A missense mutation in zinc finger homeobox-3 (ZFHX3) impedes growth and alters metabolism and hypothalamic gene expression in mice

Abstract

A protein altering variant in the gene encoding zinc finger homeobox-3 (ZFHX3) has recently been associated with lower BMI in a human genome-wide association study. We investigated metabolic parameters in mice harboring a missense mutation in Zfhx3 (Zfhx3^Sci/+ ) and looked for altered in situ expression of transcripts that are associated with energy balance in the hypothalamus to understand how ZFHX3 may influence growth and metabolic effects. One-year-old male and female Zfhx3^Sci/+ mice weighed less, had shorter body length, lower fat mass, smaller mesenteric fat depots, and lower circulating insulin, leptin, and insulin-like growth factor-1 (IGF1) concentrations than Zfhx3^+/+ littermates. In a second cohort of 9-20-week-old males and females, Zfhx3^Sci/+ mice ate less than wildtype controls, in proportion to body weight. In a third cohort of female-only Zfhx3^Sci/+ and Zfhx3^+/+ mice that underwent metabolic phenotyping from 6 to 14 weeks old, Zfhx3^Sci/+ mice weighed less and had lower lean mass and energy expenditure, but fat mass did not differ. We detected increased expression of somatostatin and decreased expression of growth hormone-releasing hormone and growth hormone-receptor mRNAs in the arcuate nucleus (ARC). Similarly, ARC expression of orexigenic neuropeptide Y was decreased and ventricular ependymal expression of orphan G protein-coupled receptor Gpr50 was decreased. We demonstrate for the first time an energy balance effect of the Zfhx3^Sci mutation, likely by altering expression of key ARC neuropeptides to alter growth, food intake, and energy expenditure.

Keywords: Gpr50; IGF1; appetite; atbf1; energy balance; growth; mouse model; somatostatin; zfhx3.

FIGURE 1

One year old Zfhx3 ^Sci/+ mice are smaller than wildtype littermates. Male and female mice weigh less (A) and have shorter nose‐tail body length (B) and lower lean (C) and fat mass (E) although this is proportionate to body weight (D, F). Body weight/body length² was significantly lower in Zfhx3 ^Sci/+ mice (G). Fat depot weights at dissection did not differ by genotype, except for mesenteric white adipose, which weighed less in Zfhx3 ^Sci/+ mice (H–L). Plotted are mean ± SEM with individual values. Statistical comparison is by 2‐way ANOVA, with overall results and within‐sex comparisons indicated on the graphs. ****p < .0001, ***p < .001, **p < .01, *p < .05. BW, body weight; gWAT, gonadal white adipose tissue; iBAT, interscapular brown adipose tissue; iWAT, inguinal white adipose tissue; mWAT, mesenteric white adipose tissue; prWAT, perirenal white adipose tissue.

FIGURE 2

One year old Zfhx3 ^Sci/+ mice have lower circulating concentrations of metabolic hormones. Circulating leptin was lower in fasted female Zfhx3 ^Sci/+ mice (A), however this is proportionate to fat mass (g, EchoMRI; B). Circulating insulin (C) and insulin‐like growth factor‐1 (IGF1, D) were also lower in Zfhx3 ^Sci/+ mice. Plotted are mean ± SEM with individual values. Statistical comparison is by 2‐way ANOVA, with overall results and within‐sex comparisons indicated on the graphs. ****p < .0001, ***p < .001, **p < .01, *p < .05.

FIGURE 3

Male and female Zfhx3 ^Sci/+ mice eat less than wildtype littermates. Mice were pair‐housed (genotype and sex matched) and food weighed weekly (A; n = 5–6 cages). Male mice ate more than female mice from the second week of the experiment, and Zfhx3 ^+/+ ate more than Zfhx3 ^Sci/+ mice from the second week of the experiment. Weekly body weight revealed that the Sci mutation already inhibited weight gain from 9 weeks of age overall, with female Zfhx3 ^Sci/+ mice weighing less than wildtype littermates from 10 weeks old onwards (B; n = 10–12). Although the genotype effect was not maintained from 14 weeks old onwards, a significant interaction could be detected at 15 and 18 weeks old. When energy intake was calculated per gram bodyweight (paired mouse weight) no significant differences were detected (C). Plotted are mean ± SEM, statistical comparison is by 2‐way ANOVA per timepoint, and within‐sex differences are indicated on the graph. ♀ p < .05 Zfhx3 ^Sci/+ vs Zfhx3 ^+/+ female, ♂p < .05 Zfhx3 ^Sci/+ vs Zfhx3 ^+/+ male, ***p < .001, **p < .01, ^† p < .05 between sexes, ^††† p < .001 between sexes, ^# p < .05 genotype—sex interaction. Full statistics are listed in Table S1.

FIGURE 4

Early weight gain of female Zfhx3 ^Sci/+ mice is lower than wildtype littermates, and driven by lower lean mass. The Sci mutation in female mice inhibits overall weight gain (A), composed of lean mass (B), but no differences in fat mass (C) or % fat mass (D). At 12 weeks old, these mice underwent indirect calorimetry and the second 24 h of measurements were compared to reveal a trend for lower 24 h RER and significantly reduced dark phase RER (E, G), while energy expenditure was reduced in Zfhx3 ^Sci/+ mice (time course, F; lean mass ANCOVA adjusted, H). Plotted are mean ± SEM (A–H) with individual values (G‐H). Statistical comparisons are by mixed effects model (REML) with Šídák's multiple comparisons (A–D), RM‐2‐way ANOVA (E, F) and unpaired t‐tests (G, H). A–D, n = 21–23; E–H n = 15–19. ****p < .0001, ***p < .001, **p < .01, *p < .05, NSD, no significant differences.

FIGURE 5

Hypothalamic expression of growth axis component genes is altered by the Sci mutation. Expression of growth hormone releasing hormone (Ghrh) is lower in Zfhx3 ^Sci/+ mice arcuate nucleus (ARC, A) but not posterior paraventricular nucleus (PoPVN, D). Expression of somatostatin (Sst) is increased in Zfhx3 ^Sci/+ mice in the ARC (B) but not periventricular nucleus (PeVN, E). Growth hormone receptor (Ghr) expression was lower in Zfhx3 ^Sci/+ mice and at zeitgeber time (ZT) 3 in the ARC (C) but expression did not differ between groups in the paraventricular nucleus (PVN, F). Plotted are mean ± SEM with individual values overlaid. Example images are shown beneath each plot. Comparisons are by 2‐way ANOVA with Šídák's multiple comparisons tests indicated where appropriate. *p < .05, **p < .01, NSD, no significant differences. N = 10–12.

FIGURE 6

Additional hypothalamic gene expression is altered by the Sci mutation. Expression of Bmal1 in the suprachiasmatic nucleus (SCN) was increased at ZT 15 overall, with an interaction between ZT and genotype, and reduced expression in Zfhx3 ^Sci/+ mice at ZT 15 only (A). Expression of Pomc in the arcuate nucleus (ARC) approached a significant reduction in Zfhx3 ^Sci/+ mice (B), and ARC expression of Npy was overall reduced in Zfhx3 ^Sci/+ mice (C). Expression of Gpr50 was increased overall in the ventricular ependymal layer, with pairwise comparisons reaching significance at ZT 15 (D). Plotted are mean ± SEM with individual values overlaid. Representative images are shown beneath each plot. Comparisons are by 2‐way ANOVA with Šídák's multiple comparison tests indicated where appropriate. Npy data were log transformed for statistical comparison. *p < .05, **p < .01, ***p < .001. NSD, no significant differences. N = 10–12.