Regulation of redox homeostasis by ATF4-MTHFD2 axis during white adipose tissue browning

Abstract

Maintaining redox balance is crucial for mitochondrial homeostasis. During browning of white adipocytes, both the quality and quantity of mitochondria undergo dramatic changes. However, the mechanisms controlling the redox balance in the mitochondria during this process remain unclear. In this study, we demonstrate that thermogenic activation occurs before mitochondrial biogenesis during cold-induced browning of inguinal white adipose tissue (iWAT) and is accompanied by increased mitochondrial stress and integrated stress response (ISR) signaling. Specifically, cold exposure enhances the expression of ATF4, an ISR effector. Adipocyte-specific deletion of ATF4 results in increased energy expenditure, but paradoxically leads to a lower core body temperature, and heightened pro-inflammation in iWAT after cold exposure, which is restored by the antioxidant, MitoQ. Mechanistically, ATF4 regulates the redox balance through MTHFD2, an enzyme involved in mitochondrial redox homeostasis by NADPH generation. Cold exposure upregulates MTHFD2 expression in an ATF4-dependent manner, and its inhibition by DS18561882 in vivo leads to impaired cold-induced mitochondrial respiration similar to the effects of ATF4 loss. These findings suggest that ATF4 is essential for redox balance via MTHFD2, thereby affecting tissue homeostasis during iWAT browning.

Fig. 1

Cold exposure activates the ATF4-mediated ISR, which is essential for maintaining systemic energy expenditure. (a–c) Representative immunoblot analysis of phosphorylation of eIF2α (left) and densitometric quantification of phosphorylated eIF2α normalized to total eIF2α (right) (a), quantitative RT-PCR analysis of Atf4(b) and Fgf21 and Chop(c) in iWAT of 10-week-old C57BL/6 mice housed in RT or 5 °C for the indicated times (n = 5 per time point). Quantitative RT-PCR analysis of Atf4(d) and immunoblot analysis of ATF4 and phosphorylated eIF2α (left), with densitometric quantifications of ATF4 normalized to HSP90, and phosphorylation of eIF2α normalized to total eIF2α (right) (e) in white adipocytes treated with CL-316,243 for 6 h (n = 3 per treatment). (f) Graphical scheme of the generation of adipocyte-specific ATF4 knockout mice. (g and h) Assessment of oxygen consumption (n = 6 per genotype) (g) and core body temperature (n = 5 for Atf4^WT; n = 4 for Atf4^AKO) (h) in 10-week-old Atf4^WT and Atf4^AKO mice housed at different ambient temperatures. (i) Assessment of cold-induced changes of oxygen consumption in 10-week-old Atf4^WT and Atf4^AKO mice (n = 3 per genotype). (j) Oxygen consumption rate in mitochondria isolated from iWAT as measured using the Seahorse XF96 Extracellular Flux analyzer (n = 6 per each group). Data are presented as mean ± SEM. ∗∗∗p < 0.005, ∗∗p < 0.01, ∗p < 0.05.

Fig. 2

Increased energy expenditure in Atf4^AKO mice is independent of beige adipogenesis. (a–d) Representative images of Hematoxylin & Eosin (H&E) staining (scale bar, 60 μm) (a), immunoblot analysis for mitochondria OXPHOS complexes and UCP1 (left), and densitometric quantification of OXPHOS complexes and UCP1 normalized to HSP90 (right) (n = 3 for each genotype housed in RT, n = 4 for each genotype housed in 5 °C) (b), quantitative RT-PCR analysis of Ucp1, Cox8b, and Pgc1a (n = 5 per genotype) (c), and quantitative RT-PCR analysis of OXPHOS complexes encoded genes (n = 5 per genotype) (d) in iWAT from 10-week-old C57BL/6 mice housed at RT or 5 °C for 3 days. (e) Oil red O (ORO) staining of beige adipocyte differentiated from SVCs isolated from iWAT of Atf4^WT and Atf4^AKO mice (scale bar, 70 μm; n = 3 per genotype). (f) Quantitative RT-PCR analysis of Fabp4, Ucp1, and Prdm16 during beige adipogenesis of Atf4^WT and Atf4^AKO SVCs (n = 3 per genotype and time point). (g) Oxygen consumption rate of Atf4^WT and Atf4^AKO beige adipocytes, as measured using the Seahorse XF96 Extracellular Flux analyzer. Data are presented as mean ± SEM. ∗∗∗p < 0.005, ∗∗p < 0.01, ∗p < 0.05.

Fig. 3

ATF4 regulates macrophage homeostasis and inflammatory states in iWAT during cold exposure. (a) Volcano plot analysis of differentially expressed genes from Metabolic Pathways Panel in iWAT from 10-week old Atf4^WT and Atf4^AKO mice housed at 5 °C for 3 days. Genes with a fold change >1.5 (red) or < - 1.5 (blue) and a P-value <0.05 are highlighted. (b) Relative normalized mRNA counts of Il6 and Tnf from NanoString analysis. (c–g) Quantitative RT-PCR analysis of pro-inflammatory cytokines (Il1b, Tnfa, and Il6) (c), representative flow cytometry (d), M1 to M2 macrophage ratio (e), quantitative RT-PCR analysis of M1 macrophage markers (iNos and Itgax) (f), and M2 macrophage markers (Arg1, Cd206, Mgl1) (g) in iWAT from 10-week-old Atf4^WT and Atf4^AKO mice housed at RT or exposed to cold (5 °C) for 3 days (n = 5 per genotype). Data are presented as mean ± SEM. ∗∗∗p < 0.005, ∗∗p < 0.01, ∗p < 0.05.

Fig. 4

ATF4 maintains mitochondrial ROS levels in adipocytes. (a) Immunoblot analysis of 4-HNE (left) and densitometric quantification for 4-HNE normalized to HSP90 (right) in iWAT from 10-week-old Atf4^WT and Atf4^AKO mice exposed to cold (5 °C) for 24 h (n = 5 for Atf4^WT; n = 4 for Atf4^AKO). (b–c) Assessment of oxygen consumption (b) and core body temperature (c) in 10-week-old Atf4^WT and Atf4^AKO mice injected with MitoQ (5 mg/kg) during cold exposure (5 °C) (n = 4 per group). (d and e) Quantitative RT-PCR analysis of M1 macrophage markers (iNos and Itgax) (d) and pro-inflammatory cytokines (Il1b, Tnfa, and Il6) (e) in iWAT from 10-week-old Atf4^WT and Atf4^AKO mice exposed to cold (5 °C) for 3 days with or without injection of MitoQ (5 mg/kg) (n = 6 for Atf4^WT Vehicle treatment; n = 5 for Atf4^WT MitoQ treatment; n = 5 for Atf4^AKO Vehicle treatment; n = 6 for Atf4^AKO MitoQ treatment). (f–i) Fluorescence images and quantification of DCFDA staining (f) and MitoSox staining (scale bar, 100 μm; n = 3 per group) (g), immunoblot analysis of 4-HNE (left) and densitometric quantification of 4-HNE normalized to HSP90 (right) (n = 3 per group) (h), and quantitative RT-PCR analysis of pro-inflammatory cytokines (Il1b, Tnfa, and Il6) (n = 4 per group) (i) in Atf4^WT and Atf4^AKO white adipocytes treated with control or CL316,243 for 6 h. Data are presented as mean ± SEM. ∗∗∗p < 0.005, ∗∗p < 0.01, ∗p < 0.05.

Fig. 5

ATF4 regulates MTHFD2 expressions and maintains redox balance. (a) Transcriptomic analysis of iWAT from Atf4^WT and Atf4^AKO mice exposed to cold (5 °C) for 24 h. RNA-seq results are presented as a volcano plot. Genes with fold change >1.5 (red) or < −1.5 (blue) and p-value <0.05 are highlighted. (b) Venn diagram of genes from differentially expressed genes (DEG), Mitocarta 3.0, and ChIP-Atlas lists. Overlapping genes from the three lists are labeled in red. (c) Transcription factor enrichment analysis of promoter regions of genes labeled in red from Venn diagram shows significant enrichment of ATF4 binding sites in Mthfd2. (d) ChIP assay performed in differentiated white adipocytes; Fgf21 and Ddit3 were used as well-known ATF4 target genes. (e–f) Quantitative RT-PCR analysis of Mthfd2 (n = 5 per genotype) (e) and immunoblot analysis of MTHFD2 (left) and densitometric quantification of MTHFD2 normalized to HSP90 (right) (n = 3 per genotype housed at RT; n = 4 per genotype exposed to cold) (f) in iWAT from 10-week-old Atf4^WT and Atf4^AKO mice housed at RT or exposed to cold (5 °C) for 3 days. (g–h) Ratios of NADP+/NADPH (g) and GSH/GSSG (h) in iWAT from 10-week-old Atf4^WT and Atf4^AKO mice exposed to cold (5 °C) for 3 days (n = 4 for Atf4^WT; n = 5 for Atf4^AKO). Data are presented as mean ± SEM. ∗∗∗p < 0.005, ∗∗p < 0.01, ∗p < 0.05.

Fig. 6

Chemical inhibition of MTHFD2 leads to impaired energy homeostasis and inflammatory state in vivo. (a) Graphical scheme of DS18561882 oral gavage administration (100 mg/kg). (b–e) Body weight (b) and tissue weight (c) before and after cold exposure, assessment of oxygen consumption (d), and core body temperature (e) in 10-week-old C57BL/6 mice administered vehicle or DS18561882 (100 mg/kg) (n = 4 for vehicle; n = 5 for DS18561882). (f–k) Ratios of NADP+/NADPH (f) and GSH/GSSG (g), quantitative RT-PCR analysis of Atf4, Ucp1, and Mthfd2(h), pro-inflammatory cytokines (i), M1-macrophage markers (j), and M2-macrophage markers (k) in iWAT from 10-week-old C57BL/6 mice administered vehicle or DS18561882 (100 mg/kg) (n = 4 for vehicle; n = 5 for DS18561882). Data are presented as mean ± SEM. ∗∗∗p < 0.005, ∗∗p < 0.01, ∗p < 0.05.

Fig. 7

Proposed model for the role of ATF4 during cold-induced browning. ATF4 is rapidly activated in iWAT in response to acute cold exposure, where it plays a pivotal role in regulating MTHFD2 expression to maintain mitochondrial redox homeostasis. This regulation is essential for beige adipocyte function, facilitating systemic energy expenditure and preserving adipose tissue homeostasis during the browning of iWAT.