HDAC5 controls a hypothalamic STAT5b-TH axis, the sympathetic activation of ATP-consuming futile cycles and adult-onset obesity in male mice

Abstract

With age, metabolic perturbations accumulate to elevate our obesity burden. While age-onset obesity is mostly driven by a sedentary lifestyle and high calorie intake, genetic and epigenetic factors also play a role. Among these, members of the large histone deacetylase (HDAC) family are of particular importance as key metabolic determinants for healthy ageing, or metabolic dysfunction. Here, we aimed to interrogate the role of class 2 family member HDAC5 in controlling systemic metabolism and age-related obesity under non-obesogenic conditions. Starting at 6 months of age, we observed adult-onset obesity in chow-fed male global HDAC5-KO mice, that was accompanied by marked reductions in adrenergic-stimulated ATP-consuming futile cycles, including BAT activity and UCP1 levels, WAT-lipolysis, skeletal muscle, WAT and liver futile creatine and calcium cycles, and ultimately energy expenditure. Female mice did not differ between genotypes. The lower peripheral sympathetic nervous system (SNS) activity in mature male KO mice was linked to higher dopaminergic neuronal activity within the dorsomedial arcuate nucleus (dmARC) and elevated hypothalamic dopamine levels. Mechanistically, we reveal that hypothalamic HDAC5 acts as co-repressor of STAT5b over the control of Tyrosine hydroxylase (TH) gene transactivation, which ultimately orchestrates the activity of dmARH dopaminergic neurons and energy metabolism in male mice under non-obesogenic conditions.

Keywords: Adult-onset obesity; Brown fat thermogenesis; Dopamine; Futile ATP-consuming cycles; Histone deacetylase 5; Hypothalamus.

Graphical abstract

Figure 1

Chow-fed male HDAC5-KO mice develop adult-onset obesity. Chow-fed male WT and HDAC5-KO mice were evaluated for changes in (A) body weight and (B) body composition over a period of 7 months. At 6 months of age, fasting plasma was collected to assess the (C) HOMA-IR, (D) leptin, (E) cholesterol, (F) non-esterified free fatty-acids (NEFA) and (G) triglyceride (TAG) levels. At 3 months of age, WT and HDAC5-KO males were subjected to 72 h of combined indirect calorimetry to assess (H) temporal changes in energy expenditure (EE), (I) average total EE values in correlation to body weight, (J) average light and dark phase EE, temporal changes and average light and dark phase values for (K,L) locomotor activity, (M,N) respiratory exchange ratio (RER) and (O,P) food intake. Values represent means ± SEM. Statistical analysis were done by either two-way ANOVA with Bonferroni post-hoc test (A,B,K,M,O), two-tailed unpaired Students' t-tests (C-G,J,L,N,P) or ANCOVA with body weight as co-variate (H–J). ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001 and ∗∗∗∗p < 0.0001.

Figure 2

Impaired brown adipose tissue thermogenesis in male HDAC5-KO mice. BAT function in male WT and HDAC5-KO mice at the age of 6 months was assessed by (A) infrared thermography depicting (B) skin temperature above the intrascapular BAT. Subsequent (C) Western Blot and (D) densitometric analysis of UCP1 protein levels and HPLC-ECD based quantification of (E) norepinephrine and (F) epinephrine. Additional cohorts of male WT and HDAC-5 KO mice at the age of 3 months were exposed for 5 h to ambient temperature (23 °C) or cold exposure (4 °C) to assess (G) average energy expenditure. As surrogate parameter for cold-induced shivering, we moreover recorded the (H) cumulative fine movement of mice subjected to 5 h of cold (4 °C). Values represent means ± SEM. Statistical analysis were done by two-tailed unpaired Students' t-test (B–F,H), or two-way ANOVA with Bonferroni post-hoc test (G). ∗p < 0.05, ∗∗p < 0.01 and ∗∗∗∗p < 0.0001. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article).

Figure 3

Impaired adrenergic lipolysis in white adipose tissue of HDAC5-KO males. Epididymal WAT of HDAC5-KO and WT mice was subjected to (A) HPLC-ECD-based detection of norepinephrine levels and (B) Western Blotting and (C) densitometric analyses for total HSL (normalized to β-actin) and pHSL (normalized to total HSL). (D) Quantitative PCR to assess changes in mRNA levels of genes critically involved in adrenergic signaling and lipolysis; hormone-sensitive lipase (Hsl), adipocyte triacyl glyceride lipase (Atgl), lipoprotein lipase (Lpl), beta2-, beta3-and alpha2-adrenergic receptors (Adrab2, Adrab3, & Adra2a), nuclear receptor subfamily 4 group A member 3 (Nr4a3). Values represent means ± SEM. Statistical significance was determined using (A) two-tailed unpaired Students' t-test (A,C) with significance levels indicated as ∗ p < 0.05, ∗∗p < 0.01 and ∗∗∗p < 0.001, or (D) multiple t-testing following the two-stage step-up procedure of Benjamini, Krieger and Yekutieli to control the false discovery rate (FDR). Symbols indicate the following adjusted q-value thresholds: q < 0.1 (^§), q < 0.05 (^§§), and q < 0.001 (^§§§§). The desired FDR threshold was set at 10%.

Figure 4

Persistent activation of dopaminergic neurons in the dorsomedial ARH of HDAC5-KO males drives hypothalamic dopamine levels. (A) Co-immunostainings for TH and cFOS in the paraventricular hypothalamus (PVH) from HDAC5-KO and WT mice were assessed for (B) tyrosine hydroxylase (TH) fluorescence intensities (FI) and numbers of cFOS + neurons within the (C) PVN or (D) additional areas such as the ventromedial (VMH, dorsomedial (DMH) or lateral hypothalamus (LH). (E) Dopamine levels were assessed by HPLC-ECD in the hypothalamus (Hyp), prefrontal cortex (PFC) and striatum. (F) Co-immunostainings for TH and cFOS in the hypothalamic arcuate nucleus (dmARH) revealed elevated numbers of (G) TH, cFOS and TH-cFOS double-positive neurons. (H) TH immunostaining and (I) fluorescence intensity in the median eminence (ME). Prolactin (J) mRNA levels in the pituitary and (K) plasma concentrations. (L-O) TH and cFOS immunostainings in 7, 45 and 180 days-old WT and HDAC5 KO males were assessed for (L) the number of TH + neurons, (M) the TH mean fluorescence intensity (MFI), (N) the number of cFOS + neurons and (O) the number of TH-cFOS double-positive neurons. Values represent means ± SEM. Statistical analysis were done by two-tailed unpaired Students' t-tests. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001 and ∗∗∗∗p < 0.0001.

Figure 5

Unperturbed hypothalamic dopamine tone, systemic metabolism and BAT thermogenesis in HDAC5-KO females. (A) Co-immunostainings for TH and cFOS in the hypothalamic arcuate nucleus (ARH) of HDAC5-KO and WT female mice were evaluated for the numbers of (B) tyrosine hydroxylase (TH) and cFOS + neurons. (C) Plasma levels of prolactin. (D) Average number of pups and sex distribution per litter. (E) Body weight and (F) body composition over a period of 6 months. (G) fasting blood glucose levels and (H) infrared thermographic images depicting heat production with the respective quantification of (I) surface skin temperature above the intrascapular BAT region. Values represent means ± SEM. Statistical analysis were done by two-tailed unpaired Students' t-tests (B,C,G,I) or two-way ANOVA with Bonferroni post-hoc test (D–F).

Figure 6

HDAC5 interacts with STAT5b to drive its nuclear translocation and Th transactivation. Micropunches from the ARH region of male WT and HDAC5-KO mice were subjected to (A) Western Blotting for HDAC5 (total) and STAT5b and ERK (total & phospho: p), or (B) co-immunoprecipitation using anti-STAT5b and Western blotting detection of HDAC5, lysin acetylation (K–Ac), Stat5b, IgG or beta-actin. (C) Western blotting of nuclear and cytosolic fractions depicting the subcellular locations of HDAC5 and STAT5b, with (D) densitometric analyses for STAT5b normalized to the housekeeping proteins H3 in the nucleus and GAPDH in the cytosol of male WT and HDAC5-KO hypothalami. (E) Following high throughput ChIPmentation with an anti-STAT5 antibody, binding of STAT5 to the promotor region of Th was assessed by qPCR. (F,G) Expression levels of the STAT5 target genes Th and Prlr, as well as neurotransmitter-linked Slc6a3, Gad1 and Gad2 in TH expressing cells of the ARH of WT and HDAC5-KO males were assessed by the (F) selective enrichment of TH+/NeuN- cells using FACS, and (G) qPCR with Malat1 as housekeeping gene. Values represent means ± SEM. Statistical analysis were done by (D, E) Students' t-tests (∗p < 0.05) or (G) multiple t-testing following the two-stage step-up procedure of Benjamini, Krieger and Yekutieli and a FDR of 10% (^§q < 0.1, ^§§q < 0.5).