Glycogen distribution in the microwave-fixed mouse brain reveals heterogeneous astrocytic patterns

Abstract

In the brain, glycogen metabolism has been implied in synaptic plasticity and learning, yet the distribution of this molecule has not been fully described. We investigated cerebral glycogen of the mouse by immunohistochemistry (IHC) using two monoclonal antibodies that have different affinities depending on the glycogen size. The use of focused microwave irradiation yielded well-defined glycogen immunoreactive signals compared with the conventional periodic acid-Schiff method. The IHC signals displayed a punctate distribution localized predominantly in astrocytic processes. Glycogen immunoreactivity (IR) was high in the hippocampus, striatum, cortex, and cerebellar molecular layer, whereas it was low in the white matter and most of the subcortical structures. Additionally, glycogen distribution in the hippocampal CA3-CA1 and striatum had a 'patchy' appearance with glycogen-rich and glycogen-poor astrocytes appearing in alternation. The glycogen patches were more evident with large-molecule glycogen in young adult mice but they were hardly observable in aged mice (1-2 years old). Our results reveal brain region-dependent glycogen accumulation and possibly metabolic heterogeneity of astrocytes. GLIA 2016;64:1532-1545.

Keywords: aging; astrocytes; focused microwave irradiation; glycogen; immunohistochemistry.

Figure 1

Glycogen IHC using IV58B6 and ESG1A9 antibodies. (A) Schematic illustration of antibody affinity to glycogen. (B) Comparison of glycogen IHC using two different euthanasia methods: focused microwave irradiation‐assisted fixation (see methods) vs. perfusion fixation. Both IV58B6 (IV) and ESG1A9 (ESG) antibodies were subjected to IHC of the mouse dorsal hippocampal CA1. CC corpus callosum, SO stratum oriens, SP stratum pyramidale, SR stratum radiatum. (C) Magnified views of glycogen IHC in CA1 stratum radiatum from doted red box in (B) with the same arrangement as in (B). (D) Comparison of whole brain glycogen content by biochemical assay. (mean ± SEM, N = 6, ***P < 0.001, unpaired t‐test). (E) PAS staining of microwave‐fixed (upper) and perfusion‐fixed (lower) hippocampal CA1 sections. Dotted regions are magnified in the right panels. Arrowheads points to glycogen‐like granules which appear as more intense spots in the PAS staining. (F) Amylase‐digested hippocampal CA1 sections are not labeled with either IV58B6 or ESG1A9. Scale bars: (B) 200 μm; (C) 50 μm; (E) 200 μm; (left) 20 μm (right); (F) 100 μm. [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]

Figure 2

Brain‐wide distribution of glycogen in the brain. (A) Coronal sections of glycogen IHC by IV58B6 (left) and ESG1A9 (right). (B) Sagittal sections of glycogen IHC by IV58B6 (upper) and ESG1A9 (lower). Relative IV58B6 IR in various brain regions of the brain was evaluated for coronal (C) and sagittal (D) sections (mean ± SEM). These comparisons are pictorially represented in (E) and (F). (G) Glycogen assay for microwave‐fixed cortex, hippocampus, striatum, and cerebellum. Note that the cerebellar assay contains both the cerebellar cortex and white matter (mean ± SEM, N = 7, **P < 0.01, ***P < 0.001, one‐way ANOVA with Tukey HSD test). Fluorescence image of a PAS‐stained section (H) is compared with MBP IHC (I), which labels myelin. Scale bars: (A), (B), (H), and (I) 1 mm. [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]

Figure 3

Punctate glycogen IR in various brain regions. (A) Confocal images of various brain regions showing glycogen IR with IV58B6 (left) and ESG1A9 (right) antibodies. (B) Double IHC of IV58B6 with cell‐type specific markers for astrocytes (GFAP), microglia (Iba1), or neurons (NeuN) in the CA1 s.r. (C) Double IHC of ESG1A9 with cell‐type specific markers, as in (B). Cell‐type dependent localization of glycogen IR by IV58B6 (D, F, and H) or ESG1A9 (E, G, and I) was examined by computing the area fraction of glycogen IR to a respective cell‐type specific marker (astrocyte, GFAP; microglia, Iba1; neuron, NeuN; oligodendrocyte, MBP) in the CA1 s.r. (D and E), cerebral cortex (F and G), and corpus callosum (H and I) (mean ± SEM, N = 7–11 animals, n = 140 images, *P < 0.05, **P < 0.01, ***P < 0.001, one‐way ANOVA with Tukey HSD test). Scale bars: (A) 60 μm; (B and C) 30 μm.

Figure 4

Subcellular distribution of glycogen in cortical and hippocampal astrocytes. (A and B) Confocal images of IV58B6 (upper row) and ESG1A9 (lower row) glycogen immunofluorescence with EYFP‐expressing astrocytes in Layer 1 of the parietal cortex (A) and hippocampus (B). EYFP expression in a sparse population of astrocytes was induced by tamoxifen administration to GLAST‐CreERT × Rosa26‐LSL‐EYFP mice. Scale bars: (A and B) 20 μm. (C–F) Subcellular glycogen distribution in EYFP‐labeled cortical astrocytes were analyzed by two methods; glycogen puncta density (particles/μm²; C and D) and area fraction of glycogen immunofluorescence (%; E and F). Blue and red bar graphs represent analyses from IV58B6 and ESG1A9 data, respectively. (mean ± SEM, N = 5 animals, n = 45 images, *P < 0.05, **P < 0.01, ***P < 0.001, one‐way ANOVA with Tukey HSD test). (G–J) Subcellular of glycogen distribution analyses in the hippocampus, performed in the same manner as (C–F) (mean ± SEM, N = 5 animals, n = 47 images, *P < 0.05, **P < 0.01, ***P < 0.001, one‐way ANOVA with Tukey HSD test).

Figure 5

Glycogen distribution in the cerebral cortex. (A) IV58B6 glycogen IHC (middle) of the parietal association cortex is displayed with NeuN IHC of the contiguous section for neuronal cell body distribution (left). Pseudocolored image of the IV58B6 labeling is shown on the right. (B) Comparison of IV58B6 signal intensity across cortical layers of the parietal association cortex (mean ± SEM, N = 9, *P < 0.05, **P < 0.01, ***P < 0.001, one‐way ANOVA with Tukey HSD test). (C) Comparison of cytochrome c oxidase staining (CO) (left, brightness and contrast adjusted) and IV58B6 IHC (middle) for the barrel area of the primary somatosensory cortex demonstrates a lack of distinct Layer 4 (between red dotted lines) barrel structure in glycogen distribution. Pseudocolored image of the IV58B6 labeling is shown on the right. (D) IV58B6 IR patterns in the primary visual cortex, primary motor cortex, and prefrontal cortex (PrF). There is a general tendency that superficial layers yield higher signals than deep layers, as shown in (B) for the parietal association cortex. Scale bars: (A, C, and D) 100 μm.

Figure 6

Stratified and patchy distribution of glycogen in the hippocampus. Glycogen IHC by IV58B6 (A) and ESG1A9 (B) in the hippocampal CA1 and dentate gyrus. In the merged panel (right), NeuN (left) and glycogen (middle) signals are displayed in red and green, respectively. SO stratum oriens, SP stratum pyramidale, SR stratum radiatum, SLM stratum lacunosum‐moleculare, DGML‐U dentate gylus molecular layer upper blade, DGGL‐U dentate gylus granular layer upper blade, HL hilus, DGML‐L dentate gylus molecular layer lower blade, DGGL‐L dentate gylus granular layer lower blade. Layer‐dependent glycogen IR signal profile by IV58B6 (C) and ESG1A9 (D). Signal strength of glycogen IHC was quantified (N = 5) for various layers of the hippocampus by taking the CA1 pyramidal cell layer as reference (mean ± SEM). (E and F) Images from CA1 s.r. with IV58B6 (left) and ESG1A9 (right) are displayed for autocorrelation analysis. 2D‐autocorrelogram is shown on the right side of each micrograph. High and low correlation areas appear in alternation, indicating patchy distribution. Patchy distribution is more evident with ESG1A9. Images were taken from three individual animals. Green and yellow circles correspond to “on‐patch” and “off‐patch” areas, respectively. (G) Patchy glycogen distribution and 2D‐autocorrelogram in the striatum. (H) Similar analysis for L2/3 of the cerebral cortex, where patchy distribution is not observed. Images in (A) and (B) are in the same scale. Histological images (upper row) in (E–H) are in the same scale and numbers on the axes of 2D‐autocorrelograms (lower row) are in μm. Scale bars: (A and B) 100 μm; (H) 50 μm.

Figure 7

Patchy structure is composed of individual astrocytes. (A) ESG1A9 glycogen IHC (green) in the hippocampal CA1 s.r. was counterstained with GFAP (red). Some astrocytes are distinctively low in ESG1A9‐labeled glycogen, as marked with yellow circles. (B) Example of an ESG1A9‐rich astrocyte. GLAST‐CreERT × Rosa26‐LSL‐EYFP mouse is used to induce EYFP expression in a sparse population of astrocytes. Green line represents the boarder of the EYFP‐expressing and ESG1A9‐rich astrocyte. (C) Example an ESG1A9‐poor astrocyte from the same mouse as in (B). Yellow line represents the boarder of the EYFP‐expressing and ESG1A9‐poor astrocyte. Glycogen accumulation depends on individual astrocytes and does not depend on EYFP expression. Images in (B) and (C) are in the same scale. Scale bars: (A) 100 μm and (C) 20 μm.

Figure 8

Diminishment of patchy glycogen distribution by aging. (A) IV58B6 and ESG1A9 glycogen IHC of the hippocampus from young adult (9–12 weeks old) and aged (1 year old) mice is displayed. Patchy distribution is evident in the young adult mouse whereas IV58B6 signal pattern is smoother and ESG1A9 signal is largely diminished in the aged mouse. Additionally, abnormal glycogen accumulation, known as polyglucosan body, is observable in the aged mouse (arrows). (B) Images of the striatal glycogen IHC from the same animals as in (A). ESG1A9 glycogen IHC of the hippocampus from even older mice [(C) 1.5 years old and (D) 2 years old]. The images from aged mice show clear diminishment of the patchy distribution and more frequent occurrence of polyglucosan with aging. (E) 2D‐autocorrelation in the CA1 s.r. was computed for a young adult mouse and a 1.5‐year‐old mouse. Numbers on the axes of 2D‐autocorrelograms are in μm. Scale bars: All scale bars are 500 μm. IV58B6 in aged (2 years old) was uniformly stained similar to (A, upper right panel). (F) Biochemical assessment of glycogen was performed from young adult and aged (2 years old) mouse hippocampus. In both age groups, similar glycogen amounts were measured. [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]