Adipocyte iron levels impinge on a fat-gut crosstalk to regulate intestinal lipid absorption and mediate protection from obesity

Abstract

Iron overload is positively associated with diabetes risk. However, the role of iron in adipose tissue remains incompletely understood. Here, we report that transferrin-receptor-1-mediated iron uptake is differentially required for distinct subtypes of adipocytes. Notably, adipocyte-specific transferrin receptor 1 deficiency substantially protects mice from high-fat-diet-induced metabolic disorders. Mechanistically, low cellular iron levels have a positive impact on the health of the white adipose tissue and can restrict lipid absorption from the intestine through modulation of vesicular transport in enterocytes following high-fat diet feeding. Specific reduction of adipocyte iron by AAV-mediated overexpression of the iron exporter Ferroportin1 in adult mice effectively mimics these protective effects. In summary, our studies highlight an important role of adipocyte iron in the maintenance of systemic metabolism through an adipocyte-enterocyte axis, offering an additional level of control over caloric influx into the system after feeding by regulating intestinal lipid absorption.

Keywords: Fpn1; adipocyte; diabetes and obesity; enterocyte; intestine; iron; lipid absorption; transferrin receptor 1; vesicular transport; white adipose tissue.

Figure 1.. The Contribution of TFRC Mediated Iron Transport during Adipocytes Development and Maturation

(A) mRNA levels of Tfrc in different fat pads of WT mice during development. n=5. (B) mRNA levels of Tfrc in purified adipocytes fraction and SVF from fat pads of 2-month-old WT mice. n=4. (C) Relative mRNA levels of Tfrc in purified adipocytes fraction from fat pads of 2-month-old Tfrc^fl/fl and Tfrc^AKO mice. n=3. (D) DAB iron staining of BAT harvested from Tfrc^fl/fl and Tfrc^AKO mice at indicated age. Scale bar, 50 μM. Representative of 3 biological replicates. (E) Basal OCR of diced fat pads of 2-month-old Tfrc^fl/fl and Tfrc^AKO mice. n=5. (F) Photograph of fat pads harvested from Tfrc^fl/fl and Tfrc^AKO mice at indicated age. Representative of at least 5 biological replicates. (G) H&E staining of fat pads harvested from Tfrc^fl/fl and Tfrc^AKO mice at indicated age. Scale bar, 100 μM. Representative of 3 biological replicates. (H) Overview of Tfrc^iAKO mice model. (I) Photograph of fat pads of 2-month-old iCon and Tfrc^iAKO mice (Dox diet: E0-2 months). Representative of at least 5 biological replicates. (J) Photograph of fat pads of 2-month-old iCon and Tfrc^iAKO mice (Dox diet: E13.5-2 months). Representative of at least 5 biological replicates. (K) Photograph of fat pads of 2-month-old iCon and Tfrc^iAKO mice (Dox diet: 3 weeks-2 months). Representative of at least 5 biological replicates. (L) Photograph of fat pads of 4-month-old iCon and Tfrc^iAKO mice (Dox diet: 2 months-4 months). Representative of at least 5 biological replicates. (M) Photograph of fat pads of 2-month-old iCon and Tfrc^iAKO mice (Dox diet: E13.5-P21). Representative of at least 5 biological replicates. (N) Photograph of fat pads of 2-month-old Tfrc^fl/fl and Tfrc^AKO mice (T_N: E13.5-2 months). Representative of at least 5 biological replicates. Data are presented as mean ± SD. Two-tailed Student’s t-test. *P < 0.05, **P < 0.01. Results were confirmed in at least 2 independent experiments.

Figure 2.. Tfrc^AKO Mice Displayed Improved Metabolic Phenotype under HFD Challenge at Both Room Temperature and T_N Status

(A) Representative photograph of Tfrc^fl/fl and Tfrc^AKO mice after 10 weeks of HFD feeding. Representative of 5 biological replicates. (B) Body weight of Tfrc^fl/fl and Tfrc^AKO mice following HFD feeding. n=6. (C-F) Body mass (C), fat pads weight (D), photograph of fat pads (E), and fasting and random blood glucose (F) of Tfrc^fl/fl and Tfrc^AKO mice after 10 weeks of HFD feeding. n=6. (G-I) Oral glucose tolerance test (OGTT) (G), insulin tolerance test (ITT) (H), and serum insulin levels during OGTT (I) of Tfrc^fl/fl and Tfrc^AKO mice after 10 weeks of HFD feeding. n=6. (J-N) Fasting serum triglycerides (J), NEFA (K), cholesterol (L), and random serum adiponectin (M) and leptin (N) levels of Tfrc^fl/fl and Tfrc^AKO mice after 10 weeks of HFD feeding. n≥6. (O-Q) Photograph of liver (O), H&E staining of liver (P), H&E staining of fat pads (Q) harvested from Tfrc^fl/fl and Tfrc^AKO mice after 10 weeks of HFD feeding. Scale bar, 100 μM. Representative of 3 biological replicates. (R-T) Tfrc^fl/fl and Tfrc^AKO mice were exposed to T_N (from E13.5, and thereafter), and fed on HFD from 2-month-old under T_N. Body weight (R), weight of fat pads (S), and representative photograph of fat pads (T) of Tfrc^fl/fl and Tfrc^AKO mice were presented after 10 weeks of HFD Feeding. n=8. (U-V) OGTT (U) and serum insulin levels during glucose tolerance test (V) of Tfrc^fl/fl and Tfrc^AKO mice described in R-T. n=8. (W) H&E staining of liver and fat pads of Tfrc^fl/fl and Tfrc^AKO mice described in R-T. Scale bar, 100 μM. Representative of 3 biological replicates. Data are presented as mean ± SD. Two-way ANOVA with Dunnett’s test (B, G, H, I, U, V) or Two-tailed Student’s t-test (C, D, F, J-N, R, S). *P < 0.05, **P < 0.01. The weight of BAT includes the weight of BAT and the WAT connected to BAT. Results were confirmed in 3-5 independent experiments.

Figure 3.. Specific Loss of Tfrc in BAT Does Not Protect Mice from HFD-induced Metabolic Disorder

(A) Overview of Tfrc^UKO mice model. (B) Relative mRNA levels of Tfrc in purified adipocytes fraction from BAT of 2-month-old Tfrc^fl/fl and Tfrc^UKO mice. n=4-5. (C) Photograph of BAT harvested from 2-month-old Tfrc^fl/fl and Tfrc^UKO mice with 2 more months of HFD feeding. Representative of 3 biological replicates. (D) Body weight of Tfrc^fl/fl and Tfrc^UKO mice during HFD feeding. n=5. (E) Fasting and random blood glucose of Tfrc^fl/fl and Tfrc^UKO mice after 2 months of HFD feeding. n=5. (F-H) OGTT (F), ITT (G), and serum insulin levels during OGTT (H) of Tfrc^fl/fl and Tfrc^UKO mice after 2 months of HFD feeding. n=5. Data are presented as mean ± SD. Two-tailed Student’s t-test (B, E) or Two-way ANOVA with Dunnett’s test (D, F-H). *P < 0.05, **P < 0.01. Results were confirmed in 2 independent experiments.

Figure 4.. Tfrc^AKO Mice Show Less Intestinal Lipid Absorption compared to Tfrc^fl/fl Mice

(A-C) Realtime food intake (A), accumulated food intake (B) and energy expenditure (C) in 2-month-old Tfrc^fl/fl and Tfrc^AKO mice following 9 weeks of HFD. n=6. (D) Feed efficiency of Tfrc^fl/fl and Tfrc^AKO mice during HFD feeding. n=6. (E) Serum triglycerides levels of fasted Tfrc^fl/fl and Tfrc^AKO mice (HFD 8 weeks) before and after refeed. n≥9. (F) Post gavage (olive oil) serum triglycerides levels in HFD-fed Tfrc^fl/fl and Tfrc^AKO mice pre-treated with tyloxapol. n=5. (G-I) Daily feces weight (G), fecal energy content (H), and daily fecal calorie content (I) of Tfrc^fl/fl and Tfrc^AKO mice after 8 weeks of HFD feeding. n≥9. (J) Feces from HFD-fed Tfrc^fl/fl and Tfrc^AKO mice were collected to tubes with DNA/RNA shield buffer and put at 4°C for 6 hours. Representative of 3 biological replicates. (K-N) Radioactivity of ³H and ¹⁴C in blood and feces of HFD-fed Tfrc^fl/fl and Tfrc^AKO mice following ³H-Triolein and ¹⁴C-Palmitate gavage. n=6. (O-P) Oil-Red-O staining of villus of proximal small intestine of fasted Tfrc^fl/fl and Tfrc^AKO mice (HFD 8 weeks) after 1 hour refeed (O). Scale bar, 50 μM. Quantification of the oil red signal in villus (P). Representative of 3 biological replicates. (Q-R) Electron microscopy images for lipid droplets in enterocytes of proximal small intestine of fasted Tfrc^fl/fl and Tfrc^AKO mice (HFD 8 weeks) after 1 hour refeed. (Q). Scale bar, 6 μM. Quantification of the vesicle area in villus (R). Representative of 3 biological replicates. (S) Relative mRNA levels of vesicle transport related genes in small intestine of Tfrc^fl/fl and Tfrc^AKO mice after 10 weeks of HFD feeding. n=5. (T) Model for fat transplantation. (U) Body weight of recipient mice (Tfrc^AKO) received fat from either Tfrc^fl/fl or Tfrc^AKO mice following HFD feeding. n=5. (V-Y) Daily food intake (V), feces weight (W), fecal energy content (X), and daily fecal calorie content (Y) of the recipient mice after 4 weeks of HFD feeding. n=5. Data are presented as mean ± SD. Two-tailed Student’s t-test (D, E, G-I, M, N, P, R, S, V-Y) or Two-way ANOVA with Dunnett’s test (F, K, L, U). *P < 0.05, **P < 0.01. Results were confirmed in 2-4 independent experiments.

Figure 5.. The Beneficial Effects of Lowering Iron in Adipocytes upon HFD-challenge

(A-C) Hierarchical clustering of transcriptional profiles of gWAT obtained from Tfrc^fl/fl and Tfrc^AKO mice after 10 weeks of HFD feeding (A). Labels identify gene clusters showing enrichment GO analyses (B-C). n=3, each sample combined from 2 mice. FC, fold change. (D) Weight of mWAT in Tfrc^fl/fl and Tfrc^AKO mice after 10 weeks of HFD feeding. n=6. Results were confirmed in at least 3 independent experiments. (E-G) Volcano plot depicting the fold change and significance of differential protein abundance in sEV isolated from mWAT of Tfrc^fl/fl and Tfrc^AKO mice after 10 weeks of HFD feeding (E). GO analysis of proteins downregulated (F) and upregulated (G) more than two-fold in sEV of Tfrc^AKO mice. n=2, each sample combined from 3-4 mice. (H-I) F2-Isoprostane levels (H) the expression of triglycerides synthesis related genes (I) in gWAT of Tfrc^fl/fl and Tfrc^AKO mice after 10 weeks of HFD feeding. n=5. (J-L) Profiling of sphingolipids (J), serum phospholipids (K), and fatty acids (L) in serum of Tfrc^fl/fl and Tfrc^AKO mice after 10 weeks of HFD feeding. n=6. (M) Scatter plot of differential protein levels upon fold changes in medium of fully differentiated Tfrc^fl/fl and Tfrc^AKO adipocytes, or Adipoq-Cre adipocytes with AAV mediated specific overexpression of AAV-eGFP or AAV-Fpn1^C326S. Representative of 2-3 biological replicates. (N-O) Photograph of small intestine (N) and H&E staining of proximal small intestine (O) harvested from Tfrc^fl/fl and Tfrc^AKO mice after 4 months of HFD feeding. Scale bar, 200 μM. Representative of at least 15 biological replicates. Results were confirmed in 3 independent experiments. (P) Phylum level and class level comparison of gut microbiota proportional abundance in feces of Tfrc^fl/fl (n=5) and Tfrc^AKO (n=4) mice after 2 months of HFD feeding. Data are presented as mean ± SD. Two-tailed Student’s t-test. *P < 0.05, **P < 0.01.

Figure 6.. AAV-Mediated Specific Overexpression of An Iron Exporter in Adipocytes Mitigates Metabolic Dysfunction

(A) Overview for AAV-mediated specific overexpression of eGFP or Fpn1^C326S in adipose tissue of adult mice. (B) Body weight of Adipoq-Cre mice received either AAV-eGFP or AAV-Fpn1^C326S following HFD feeding. n=6. (C-E) Body mass (C), weight of fat pads (D), representative photograph of fat pads (E) of Adipoq-Cre mice received either AAV-eGFP or AAV-Fpn1^C326S, after 4 weeks HFD feeding. n=6. (F-H) OGTT (F), serum insulin levels during OGTT (G), fasting serum triglycerides, NEFA, and cholesterol (H) of Adipoq-Cre mice received either AAV-eGFP or AAV-Fpn1^C326S, after 4 weeks HFD feeding. n=6. (I) H&E staining of liver and fat pads of Adipoq-Cre mice received either AAV-eGFP or AAV-Fpn1^C326S , after 4 weeks HFD feeding. Scale bar, 200 μM. Representative of 3 biological replicates. (J-K) Food intake (J), fecal output (K) of Adipoq-Cre mice received either AAV-eGFP or AAV-Fpn1^C326S, after 3 weeks HFD feeding. n=6. (L-M) Radioactivity of ³H and ¹⁴C in feces following ³H-Triolein and ¹⁴C-Palmitate gavage in Adipoq-Cre mice received either AAV-eGFP or AAV-Fpn1^C326S, after one week of HFD feeding. n=6. (N-Q) Body weight (N), Body mass (O), weight of fat pads (P), representative photograph of fat pads (Q) of Adipoq-Cre mice with 2 months HFD feeding, then received either AAV-eGFP or AAV-Fpn1^C326S, with continued HFD feeding for 7 weeks. n=5. (R-T) OGTT (R), serum insulin levels during OGTT (S), fasting serum triglycerides, NEFA, and cholesterol (T) of Adipoq-Cre mice with 2 months HFD feeding, then received either AAV-eGFP or AAV-Fpn1^C326S, with continued HFD feeding for 6.5 weeks. n=5. (U-W) H&E staining of liver and fat pads (U), food intake (V), fecal output (W) of Adipoq-Cre mice with 2 months HFD feeding, then received either AAV-eGFP or AAV-Fpn1^C326S , with continued HFD feeding for 7 weeks. Scale bar, 200 μM. Representative of 5 biological replicates. Data are presented as mean ± SD. Two-way ANOVA with Dunnett’s test (B, F, G, N, R, S, V) or Two-tailed Student’s t-test (C, D, H, J-M, O, P, T, W). *P < 0.05, **P < 0.01. Results were confirmed in at least 2 independent experiments.

Figure 7.. Summary of The Effects of Lowering Adipocytes Iron on Metabolism.

Selective lowering iron in white adipocytes has multiple beneficial effects in mice upon a HFD-feeding. Adipocytes iron content can be used as a sensor of WAT to activate an adipose-gut crosstalk to regulate lipid absorption.