To eat or not to eat: ontogeny of hypothalamic feeding controls and a role for leptin in modulating life-history transition in amphibian tadpoles

Abstract

Many animal life histories entail changing feeding ecology, but the molecular bases for these transitions are poorly understood. The amphibian tadpole is typically a growth and dispersal life-history stage. Tadpoles are primarily herbivorous, and they capitalize on growth opportunities to reach a minimum body size to initiate metamorphosis. During metamorphic climax, feeding declines, at which time the gastrointestinal (GI) tract remodels to accommodate the carnivorous diet of the adult frog. Here we show that anorexigenic hypothalamic feeding controls are absent in the tadpole, but develop during metamorphosis concurrent with the production of the satiety signal leptin. Before metamorphosis there is a large increase in leptin mRNA in fat tissue. Leptin receptor mRNA increased during metamorphosis in the preoptic area/hypothalamus, the key brain region involved with the control of food intake and metabolism. This corresponded with an increase in functional leptin receptor, as evidenced by induction of socs3 mRNA and phosphorylated STAT3 immunoreactivity, and suppression of feeding behaviour after injection of recombinant frog leptin. Furthermore, we found that immunoneutralization of leptin in tadpoles at metamorphic climax caused them to resume feeding. The absence of negative regulation of food intake in the tadpole allows the animal to maximize growth prior to metamorphosis. Maturation of leptin-responsive neural circuits suppresses feeding during metamorphosis to facilitate remodelling of the GI tract.

Keywords: feeding behaviour; leptin; metamorphosis; tadpole.

Figure 1.

Changes in BW, intestine weight and length, and intestinal contents in X. laevis tadpoles during metamorphosis. We analysed wet and dry tadpole BW, intestine dry weight, intestine length, intestinal contents and the ratio of intestinal contents to intestine length from early prometamorphosis (NF stage 56) through metamorphic climax (NF stage 64) as described in the Material and methods. Shown are the means ± s.e.m. (n = 6 per developmental stage). We found statistically significant changes in each of the measured parameters by one-way ANOVA (wet BW: F_4,25 = 5.78, p = 0.002; dry BW: F_4,25 = 2.961, p = 0.039; dry intestine weight: F_4,25 = 10.848, p < 0.0001; intestine length: F_4,25 = 56.287, p < 0.0001; intestinal contents: F_4,25 = 11.572, p < 0.0001; intestinal contents/intestine length: F_4,25 = 4.067, p < 0.011). Means with the same letter are not significantly different (p < 0.05; Fisher's LSD test).

Figure 2.

(a) Analysis of lep mRNA in Xenopus oocytes, and through embryogenesis, larval development and metamorphosis. We isolated total RNA from oocytes, embryos and whole tadpoles, and analysed lep mRNA by RTPCR, with cDNAs resolved on a 1% agarose gel and stained with ethidium bromide. The lep mRNA was normalized to mRNA for the reference gene rpL8. The gel shown is representative of 3–4 biological replicates. (b) Densitometric quantification of data shown in (a), beginning at the developmental stage (NF stage 40) when the lep transcript became detectable by RTPCR. The lep mRNA values were normalized to rpL8 mRNA. Band densities were quantified using ImageJ software. Shown are the means ± s.e.m. (n = 3–4 per developmental stage). We observed statistically significant changes in lep mRNA band density throughout metamorphosis (F_7,20 = 11.419, p < 0.0001; ANOVA). Means with the same letter are not significantly different (p < 0.05; Fisher's LSD test). (c) Food restriction of NF stage 50 tadpoles reduced in steady-state lep mRNA in fat and liver (*p < 0.05, Student's unpaired t-test). Shown are the means ± s.e.m. (n = 8 per treatment). As we showed previously [14], the majority of lep mRNA in the tadpole body is contributed by the adipose tissue. (d) Changes in lepr mRNA throughout metamorphosis in three different brain regions of X. laevis tadpoles, measured by RTqPCR. We saw statistically significant changes during metamorphosis in lepr mRNA in the POA/hypothalamus during metamorphosis (F_4,20 = 9.975, p < 0.0001 ANOVA), but not in the olfactory bulbs/forebrain or hindbrain/spinal cord regions. Shown are the means ± s.e.m. (n = 5 per developmental stage).

Figure 3.

Ontogeny of functional leptin signalling in tadpole brain measured by changes in socs3 mRNA and food intake after i.c.v. injection of rxLeptin. (a) Levels of socs3 mRNA 2 h after i.c.v. injection of 0.6% saline or rxLeptin (20 ng g⁻¹ × BW) in tadpole POA/hypothalamus at different stages of metamorphosis. We extracted total RNA from tadpole brain and measured socs3 mRNA by SYBR Green RTqPCR, and normalized it to the level of the reference gene rpL8, which was not affected by rxLeptin injection (data not shown). Bars represent the mean ± s.e.m., n = 4–5 per developmental stage. There were significant effects of developmental stage (F_4,36 = 10.48, p < 0.0001) and treatment (F_1,36 = 64.602, p < 0.0001), and developmental stage by treatment interaction (F_4,36 = 6.086, p = 0.001; two-way ANOVA). Asterisks indicate statistically significant differences between saline and rxLeptin-injected groups within a developmental stage (p < 0.0001, Fisher's LSD). (b) The anorexigenic action of leptin develops during tadpole metamorphosis. We gave i.c.v. injections of 0.6% saline or rxLeptin (20 ng g⁻¹ × BW) at two stages of tadpole metamorphosis, premetamorphosis (NF stage 50) or late prometamorphosis (NF stage 58) and measured feeding behaviour as described in the Material and methods. We conducted 3–4 feeding trials per developmental stage and injection treatment, with each trial having four tanks per treatment (10 tadpoles per tank; the replicate was feeding trial; points represent the mean ± s.e.m.; n = 3–4 per time point). Tadpoles were randomly distributed throughout the tank before addition of food to quadrant A. They remained randomly distributed over the 15 min measurement period if no food was added (the positions of animals in the tank were scored at 1 min intervals) (p > 0.05; linear regression: least squares). After the addition of food to quadrant A, there were more tadpoles in this quadrant over the 15 min measurement period in the saline-injected NF stage 50 (p < 0.0001) and NF stage 58 (p < 0.0001), and rxLeptin-injected NF stage 50 (p < 0.0001) treatments compared to the no food treatments (see electronic supplementary material, figure S7). By contrast, rxLeptin treatment of NF stage 58 tadpoles strongly suppressed the directed movement into quadrant A after the addition of food; that is, there was no difference between the no food and food added curves (p = 0.897).

Figure 4.

Passive immunization with antiserum to rxLeptin caused metamorphic climax stage tadpoles to resume feeding. We gave NF stage 58 tadpoles four intraperitoneal injections (10 µl each) of rabbit pre-immune serum or rabbit antiserum to rxLeptin on days 1, 3, 5 and 7. During the experiment tadpoles were fed both tadpole food suspension and metamorphic frog pellets ad libitum, then the animals were killed on day 7, 4 h after the last injection. We dissected the entire GI tract (stomach and intestines) and measured the ratio of GI tract contents to final dry GI tract weight as described in the Material and methods. Bars indicate the mean + s.e.m., n = 8 per treatment. Asterisks indicate p < 0.05 (Student's unpaired t-test).