Differential contribution of POMC and AgRP neurons to the regulation of regional autonomic nerve activity by leptin

Abstract

Objectives: The autonomic nervous system is critically involved in mediating the control by leptin of many physiological processes. Here, we examined the role of the leptin receptor (LepR) in proopiomelanocortin (POMC) and agouti-related peptide (AgRP) neurons in mediating the effects of leptin on regional sympathetic and parasympathetic nerve activity.

Methods: We analyzed how deletion of the LepR in POMC neurons (POMC^Cre/LepR^fl/fl mice) or AgRP neurons (AgRP^Cre/LepR^fl/fl mice) affects the ability of leptin to increase sympathetic and parasympathetic nerve activity. We also studied mice lacking the catalytic p110α or p110β subunits of phosphatidylinositol-3 kinase (PI3K) in POMC neurons.

Results: Leptin-evoked increase in sympathetic nerve activity subserving thermogenic brown adipose tissue was partially blunted in mice lacking the LepR in either POMC or AgRP neurons. On the other hand, loss of the LepR in AgRP, but not POMC, neurons interfered with leptin-induced sympathetic nerve activation to the inguinal fat depot. The increase in hepatic sympathetic traffic induced by leptin was also reduced in mice lacking the LepR in AgRP, but not POMC, neurons whereas LepR deletion in either AgRP or POMC neurons attenuated the hepatic parasympathetic nerve activation evoked by leptin. Interestingly, the renal, lumbar and splanchnic sympathetic nerve activation caused by leptin were significantly blunted in POMC^Cre/LepR^fl/fl mice, but not in AgRP^Cre/LepR^fl/fl mice. However, loss of the LepR in POMC or AgRP neurons did not interfere with the ability of leptin to increase sympathetic traffic to the adrenal gland. Furthermore, ablation of the p110α, but not the p110β, isoform of PI3K from POMC neurons eliminated the leptin-elicited renal sympathetic nerve activation. Finally, we show trans-synaptic retrograde tracing of both POMC and AgRP neurons from the kidneys.

Conclusions: POMC and AgRP neurons are differentially involved in mediating the effects of leptin on autonomic nerve activity subserving various tissues and organs.

Keywords: Autonomic nervous system; Cardiovascular regulation; Energy homeostasis; Leptin.

Graphical abstract

Figure 1

LepR deletion from POMC or AgRP neurons prevents leptin-induced STAT3 phosphorylation. A-B) In control mice, ICV leptin treatment increased STAT3 phosphorylation in both POMC (A) and AgRP (B) neurons. C-D) Loss of LepR from POMC (C) or AgRP (D) neurons abolished leptin-induced phosphorylation of STAT3 in those neurons. Images acquired at 20× magnification. Scale bar, 100 μm. E-F) Relative to littermate controls, male and female mice lacking the LepR in POMC (E, n = 7–19 male, 9–23 female LepR^fl/fl; n = 16–35 male, 10–27 female POMC^Cre/LepR^fl/fl) or AgRP (F, n = 7–9 male, 9–10 female LepR^fl/fl; n = 12–21 male, 13–17 female AgRP^Cre/LepR^fl/fl) neurons display increased body weight.

Figure 2

Contrasting effects of LepR ablation in POMC and AgRP neurons on leptin control of BAT and WAT SNA. A–B) Leptin-induced increase in BAT SNA is partially attenuated in both POMC^Cre/LepR^fl/fl and AgRP^Cre/LepR^fl/fl mice (n = 11 leptin treated LepR^fl/fl mice; n = 9 leptin treated POMC^Cre/LepR^fl/fl mice; n = 9 leptin treated AgRP^Cre/LepR^fl/fl mice; n = 9 vehicle treated LepR^fl/fl mice; n = 7 vehicle treated POMC^Cre/LepR^fl/fl mice; n = 5 vehicle treated AgRP^Cre/LepR^fl/fl mice). C–D) Sympathetic nerve activation to inguinal WAT elicited by ICV leptin is impaired in AgRP^Cre/LepR^fl/fl mice, but not POMC^Cre/LepR^fl/fl mice (n = 15 leptin treated LepR^fl/fl mice; n = 13 leptin treated POMC^Cre/LepR^fl/fl mice; n = 17 leptin treated AgRP^Cre/LepR^fl/fl mice; n = 5 vehicle treated LepR^fl/fl mice; n = 6 vehicle treated POMC^Cre/LepR^fl/fl mice; n = 6 vehicle treated AgRP^Cre/LepR^fl/fl mice). Bar graphs represent the averages of the last hour of SNA recordings. Comparisons were made using 2-way ANOVA with or without repeated measures, *p < 0.05 vs genotype matched vehicle, †p < 0.05 vs leptin-LepR^fl/fl controls.

Figure 3

Loss of LepR in POMC and AgRP neurons affects differentially the effects of leptin on hepatic autonomic nerve activity. A–B) Leptin stimulation of hepatic SNA was partially blunted in AgRP^Cre/LepR^fl/fl mice, but not POMC^Cre/LepR^fl/fl mice (n = 12 leptin treated LepR^fl/fl mice; n = 9 leptin treated POMC^Cre/LepR^fl/fl mice; n = 6 leptin treated AgRP^Cre/LepR^fl/fl mice; n = 3 vehicle treated LepR^fl/fl mice; n = 3 vehicle treated POMC^Cre/LepR^fl/fl mice; n = 9 vehicle treated AgRP^Cre/LepR^fl/fl mice). C–D) The increase in hepatic PSNA evoked by leptin was abolished in both POMC^Cre/LepR^fl/fl and AgRP^Cre/LepR^fl/fl mice (n = 15 leptin treated LepR^fl/fl mice; n = 13 leptin treated POMC^Cre/LepR^fl/fl mice; n = 17 leptin treated AgRP^Cre/LepR^fl/fl mice; n = 5 vehicle treated LepR^fl/fl mice; n = 6 vehicle treated POMC^Cre/LepR^fl/fl mice; n = 6 vehicle treated AgRP^Cre/LepR^fl/fl mice). Bar graphs represent the averages of the last hour of SNA and PNSA recordings. Comparisons were made using 2-way ANOVA with or without repeated measures, *p < 0.05 vs genotype matched vehicle, †p < 0.05 vs leptin-LepR^fl/fl controls.

Figure 4

The LepR on POMC, but not AgRP, neurons is required for leptin activation of lumbar and splanchnic SNA, but not adrenal SNA. A–D) ICV leptin (2 μg) increased lumbar (A-B, n = 13 leptin treated LepR^fl/fl mice; n = 8 leptin treated POMC^Cre/LepR^fl/fl mice; n = 8 leptin treated AgRP^Cre/LepR^fl/fl mice; n = 5 vehicle treated LepR^fl/fl mice; n = 6 vehicle treated POMC^Cre/LepR^fl/fl mice; n = 5 vehicle treated AgRP^Cre/LepR^fl/fl mice), and splanchnic (C-D, n = 12 leptin treated LepR^fl/fl mice; n = 9 leptin treated POMC^Cre/LepR^fl/fl mice; n = 6 leptin treated AgRP^Cre/LepR^fl/fl mice; n = 3 vehicle treated LepR^fl/fl mice; n = 4 vehicle treated POMC^Cre/LepR^fl/fl mice; n = 5 vehicle treated AgRP^Cre/LepR^fl/fl mice) SNA in control and AgRP^Cre/LepR^fl/fl mice; but not in POMC^Cre/LepR^fl/fl mice. E–F) The leptin-evoked increase in SNA subserving the adrenal gland was not affected in POMC^Cre/LepR^fl/fl and AgRP^Cre/LepR^fl/fl mice (n = 14 leptin treated LepR^fl/fl mice; n = 9 leptin treated POMC^Cre/LepR^fl/fl mice; n = 7 leptin treated AgRP^Cre/LepR^fl/fl mice; n = 3 vehicle treated LepR^fl/fl mice; n = 3 vehicle treated POMC^Cre/LepR^fl/fl mice; n = 3 vehicle treated AgRP^Cre/LepR^fl/fl mice). Bar graphs represent the averages of the last hour of SNA recordings. Comparisons were made using 2-way ANOVA with repeated or without measures, *p < 0.05 vs genotype matched vehicle, †p < 0.05 vs leptin-LepR^fl/fl controls.

Figure 5

LepR ablation from POMC, but not AgRP, neurons effects selectively leptin control of renal SNA. A–C) Leptin administered ICV (2 μg, A-B, n = 13 leptin treated LepR^fl/fl mice; n = 8 leptin treated POMC^Cre/LepR^fl/fl mice; n = 8 leptin treated AgRP^Cre/LepR^fl/fl mice; n = 5 vehicle treated LepR^fl/fl mice; n = 6 vehicle treated POMC^Cre/LepR^fl/fl mice; n = 5 vehicle treated AgRP^Cre/LepR^fl/fl mice) or intravenously (120 μg, C, n = 16 LepR^fl/fl mice; n = 8 POMC^Cre/LepR^fl/fl mice; n = 8 AgRP^Cre/LepR^fl/fl mice) increased renal SNA in control (LepR^fl/fl) and AgRP^Cre/LepR^fl/fl mice, but not in POMC^Cre/LepR^fl/fl mice. D) POMC^Cre/LepR^fl/fl mice displayed a normal increase in renal SNA in response to ICV insulin (n = 6 LepR^fl/fl mice; n = 8 POMC^Cre/LepR^fl/fl mice). The bar graph represents the averages of the last hour of SNA recordings. Comparisons were made using 2-way ANOVA with or without repeated measures, *p < 0.05 vs genotype matched vehicle, †p < 0.05 vs leptin-LepR^fl/fl controls.

Figure 6

Effects of POMC neurons-selective deletion of p110α or p110β catalytic subunits of PI3K. A–B) Deletion of the p110α (A) or p110β (B) had no effect on body weight in male or female mice fed normal chow or male mice fed a high fat diet (HFD). C–D) Leptin-induced increase in renal SNA, measured in the conscious state, was abolished in mice lacking the p110α, but not p110β, PI3K in POMC neurons (n = 9 p110α^fl/fl + p110 β^fl/fl mice; n = 6 POMC^Cre/p110α^fl/fl mice; n = 6 p110 β^fl/fl mice). The bar graph represents the averages of the last hour of SNA recordings. Comparisons were made using t-test, 1- or 2-way ANOVA with or without repeated measures, *p < 0.05 vs p110α^fl/fl + p110 β^fl/fl mice.

Figure 7

Trans-synaptic retrograde tracing of POMC and AgRP neurons from the kidneys. A–B) GFP-expressing PRV-152 pseudorabies virus injected into the left and right kidneys trans-synaptically labels, after 5 days, the ARC POMC (A) and AgRP neurons (B) in POMC^Cre/tdTomato mice and AgRP^Cre/tdTomato mice, respectively. Images acquired at 20× magnification. Scale bar, 100 μm.