and why monocyte/macrophages may show some response to leptin in some cases (such as when they are stimulated by IL-1 to produce IL-6) but not in others (such as when they are stimulated by LPS to produce TNF).

Only the dose of leptin larger than the one sufficient to cause body weight loss showed some inflammation-related activities, such as the potentiation of IL-1 induction of corticosterone and IL-6, but only to a minimal extent compared with a tenfold lower dose of LIF. This observation fails to qualify leptin as a typical cytokine of the inflammatory cascade, at variance with previous results (17).

Recently, it has been found that an OB-R isoform is expressed on very primitive hematopoietic cells (10). Consistently, leptin has been found to be able to induce in vitro proliferation, differentiation, and activation of hematopoietic cells (13). It has also been proposed that the positive correlation observed in the clinics between the white blood cell count and the percentage of body fat is mediated by leptin, because positive correlations were found between leptin plasma levels and either white blood cell counts or body fat percentages (43). The preliminary investigation here reported, however, does not support the possibility that leptin affects peripheral blood cell counts in vivo.

Perspectives