Characterization of female adipose-specific ghr knockout mice

We previously created an adipose-specific Ghr gene knockout (KO) mouse model to study the effects of adipose-specific Ghr deficiency on the growth, metabolism and tissue function of male KO mice in the F2 generation [22]. Here, we focused on female KO mice in the F2 generation and female offspring in the F3 generation (Fig. 1A). The bodyweight of female KO mice in the F2 generation had no significant different with LL mice (Fig. 1B). Compared with those in the LL mice, the weights of different adipose depots in the female KO mice were greater (Fig. 1C) (n = 7 females, multiple t test; BAT: t = 2.889, p = 0.013; sWAT: t = 7.447, p < 0.001; gWAT: t = 4.539, p < 0.001; pWAT: t = 4.144, p = 0.0013; mWAT: t = 3.380, p = 0.0023), whereas the weight of the heart was lower (n = 7 females, multiple t test; Heart: t = 3.046, p = 0.01). No change in the weight of the kidney or liver was observed (Fig. 1D) (n = 7 females, multiple t test; Kidney: t = 0.5378, p = 0.6005; Liver: t = 0.5246, p = 0.6094).The mass of fat depots and other organs were corrected by body weight (Figure S1A and S1B) (Figure S1A: n = 7 females, multiple t test; BAT: t = 2.441, p = 0.031; sWAT: t = 6.616, p < 0.001; gWAT: t = 4.028, p = 0.0012; pWAT: t = 3.793, p = 0.0025; mWAT: t = 3.244, p = 0.007) (Figure S1B: n = 7 females, multiple t test; Heart: t = 3.239, p = 0.0071; Kidney: t = 0.6048, p = 0.5566; Liver: t = 1.320, p = 0.2116). Serological analysis indicated that adipose-specific Ghr deficiency did not affect the circulating levels of GH or IGF-1 or fasting insulin in female mice (Fig. 1E-G) (n = 5 females, two-tailed unpaired t test; GH: t = 0.08473, p = 0.9346; IGF-1: t = 0.2781, p = 0.7880; Insulin: t = 1.313, p = 0.2257), but female KO mice had lower fasting blood glucose (Fig. 1H) (n = 7–9 females, two-tailed unpaired t test; t = 3.389, p = 0.005) and adiponectin levels (Fig. 1I) (n = 5 females, two-tailed unpaired t test; t = 4.950, p = 0.001). Glucose tolerance tests (GTTs) and insulin tolerance tests (ITTs) revealed no difference between female KO mice and control littermates (Fig. 1J and K). Overall, these data indicated that Ghr deficiency in female mice has no effect on body weight. However, it can increase adipose tissue mass and improve fasting blood glucose levels.

Fig. 1

Characterization of female adipose-specific Ghr knockout mice. A: Schematic view of the mouse lines generation. B: Growth curves of female KO and LL mice after 24 weeks of RC feeding (n = 7 mice/group). C: Fat mass of 24-week-old female KO and LL mice (n = 7 mice/group). D: Organ mass of 24-week-old female KO and LL mice (n = 7 mice/group). E-I: Serum GH (E), IGF-1 (F), fasting insulin (G), blood glucose (H) and adiponectin (I) levels of 24-week-old female LL and KO mice (n = 5–9 mice/group). J: Glucose tolerance tests (GTTs) of 24-week-old female KO and LL mice (n = 7 mice/group). K: Insulin tolerance tests (ITTs) of 24-week-old female KO and LL mice (n = 7 mice/group)

All the mice were studied at 24 weeks of age unless otherwise stated. Adipoq-Cre: Tg(Adipoq-cre)1Evdr (JAX stock #010803); LL: Ghrflox/flox mice; KO: adipose-specific Ghr knockout mice; L-LL: LL offspring of LL maternal parent; L-KO: KO offspring of LL maternal parent; K-LL: LL offspring of KO maternal parent; K-KO: KO offspring of KO maternal parent. BAT: brown adipose tissue; WAT: white adipose tissue; sWAT: subcutaneous WAT; gWAT: gonadal WAT; pWAT: perigonadal WAT; mWAT: mesenteric WAT. GH: growth hormone; IGF-1: insulin-like growth factor 1. All the values are presented as the means ± SEMs. The statistical significance was indicated by *p < 0.05, **p < 0.01 and ***p < 0.001.

Adipose-specific ghr deficiency impairs fertility in female KO mice

Global GH deficiency could modify follicular development, ovarian maturation, ovulation rate, sexual maturation and pregnancy. To investigate the effects of adipose-specific Ghr deficiency on reproduction, we mated male LL mice and male KO mice with female LL or female KO mice respectively. The parental ages were 8–10 weeks. Our results showed that whether mated female KO mice with male LL or male KO mice, the reproductive function and the fertility rate were reduced significantly in female KO mice, and the number of pups per litter was lowest for the KO and KO mating pairs (Two-tailed unpaired t test; ♂LL x ♀LL vs. ♂LL x ♀KO: t = 6.453, p < 0.001; ♂KO x ♀LL vs. ♂KO x ♀KO: t = 3.984, p < 0.001). However, when the female LL mice mated with male LL or male KO mice, the litter size was unchanged (Two-tailed unpaired t test; t = 1.467, p = 0.153), which indicates that adipose-specific Ghr deficiency impairs fertility in female KO mice while the reproductive ability of male KO mice was not significantly affected (Fig. 2A). When the Adipo-Cre transgene is hemizygous in KO paternal mice, the LL maternal parent may give birth to L-LL and L-KO mice, while the KO maternal parent may give birth to K-LL and K-KO mice (Fig. 1A, F2 to F3). Therefore, we mated male KO mice with female LL or female KO mice to investigate the effects of different maternal genotypes on reproduction and offspring in the subsequent experiments. We assessed the reproductive organ weight and histomorphology of the ovaries of 10-week-old female mice. The weight of the uterus did not differ between female LL and KO mice (Fig. 2B) (n = 6 female mice, two-tailed unpaired t test; t = 1.259, p = 0.2366). However, compared with those in LL mice, the weight of the ovaries in female KO mice was significantly lower (Fig. 2C and D) (n = 6 female mice, two-tailed unpaired t test; t = 2.669, p = 0.0235). Follicular quantification showed a reduced percentage of primary follicles (PF, two-tailed unpaired t test; t = 6.016, p < 0.001) and secondary follicles (SF, two-tailed unpaired t test; t = 4.333, p = 0.002) in female KO mice while the number of mature follicles (MF, two-tailed unpaired t test; t = 0.5898, p = 0.5716) and corpus luteum (CL, two-tailed unpaired t test; t = 0.1741, p = 0.8661) has no significantly difference between LL and KO mice (Fig. 2E and F). These findings indicate that fertility defects in female KO mice are likely due to underdeveloped ovaries and impaired early follicular development. We measured the serum follicle stimulating hormone (FSH), estradiol (E2) and luteinizing hormone (LH) levels in 10-week-old and 16-week-old female mice separately to see whether sex hormone affect the fertility function of KO mice. We found that the E2 and FSH levels in 10-week-old female KO mice were significantly lower than those in LL mice (Fig. 2G and H) (n = 5 females, two-tailed unpaired t test; E2: t = 2.794, p = 0.0234; FSH: t = 2.579, p = 0.0327), while LH showed no significant changes (Fig. 2I) (n = 5 females, two-tailed unpaired t test; t = 1.093, p = 0.306). Female KO mice at 16 weeks of age still have lower FSH levels than LL mice (Fig. 2K) (n = 5 females, two-tailed unpaired t test; t = 2.800, p = 0.0232) but there is no significant difference in E2 and LH levels (Fig. 2J and L) (n = 5 females, two-tailed unpaired t test; E2: t = 1.484, p = 0.1762; LH: t = 1.118, p = 0.2959). These results are consistent with the fertility rate and histological staining, indicating that a decrease in E2 and FSH levels are probably a reason for the reduction of the reproductive function and the fertility rate in female KO mice, although the exact mechanism involved in the reduction of sex hormone remains unclear.

Fig. 2

Adipose-specific Ghr deficiency impairs fertility in female KO mice. A: Evaluation of fertility in KO and LL mice. B and C: Uterine (B) and ovarian (C) weights of 10-week-old female KO and LL mice (n = 6 mice/group). D: Representative image of uterus and ovaries from 10-week-old female LL and KO mice. Scale bar: 1 cm. E: Representative photos of ovarian sections stained with H&E. F: Follicular quantification in each stage of 10-week-old female LL and KO mice. Follicles were counted on serial ovarian sections after H&E staining. G-I: Serum E2 (G), FSH (H) and LH (I) levels of 10-week-old female LL and KO mice (n = 5 mice/group). J-L: Serum E2 (J), FSH (K) and LH (L) levels of 16-week-old female LL and KO mice (n = 5 mice/group). PF: primary follicle; SF: secondary follicle; MF: mature follicle, and CL: corpus luteum. E2: estradiol; FSH: follicle stimulating hormone; LH: luteinizing hormone. All the values are presented as the means ± SEMs. The statistical significance was indicated by *p < 0.05, **p < 0.01 and ***p < 0.001.

Maternal adipose-specific ghr deficiency increases body fat mass and adipocyte size in female offspring

GH plays a pivotal role in directing postnatal growth and regulating fat metabolism. We analyzed the body weight and fat mass of the offspring of KO and LL mothers. In the postnatal growth phase, the body weight of female K-KO mice (KO offspring of KO mothers) was 15.6% lower than that of female L-KO mice (KO offspring of LL mothers) at 3 weeks of age without significant difference (Fig. 3A) (One-way ANOVA; F (3, 18) = 3.550, p = 0.0354) (post hoc comparison, Tukey’s multiple comparisons test, ***p < 0.001). The body weight of the female L-KO mice was significantly greater than that of the female L-LL mice (LL offspring of LL mothers, black asterisk) and female offspring of the KO mothers (K-LL and K-KO, orange asterisk) during postnatal growth. Although there was no difference in body weight among four groups in 24-week old (Fig. 3B) (n = 4–5, two-way ANOVA; F (3, 84) = 23.63; p < 0.0001) (Post hoc comparison, Tukey’s multiple comparisons test, *p < 0.05 from week 8). However, we found that the maternal genotype did not affect bodyweight in male offspring (Figure S2A) (Figure S2A: n = 4–5, two-way ANOVA; F (3, 89) = 0.3895; p = 0.9786) (Post hoc comparison, Tukey’s multiple comparisons test). The white adipose tissue (WAT) of the KO mice was significantly heavier than that of the LL mice in the two groups of female offspring (Fig. 3C) (n = 4–5 females, one-way ANOVA for each fat depots; BAT: F (3, 14) = 3.353, p = 0.05; sWAT: F (3, 14) = 21.23, p < 0.0001; gWAT: F (3, 14) = 11.21, p < 0.0001; pWAT: F (3, 14) = 8.327, p = 0.002; mWAT: F (3, 14) = 6.911, p = 0.004) (post hoc comparison, Tukey’s multiple comparisons test, *p < 0.05, **p < 0.01 and ***p < 0.001). Fat index corrected by body weight showed the same result (Figure S1C) (n = 4–5 females, one-way ANOVA for each fat depots; BAT: F (3, 14) = 2.572, p = 0.095; sWAT: F (3, 14) = 17.57, p < 0.0001; gWAT: F (3, 14) = 11.38, p = 0.0005; pWAT: F (3, 14) = 6.553, p = 0.0054; mWAT: F (3, 14) = 6.495, p = 0.0056) (post hoc comparison, Tukey’s multiple comparisons test, *p < 0.05, **p < 0.01 and ***p < 0.001). Histology analysis revealed a significant increase in adipocyte size in L-KO and K-KO mice (Fig. 3D). The percentage of subcutaneous WAT (sWAT) adipocytes in L-KO and K-KO mice was 49.5% and 51.6% greater, respectively, than that in littermate LL offspring (Fig. 3E) (One-way ANOVA; F (3, 551) = 203.6, p < 0.0001) (post hoc comparison, Tukey’s multiple comparisons test, ***p < 0.001). However, the percentages of gonadal WAT (gWAT) were 44.2% and 34%, respectively (Fig. 3F) (One-way ANOVA; F (3, 558) = 134.3, p < 0.0001) (post hoc comparison, Tukey’s multiple comparisons test, ***p < 0.001). Overall, these results indicated that a lack of Ghr in adipose tissue may increase fat mass and adipocyte size in mice fed regular chow, consistent with our previous study in male mice [22].

Fig. 3

Maternal adipose-specific Ghr deficiency increases body fat mass and adipocyte size in female offspring fed regular chow. A: Body weights of 3-week-old female offspring of KO and LL maternal mice (n = 5–6 mice/group). B: Growth curves of female offspring of KO and LL maternal mice during RC feeding (n = 4–5 mice/group). C: Fat tissue weight of female offspring of KO and LL maternal mice under RC feeding (n = 4–5 mice/group). D: H&E staining of WAT sections. Scale bar: 100 μm. E and F: Quantification of the adipocyte area in sWAT (F) and gWAT (G). The average area of adipocytes was analyzed using ImageJ software on H&E-stained slides from at least three fields of three mice per genotype. All the values are presented as the means ± SEMs. The statistical significance was indicated by *p < 0.05, **p < 0.01 and ***p < 0.001.

The female offspring of KO mothers are resistant to dietary obesity, which improves glucose homeostasis

In our previous study [22], we found that the absence of Ghr in fat exacerbated diet-induced obesity in male KO mice, and the maternal genotype did not affect this outcome in male offspring (Figure S2B and S2C) (Figure S2B: n = 6–8, two-way ANOVA; F (3, 125) = 12.43, p < 0.0001) (Post hoc comparison, Tukey’s multiple comparisons test, *p < 0.05 from week 20) (Figure S2C: One-way ANOVA; F (3, 24) = 9.377, p = 0.0003) (post hoc comparison, Tukey’s multiple comparisons test, *p < 0.05 and **p < 0.01). Consistent with this, when the mothers were LL mice, the body weights of the L-LL and L-KO female mice were significantly greater after with the mice were fed a HF diet, while the L-KO female mice were more obese (Fig. 4A and B). In contrast, when the matrilineal parent was a KO mouse, both K-LL and K-KO female mice were resistant to dietary obesity (Fig. 4A and B) (Fig. 4A: n = 4–5, two-way ANOVA; F(3, 135) = 106.1; p < 0.0001) (Post hoc comparison, Tukey’s multiple comparisons test, *p < 0.05 from week 12) (Fig. 4B: n = 4–5, one-way ANOVA; F (3, 15) = 18.89, p < 0.0001) (post hoc comparison, Tukey’s multiple comparisons test, *p < 0.05, **p < 0.01 and ***p < 0.001). After feeding a HF diet for 8 weeks, the fasting blood glucose levels have no significantly different in four groups (K-KO, 78.12 ± 6.44 mg/dL; K-LL, 98.1 ± 11.89 mg/dL; L-LL, 108 ± 27.9 mg/dL; L-KO, 97.56 ± 22.24 mg/dL) (Fig. 4C) (n = 4–5, one-way ANOVA; F (3, 15) = 2.057, p = 0.1491). GTTs and ITTs revealed that female K-KO mice were more responsive to glucose challenge (Fig. 4D) (n = 4–5, one-way ANOVA; F (3, 15) = 4.237, p = 0.0234) (post hoc comparison, Tukey’s multiple comparisons test, *p < 0.05) and more sensitive to insulin stimulation than L-KO mice (Fig. 4E) (n = 4–5, one-way ANOVA; F (3, 14) = 4.154, p = 0.0267) (post hoc comparison, Tukey’s multiple comparisons test, *p < 0.05). After prolonged HF diet induction for 16 weeks, the fasting blood glucose levels of K-LL and K-KO female mice were lower than those of L-LL and L-KO female mice (K-LL, 91.44 ± 16.76 mg/dL; K-KO, 95.04 ± 18.51 mg/dL; L-LL, 140.04 ± 27.58 mg/dL; L-KO, 133.2 ± 6.06 mg/dL) (Fig. 4F) (n = 4–5, one-way ANOVA; F (3, 15) = 7.361, p = 0.0029) (post hoc comparison, Tukey’s multiple comparisons test, *p < 0.05 and **p < 0.01). In addition, K-KO female mice responded better to the glucose challenge (Fig. 4G) (n = 4–5, one-way ANOVA; F (3, 15) = 5.742, p = 0.008) (post hoc comparison, Tukey’s multiple comparisons test, **p < 0.01) despite not exhibiting an advantage in insulin sensitivity (Fig. 4H) (n = 4–5, one-way ANOVA; F (3, 15) = 2.322, p = 0.1165). However, we found that the maternal genotype did not affect the fasting blood glucose levels in male offspring with HF diet (Figure S2D and S2E) (Figure S2D: n = 6 males, One-way ANOVA; F (3, 20) = 17.51, p < 0.0001) (post hoc comparison, Tukey’s multiple comparisons test, *p < 0.05, **p < 0.01 and ***p < 0.001) (Figure S2E: n = 6 males, One-way ANOVA; F (3, 20) = 7.810, p = 0.0012) (post hoc comparison, Tukey’s multiple comparisons test, *p < 0.05 and **p < 0.01). These data suggest that maternal adipose Ghr disruption has a beneficial effect only on body weight and glucose metabolism in female offspring fed a HF diet.

Fig. 4

Female offspring of KO mothers are resistant to diet-induced obesity and glucose homeostasis is improved. A: Growth curves of female offspring of KO and LL maternal mice during HF feeding (n = 4–5 mice/group). B: Increase in body weight (24-week vs. 8-week) of female offspring of KO and LL maternal mice (n = 4–5 mice/group). C: Fasting blood glucose levels of female offspring mice fed a high-fat diet for 8 weeks (n = 4–5 mice/group). D and E: GTTs (D) and ITTs (E) results for female offspring mice fed a high-fat diet for 8 weeks (n = 4–5 mice/group). Inset: The area under the curve. F: Fasting blood glucose levels of female offspring mice fed a high-fat diet for 16 weeks (n = 4–5 mice/group). G and H: GTTs (G) and ITTs (H) results for female offspring mice fed a high-fat diet for 16 weeks (n = 4–5 mice/group). Inset: The area under the curve. All the values are presented as the means ± SEMs. The statistical significance was indicated by *p < 0.05, **p < 0.01 and ***p < 0.001.

Maternal adipose-specific ghr deficiency inhibits adipose hyperplasia and alleviates hepatic steatosis and hyperlipidemia in female offspring

To evaluate the effect of maternal genotype on adipose tissue parameters, we compared fat mass among four different offspring. Consistent with the body weight, fat mass across various locations was significantly greater in the L-LL and L-KO female mice after HF induction, while fat mass was less increased in the K-LL and K-KO female mice (Fig. 5A) (n = 4–5 females, one-way ANOVA for each fat depots; BAT: F (3, 15) = 12.44, p = 0.0002; sWAT: F (3, 14) = 24.46, p < 0.0001; gWAT: F (3, 14) = 7.252, p = 0.0036; pWAT: F (3, 14) = 10.23, p = 0.0008; mWAT: F (3, 13) = 6.908, p = 0.0051) (post hoc comparison, Tukey’s multiple comparisons test, *p < 0.05, **p < 0.01 and ***p < 0.001). Because the body weights of the L-LL and L-KO female mice were significantly greater than K-LL and K-KO female mice under HF diet (Fig. 4A). We corrected the weight of fat depots with bodyweight. As shown in figure S1D, the fat index of K-LL and K-KO mice were still significantly lower than L-LL and L-KO female mice (n = 4–5 females, one-way ANOVA for each fat depots; BAT: F (3, 14) = 4.209, p = 0.0256; sWAT: F (3, 14) = 13.29, p = 0.0002; gWAT: F (3, 14) = 4.325, p = 0.0235; pWAT: F (3, 14) = 7.03, p = 0.0041; mWAT: F (3, 14) = 11.04, p = 0.0006) (post hoc comparison, Tukey’s multiple comparisons test, *p < 0.05, **p < 0.01 and ***p < 0.001). We previously showed that adipose-specific Ghr deficiency leads to lipid accumulation in brown adipose tissue (BAT) and impaired cold tolerance [22]. Here, we found lipid droplet expansion and accumulation in the BAT of female L-LL and L-KO mice; however, the degree of whitening in female K-LL and K-KO mice was markedly reduced (Fig. 5B). In addition, the expression levels of thermogenic genes, such as Ucp1 and Pgc1α, were decreased in the BAT of female L-KO HF mice. Compared to that in the offspring from the LL mothers, the expression of thermogenic genes was upregulated in female K-LL and K-KO mice (Fig. 5C and D) (One-way ANOVA; Ucp1: F (3, 8) = 53.53, p < 0.0001; Pgc1α: F (3, 8) = 9.608, p = 0.005) (post hoc comparison, Tukey’s multiple comparisons test, *p < 0.05, **p < 0.01 and ***p < 0.001). Obesity in both Ghr total knockout and adipose-specific Ghr knockout mice is attributed to fat expansion in terms of both the number and size of adipocytes [22, 32,33,34,35]. Therefore, we examined the histopathological features of sWAT and gWAT in these mice. As expected, the adipocytes of female L-KO and K-KO mice were significantly larger than those of female L-LL and K-LL mice fed a HF diet (Fig. 5E). Although fat mass was decreased in K-LL and K-KO mice, surprisingly, no significant difference in adipocyte size was observed between the offspring of the LL and KO matrilineal parents (Fig. 5E and G) (One-way ANOVA; sWAT: F (3, 420) = 33.29, p < 0.0001; gWAT: F (3, 401) = 27.66, p < 0.0001) (post hoc comparison, Tukey’s multiple comparisons test, ***p < 0.001). Thus, the reduction in adipose tissue mass in female K-LL and K-KO mice resulted from a decrease in the number of adipocytes rather than the size of adipocytes.

Fig. 5

Maternal adipose-specific Ghr deficiency inhibits adipose hyperplasia during high-fat diet feeding. A: Fat tissue weight of female offspring of KO and LL maternal mice after high-fat diet feeding for 16 weeks (n = 4–5 mice/group). B: H&E staining of BAT sections. Scale bar: 100 μm. CsD: Relative mRNA expression levels of Ucp1 (C) and Pgc1a (D) in BAT (n = 3 mice/group). E: H&E staining of sWAT and gWAT sections. Scale bar: 100 μm. F and G: Quantification of the adipocyte area in sWAT (F) and gWAT (G). Ucp1: uncoupling protein 1; Pgc1α: peroxisome proliferator-activated receptor gamma coactivator 1 alpha. The average area of adipocytes was analyzed using ImageJ software on H&E-stained slides from at least three fields of three mice per genotype. All the values are presented as the means ± SEMs. The statistical significance was indicated by *p < 0.05, **p < 0.01 and ***p < 0.001.

As a major site for lipogenesis and lipid oxidation, the liver is the central engine modulating whole-body metabolic homeostasis and is vulnerable to nutritional damage. Organ weight analysis revealed reduced weights of the liver in female K-KO mice than L-LL and L-KO mice fed a HF diet (Fig. 6A) (n = 4–5 females, one-way ANOVA F (3, 14) = 7.310, p = 0.0035) (post hoc comparison, Tukey’s multiple comparisons test, *p < 0.05 and **p < 0.01). And the liver index of female K-KO mice was also significantly lower than those of L-LL HF and K-LL HF groups (Fig. 6B) (One-way ANOVA; F (3, 14) = 4.544, p = 0.0201) (post hoc comparison, Tukey’s multiple comparisons test, *p < 0.05 and **p < 0.01). Furthermore, compared to L-LL HF group, the liver triglyceride (TG) contents were brought down significantly by 27.5% in L-KO HF, 11.7% in K-LL HF and 38.1% in K-KO HF (Fig. 6C) (One-way ANOVA; F (3, 14) = 15.81, p < 0.0001) (post hoc comparison, Tukey’s multiple comparisons test, *p < 0.05, **p < 0.01 and ***p < 0.001). Histological staining revealed features of hepatic steatosis in the HF diet-fed group. A significant increase of hepatic steatosis marked by visible lipid droplets and ballooning hepatocytes was observed in both groups fed with HFD (Fig. 6D). More visible lipid droplets were observed in female L-LL and L-KO mice. Hepatic lipid accumulation was reduced in the K-LL and K-KO mice, especially in female K-KO mice (Fig. 6D). In the oil red O staining, the lipid droplets are colored into red. Liver tissues of L-LL HF mice and L-KO HF mice both showed abundant red area, further indicating massive lipids accumulation in the liver (Fig. 6D). The release of aspartate aminotransferase (AST) and alanine aminotransferase (ALT) frequently adapted as reflection of hepatocellular integrity. We found a significant reduction in relative activity of both AST and ALT of K-KO mice (Fig. 6E and F) (One-way ANOVA; AST: F (3, 14) = 24.25, p < 0.0001; ALT: F (3, 14) = 21.01, p < 0.0001) (post hoc comparison, Tukey’s multiple comparisons test, *p < 0.05, **p < 0.01 and ***p < 0.001). The influence of maternal adipose-specific Ghr deficiency on lipid profile of female offspring was examined by measuring the serum TG and low-density lipoprotein cholesterol (LDL-C) levels under HF diet condition. The results showed that the serum TG of K-KO HF mice was significantly lower than in L-LL HF mice (Fig. 6G) (One-way ANOVA; F (3, 14) = 4.167, p = 0.0247) (post hoc comparison, Tukey’s multiple comparisons test, *p < 0.05). Similarly, the serum LDL-C levels in K-LL and K-KO mice were also significantly lower than in female offspring of LL mice under HF diet condition (Fig. 6H) (One-way ANOVA; F (3, 14) = 13.78, p = 0.0002) (post hoc comparison, Tukey’s multiple comparisons test, *p < 0.05, **p < 0.01 and ***p < 0.001), confirming a suppressed development of hyperlipidemia in the offspring of KO mice.

Fig. 6

Maternal adipose-specific Ghr deficiency alleviates hepatic steatosis and hyperlipidemia in female offspring fed a high-fat diet. A: Liver mass of female offspring mice fed a high-fat diet for 16 weeks (n = 4–5 mice/group). B: Liver indices (liver weight/body weight%) of female offspring mice fed a HF diet for 16 weeks (n = 4–5 mice/group). C: Liver TG content of female offspring mice fed a HF diet for 16 weeks (n = 4–5 mice/group). D: H&E and Oil red O staining of liver sections from female offspring mice fed a RC or HF diet for 16 weeks. Bar: 200 μm. E-H: Serum ALT (E), AST (F), TG (G) and LDL-C (H) levels of female offspring mice fed a HF diet for 16 weeks (n = 4–5 mice/group). BW: body weight; TG: triglyceride; ALT: alanine aminotransferase; AST: aspartate aminotransferase; LDL-C: low-density lipoprotein cholesterol. All the values are presented as the means ± SEMs. The statistical significance was indicated by *p < 0.05, **p < 0.01 and ***p < 0.001.

Swapping of the matrilineal parent during lactation reverses the phenotype of the female offspring

To verify that the body weight and metabolic health of the offspring are influenced by feeding from different matrilineal parents, we exchanged the mother of the first female filial generation in the early postnatal period. Weanling mice were acclimated for one week and then fed a HF diet until they were 16 weeks old (Fig. 7A). Surprisingly, when the female offspring of LL mice (L-LL and L-KO) were fed by a KO mother (L-LL-K and L-KO-K), they were also resistant to diet-induced obesity (Fig. 7B red line and Fig. 7D) (Fig. 7B: n = 4–6, two-way ANOVA; F(3, 79) = 43.02; p < 0.0001) (Post hoc comparison, Tukey’s multiple comparisons test, *p < 0.05 from week 10) (Fig. 7D: n = 4–6, one-way ANOVA; F (7, 32) = 8.907, p < 0.0001) (post hoc comparison, Tukey’s multiple comparisons test, *p < 0.05, **p < 0.01 and ***p < 0.001). K-LL-L mice (female LL offspring of KO mice fed by a LL mother) remained resistant to dietary obesity even when fed by a LL mother, but the body weight of K-KO-L mice (female KO offspring of KO mice fed by a LL mother) increased significantly (Fig. 7C) (Fig. 7C: n = 4–6, two-way ANOVA; F(3, 78) = 15.58; p < 0.0001) (Post hoc comparison, n = 4–5, Tukey’s multiple comparisons test, *p < 0.05 from week 12). Furthermore, the offspring of KO mothers exhibited reduced body weight gain (Fig. 7D, right side) (Fig. 7D: n = 4–6, one-way ANOVA; F (7, 32) = 8.907, p < 0.0001) (post hoc comparison, Tukey’s multiple comparisons test, *p < 0.05, **p < 0.01 and ***p < 0.001). To further investigate whether swapping feeding sources could influence glucose metabolism in female offspring, fasting blood glucose levels were measured. Blood glucose levels were decreased in K-KO-K mice compared with K-LL-K mice; when the offspring of a KO mouse were fed by a LL mother (K-LL-L and K-KO-L), blood glucose levels were no longer reduced (Fig. 7E) (Fig. 7E: n = 4–7, one-way ANOVA; F (7, 35) = 1.433, p = 0.2238) (post hoc comparison, Tukey’s multiple comparisons test). Furthermore, the blood glucose level in the offspring of a LL mother did not significantly differ among the different mouse groups (Fig. 7E). Consistent with these findings, the GTT and ITT showed that feeding from a KO mother significantly increased the insulin sensitivity of L-LL and L-KO offspring when challenged with a HF diet (Fig. 7F and J) (Fig. 7J: n = 4–6, one-way ANOVA; F (7, 33) = 3.369, p = 0.008) (post hoc comparison, Tukey’s multiple comparisons test, *p < 0.05), but glucose tolerance was not affected (Fig. 7G and K). The insulin sensitivity of K-LL and K-KO offspring was not affected by the genotype of their foster mothers (Fig. 7H and J). However, when the foster mother was a LL mouse, the offspring of the KO mice (K-LL-L and K-KO-L) no longer exhibited any increases in glucose tolerance as K-KO-K mice (Fig. 7I and K) (Fig. 7K: n = 4–6, one-way ANOVA; F (7, 35) = 2.503, p = 0.034) (post hoc comparison, Tukey’s multiple comparisons test, *p < 0.05).

Fig. 7

Swapping of the matrilineal parent during lactation reverses the phenotype of the female offspring. A: Schematic diagram of KO and LL mouse breeding and swapping of maternal feeding. B: Growth curves of female offspring of LL mice after switching maternal feeding (n = 4–6 mice/group). C: Growth curves of female offspring of KO mice after switching maternal feeding (n = 4–6 mice/group). D: Increase in the body weight of female offspring after switching the matrilineal parent (n = 4–7 mice/group). E: Blood glucose levels of female offspring after switching the matrilineal parent (n = 4–7 mice/group). F and G: ITTs (F) and GTTs (G) results for female offspring of LL maternal mice after swapping maternal feeding (n = 4–5 mice/group). s I: ITTs (H) and GTTs (I) results for female offspring of KO maternal mice after swapping maternal feeding (n = 5–7 mice/group). J and K: The area under the curves (AUCs) of the ITT (J) and GTT (K) (n = 4–7 mice/group). L-LL-L: LL offspring of LL maternal parent fed by LL mother; L-KO-L: KO offspring of LL maternal parent fed by LL mother; K-LL-L: LL offspring of KO maternal parent fed by LL mother; K-KO-L: KO offspring of KO maternal parent fed by LL mother; L-LL-K: LL offspring of LL maternal parent fed by KO mother; L-KO-K: KO offspring of LL maternal parent fed by KO mother; K-LL-K: LL offspring of KO maternal parent fed by KO mother; K-KO-K: KO offspring of KO maternal parent fed by KO mother. All the values are presented as the means ± SEMs. The statistical significance was indicated by *p < 0.05, **p < 0.01 and ***p < 0.001.