Visualization of spatiotemporal energy dynamics of hippocampal neurons by mass spectrometry during a kainate-induced seizure

Abstract

We report the use of matrix-assisted laser desorption/ionization (MALDI) imaging mass spectrometry combined with capillary electrophoresis (CE) mass spectrometry to visualize energy metabolism in the mouse hippocampus by imaging energy-related metabolites. We show the distribution patterns of ATP, ADP, and AMP in the hippocampus as well as changes in their amounts and distribution patterns in a murine model of limbic, kainate-induced seizure. As an acute response to kainate administration, we found massive and moderate reductions in ATP and ADP levels, respectively, but no significant changes in AMP levels--especially in cells of the CA3 layer. The results suggest the existence of CA3 neuron-selective energy metabolism at the anhydride bonds of ATP and ADP in the hippocampal neurons during seizure. In addition, metabolome analysis of energy synthesis pathways indicates accelerated glycolysis and possibly TCA cycle activity during seizure, presumably due to the depletion of ATP. Consistent with this result, the observed energy depletion significantly recovered up to 180 min after kainate administration. However, the recovery rate was remarkably low in part of the data-pixel population in the CA3 cell layer region, which likely reflects acute and CA3-selective neural death. Taken together, the present approach successfully revealed the spatiotemporal energy metabolism of the mouse hippocampus at a cellular resolution--both quantitatively and qualitatively. We aim to further elucidate various metabolic processes in the neural system.

Figure 1. CA3 cell-selective consumption of ATP and ADP during a kainate-induced seizure.

a. MALDI imaging of adenosine nucleotides in a mouse hippocampus. b. Absolute quantification of ATP, ADP, and AMP in a mouse cerebrum using CE-MS. Massive reductions in the levels of ATP and ADP, but not AMP were observed during kainate-induced seizures. (n = 4 mice for each group; *p<0.05, **p<0.01; Welch's t-test). c. Results of the relative quantification of ion intensity for ATP, ADP, and AMP calculated from the averaged mass spectra of each hippocampal sub-region obtained using MALDI imaging. The values shown are logarithmic ratios of ion intensities between sham-operated (sham) and kainate-treated mice (KA). (n = 4 mice for each group; *p<0.05, **p<0.01; Welch's t-test). d. Mapping of energy-charge index values on tissue sections. The region-specific reduction of these values in the CA3 region (arrows) suggests massive energy metabolism in CA3 neurons. e. Relative quantitative comparison of adenosine nucleotides and related metabolites using CE-MS. Each result is mapped on the metabolic pathway and clearly shows the depletion of ATP and ADP due their conversion into downstream metabolites. Colored graphs indicate significant increases (orange) and decreases (blue). (n = 4, *p<0.01, **p<0.05; Welch's t-test).

Figure 2. Enhanced energy production pathways compensate for ATP consumption.

Relative quantitative comparisons of glycolysis and TCA cycle-related metabolites using CE-MS. In addition, IMS results of key metabolites for each pathway, including fructose 1,6-bi-phosphates in glycolysis and citrate in the TCA cycle, are also presented. Each result is mapped on the metabolic pathway (n = 4; *p<0.05, **p<0.01; Welch's t-test). Colored graphs indicate significant increases (orange) and decreases (blue). Open graphs indicate “not detected.” The full name of each abbreviated metabolite is shown in Table S1.

Figure 3. Recovery of energy consumption in CA1 and CA3 cells.

The MALDI-imaging results of mouse hippocampus of control and kainate-administered mice at 30 and 180 minutes after administration. The signals of adenosine nucleotides as well as the calculated EC-index values on each pixel, were collected from CA1 and CA3 cell-layer regions and averaged. Values are presented as means (S.E.M.) Control (sham-operated, n = 3), 30 min (n = 4), and 180 min (n = 4).

Figure 4. Heterogeneous energy recovery capacity within CA3 cell populations.

Two-dimensional EC-value maps were reconstructed from MALDI-imaging data (at a spatial resolution of 22 µm) by calculating EC values at each pixel using adenosine nucleotide signals; these are shown in optical images of HE-stained tissue sections (a–c). The pixel appearance frequencies of each EC-index value range are displayed as histograms (d–f). The histograms were fitted into single- and double-peaked Gaussian distribution curves; for the 3 presented conditions, the curve with greatest correlation factor (R2) was adopted and shown.