Search
Close this search box.

Transgenic female mice producing trans 10, cis 12-conjugated linoleic acid present excessive prostaglandin E2, adrenaline, corticosterone, glucagon, and FGF21 – Scientific Reports

A shift of t10c12-CLA and changes in fatty acids compositions in Pai mice

Real-time analysis revealed that the foreign Pai gene transcribed in WAT, BAT, livers, and hypothamalus of Pai/wt and Pai/Pai mice in a dose-dependent manner (Fig. 1a). Gas chromatography analysis of FA compositions in the livers, kidneys, hearts, tibialis anterior muscle, and interscapular BAT tissues revealed that the content of t10c12-CLA had increased by 69% in the Pai/wt kidneys compared to wt littermates (p < 0.05; Fig. 1b). However, the quantities of substrate linoleic acid had decreased in the Pai/Pai livers by 23% and skeletal muscle by 34% and had increased in the Pai/wt kidneys by 36% (p < 0.05; Fig. 1b). The contents of other FAs, such as myristic (14:0), palmitic (16:0), stearic (18:0), palmitoleic (16:1n-7), cis-vaccenic (18:1n-7), oleic (18:1n-9), arachidonic (20:4n-6), and docosahexaenoic (22:6n-3) acids were altered to varying degrees in one or more tissues (Supplementary Table S3). Additionally, the content of total FAs (mg/g) had increased by 30% in the Pai/wt kidneys (p < 0.05; Fig. 1b), similar to the Pai/wt kidney of male mice3. The results suggest that the t10c12-CLA-induced changes in the content of each FA were genotype- or tissue-specific.

Figure 1
figure 1

Different mRNA levels of Pai gene (a) and contents of fatty acids (b) in the tissues from wild-type (wt) and Pai mice at the age of 11 weeks. WAT, white adipose tissues; BAT, brown adipose tissue. All fatty acid contents are listed in Supplementary Table S3. The bars represent the mean ± SD. * indicates p < 0.05; ** indicates p < 0.01; *** indicates p < 0.001; and **** indicates p < 0.0001, respectively.

No fat reduction in Pai mice

We first concentrated on the effect of t10c12-CLA on bodyweight and adiposity in Pai mice. The results showed that there was a gradual reduction of weaning weight in Pai/wt and Pai/Pai genotypes compared to wt mice (p < 0.05, Fig. 2a); however, the difference in bodyweight disappeared after five weeks of age and did not appear during the onward ages in both Pai mice (Supplementary Fig. S1), similar to the weaning weight of Pai male mice in our previous study3. Magnetic resonance imaging, dissection and histological analyses revealed that Pai/Pai mice at 11 weeks had no reduction of WAT mass and the cross-sectional area of white adipocytes (Fig. 2b–e) but exhibited generalised organomegaly, including significantly enlarged livers, spleens, hearts, and ovaries when body weight was considered (p < 0.05; Supplementary Table S4). Pai/wt mice also had no mass loss of WAT and only showed enlarged ovaries.

Figure 2
figure 2

Magnetic resonance imaging and histological and RNA expression analysis of white adipose tissues (WAT). (a) Decrease of weaning weight in Pai mice. (b) Coronal sections were obtained on the whole-body midsection for four wild-type (wt) or Pai/Pai mice at 11 weeks in each group during magnetic resonance imaging. The wt and Pai/Pai mice show similar areas of signal intensity of fatty tissue in the subcutaneous and intra-abdominal regions. (c) There is no difference in the mass of WAT among wt, Pai/wt, and Pai/Pai mice. (d) Imaging of hematoxylin–eosin staining, (e) Analysis of cross-sectional area per cell, and (f) Relative expression of mRNA in the WAT. Ampk AMP-activated protein kinase; Atgl adipose triglyceride lipase; Cd cluster of differentiation; Cebp CCAAT/enhancer-binding-protein beta; Cgi58 comparative gene identification 58; Chrebp carbohydrate response element binding protein; Cpt1a carnitine palmitoyltransferase I-a; Fasn fatty acid synthase; Fgf21 fibroblast growth factor 21; Foxc2 forkhead box protein c2; G6p glucose-6-phosphatase; Glut4 glucose transporter type 4; Hsl hormone-sensitive lipase; Igfbp1 insulin-like growth factor binding protein 1; Insr insulin receptor; Irs1 insulin receptor substrate-1; Irs2 insulin receptor substrate-2; Lcad long-chain acyl-CoA dehydrogenase; Lpl lipoprotein lipase; Mcad medium-chain acyl-CoA dehydrogenase; Mgl monoacylglycerol lipase; Nr3c1 nuclear receptor subfamily 3 group C member 1; Pai propionibacterium acnes isomerase; Pepck phosphoenolpyruvate carboxykinase; Pgc1⍺ Ppar-gamma coactivator 1 alpha; Pi3k phosphoinositide 3-kinase; Plin1 perilipin 1a; Ppar-γ peroxisome proliferator-activated receptor-γ; Prdm16 PR domain containing 16; Ucp1 uncoupling protein 1. The bars represent the mean ± SD. * indicates p < 0.05; ** indicates p < 0.01; *** indicates p < 0.001; and **** indicates p < 0.0001, respectively.

To investigate in-depth changes in WAT, RNA levels of 33 essential genes were determined in Pai/Pai mice and the total RNA levels of 14 (42%) genes, including seven up-regulated and seven down-regulated, were modified among them (Fig. 2f). Briefly, the seven up-regulated genes were Fasn, Peroxisome proliferator-activated receptor (Ppar)-γ, Hsl, glucose-6-phosphatase (G6p), F4/80, Adiponectin, and nuclear receptor subfamily 3 group C member 1 (Nr3c1, which is the corticosterone receptor). The seven down-regulated genes were lipoprotein lipase (Lpl), Cpt1a, perilipin 1a (Plin1), PR domain containing 16 (Prdm16), insulin receptor (Insr), glucose transporter type 4 (Glut4), and carbohydrate response element binding protein (Chrebp). In Pai/wt adipocytes, the RNA levels of six (43%) genes were changed among 14 examined genes. There were up-regulated Ppar-γ, Nr3c1, F4/80, and Leptin, and down-regulated Cpt1a and Chrebp (Fig. 2f). These aberrant gene expressions in white adipocytes suggest that t10c12-CLA can affect a series of metabolic processes, such as lipid metabolism, lipid and glucose uptake, energy expenditure, stimuli response and glucose homeostasis.

Enhanced lipolysis and thermogenesis in BAT of Pai mice

To clarify the effect of t10c12-CLA on BAT thermogenesis, we measured the BAT features of Pai mice. Magnetic resonance imaging and dissection analysis did not reveal any mass changes of BAT in both Pai mice (Fig. 3a,b). Hematoxylin–eosin staining showed Pai/Pai adipocytes with uniform and small-sized lipid droplets and Pai/wt adipocytes with irregular-sized lipid droplets compared to wt adipocytes with uniform and large-sized lipid droplets (Fig. 3c). The reduced cross-sectional areas per adipocyte indicated that the volumes of brown adipocytes decreased significantly in both Pai genotypes (p < 0.05; Fig. 3d).

Figure 3
figure 3

Aspects of brown adipose tissues (BAT). (a) Axial sections of the interscapular BAT show similar areas of grey signal intensity on magnetic resonance images (brown arrows) between four wild-type (wt) and four Pai/Pai mice at 11 weeks. (b) There is no BAT mass difference among wt, Pai/wt, and Pai/Pai mice. (c, d) Analysis of hematoxylin–eosin staining shows that the sizes of lipid droplets become small in the Pai/wt or Pai/Pai adipocytes, and the cross-sectional area per cell is decreased in Pai/wt and Pai/Pai mice. (e) Analyses of the relative expression of mRNAs of critical genes in wt and Pai/Pai mice. (f, g) Western blot analysis of proteins and their relative intensities in wt and Pai/Pai mice. Each membrane was cut into 2–3 parts, and each portion was then hybridised with a corresponding antibody in the western blot. All original, replicated blots were provided in the Supplementary information file 1. Acly ATP citrate lyase; Ampk AMP-activated protein kinase; Atgl adipose triglyceride lipase; Cpt1b carnitine palmitoyltransferase I-b; Fasn fatty acid synthase; Fgf21 fibroblast growth factor 21; Irs1 insulin receptor substrate-1; Mgl monoacylglycerol lipase; Nr3c1 nuclear receptor subfamily 3 group C member 1; pAMPK phosphorylated AMPK; Pgc1⍺ Ppar-gamma coactivator 1 alpha; pHSL phosphorylated hormone-sensitive lipase; Plin1 perilipin 1a; Ppar-γ and peroxisome proliferator-activated receptor-γ and -δ; Prdm16 PR domain containing 16; Ucp1 and 2 uncoupling protein 1 and 2. The bars represent the mean ± SD. * indicates p < 0.05; ** indicates p < 0.01; and *** indicates p < 0.001, respectively.

RNA analysis of Pai/Pai BAT showed that the transcriptional levels of eight genes were changed among 18 tested genes, consisting of up-regulated Fasn, Atgl, Plin1, Fgf21, and Ampk, as well as down-regulated Hsl, monoacylglycerol lipase (Mgl), and ppar-γ coactivator 1 alpha (Pgc1α) (p < 0.05; Fig. 3e). In contrast, the other critical genes involved in lipolysis and thermogenesis, such as Ppar-γ, Ucp1, Ucp2, Prdm16, and Ppar-δ, remained unchanged RNA levels. Western blot analysis revealed over-expressed AMPK, pHSL, CPT1B, UCP1, UCP2, and as well as down-expressed pAMPK in Pai/Pai BAT (p < 0.05; Fig. 3f,g). These results indicate that the t10c12-CLA can stimulate lipolysis, beta-oxidation, and thermogenesis of BAT.

Hepatic steatosis in Pai/Pai mice

Whether t10c12-CLA leads to hepatic steatosis is a conflict in the previous studies. In the current study, Pai/Pai mice exhibited hepatic hypertrophy, unchanged TC concentrations, and increased TGs levels in the livers (p < 0.05; Fig. 4a–c). Histological staining showed profound fat accumulation (~ 2.3 fold) in swollen hepatocytes (p < 0.05; Fig. 4d–g), indicating hepatic steatosis in Pai/Pai mice. However, the above parameters of Pai/wt livers remained unchanged compared to wt samples. The results suggest that hepatic steatosis is associated with high doses of t10c12-CLA.

Figure 4
figure 4

Histological and lipid analysis of livers. (a) Dissection analysis shows an increased liver mass in Pai/Pai mice. (b, c) ELISA analyses indicate normal levels of hepatic total cholesterol (TC) and increased levels of triglycerides (TGs) in Pai/Pai livers. (d, e) Analyses of hematoxylin–eosin staining show the abnormal morphology and hepatocyte oedema and the enlarged cross-sectional area per hepatocyte in Pai/Pai mice. (f, g) Analyses of oil red staining show hepatic lipid accumulation in Pai/Pai mice. The bars represent the mean ± SD. * indicates p < 0.05 and *** indicates p < 0.001, respectively.

RNA analysis of Pai/Pai livers revealed that 25 (50%; Fig. 5a) genes significantly (p < 0.05) altered their RNA levels among 50 tested genes, Briefly, genes involved in lipid metabolism included three up-regulated Fasn, Mgl, and comparative gene identification 58 (Cgi58) and eight down-regulated genes Ppar-γ, diacylglycerol acyltransferase (Dgat) 1, Dgat2, Atgl, Lpl, Cpt1a, medium-chain acyl-CoA dehydrogenase (Mcad), and acyl-CoA oxidase (Acox1). Genes involved in the sterol pathway included up-regulated sterol regulatory element binding protein (Srebp) 1c and insulin-induced gene (Insig) 1, as well as down-regulated Srebp1a, Insig2, HMG-CoA reductase (Hmgcr), and LDL receptor (Ldlr). Moreover, genes related to insulin/insulin-like growth factor (IGF) signalling and glucose metabolism included up-regulated forkhead box protein a2 (Foxa2), as well as down-regulated IGF bind protein 1 (Igfbp1), Glut4, Fgf21, G6p, and phosphoenolpyruvate carboxykinase (Pepck). Additionally, two NADPH-producing enzymes, Malic and 6-Phosphogluconate dehydrogenase (Pgd), were up-regulated, while the inflammatory factor cluster of differentiation 11c (Cd11c) was down-regulated in the Pai/Pai livers. Western blot analysis revealed that the protein levels of pAMPK and FASN were reduced, and CPT1A was increased, while AMPK, ATGL, CPA1B, and FGF21 levels remained unchanged (Fig. 5b,c).

Figure 5
figure 5

Analysis of gene expression in the livers. (a) Analyses of the relative expression of mRNAs of critical genes in wild-type, Pai/wt, and Pai/Pai mice. (b, c) Western blot analysis of proteins and their relative intensities in wild-type and Pai/Pai mice. Each membrane was cut into 2–3 parts, and each portion was then hybridised with a corresponding antibody in the western blot. All original, replicated blots were provided in the Supplementary information file 1. Acca2 Acetyl-CoA acyltransferase 2; Acox Acyl-CoA oxidase; Adcy3 adenylate cyclase 3; Agpat2 1-acylglycerol-3-phosphate O-acyltransferase 2; Ampk AMP-activated protein kinase; Atgl adipose triglyceride lipase; Cd cluster of differentiation; Cebp CCAAT/enhancer-binding-protein beta; Cgi58 comparative gene identification 58; Chrebp carbohydrate response element binding protein; Cpt1a carnitine palmitoyltransferase I-a; Dgat diacylglycerol acyltransferase; Fasn fatty acid synthase; Fgf21 fibroblast growth factor 21; Foxa2 forkhead box protein a2; G6p glucose-6-phosphatase; G6pd glucose-6-phosphate dehydrogenase; GAPDH glyceraldehyde-3-phosphate dehydrogenase; Glut4 glucose transporter type 4; Gpat1 glycerol-3-phosphate acyltransferase 1; Hmgcr HMG-CoA reductase; Hsl hormone-sensitive lipase; Htgl hepatic triglyceride lipase; Igf1 insulin-like growth factor-1; Igfbp1 IGF binding protein 1; Insig 1 and 2 insulin induced gene 1 and 2a; Insr insulin receptor; Irs insulin receptor substrate; Lcad long-chain acyl-CoA dehydrogenase; Lchad long-chain 3-hydroxyacyl-CoA dehydrogenase; Ldlr LDL receptor; Lkb1 liver kinase B1; Lpl lipoprotein lipase; Lxr liver X receptor; Mcad medium-chain acyl-CoA dehydrogenase; Mgl monoacylglycerol lipase; Nr3c1 nuclear receptor subfamily 3 group C member 1; Pai propionibacterium acnes isomerase; Pepck phosphoenolpyruvate carboxykinase; Pgc1⍺ Ppar-gamma coactivator 1 alpha; Pgd 6-phosphogluconate dehydrogenase; Pi3k phosphoinositide 3-kinase; Plin1 perilipin 1a; Ppar-γ peroxisome proliferator-activated receptor-γ; Prdm16 PR domain containing 16; Scap SREBP cleabage-activating protein; Srebp sterol regulatory element binding protein; Scd1 stearoyl-CoA desaturase 1; Ucp1 uncoupling protein 1. The bars represent the mean ± SD; * indicates p < 0.05; ** indicates p < 0.01; and *** indicates p < 0.001, respectively.

RNA analysis of Pai/wt livers showed that 32 genes had similar transcription levels to Pai/Pai samples, and 13 (33%) of them were markedly (p < 0.05) different from those in wt samples among 40 tested genes (Fig. 5a). To be specific, these 13 genes included four up-regulated (Mgl, Srebp1c, Insig1, and Foxa2) and nine down-regulated (Lpl, Cpt1a, Mcad, Srebp1a, Insig2, Igfbp1, G6p, Pepck, and Cd11c). Furthermore, RNA results indicated that the mRNA levels of almost all genes in the Pai/wt sample were intermediate between those of wild-type and Pai/Pai mice, and the RNA levels of most genes decreased in Pai’s liver, suggesting that the 10c12-CLA attenuates glucose and lipid metabolism in a dose-dependent manner.

Changes in hormones, triglycerides and inflammatory factors in Pai mice

Although the circulating levels of TC, FFAs, and HDL showed no differences between Pai mice and their wt littermates, the levels of plasma TGs were significantly elevated in Pai/Pai mice compared to wt or Pai/wt mice (p < 0.05, Fig. 6), indicating t10c12-CLA-induced hypertriglyceridemia in Pai/Pai mice. However, the concentrations of circulating glucose, insulin, leptin, and ghrelin (Fig. 6), as well as small intestine length (Supplementary Table S4), showed no differences (p > 0.05) between Pai mice and their wt littermates, suggesting no effect of t10c12-CLA on energy intake.

Figure 6
figure 6

Comparisons of circulating factors in wild-type and Pai mice. Blood samples were collected from non-fasted mice at the age of 11 ~ 15 weeks. The bars represent the mean ± SD. * indicates p < 0.05; ** indicates p < 0.01; and *** indicates p < 0.001, respectively.

Some critical hormones related to lipid metabolism and inflammation were also investigated. Serum concentrations of PGE2 (109%), glucagon (twofold), corticosterone (116%), adrenaline (4.5-fold), TNFα (threefold), CRP (threefold), and IL-6 (2.5-fold) were significantly (p < 0.05) increased in Pai/Pai mice when compared to their wt littermates. However, only adrenaline (3.5-fold), TNFα (2.5-fold), and IL-6 (twofold) were significantly (p < 0.05) elevated in Pai/wt mice. Additionally, the levels of all the above factors in Pai/wt mice were intermediate between those of wild-type and Pai/Pai mice, suggesting that the effect of 10c12-CLA on hormone overproduction is dose-dependent. The only exception is FGF21, which was exclusively elevated by 9% in Pai/wt, not in Pai/Pai female mice (Fig. 6), similar to the previous result in male mice3, suggesting that FGF21 is sensitive to low-dose, not high-dose t10c12-CLA. Moreover, there was no increase in blood levels of lactic acid dehydrogenase in either Pai mice, indicating no cellular injury caused by the PAI protein.

Both glucose and insulin tolerance tests showed no difference in the circulating glucose concentrations or the area under the curve between wt and two Pai genotypes (Fig. 7a,b). Interestingly, the fasted glucose levels before glucose injection were significantly (p < 0.05) higher in Pai/wt mice than in wt or Pai/Pai mice (Fig. 7a). So we further measured blood glucose levels during a 24-h fasting duration to investigate this discrepancy. We found that Pai/wt mice continuously maintained higher glucose levels than their wt littermates (p < 0.05; Fig. 7c). This suggests that glucose stability in starved Pai/wt mice is sensitive to low doses of t10c12-CLA, which may be associated with the beneficial insulin-sensitising and glucose-lowering effects of FGF2119.

Figure 7
figure 7

Comparisons of glucose (a) and insulin (b) tolerance tests and dynamic glucose levels (c) in wild-type (wt) and Pai mice. (a) Absolute blood glucose levels and the area under the curve. (b) Blood glucose levels relative to initial values and the area under the curve. (c) Dynamic blood glucose levels and the area under the curve are measured in fasting mice starved from Zeitgeber time 0 to 24. Zeitgeber times 0 and 12 are lights-on and -off times, respectively. The bars represent the mean ± SD. † indicates p < 0.05 between Pai/wt and wt or Pai/Pai mice; # indicates p < 0.05 between wt and Pai/Pai mice; * indicates p < 0.05 between wt and Pai/wt mice.

Less heat release, oxygen consumption, and physical activities in Pai mice

To investigate the potential impact of t10c12-CLA on energy homeostasis, energy metabolism and activity parameters were assessed using metabolic cages, considering body weight. Compared to their wt littermates, both Pai/wt and Pai/Pai mice had no differences in body weight gain, food and water intake (Fig. 8a–c), total distance travelled during the whole 72-h observation (Fig. 8d), and respiratory exchange ratio (Fig. 8e,f). However, the following parameters showed varying differences.

Figure 8
figure 8

Abnormal energy metabolism and activities in Pai mice. Comparison of body weight gain during a 72-h period (a), food intake (b), water intake (c), total distance travelled during a 72-h period (d), respiratory exchange ratio (RER) and RER time course (e, f), distance travelled during a 39-min sampling duration (g), speed (h), locomotor activities in the centre of the cage and its time course (i, j), central ambulating and stereotypic activities (k, l), locomotor activities in the margin or corners of the cage and its time course (m, n), marginal ambulating and stereotypic activities (o, p), O2 consumption (q), CO2 production (r), heat release and its time course (s, t) among wild-type, Pai/wt, and Pai/Pai mice at 11 weeks. The lights turned on and off at 7:00 am and 7:00 pm. The data are normalised to lean body weight. The bars represent the mean ± SD. * indicates p < 0.05; ** indicates p < 0.01; and *** indicates p < 0.001, respectively.

The Pai/Pai mice covered a longer distance per sampling duration (~ 44% longer) and exhibited 47.2% higher speed during the light phase (p < 0.05; Fig. 8g,h) but did not reduce their total locomotor activities in either the centre (Fig. 8i–l) or margin/corners (Fig. 8m–p). Nevertheless, during both the light (p < 0.05) and dark (p > 0.05) periods, they consumed 7.1% and 7.6% less O2 (Fig. 8q) and produced 9.3% and 5.9% less CO2 (Fig. 8r) as well as 7.6% and 6.9% less heat (Fig. 8s,t), respectively. The results suggest that Pai/Pai mice simultaneously reduced oxygen consumption and heat release during the light periods while maintaining normal locomotor activities.

On the other hand, Pai/wt females did not reduce their central activities (p > 0.05) but displayed 29.9% fewer marginal activities, such as marginal ambulating and stereotypic activities during the dark period (p < 0.05). Moreover, during both the light (p < 0.05) and dark (p < 0.05) periods, they consumed 9.0% and 12.4% less O2, produced 9.9% and 12.0% less CO2, as well as 8.9% and 10.7% less heat, respectively, suggesting that Pai/wt mice simultaneously reduced oxygen consumption and heat release during the whole day while exhibiting fewer marginal activities.

Abnormal gene transcription in the hypothalamus of Pai mice

To determine whether t10c12-CLA affected the hypothalamus, critical hypothalamic genes and proteins were also measured. The mRNA levels of the leptin receptor and agouti-related peptide increased, and the transcription levels of the ghrelin receptor and Orexins decreased in the Pai/wt or Pai/Pai hypothalamus compared to wt mice (Supplementary Fig. S2). Western blot analysis had not revealed any changes in AMPK, phosphorylated AMPK, and glucose-related protein 78 (GRP78) in the Pai/Pai hypothalamus; and the ratio of pAMPK/AMPK had no difference between wt and Pai/Pai samples (p > 0.05; Supplementary Fig. S2). These findings suggest that the central regulation of energy intake may only be affected at the transcriptional level.