Annonaceous acetogenins mimic AA005 targets mitochondrial trifunctional enzyme alpha subunit to treat obesity in male mice

Abstract

Obesity and related diseases pose a major health risk, yet current anti-obesity drugs inadequately addressing clinical needs. Here we show AA005, an annonaceous acetogenin mimic, resists obesity induced by high-fat diets and leptin mutations at non-toxic doses, with the alpha subunit of the mitochondrial trifunctional protein (HADHA) as a target identified through proteomics and in vitro validation. Pharmacokinetic analysis shows AA005 enriches in adipose tissue, prompting the creation of adipose-specific Hadha-deficient mice. These mice significantly mitigate diet-induced obesity, echoing AA005's anti-obesity effects. AA005 treatment and Hadha deletion in adipose tissues increase body temperature and energy expenditure in high-fat diet-fed mice. The beneficial impact of AA005 on obesity mitigation is ineffective without uncoupling protein 1 (UCP1), essential for thermogenesis regulation. Our investigation shows the interaction between AA005 and HADHA in mitochondria, activating the UCP1-mediated thermogenic pathway. This substantiates AA005 as a promising compound for obesity treatment, targeting HADHA specifically.

Fig. 1. AA005 resists obesity and its metabolic disorders in diet-induced obese mice.

a Chemical structure of AA005. b Weight gain of male littermates fed HFD without or with AA005 (1.7 mg/kg) was recorded for 22 weeks (n = 9 for each group). c Representative photograph of vehicle and AA005-treated HFD-induced obese mice at experimental endpoint. Scale bar: 1 cm. d Fat weight was measured after 22 weeks of treatment (n = 9). e Fat weight of inguinal white adipose tissue (iWAT), abdominal subcutaneous white adipose tissue (asWAT), epididymal white adipose tissue (eWAT), retroperitoneal white adipose tissue (rWAT), mesenteric white adipose tissue (mWAT), and brown adipose tissue (BAT) in vehicle and AA005-treated obese mice (n = 9). f, Representative images of haematoxylin and eosin (H&E) stained of iWAT, eWAT, and BAT. iWAT, n = 79 (Vehicle) and 163 (AA005) adipocytes per group; eWAT, n = 90 (Vehicle) and 142 (AA005) adipocytes per group; BAT, n = 541 (Vehicle) and 392 (AA005) LDs per group. Scale bar, 50 μm. Glucose tolerance test (GTT) (g), area under the curve (AUC) of GTT (h), insulin tolerance test (ITT) (i), and AUC of ITT (j) in vehicle and AA005-treated obese mice (n = 9). k Body temperature was compared between the two groups (n = 5 per group). Three different biological replicates were tested. Mice at the end of week 21 were performed metabolic cage assays for energy expenditure (l, m), O₂ consumption (n, o), and CO₂ production (p, q) (n = 5). Data are presented as mean ± SEM. Unpaired, two-tailed Student’s t test. *P < 0.05, **P < 0.01, ***P < 0.001, NS indicates not significant. Source data are provided as a Source Data file.

Fig. 2. AA005 ameliorates obesity independent of leptin.

a Plasma leptin levels in vehicle and AA005-treated mice fed CD or HFD (n = 5). b Plasma leptin levels in wild-type (WT) and ob/ob mice treated with vehicle or AA005 (n = 5). c Body weight of WT and ob/ob male littermates treated with vehicle or AA005 (1.7 mg/kg) (n = 9 for WT mice, n = 8 for ob/ob mice for each group). d The average daily food intake of each mouse was recorded (n = 9 for WT mice, n = 8 for ob/ob mice for each group). e A representative photograph of WT and ob/ob mice treated with vehicle or AA005 at experimental endpoint. Scale bar, 1 cm. f Fat weight of iWAT, asWAT, eWAT, rWAT, mWAT, and BAT in WT and ob/ob mice treated with vehicle or AA005 (n = 9 for WT mice, n = 8 for ob/ob mice). g Fat and lean mass (n = 9 for WT mice, n = 8 for ob/ob mice for each group). h Representative images of H&E stained of iWAT, eWAT, and BAT. iWAT, n = 136 (ob/ob;vehicle) and 221 (ob/ob; AA005) adipocytes per group; eWAT, n = 93 (ob/ob;vehicle) and 165 (ob/ob; AA005) adipocytes per group; BAT, n = 546 (ob/ob;vehicle) and 384 (ob/ob; AA005) LDs per group;. Scale bar, 50 μm. Glucose tolerance test (GTT) (i), area under the curve (AUC) of GTT (j), insulin tolerance test (ITT) (k), and AUC of ITT (l) in vehicle and AA005-treated WT and ob/ob mice (n = 9 for WT mice, n = 8 for ob/ob mice for each group). Data are mean ± SEM. Unpaired, two-tailed Student’s t test. *P < 0.05, **P < 0.01, ***P < 0.001, NS indicates not significant. Source data are provided as a Source Data file.

Fig. 3. AA005 targets the mitochondrial trifunctional enzyme α subunit (HADHA).

a Preadipocyte 3T3-L1 cells were induced to undergo adipogenesis and concomitantly treated with different concentrations of AA005 on day 8. The cells were stained with Oil-Red-O. Scale bar: 10 μm. Three different biological replicates were tested. b Workflow illustrating the use of biotin-tagged AA005. c On day 8 after induction of adipogenesis concomitant treatment with biotin or biotin-AA005, 3T3-L1 cells were stained with Oil-Red-O. Scale bar, 10 μm. Three different biological replicates were tested. d Immunofluorescence imaging showing the localization of biotin-AA005 and HADHA in cells. Mitotracker was used to label mitochondria, and DAPI was used to label nuclei. Three different biological replicates were tested. Scale bar, 10 μm. e Schematic representation of the discovery of interacting proteins of AA005 using chemical proteomics. f Mitochondrial fractions of 3T3-L1 cells were incubated with biotin or biotin-AA005 in the presence or absence of a 5- or 10-fold excess of unlabeled AA005. The proteins were then pulled down with streptavidin-agarose, and western blotting was performed to detect HADHA. Three different biological replicates were tested. g Structural overview of the predicted HADHA–AA005 complex model. The docking result was performed by Swiss-docking method. The protein is shown in cartoon, and AA005 is shown in a stick representation. Residues around the predicted binding pocket were highlighted with a red rectangle. On the right is the zoomed-in view of the predicted HADHA–AA005 interface. Potential key interface residues (V174, L175, L202, T203, R205, T294, Q293, E500, K505, M506, P581, V582, E590, G708, G709, and L708) in HADHA were colored in green and labeled. h Microscale thermophoresis (MST) binding assay determined the K_d value (WT, K_d = 1.1 μΜ; Mut-V174, K_d = 9.6 μΜ; L175, K_d = 20 μΜ; L202, K_d = 6.1 μΜ) for the binding of AA005 toward wild and mutant HADHA. Three different biological replicates were tested. i Immunoprecipitation analysis of the physical interaction between wild and mutant HADHA and biotin-AA005. Three different biological replicates were tested. j On day 8 after induction of adipogenesis, 3T3-L1 cells with normal HADHA expression (3T3-L1-NC) or cells with Hadha knockdown (3T3-L1-siHadha) were stained with Oil-Red-O. Scale bar: 10 μm. Three different biological replicates were tested. Data are mean ± SEM. Source data are provided as a Source Data file.

Fig. 4. AA005 resists HFD-induced obesity in mice by targeting HADHA.

a Heatmap and histogram depicting the tissue distribution of AA005 in mice at various time points after AA005 administration (n = 3 per group). b Weight gain of male littermates fed a high-fat diet (HFD) in Hadha^fl/fl and Hadha^ΔAdipo (adipose Hadha deficiency) mice was recorded for 18 weeks (n = 6). c Representative photograph of Hadha^fl/fl and Hadha^ΔAdipo mice fed HFD at experimental endpoint. Scale bar, 1 cm. d Fat and lean mass were recorded after 18 weeks treatment (n = 5). e Fat weight of iWAT, asWAT, eWAT, rWAT, mWAT, and BAT in Hadha^fl/fl and Hadha^ΔAdipo mice fed HFD (n = 6). Glucose tolerance test (GTT) (f), and insulin tolerance test (ITT) (g) performed on Hadha^fl/fl and Hadha^ΔAdipo mice fed HFD (n = 6 for each group). h Comparison of body temperature between the two groups (n = 5 per group). Three different biological replicates were tested. Mice at the end of week 17 were performed metabolic cage assays for comparison of 24 h energy expenditure between Hadha^fl/fl and Hadha^ΔAdipo mice fed HFD (i) (n = 5 per group); The adjusted means of energy expenditure in the two groups analysed by ANCOVA. (j) (n = 5); O₂ consumption (k, l) and CO₂ production (m, n) measured during a 24 h period (k, m) (n = 5); Average values for the light and dark periods, respectively (l, n) (n = 5). o, Weight gain recorded over 20 weeks (n = 6) in male littermates fed HFD with or without AA005 treatment in Hadha^ΔAdipo mice. Glucose tolerance test (GTT) (p), area under the curve (AUC) of GTT (q), insulin tolerance test (ITT) (r), and AUC of ITT (s) performed on vehicle and AA005-treated Hadha^ΔAdipo mice (n = 6). Data are presented as mean ± SEM. Unpaired, two-tailed Student’s t test. *P < 0.05, **P < 0.01, ***P < 0.001, NS indicates not significant. Source data are provided as a Source Data file.

Fig. 5. AA005 treated mice and Hadha^ΔAdipo mice exhibit similar alterations in cardiolipin profiling.

a The protein expression and quantification analysis of HADHA in 3T3L1 cells after treatment with AA005 for different days (n = 3). b The protein expression and quantification analysis of HADHA in iWAT from vehicle and AA005-treated mice fed CD or HFD (n = 3). c Schematic presentation of HADHA structure and fatty acid β-oxidation pathway. Left: HADHA structure (from PDB code: 5ZQZ) is colored according to domain architecture (N-terminal ECH domain, middle core linker, C-terminal HACD domain). ECH domain (residues 37-333) is colored hot pink; middle core linker (residues 334-349) is colored coral; HACD domain (residues 350-763) is colored cornflower blue. Right: Schematic diagram of fatty acid β-oxidation process. ECH domain catalyzes the second step hydration reaction of FAO, while HACD domain catalyzes the third step dehydrogenation reaction. In 3T3L1 cells, the activities of ECH (d), HACD (e), and KACT (f) enzymes were detected after treatment with AA005 and inducer for 8 days (n = 6 per group). Three different biological replicates were tested. g Structure of HADHA was obtained from human mitochondrial trifunctional protein (TFP) (PDB code: 5ZQZ). MLCLAT-1 structure model was predicted by I-TASSER server and colored pink. The region corresponding to MLCLAT-1 in HADHA was colored light blue, and the extra N-terminal 190 residues were colored green-yellow respectively. h Schematic diagram of how HADHA works to remodel CL. i Relative mRNA levels of Mlclat-1 and Tafazzin genes in BAT of vehicle or AA005-treated HFD-fed mice (n = 5). j The protein expression and quantification analysis of HADHA and MLCLAT-1 in BAT of vehicle or AA005-treated HFD-fed mice (n = 3). k Relative mRNA levels of Mlclat-1 and Tafazzin genes in BAT of Hadha^fl/fl and Hadha^ΔAdipo mice fed HFD. (n = 5). l The protein expression and quantification analysis of HADHA and MLCLAT-1 in BAT of Hadha^fl/fl and Hadha^ΔAdipo mice fed HFD. (n = 3). m Heatmaps of saturation immature cardiolipin (CL_Sat) (upper) and unsaturation mature cardiolipin (CL_Unsat) (lower) species level in BAT. (n = 5 per group, Data are presented as mean ± SEM. Unpaired, two-tailed Student’s t test. *P < 0.05, **P < 0.01, ***P < 0.001). Source data are provided as a Source Data file.

Fig. 6. The Ucp1-mediated thermogenic pathway is activated in AA005-treated mice and Hadha^ΔAdipo mice.

a Heatmap generated by RNA-seq analysis displaying differentially regulated genes between AA005-treated and Hadha^ΔAdipo mice fed HFD in iWAT (n = 5). Red and blue colors indicate upregulated or downregulated genes. b Significant enrichment in KEGG terms. Bars indicate the number or proportions of genes within each functional class, ranked based on protein coverage. c Heatmap showing the mean fold-change of the indicated metabolic processes in vehicle control, AA005-treated and Hadha^ΔAdipo mice fed HFD. Red and blue colors indicated upregulated or downregulated genes. d Hadha as the central node in the network of gene interactions with thermogenesis pathways. The protein expression and quantification analysis of UCP1, p-PKA, and p-HSL in BAT tissue derived from vehicle- or AA005-treated mice fed HFD for 22 weeks (e), as well as Hadha^fl/fl or Hadha^ΔAdipo mice fed HFD for 18 weeks (f) (n = 3). Actin was used as the loading control. g–l Vehicle- or AA005-treated mice fed CD subjected to cold room stress before analysis. Rectal temperature of vehicle- or AA005-treated mice (4 °C acute cold in (g) programmed cooling to 6 °C chronic cold in (j), n = 5 per group). Representative western blot and quantification of UCP1 and PKA activation in the BAT (acute cold in (i), chronic cold in (l)) (n = 3). Representative images of haematoxylin and eosin (H&E) stained of iWAT and BAT (acute cold in (h), iWAT, n = 248 (vehicle) and 301 (AA005) adipocytes per group; BAT, n = 287 (vehicle) and 193 (AA005) LDs per group, chronic cold in (k), iWAT, n = 100 (vehicle) and 309 (AA005) adipocytes per group; BAT, n = 600 (vehicle) and 478 (AA005) LDs per group; Scale bar, 50 μm. Data are presented as mean ± SEM. Unpaired, two-tailed Student’s t test. *P < 0.05, **P < 0.01, ***P < 0.001). Source data are provided as a Source Data file.

Fig. 7. The Ucp1-mediated thermogenic pathway contributes to the anti-obesity effect of AA005.

a PCR confirmation of Ucp1 allele deletion in Ucp1^−/− mice. Three different biological replicates were tested. b Weight gain recorded over 20 weeks in male littermates fed HFD with or without AA005 treatment in Ucp1^−/− mice (n = 6 for vehicle-treated Ucp1^−/− mice, n = 7 for AA005-treated Ucp1^−/− mice for each group). c Representative photograph of vehicle- or AA005-treated Ucp1^−/− mice fed HFD at experimental endpoint. Scale bar, 1 cm. d The average food intake of each mouse for 1 week was recorded (n = 6 for vehicle-treated mice, n = 7 for AA005-treated mice). e Fat and lean mass were recorded after 20 weeks treatment (n = 6 for vehicle-treated mice, n = 7 for AA005-treated mice). f Fat weight of iWAT, asWAT, eWAT, rWAT, mWAT, and BAT in Ucp1^−/− mice treated with vehicle or AA005 (n = 6 for vehicle-treated mice, n = 7 for AA005-treated mice). Glucose tolerance test (GTT) (g), area under the curve (AUC) of GTT (h), insulin tolerance test (ITT) (i), and AUC of ITT (j) in vehicle and AA005-treated Ucp1^−/− mice fed HFD (n = 6 for vehicle-treated mice, n = 7 for AA005-treated mice). k Representative images of haematoxylin and eosin (H&E) stained of iWAT, eWAT, and BAT. iWAT, n = 107 (Vehicle) and 114 (AA005) adipocytes per; eWAT, n = 115 (Vehicle) and 108 (AA005) adipocytes per group; BAT, n = 473 (Vehicle) and 474 (AA005) LDs per group. Scale bar, 50 μm. l AA005 targets HADHA to counteract obesity by inhibiting cardiolipin maturation and enhancing Ucp1-mediated thermogenesis pathway. AA005 demonstrates the ability to induce a brown-like transformation in adipose tissue and enhance energy expenditure in obese mice, thereby potentially exerting a pharmacological effect in combating obesity and improving glucose metabolism. Mechanistically, AA005 selectively enters adipocyte mitochondria and targets the binding site on HADHA, leading to inhibition of cardiolipin maturation in mitochondria. Consequently, AA005 alters the expression profile of cardiolipin. Notably, the presence of immature cardiolipin activates the Ucp1-mediated thermogenic pathway, consequently promoting mitochondrial thermogenesis. Data are presented as mean ± SEM. Unpaired, two-tailed Student’s t test. NS indicates not significant. Source data are provided as a Source Data file.