Metformin directly suppresses atherosclerosis in normoglycaemic mice via haematopoietic adenosine monophosphate-activated protein kinase - PubMed (original) (raw)
Figure 4
Metformin induces Mhem-like atheroprotective genes in hMDM. (A) Metformin induces HMOX1, APOE, ABCA1, and IGF1 (n = 8 different normal human donors). Gene induction was measured by RT-qPCR (Methods) by the ΔΔCt method with HPRT as housekeeping control. Equivalence of HPRT, GAPDH, and ACTAB as housekeepers was established. Mean ± SE, *each induction was statistically significant (ANOVA, P < 0.05, with Holm–Sidak correction, after normality testing by Shapiro–Wilk). _X_-axes, time in days (log scale). (B) hMDM were transfected with si-RNA to knock down ATF1. _X_-axis, si-RNA transfection of hMDM; _Y_-axis, ATF1 mRNA levels by RT-qPCR (Methods) by the ΔΔCt method with HPRT as housekeeping control, mean ± SE, *P < 0.05, Student’s _t_-test, n = 5 different human donors. (C) Metformin induces mRNA for HMOX1 in hMDM in control-transfected macrophages, but not in ATF1-siRNA-transfected macrophages (mean ± SE *Student’s _t_-test P < 0.05, n = 5 different human donors). (D) Metformin induces binding of p-ATF1 to the HMOX1 enhancer at the CRE site at −4200 bp in hMDM. Chromatin immunoprecipitation was performed as before, HMOX1 enhancer enrichment quantified by qPCR as before and expressed as fold enrichment and normalised to the IgG control (mean ± SE, *Student’s _t_-test P < 0.05, n = 5 different human donors). (E) FASL expression in hMDM at culture Day 8, measured by RT-qPCR. _Y_-axis, FASL fold induction measured by the ΔΔCt method (mean ± SE, n = 8 different human donors, *ANOVA P < 0.05, with Holm–Sidak correction and normality testing by Shapiro–Wilk). (F) CCR5 expression is induced in hMDM by loading with OxLDL, and reversed by metformin treatment. _Y_-axis, CCR5 fold induction by qRT-PCR and ΔΔCt method (*P < 0.05, _t_-test, n = 4 different human donors). (G) Metformin-enhanced phagocytosis in hMDM. _Y_-axis, phagocytic index for each phagocytosable particle, defined as number of particles left after washing divided by number of phagocytes (counted by nuclear DAPI stain). Expressed as mean ± SE n = 9, *passes significance, P < 0.002 paired Students’ _t_-test (beads), P < 0.002 Wilcoxon (ApC), P < 0.03 Wilcoxon (OxLDL). (H) Principal component analysis of microarray analysis of metformin versus vehicle control treatment of cultured hMDM (n = 4 different human donors) with arrows showing a consistent small effect of metformin in the same direction in each of the donors. Each donor is represented by a pair of points. (I) Hierarchical clustering and heat-map of microarray data of hMDM metformin treatment. The donors tended to cluster together. The genes that were significantly modulated by metformin were assessed in greater depth. (J) Venn diagram comparing genes modulated by heme, metformin, or both (intersection) on microarray analysis of over-represented TFBS in each subset of genes indicated likely differences in signalling and transcription factor usage. (K) Classification of top metformin-modulated genes by Gene Ontology from microarray analysis. Many of the genes encoded enzymes in intermediary metabolism, cell signalling, or leucocyte activation. The largest single category was unknown function. (L) Metformin induces NAMPT. NAMPT was one of the few unanticipated metformin-modulated protein-coding genes and was validated by qPCR. _Y_-axis: fold induction by RT-qPCR and −2ΔΔCt method (mean ± SE *P<0.05, ANOVA, Holm–Sidak adjustment n = 8 different human donors, normality tested by Shapiro–Wilk).