In Vivo Reductionist Approach Identifies miR-15a Protecting Mice From Obesity

Abstract

Obesity is a growing medical and social problem worldwide. The control of energy homeostasis in the brain is achieved by various regions including the arcuate hypothalamic nucleus (ARH). The latter comprises a number of neuronal populations including the first order metabolic neurons, appetite-stimulating agouti-related peptide (AgRP) neurons and appetite-suppressing proopiomelanocortin (POMC) neurons. Using an in vivo reductionist approach and POMC^Cre-dependent CRISPR-Cas9, we demonstrate that miR-15a-5p protects from obesity. Moreover, we have identified Bace1, a gene previously linked to energy metabolism imbalance, as a direct target of miR-15a-5p. This work warrants further investigations of non-coding RNA-mediated regulation of energy homeostasis and might contribute to the development of novel therapeutic approaches to treat metabolic diseases.

Keywords: Dicer; energy homeostasis; hypothalamus; in situ CRISPR/Cas9 knock-out; mice; microRNA; obesity.

Figure 1

Reductionist approach to identify microRNAs attenuating obesity phenotype caused by neuronal Dicer1 depletion. (A) Scheme of the experiment to deliver liposomal formulations of microRNA mimics mixtures into the arcuate hypothalamic nucleus (ARH) of mice with a depletion of Dicer1 in the forebrain neurons. (B) Body weight dynamics in DicerCKO mice injected by the indicated LNA-stabilized microRNAs mimics Group 1 and Group 2 or scrambled oligonucleotides, and control mice (n = 9, 4, 9 and 5, respectively). (C,D) Weight of perigonadal (C) and perirenal (D) fat pads from DicerCKO mice injected by the microRNAs mimics Group 1 or scrambled oligonucleotides, and control mice (n = 9, 9 and 5, respectively). Error bars represent standard error of means (SEM). *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001 vs. groups indicated by respective colors, as assessed by 2-way (B, F_{(23, 438)} = 44.9, F_{(3, 438)} = 323.4 and F_{(69, 438)} = 6.3 for time, experimental group and interaction factors) or 1-way (C, D, F_{(2, 20)} = 21.47 and 24.57, respectively) ANOVA followed by Holm-Šídák pairwise comparison tests.

Figure 2

Stabilized miR-15a-5p mimic attenuates hyperphagia and obesity in DicerCKO mice. (A) Body weight dynamics in DicerCKO females injected by miR-27b-3p, miR-320-3p, miR-15a-5p LNA-stabilized mimics or scrambled nucleotides and control mice injected to the arcuate hypothalamic nucleus (ARH) by scrambled oligonucleotides (n = 5, 5, 5, 4, 9, respectively). A separate experiment was done to deliver high dose of miR-15a-5p mimics to ARH of DicerCKO or controls mice (n = 5). (B–D) Perigonadal (B), perirenal (C) fat pad and food intake (D) analyses in Control - Scrambled, DicerCKO - Scrambled and DicerCKO - miR-15a-5p mice (n = 9, 4, 5, respectively). Error bars represent SEM. *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001 vs. groups indicated by respective colors, as assessed by 2-way (A, F_{(23, 529)} = 466.9, F_{(23, 529)} = 60.48 and F_{(92, 529)} = 29.66 for time, experimental group and interaction factors; D, F_{(5, 55)} = 912.4, F_{(11, 55)} = 18.44 and F_{(10, 55)} = 15.71 for time, experimental group and interaction factors) or 1-way (B, C, F_{(2, 15)} = 46.84 and 80.56, respectively) ANOVA followed by Holm-Šídák pairwise comparison tests.

Figure 3

POMC^Cre-restricted depletion of miR-15a causes obesity in mice. (A) Scheme of the experiment to inactivate the miR-15a/16-1 cluster or all miR-15 family microRNAs by injection of single guide RNA-equipped recombinant adeno-associated virus (rAAV) to the arcuate hypothalamic nucleus of POMC^CreCas9 males or females, respectively. (B) On-target effectivity validation of single guide RNAs to targeting the miR-15 family (n = 5, 5, 5, 4, 5, 5, respectively for the indicated comparison groups). (C) Microphotographs of the arcuate hypothalamic nucleus stained by in situ hybridization probe against miR-15a-5p, anti-EGFP antibody labelling POMC^Cre-GFP neurons and anti-mCherry antibody staining sgRNA-equipped rAAV-transduced cells. POMC^Cre-GFP neurons effectively transduced (POMC^CreGFP⁺mCherry⁺) or not transduced (POMC^CreGFP⁺mCherry^-) by rAAVs are indicated by red triangles or yellow arrows, respectively. (D-G) Body weight dynamics (D), hematoxylin and eosin staining (E) and weight of perirenal fat tissue (F), perigonadal fat pad weight (G) in POMC^Cre –miR-15/16CKO and control female mice (n = 4 and 5, respectively). (H, I) Representative photographs (H) and weight gain dynamics (I) in POMC^Cre –miR-15aCKO male mice and controls (n = 8 and 10, respectively). Error bars represent SEM. *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001 as assessed by unpaired two-tailed Student’s t-test (B, F, G) or 2-way ANOVA (D, F_{(18, 126)} = 26.28, F_{(7, 126)} = 47.22 and F_{(18, 126)} = 4.31 for time, experimental group and interaction factors; I, F_{(29, 413)} = 10.77, F_{(1, 413)} = 309.67 and F_{(29, 413)} = 1.32 for time, experimental group and interaction factors) followed by Holm-Šídák pairwise comparison tests.

Figure 4

Bace1 is a target of miR-15a-5p. (A, B) Transcriptomics profiling of the arcuate hypothalamic nucleus (see Figure 2 ) reveals genes (indicated by blue circles, Table S4 ) that are both significantly up-regulated in DicerCKO - Scrambled vs. Control - Scrambled mice and significantly down-regulated in DicerCKO - miR-15a-5p vs. DicerCKO - Scrambled mice (A) and Bace1 among the latter genes (panel B and indicated with a red circle in A) that demonstrated the highest significance of up-regulation in DicerCKO - Scrambled group (n = 3). Black circle indicates the central point of the graph where relative expression vs. any group is equal. Acronyms of 15 genes with the highest fold change of up-regulation in DicerCKO - Scrambled vs. Control - Scrambled group are depicted (see Table S4 for more details). (C) Validation of Bace1 as a target of the mmu-miR-15a-5p by dual-luciferase assay (n = 5, 5, 4, respectively for Scrambled treated, empty vector and miR-15a-5p-treated groups). Error bars represent SEM. **p < 0.01; ***p < 0.001; ****p < 0.0001 as assessed by 1-way (F_{(2, 6)} = 87.81 and F_{(2, 11)} = 21.1 for (B, C), respectively) ANOVA followed by Holm-Šídák pairwise comparison tests.