Arcuate nucleus proopiomelanocortin neurons mediate the acute anorectic actions of leukemia inhibitory factor via gp130

Abstract

The proinflammatory cytokine leukemia inhibitory factor (LIF) is induced in disease states and is known to inhibit food intake when administered centrally. However, the neural pathways underlying this effect are not well understood. We demonstrate that LIF acutely inhibits food intake by directly activating pro-opiomelanocortin (POMC) neurons in the arcuate nucleus of the hypothalamus. We show that arcuate POMC neurons express the LIF-R, and that LIF stimulates the release of the anorexigenic peptide, alpha-MSH from ex vivo hypothalami. Transgenic mice lacking gp130, the signal transducing subunit of the LIF-R complex, specifically in POMC neurons fail to respond to LIF. Furthermore, LIF does not stimulate the release of alpha-MSH from the transgenic hypothalamic explants. These findings indicate that POMC neurons mediate the acute anorectic actions of central LIF administration and provide a mechanistic link between inflammation and food intake.

Figure 1

Hypothalamic LIF expression in response to LPS and IL-1β. A, Relative quantity of hypothalamic LIF mRNA after ip LPS (100 μg/kg) compared with vehicle in mice. LPS transiently induces hypothalamic LIF expression at 1 and 4 h after injection, although this effect is abolished at 8 h. B, Central (i.c.v.) IL-1β (10 ng) potently induces hypothalamic LIF expression 8 h after administration. Data normalized to vehicle at each time point. Results expressed as mean ± sem. Statistics calculated by two-way ANOVA followed by post hoc analysis using a Bonferroni corrected t test (A) or two-tailed Student’s t test (B) (**, P < 0.01 vs. vehicle; ***, P < 0.001 vs. vehicle).

Figure 2

Effects of central LIF on food intake in WT mice. A, Cumulative food intake after i.c.v. bolus injection of LIF (100 ng) vs. vehicle in wild-type mice. Mice fasted overnight and pretreated with ipketorolac (40 mg/kg). B, Cumulative food intake in LIF-treated mice expressed as percent reduction compared with vehicle. Inhibition of food intake was significant at 2 and 4 h posttreatment. Data are expressed as mean ± sem. Statistics calculated two-way ANOVA followed by post hoc analysis using a Bonferroni corrected t test (**, P < 0.01 vs. vehicle).

Figure 3

LIF activates POMC-EGFP neurons in the hypothalamus. A, D, and G, POMC-EGFP expression in the arcuate nucleus. B, Low cFos IR (red) in vehicle-treated mice (n = 3). C, Few POMC neurons show nuclear cFos IR after vehicle treatment. E and H, cFos IR is increased in LIF-treated (10 ng, n = 4) mice. F and I, LIF increases cFos IR in POMC neurons. G–I, Enlargement of D–F shown for clarity (white boxes indicate regions of enlargement). J, LIF increases cFos IR per section. K, Approximately 30% of POMC neurons contain cFos immunoreactivity after either 10 or 100 ng LIF treatment. Data are expressed as mean ± sem, and statistics calculated by one-way ANOVA followed by a post hoc analysis using a Bonferroni corrected t test (***, P < 0.001 vs. vehicle). Scale bars, 100 mm (A–F) and 50 mm (G–I). 3V, Third ventricle.

Figure 4

Expression of LIF-R in the hypothalamus. A, Representative dark-field photomicrograph showing expression of LIF-R mRNA (silver grain clusters) in the ARC of WT rats. B, Double-label in situ hybridization showing expression of LIF-R (silver grain clusters) by cells expressing POMC mRNA (red precipitate). Arrows point to POMC neurons that coexpress LIF-R mRNA. Open arrowheads signify POMC neurons that do not express LIF-R mRNA. Filled arrowheads denote cells that express LIF-R mRNA but not POMC mRNA. Scale bars, 100 mm (A) and 25 mm (B). 3V, Third ventricle.

Figure 5

LIF stimulates α-MSH release from murine hypothalamic explants. LIF (50 nm, n = 10) increased α-MSH release in vitro by approximately 2.5-fold over vehicle (n = 9). Results are expressed as mean ± sem. Statistics calculated by two-tailed Student’s t test (***, P < 0.001 vs. vehicle).

Figure 6

LIF-induced pSTAT3 IR in WT and gp130^flox/flox POMC neurons. A, E, I, and M, ACTH IR in the ARC of WT and gp130^flox/flox mice was used to identify POMC neurons. B and J, pSTAT3 IR is low in vehicle-treated WT (n = 5) and gp130^flox/flox (n = 4) mice. C, D, K, and L, Few POMC neurons show pSTAT3 IR in vehicle-treated mice of both genotypes. F and N, LIF treatment (50 ng) increased pSTAT3 IR in WT (n = 6) and gp130^flox/flox (n = 5) mice 30 min after i.c.v. injection. G and H, LIF treatment increases pSTAT3 IR in WT POMC neurons. O and P, Despite increased total pSTAT3 IR, few gp130^flox/flox POMC neurons show pSTAT3 IR after LIF treatment. D, H, L, and P, Enlargement of C, G, K, and O for clarity (white boxes denote area of enlargement − area chosen to maximize number of POMC neurons in panel). Q, pSTAT3 IR per section is increased by LIF treatment in both genotypes. R, LIF treatment increases the number of POMC neurons containing pSTAT3 IR in WT but not gp130^flox/flox mice. Data are expressed as mean ± sem. Statistics calculated by two-way ANOVA followed by a post hoc analysis using a Bonferroni corrected t test (a, P < 0.001 vs. WT/vehicle and gp130^flox/flox/vehicle; b, P < 0.001 vs. WT/vehicle, P < 0.01 vs. gp130^flox/flox/vehicle; ***, P < 0.001 vs. gp130^flox/flox/LIF). Scale bars, 100 mm (A–C, E–G, I–K, M–O) and 50 mm (D, H, L, P). 3V, Third ventricle.

Figure 7

gp130^flox/flox mice are resistant to both LIF-induced α-MSH release and anorexia. A, LIF-induced α-MSH release from WT and gp130^flox/flox murine hypothalamic explants. LIF (50 nm, n = 5) increases α-MSH release from WT hypothalami compared with vehicle (n = 7). LIF does not increase α-MSH release from gp130^flox/flox hypothalami (n = 8) compared with vehicle (n = 5). Data shown as mean ± sem and statistics calculated by two-way ANOVA followed by a post hoc analysis using a Bonferroni corrected t test (*, P < 0.05 vs. LIF-treated gp130^flox/flox). B, LIF-induced anorexia is attenuated in gp130^flox/flox mice. The i.c.v. LIF (50 ng, n = 7) reduced food intake in fasted, ketorolac (40 mg/kg) treated mice at 2 and 4 h after injection compared with vehicle (n = 5). This effect was diminished in gp130^flox/flox mice (LIF, n = 6; vehicle, n = 6). Results normalized to percentage of vehicle food intake and expressed as mean ± sem. Statistics calculated by two-way ANOVA followed by a post hoc analysis using a Bonferroni corrected t test (*, P < 0.05; **, P < 0.01 vs. gp130^flox/flox). No statistically significant differences between genotypes were found at any time after vehicle treatment (P > 0.05).