Transgenic mice overexpressing amyloid precursor protein exhibit early metabolic deficits and a pathologically low leptin state associated with hypothalamic dysfunction in arcuate neuropeptide Y neurons

Abstract

Weight loss is a prominent early feature of Alzheimer's disease (AD) that often precedes the cognitive decline and clinical diagnosis. While the exact pathogenesis of AD remains unclear, accumulation of amyloid-β (Aβ) derived from the amyloid precursor protein (APP) in the brain is thought to lead to the neuronal dysfunction and death underlying the dementia. In this study, we examined whether transgenic mice overexpressing the Swedish mutation of APP (Tg2576), recapitulating selected features of AD, have hypothalamic leptin signaling dysfunction leading to early body weight deficits. We found that 3-month-old Tg2576 mice, before amyloid plaque formation, exhibit decreased weight with markedly decreased adiposity, low plasma leptin levels, and increased energy expenditure without alterations in feeding behavior. The expression of the orexigenic neuropeptide Y (NPY) in the hypothalamus to the low leptin state was abnormal at basal and fasting conditions. In addition, arcuate NPY neurons exhibited abnormal electrophysiological responses to leptin in Tg2576 hypothalamic slices or wild-type slices treated with Aβ. Finally, the metabolic deficits worsened as Tg2576 mice aged and amyloid burden increased in the brain. These results indicate that excess Aβ can potentially disrupt hypothalamic arcuate NPY neurons leading to weight loss and a pathologically low leptin state early in the disease process that progressively worsens as the amyloid burden increases. Collectively, these findings suggest that weight loss is an intrinsic pathological feature of Aβ accumulation and identify hypothalamic leptin signaling as a previously unrecognized pathogenic site of action for Aβ.

Keywords: Alzheimer's disease; NPY; amyloid; hypothalamus; leptin.

Figure 1.

Tg2576 mice develop amyloid plaques as the mice age. Representative immunohistochemistry staining of coronal mouse brain sections using a mouse monoclonal antibody against Aβ (4G8, 1:2000). The 3-month-old (3M) Tg2576 mice have no visible amyloid plaques. The 7-month-old (7M) Tg2576 mice have rare, sporadic amyloid plaques. The 14-month-old (14M) Tg2576 mice have widespread, markedly larger amyloid plaques. No amyloid plaques were visible throughout the brain of 14-month-old (14M) WT mice. Scale bar, all images: 100 μm.

Figure 2.

Tg2576 mice have decreased adiposity and low plasma leptin levels before amyloid plaque formation. A, Body weight curve showing significantly lower body weights in female Tg2576 mice compared with WT mice (n = 4–10 per group). B, Increasing brain Aβ1–42 burden correlates with worsening body weight deficits (% body weight of age-matched WT littermates) in Tg2576 mice (n = 2 or 3 per group for brain amyloid measurements; n = 4–10 per age for body weight measurements). C, Representative T1-weighted MR image showing decreased adipose tissue mass (bright, hyperintense signal) in 3-month-old female Tg2576 mice compared with WT mice. D, E, Markedly decreased adiposity in 3-month-old Tg2576 mice as demonstrated by MR spectroscopy body composition analysis (n = 5 per group). F, Tg2576 mice have significantly decreased plasma leptin levels (n = 12–14 per group). G, H, Basal plasma glucose and insulin levels are comparable in Tg2576 and WT mice (n = 5–18 per group). *p < 0.05, from respective WT by ANOVA for repeated measurements (body weight) or t test for two group comparisons. **p < 0.01, from respective WT by t test for two group comparisons.

Figure 3.

Low body weight in Tg2576 mice is not caused by alterations in feeding behavior but by increased energy expenditure. A, Daily food intake in 3-month-old Tg2576 mice was the same as WT mice (n = 6 or 7 per group). B, Energy expenditure was increased in Tg2576 mice throughout the light and dark cycle as measured by increased oxygen consumption (VO₂) in an indirect calorimetry system (Oxymax, Columbus Instruments) (n = 5 per group). C, Resting metabolic rate, as measured by VO₂ when the mice were not moving, was significantly increased in Tg2576 mice (n = 5 per group). **p < 0.01, from respective WT mice by t test.

Figure 4.

Tg2576 mice have abnormal hypothalamic responses to the low basal plasma leptin levels. A, In the hypothalamus of 3-month-old Tg2576 mice, high levels of human APP_swe RNA are present. B, No changes in hypothalamic leptin receptor (LepR) RNA expression in Tg2576 mice were seen. C, D, Unchanged hypothalamic NPY and AgRP RNA levels despite the low plasma leptin levels. E, F, An appropriate decrease in hypothalamic POMC and CART RNA levels was observed. All values were normalized with hypoxanthine-guanine phophoribosyltransferase RNA levels and expressed as relative values compared with WT mice. For all, n = 13 per group. **p < 0.01, from respective WT mice by t test. ***p < 0.001, from respective WT mice by t test.

Figure 5.

Tg2576 mice have abnormal hypothalamic responses to fasting. A, B, Comparable rate of body weight loss after a 48 h fast was seen in 3-month-old Tg2576 and WT mice. C, The 48 h fast caused an appropriate lowering of plasma leptin levels in WT mice, but Tg2576 mice had persistently low plasma leptin levels that remained unchanged after the fast. D, E, Comparable responses in blood glucose and plasma insulin levels were seen in WT and Tg2576 mice after a 48 h fast. F, Persistently elevated human APP_swe RNA levels in Tg2576 mice were unchanged after a 48 h fast. G–J, Markedly attenuated hypothalamic NPY (G) and AgRP (H) transcriptional responses, but relatively intact POMC (I) and CART (J) transcriptional responses were found in Tg2576 mice after a 48 h fast. For all, n = 9–12 per group. *p < 0.05, from respective WT mice by t test for two group comparison or ANOVA followed by Tukey post hoc test for multigroup comparisons. **p < 0.01, from respective WT mice by t test for two group comparison or ANOVA followed by Tukey post hoc test for multigroup comparisons. ***p < 0.001, from respective WT mice by t test for two group comparison or ANOVA followed by Tukey post hoc test for multigroup comparisons.

Figure 6.

Leptin-mediated hyperpolarization of hypothalamic arcuate NPY neurons is absent in Tg2576 mice. A, Leptin (100 nm) hyperpolarized arcuate NPY neurons in WT but not in Tg2576 hypothalamic slices. Arcuate nucleus from hypothalamic slices of NPY-GFP mice (3–5 weeks old) with or without the APP_swe transgene were used for whole-cell patch-clamp recordings. Representative traces are shown 5 min after leptin application. B, Changes in membrane potentials determined from the recordings of arcuate NPY neurons after application of leptin (100 nm) in WT and Tg2576 hypothalamic slices. n = 8–10 cells recorded. Spontaneous spike frequency in arcuate NPY neurons after acute application of leptin remained unchanged in both WT (WT + Vehicle: 8.4 0.3 Hz, WT + leptin: 7.9 ± 0.6 Hz, p = 0.52, n = 3 cells per group) and Tg2576 hypothalamic slices (Tg2576 + Vehicle: 1.9 ± 0.1 Hz, Tg2576 + leptin: 2.6 ± 0.3 Hz, p = 0.1015, n = 3 cells per group). C, Ghrelin (100 nm) strongly depolarized arcuate NPY neurons in WT but not in Tg2576 hypothalamic slices. Representative traces are shown 5 min after ghrelin application. The increase in resting potential is associated with a reduction in the amplitude of the action potentials. D, Changes in membrane potentials determined from the recordings of arcuate NPY neurons after application of ghrelin (100 nm) in WT and Tg2576 hypothalamic slices. n = 7 or 8 cells recorded. Ghrelin (100 nm) transiently increased spontaneous spike frequencies in the first 15 s after acute application of ghrelin (WT + Vehicle: 6.1 ± 0.8 Hz, WT + ghrelin: 14 ± 0.6 Hz, p < 0.05, n = 5 cells each), whereas no effects on spike frequencies were seen in Tg2576 hypothalamic slices (Tg2576 + Vehicle: 2.8 ± 0.3 Hz, Tg2576 + ghrelin: 2.5 ± 0.1 Hz, p = 0.373, n = 6 cells each). **p < 0.01, from WT cells by t test.

Figure 7.

Leptin-mediated hyperpolarization of hypothalamic arcuate NPY neurons is absent after application of oligomeric Aβ1–42. A, Application of oligomeric Aβ1–42 (100 nm) to WT hypothalamic slices depolarized arcuate NPY neurons. The increase in resting potential is associated with a reduction in the amplitude of the action potentials. Brain slices containing the hypothalamic arcuate nucleus from NPY-GFP mice (3–5 weeks old) were used for whole-cell patch-clamp recordings. The recordings show the duration before and after oligomeric Aβ1–42 application. B, Oligomeric Aβ1–42 (10, 30, and 100 nm) depolarized arcuate NPY neurons in a dose-dependent fashion (n = 7–25 cells recorded). C, D, Exposure to oligomeric Aβ1–42 (100 nm) inhibited the electrophysiological responses to both leptin (100 nm) (C) and ghrelin (100 nm) (D) in arcuate NPY neurons. Representative traces are shown. E–G, Changes in membrane potentials determined from the recordings of arcuate NPY neurons after leptin (100 nm) (E) and ghrelin (100 nm) (F, G) application to WT hypothalamic slices treated with oligomeric Aβ1–42 (100 nm). n = 4–8 cells recorded. Although membrane potentials in arcuate NPY neurons were readily affected by oligomeric Aβ1–42, spontaneous spike frequency remained unchanged under all conditions tested (WT + Vehicle: 8.4 ± 0.5 Hz, WT + Aβ1–42: 7.6 ± 0.2 Hz, WT + Aβ1–42 + Leptin: 8.1 ± 0.3 Hz, WT + Aβ1–42 + Ghrelin: 7.9 ± 0.2 Hz, p > 0.05 for all; n = 3–5 cells per group). **p < 0.01, from WT cells by t test.

Figure 8.

Aged Tg2576 mice with prominent amyloid plaques have worsening leptin-associated metabolic deficits compared with young Tg2576 mice with no amyloid plaques. A, Representative T1-weighted MR images showing an increase in adipose tissue (bright, hyperintense signal) as WT mice age from 3 months (3M) to 14 months (14M), whereas aged Tg2576 mice continue to have markedly lower overall body adiposity. B, The 14-month-old Tg2576 mice continue to have low adiposity compared with similarly aged WT mice as demonstrated by quantitative body composition analysis by MR spectroscopy (n = 6 per group). C, Increasing relative disparity in plasma leptin levels in 14-month-old (14M) Tg2576 mice to 14-month-old WT mice compared with plasma leptin levels in 3-month-old mice. D, The 14-month-old Tg2576 mice have significantly lower plasma insulin levels compared with 14-month-old WT mice. No significant difference in insulin levels was seen at 3 months of age. For all, n = 4–15 per group. *p < 0.05, from respective WT or group comparison by t test for two group comparisons or ANOVA followed by Tukey post hoc test for multigroup comparisons. **p < 0.01, from respective WT or group comparison by t test for two group comparisons or ANOVA followed by Tukey post hoc test for multigroup comparisons. ***p < 0.001, from respective WT or group comparison by t test for two group comparisons or ANOVA followed by Tukey post hoc test for multigroup comparisons.

Figure 9.

Schematic demonstrating a proposed pathophysiological mechanism for the early weight loss and low plasma leptin levels in young preamyloid plaque Tg2576 mice. In this model, excess Aβ causes dysfunction in hypothalamic arcuate NPY neurons and possibly other leptin-responsive neurons to disrupt the efferent signal regulating body weight. This ultimately results in decreased adiposity and a state of pathologically low levels of plasma leptin, which may decrease leptin's potential procognitive and neuroprotective effects against Aβ toxicity in the cortex and hippocampus. The mouse brain figure was adapted from a mouse brain atlas (Paxinos and Franklin, 2004).