Metformin Reduces Repeat Mild Concussive Injury Pathophysiology

Abstract

Mild traumatic brain injury (mTBI) can initiate complex pathophysiological changes in the brain. Numerous cellular and molecular mechanisms underlying these pathologic changes are altered after injury, including those involved in energy utilization, excitotoxicity, ionic disturbances, and inflammation. It is thought that targeting multiple mechanisms may be necessary to produce meaningful reductions in brain pathology and to improve outcome. Previous studies have reported that the anti-diabetic drug metformin can also affect inflammatory, cell survival, and metabolic outcomes, possibly by acting on multiple targets and/or pathways. We therefore questioned whether metformin treatment can reduce pathology after repeat mild closed head injury (rmCHI) in male C57Bl/6 mice. We found that metformin, administered acutely after each head impact, resulted in markedly reduced white matter damage, astrogliosis, loss of hippocampal parvalbumin neurons, and improved mitochondrial function. In addition, both motor and cognitive functions were significantly improved when tested after discontinuation of the treatment. These studies suggest that metformin may be beneficial as a treatment for repeat concussion.

Keywords: axonal injury; cognitive dysfunction; mild traumatic brain injury; oxygen consumption rate; repeat concussion; tissue respiration.

Figure 1.

Postinjury metformin administration reduced rmCHI-triggered axonal damage and astrogliosis. A, Schematic showing the injury and drug treatment paradigm. Representative photomicrographs showing (B) APP accumulation and (C) enhanced silver impregnation in the corpus callosum (cc) and hippocampal commissure (hc) of a sham, a rmCHI+Veh, and a (rmCHI+Met) mouse. D, Summary graph showing that rmCHI significantly increased silver impregnation in the cc and hc, which was reduced in animals treated with metformin. E, Representative images of glial fibrillary acidic protein (GFAP) immunoreactivity showing increased reactive astrogliosis in the cc and hc of a rmCHI+Veh mouse, and reduced staining in injured a rm+Met mouse, as compared with sham. F, Summary data for GFAP immunoreactivity in the cc+hc fiber tracts. G, Images of hippocampi and parietal cortex from a sham, an rmCHI+Veh mouse, and an rmCHI+Met mouse demonstrating GFAP immunoreactivity. H, Summary data for GFAP immunoreactivity in the hippocampus. Data are presented as the mean ± SEM; *p < 0.05 by one-way ANOVA.

Figure 2.

Metformin treatment preserves hippocampal parvalbumin-expressing cells after rmCHI. A, Montage images of coronal mouse brain sections and hippocampi from sham, two-week postinjury rmCHI mouse treated after each injury with vehicle (rmCHI+Veh), and a two-week postinjury rmCHI mouse treated after each injury with metformin (rmCHI+Met). Sections were immunostained with antibodies against (A, B) the neuronal marker NeuN, (C) the synaptic vesicle protein Vesicular glutamate transporter 1 (vGlut1), and (D) the synaptic vesicle protein Vesicular glutamate transporter 2 (vGlut2). E, Representative capillary western analyses and summary data showing that neither rmCHI nor postinjury metformin treatment had any demonstrable effect on the hippocampal levels of postsynaptic density protein 95 (PSD95), synaptogyrin 1 (Syngr1), neuronal synaptobrevin (n-Syb), or synaptic vesicle protein 2 (SV2). Representative photomicrographs showing (F) doublecortin (DCX) and (G) parvalbumin (PV) immunoreactivity in the hippocampus of sham, rmCHI+Veh, and rmCHI+Met mice. H, Summary data (n = 4/group) showing that rmCHI had no statistical effect on the number of doublecortin-positive newborn neurons in the hippocampus when examined 3 d after injury. I, Summary data (n = 3/group) showing that rmCHI reduced the number of PV-positive interneurons in all hippocampal regions examined. Metformin treatment reversed the loss of PV-positive interneurons in the CA1, CA2, and CA3 regions, and attenuated the loos in the DG/hilus. dg: dentate gyrus; *p < 0.05 by one-way ANOVA.

Figure 3.

Postinjury metformin administration reduced mitochondrial proton leak and increased ATP-linked respiration in the hippocampus after rmCHI. A, Timeline for injury (once daily for 4 d), metformin administration, and respiration measurements either 2 h or 7 d after the final injury was given on the fourth day. B, Drawing of a coronal brain section showing the relative positions of the tissue punches taken for cortical and hippocampal tissue respiration measurements. C, A representative trace of the oxygen consumption rate (OCR) showing the different phases of mitochondrial respiration that can be identified using inhibitors. Changes in OCR resulting from the addition of the various mitochondrial inhibitors/uncouplers is used to determine the different components of mitochondrial (and nonmitochondrial) respiration. Summary results showing the effect of rmCHI and metformin treatment on the various aspects of mitochondrial respiration in the hippocampus at (D) 2 h and (E) 7 d postinjury. Summary results showing the effects of rmCHI and metformin treatment on mitochondrial respiration in the cortex at (F) 2 h and (G) 7 d postinjury. Data are presented as the mean ± SEM; *p < 0.05 by one-way ANOVA.

Figure 4.

Metformin activates pathways involved in mitochondrial fission. A, Summary data showing that ≥2 mm metformin present in the reaction buffer significantly inhibits basal respiration of isolated hippocampal mitochondrial; *p < 0.05 by one-way ANOVA compared with vehicle (0 mm metformin). B, Addition of ADP, which requires an intact proton gradient to stimulate oxygen consumption, had no effect on mitochondria treated with ≥4 mm metformin. Trifluoromethoxy carbonylcyanide phenylhydrazone (FCCP), which uncouples oxygen consumption from ATP synthesis by disrupting the proton gradient, stimulated OCR at all metformin concentrations. C, Representative HPLC chromatographs of cortical tissue extracts in the absence (blue) and presence (red) of metformin. D, Summary data showing the concentration of metformin in various tissues detected 6 h after intraperitoneal administration of 250 mg/kg to mice. Data are presented as the mean ± SEM. E, Representative capillary western analyses and summary data (n = 5/group) showing that neither rmCHI nor metformin treatment altered the expression of components of the electron transport system 2 h after the last injury. ATP5a: ATP synthase F1 subunit α; CIII core 2: cytochrome b-c1 complex subunit 2; CIV-COX1: mitochondrially encoded cytochrome c oxidase I. F, Representative capillary western analyses and summary data showing that rmCHI significantly decreases Mff phosphorylation and increases Opa1 levels compared with sham controls. Metformin treatment restored Mff phosphorylation and increased Drp1 levels. Drp1: dynamin 1-like protein; Mff: mitochondrial fission factor; Opa1: optic atrophy 1; Mfn2: mitofusin-2; *p < 0.05 by one-way ANOVA.

Figure 5.

Postinjury administration of metformin after rmCHI reduced vestibulomotor, motor, and cognitive dysfunction. A, Time line showing the experimental design. Sham, rmCHI+Veh, and rmCHI+Met (250 mg/kg) animals were tested immediately after injury for (B) duration of apnea, (C) suppression of tail pinch reflex, and (D) latency to regain righting response. Vestibulomotor and motor performances of sham, rmCHI+Veh, and rmCHI+Met mice were tested using the (E) beam balance and (F) foot fault (grid walking) tasks. On days 8–11, animals were tested for cognitive performance in the NOR task. G, Percent time exploring the two identical objects (F1 and F2) used for familiarization in the NOR task. All groups equally explored both objects. H, When tested for their object memory 24 h later, both sham and rmCHI+Met mice spent significantly more time exploring the novel object (N) rather than the familiar one (F), indicating intact recognition memory. rmCHI+Veh mice spent more time with the familiar object, an indication of perseveration. Performance of (I) sham, (J) rmCHI+Veh, and (K) rmCHI+Met in the context discrimination task. L, Comparison of the freezing differentials between “shock” and “safe” contexts in the rmCHI+Veh versus rmCHI+Met mice. Data are presented as the mean ± SEM; ‡p < 0.05 by two-way repeated measures ANOVA; *p < 0.05 by paired t test.