Analysis of the Glucose-Dependent Transcriptome in Murine Hypothalamic Cells

Abstract

Glucose provides vital energy for cells and contributes to gene expression. The hypothalamus is key for metabolic homeostasis, but effects of glucose on hypothalamic gene expression have not yet been investigated in detail. Thus, herein, we monitored the glucose-dependent transcriptome in murine hypothalamic mHypoA-2/10 cells by total RNA-seq analysis. A total of 831 genes were up- and 1390 genes were downregulated by at least 50%. Key genes involved in the cholesterol biosynthesis pathway were upregulated, and total cellular cholesterol levels were significantly increased by glucose. Analysis of single genes involved in fundamental cellular signaling processes also suggested a significant impact of glucose. Thus, we chose ≈100 genes involved in signaling and validated the effects of glucose on mRNA levels by qRT-PCR. We identified Gnai1-3, Adyc6, Irs1, Igfr1, Hras, and Elk3 as new glucose-dependent genes. In line with this, cAMP measurements revealed enhanced noradrenalin-induced cAMP levels, and reporter gene assays elevated activity of the insulin-like growth factor at higher glucose levels. Key data of our studies were confirmed in a second hypothalamic cell line. Thus, our findings link extra cellular glucose levels with hypothalamic lipid synthesis and pivotal intracellular signaling processes, which might be of particular interest in situations of continuously increased glucose levels.

Keywords: RNA-seq; cAMP; cholesterol; glucose; hypothalamus; insulin-like growth factor; noradrenalin; serum response element.

Figure 1

Volcano plot of RNA-seq data obtained with mHypoA-2/10 cells at low and high glucose levels. Cells were stimulated according to protocol 1 illustrated in Section 2.5. Only genes with RPKM of ≥1 for at least one condition (11,062 genes) are shown. The number of genes with a fold change of ǀFCǀ ≥ 3, ǀFCǀ ≥ 3, or ǀFCǀ ≥ 1.5 are given. For each condition, one single experiment performed in triplicate was analyzed as described under experimental procedures.

Figure 2

Canonical signaling pathway analysis of RNA-seq data obtained with mHypoA-2/10 cells at low and high glucose levels. For details, see Section 2.3.

Figure 3

Glucose induced RNA expression of both Srebf genes and of key enzymes involved in cholesterol or fatty acid biosynthesis in mHypoA-2/10 cells. Cells were stimulated according to protocol 1 illustrated in Section 2.5 and RPKM determined by RNA-seq. For each condition, triplicates from one single experiment were analyzed as described under experimental procedures. In (A), data for 3-hydroxy-3-methylglutaryl-CoA synthase 1 (Hmgcs1), 3-hydroxy-3-methylglutaryl-CoA reductase (Hmgcr), mevalonate kinase (Mvk), farnesyl-diphosphate farnesyltransferase-1 (Fdft1), lanosterol 14α-demethylase (Cyp51), hydroxysteroid 17-β dehydrogenase-7 (Hsd17b7), NAD(P)-dependent steroid dehydrogenase-like (Nsdhl), 24-dehydrocholesterol reductase (Dhcr24), and 7-dehydrocholesterol reductase (Dhcr7) are shown. In (B), data for the ATP citrate lyase (Acly), acetly-Coa carboxylase-α (Acaca), fattay acid synthase (Fasn), and stearoyl-Coa desaturase (Scd1) are presented. In (C), data for the sterol regulatory element-binding protein gene-1 (Srebf1) and -2 (Srebf2) are presented. Data are expressed as the mean ± S.E.M. of one single experiment (N = 1) performed in triplicate. Asterisks indicate significant differences calculated on the basis of Bonferroni-corrected p-values of the entire data set obtained by RNA-seq. *** p < 0.001.

Figure 4

Glucose-induced SREBP protein expression and cholesterol synthesis via CRTC-2 in mHypoA-2/10 cells. Cells were stimulated according to protocol 2 illustrated in Section 2.5. (A,B) mHypoA-2/10 cells were transfected with a control or a specific siRNA against CRTC-2. (A) SREBP-1 or -2 expression was analyzed by whole-cell ELISA. (B) Cholesterol levels in the supernatant were detected after incubation of the cells with 20 mM β-cyclodextrin for 4 h. (C) Cholesterol levels were determined in the supernatant of GT1-7 cells after incubation of cells with 20 mM β-cyclodextrin for 4 h. Data of 3 independent (N = 3) experiments performed in quadruplicate are shown as the mean ± S.E.M. Asterisks indicate significant differences based on two-sample t-tests (C) or two-way ANOVA followed by Tukey’s post hoc test (A,B). * p < 0.05, ** p < 0.01, and *** p < 0.001.

Figure 5

Effects of glucose on mRNA expression of single genes involved in intracellular signaling processes in mHypoA-2/10 cells. mHypoA-2/10 cells were stimulated according to protocol 1 illustrated in Section 2.5 and RPKM determined by RNA-seq. For each condition, triplicates from one single experiment (N = 1) were analyzed as described under experimental procedures. RPKM values for 99 genes are given. Genes marked in red were significantly upregulated, and those marked blue were downregulated by glucose (Bonferroni-corrected p-value of ≤0.05). Black genes were not significantly affected by glucose. In the lower panel are genes with RPKM between 1 and 10, in the middle panel between 10 and 40, and in the upper panel between 40 and 400.

Figure 6

Heat map of glucose-dependent gene expression in mHypoA-2/10 cells combining RNA-seq and qRT-PCR data in mHypoA-2/10 cells. mHypoA-2/10 cells were stimulated accordingly to protocol 1 illustrated in Section 2.5 and RNA levels determined by RNA-seq or qRT-PCR with Sdha or Tbp as a reference gene. Reddish colors indicate gene induction; blueish colors indicate depression. For qRT-PCR experiments, data of 3 independent (N = 3) experiments performed in triplicate are shown. Data are also provided in Table S2.

Figure 7

Glucose enhanced IRS-1 expression via CRTC-2 and IGF- but not EGF-induced activation of an SRF/TCF-dependent luciferase reporter gene. mHypoA-2/10 cells were stimulated accordingly to protocol 1 (A) or 2 (B–F) illustrated in Section 2.5. Gene expression was determined by RNA-seq in (A) or by qRT-PCR in (B). (A) For each condition, triplicates were analyzed for Irs1 or -2 expression. (B) Irs-1 and -2 expression was analyzed with Sdha as the reference in 3 independent experiments performed in triplicate (mean ± S.E.M.). Data obtained at 0.1 mM glucose were set to 1.0. In (A), asterisks indicate significant differences calculated on the basis of Bonferroni-corrected p-values of the entire data set obtained by RNA-seq. In (B), asterisks indicate significant differences according to two-way ANOVA followed by Tukey’s post hoc test. (C) IRS-1 expression was analyzed after transfection of a control or a specific siRNA against CRTC-2, IRS-1 expression by ELISA. Data of 3 independent experiments (N = 3) performed in quadruplicate are shown as the mean ± S.E.M. Asterisks indicate significant differences according to two-way ANOVA followed by Tukey’s post hoc test. (D,E) Cells were transfected with an SRF/TCF-dependent-reporter gene construct, cultured for 24 h either with 0.1 or with 2.5 mM glucose, stimulated or not with IGF or EGF (both 100 nM) for 4 h and with luciferase activity determined. In (D), basal reporter activity is shown normalized to the basal obtained at 0.1 mM (100%). In (E), the corresponding basal value was set to 100% and GF-induced reporter activation was calculated as the percentage of basal. Data of 4 independent experiments (N = 4) performed in triplicate are expressed as the mean ± S.E.M. Asterisks indicate significant differences according to two-sampled t-tests (D) or two-way ANOVA followed by Tukey’s post-hoc test (E). In (F), GT1-7 cells were transfected with an SRF/TCF-dependent-reporter gene construct, cultured for 24 either with 0.1 or with 2.5 mM glucose, stimulated or not with IGF or EGF (both 100 nM) for 4 h and with luciferase activity determined. Data of 4 independent experiments (N = 4) performed in triplicate are expressed as the mean ± S.E.M. Asterisks indicate significant differences according to two-way ANOVA followed by Tukey’s post hoc test. ** p < 0.01, and *** p < 0.001.

Figure 8

Glucose enhanced AC-6 expression and NA-induced cAMP accumulation in hypothalamic cells. mHypoA-2/10 cells were stimulated accordingly to protocol 1 (A) or 2 (B–F) illustrated in Section 2.5. RPKM were determined by RNA-seq in (A). For each condition, triplicates were analyzed as described under experimental procedures. Only genes with at least one RPKM ≥ 1 were considered. Data are expressed as the mean ± S.E.M. Asterisks indicate significant differences calculated on the basis of Bonferroni-corrected p-values of the entire data set obtained by RNA-seq. (B) After transfection of a control or a specific siRNA against CRTC-2, AC-6 expression was determined by ELISA. Data of 3 independent experiments performed in quadruplicates are shown as the mean ± S.E.M. Asterisks indicate significant differences according to two-way ANOVA followed by Tukey’s post-hoc test. (C–F) Cells were labeled with [³H]-adenine, cultured for 24 h either with 0.1 or with 2.5 mM glucose, stimulated or not with 10 µM FSK or noradenaline (NA) for 20 min, and [³H]-cAMP and [³H]-ATP were determined. (C) Basal [³H]-ATP and (D) basal [³H]-cAMP levels are shown. In (E), ligand-induced signals of [³H]-cAMP signals are shown as x-fold over basal. In (F), ligand-induced signals at 0.1 mM glucose were set to 100%, and ligand-induced signals at 2.5 mM were calculated as the percentage. Data of 4 independent experiments (N = 4) performed in triplicate are expressed as the mean ± S.E.M. Asterisks indicate significant differences according to two-sampled t-tests (D,F) or two-way ANOVA followed by Tukey’s post hoc test (E). * p < 0.05, ** p < 0.01, and *** p < 0.001.