Quantification and correlation of amyloid-β plaque load, glial activation, GABAergic interneuron numbers, and cognitive decline in the young TgF344-AD rat model of Alzheimer's disease

Abstract

Background: Animal models of Alzheimer's disease (AD) are essential tools for investigating disease pathophysiology and conducting preclinical drug testing. In this study, we examined neuronal and glial alterations in the hippocampus and medial prefrontal cortex (mPFC) of young TgF344-AD rats and correlated these changes with cognitive decline and amyloid-β plaque load.

Methods: We compared TgF344-AD and non-transgenic littermate rats aged 7-8 months of age. We systematically quantified β-amyloid plaques, astrocytes, microglia, four different subtypes of GABAergic interneurons (calretinin-, cholecystokinin-, parvalbumin-, and somatostatin-positive neurons), and newly generated neurons in the hippocampus. Spatial learning and memory were assessed using the Barnes maze test.

Results: Young TgF344-AD rats had a large number of amyloid plaques in both the hippocampus and mPFC, together with a pronounced increase in microglial cell numbers. Astrocytic activation was significant in the mPFC. Cholecystokinin-positive cell numbers were decreased in the hippocampus of transgenic rats, but calretinin-, parvalbumin-, and somatostatin-positive cell numbers were not altered. Adult neurogenesis was not affected by genotype. TgF344-AD rats had spatial learning and memory impairments, but this cognitive deficit did not correlate with amyloid plaque number or cellular changes in the brain. In the hippocampus, amyloid plaque numbers were negatively correlated with cholecystokinin-positive neuron and microglial cell numbers. In the mPFC, amyloid plaque number was negatively correlated with the number of astrocytes.

Conclusion: Pronounced neuropathological changes were found in the hippocampus and mPFC of young TgF344-AD rats, including the loss of hippocampal cholecystokinin-positive interneurons. Some of these neuropathological changes were negatively correlated with amyloid-β plaque load, but not with cognitive impairment.

Keywords: Barnes maze; CCK+ interneurons; astrocyte; cell number; cholecystokinin; hippocampus; medial prefrontal cortex; microglia.

Figure 1

Amyloid β plaques in the hippocampus and prefrontal cortex of TgF344-AD rats. (A) Representative images showing immunolabeled β-amyloid plaques in the hippocampus of wild-type (WT/WT) and transgenic (TG/TG) rats. (B) Quantification of plaque numbers confirmed an even distribution of amyloid β plaques in all hippocampal subareas (DG, dentate gyrus; CA, Cornu Ammonis). (C) Aβ plaques in the medial prefrontal cortex. (D) Similar to the hippocampus, plaque densities showed an equal distribution in all subareas of the medial prefrontal cortex (anterior cingulate (aCg), pre-limbic (PrL), and infralimbic (IL) cortices). Statistical analysis: Two-way ANOVA (genotype × brain area) followed by Šídák’s multiple comparisons post-hoc test (****p < 0.0001). Scale bars represent 500 μm for all images.

Figure 2

Glial activation surrounding β-amyloid plaques. (A) β-amyloid plaques (red) surrounded by activated Iba-1+ microglia (green) in the hippocampal dentate gyrus. (B) The same phenomenon is observed at a higher magnification. Iba-1+ microglial processes (green) are oriented towards β-amyloid plaques (red). Cell nuclei were labeled with DAPI (blue). (C) β-amyloid plaques (red) surrounded by Iba-1+ microglia (white) and GFAP+ astrocytes (green) in the hippocampal dentate gyrus. (D) The same hilar area where GFAP+ astrocytes surrounded the β-amyloid plaques (red). (E) Enlarged detail of D displaying GFAP+ astrocytes encircling the plaques (red). Scale bars: 100 μm on A, C, D, E and 50 μm on B. gcl, granule cell layer; str. mol., stratum moleculare.

Figure 3

TgF344-AD rats had pronounced activation of Iba-1+ glia in both the hippocampus and mPFC. (A) Representative images of Iba-1+ cells in the hippocampi of wild-type and transgenic rats. (B) Quantitative data for Iba-1+ cells in the hippocampus. TgF344-AD rats had a significantly higher number of Iba-1+ cells in the hippocampus. Cell numbers indicate cell counts from both hemispheres. Statistical analysis: unpaired Student’s t-test; ****p < 0.0001. (C) Iba-1+ cells in the mPFC. (D) TgF344-AD rats have a significantly higher number of Iba-1+ cells in the mPFC. Statistical analysis: unpaired Student’s t-test; ** p = 0.001. (E) A high magnification image of Iba-1+ cells in the mPFC of a wild-type rat. (F) High magnification image of activated Iba-1+ cells in the mPFC of a TgF344-AD rat. gcl, granule cell layer; str. mol., stratum moleculare. Scale bars represent 200 μm in (A,C), and 50 μm in (E,F).

Figure 4

TgF344-AD rats had GFAP+ gliosis in the mPFC. (A) Representative images of GFAP+ astrocytes in the hippocampi of wild-type and transgenic rats. (B) Quantitative data for GFAP+ cells in the hippocampus. In the hippocampus, GFAP+ cell numbers were not affected by genotype. Cell numbers indicate cell counts from both hemispheres. (C) GFAP+ glia in the mPFC. (D) In the mPFC, TgF344-AD rats had significantly higher density of GFAP+ astrocytes. Statistics: unpaired Student’s t-test, *** p < 0.0004. (E) A high magnification image of GFAP+ cells in the mPFC of a wild-type rat and activated GFAP+ cells in the mPFC of a TgF344-AD rat (F). gcl, granule cell layer; str. mol., stratum moleculare. Scale bars represent 200 μm in (A,C), and 50 μm in (E,F).

Figure 5

Parvalbumin-positive interneurons were not affected by the genotype. Representative images of hippocampal interneurons expressing parvalbumin (A), and the corresponding cell quantification data on cell numbers (B). Parvalbumin-positive neurons in the mPFC (C), and PV+ cell densities in the anterior cingulate (aCg), prelimbic (PrL) and infralimbic (IL) cortices (D). The number of PV+ neurons was not altered in the transgenic rats. (E) A high magnification image of PV+ interneurons in the dentate gyrus of a wild-type rat and PV+ cells in the dentate gyrus of a TgF344-AD rat (F). DG, dentate gyrus; gcl, granule cell layer; str. mol., stratum moleculare; aCg, anterior cingulate; IL, infralimbic; mPFC, medial prefrontal cortex; PrL, prelimbic. Scale bars represent 500 μm in A, 200 μm in (C), and 50 μm in (E,F).

Figure 6

The number of calretinin-positive neurons were not affected by genotype. Calretinin immunopositive cells in the hippocampus (A), and CR+ cell numbers in the hippocampal subareas (B). Calretinin immunoreactive neurons in the mPFC (C), and CR+ cell densities in the mPFC (D). The number of CR+ neurons was not altered in the transgenic rats. (E) A high magnification image of CR+ interneurons in the dentate gyrus of a wild-type rat. (F) A high magnification image of CR+ neurons in the dentate gyrus of a TgF344-AD rat. DG, dentate gyrus; gcl, granule cell layer; str. mol., stratum moleculare; aCg, anterior cingulate; IL, infralimbic; mPFC, medial prefrontal cortex; PrL, prelimbic. Scale bars represent 500 μm in A, 200 μm in (C), and 50 μm in (E,F).

Figure 7

Somatostatin-positive neurons were not altered by genotype. Somatostatin-positive neurons in the hippocampus (A), and the corresponding cell quantification data on hippocampal SST+ cell numbers (B). Somatostatin-positive cells in the mPFC (C), and a graph depicting SST+ cell densities in the anterior cingulate, prelimbic and infralimbic cortices (D). The number of SST+ neurons was not altered in the transgenic rats. (E) A high magnification image of SST+ interneurons in the dentate gyrus of a wild-type rat. (F) A high magnification image of SST+ neurons in the dentate gyrus of a TgF344-AD rat. DG, dentate gyrus; gcl, granule cell layer; str. mol., stratum moleculare; aCg, anterior cingulate; IL, infralimbic; mPFC, medial prefrontal cortex; PrL, prelimbic. Scale bars represent 500 μm in (A), 200 μm in (C), and 50 μm in (E,F).

Figure 8

Cholecystokinin-positive neurons in the hippocampus and prefrontal cortex. (A) Representative images of CCK+ GABAergic neurons in the hippocampus of wild-type rats. (B) CCK+ interneurons in the hippocampus of transgenic rats. (C) Higher magnification images of CCK+ neurons in the dentate gyrus of wild-type rats. CCK + cells were clearly identifiable. (D) Higher magnification images of CCK+ neurons in the dentate gyrus of transgenic rats. (E) CCK+ cells in the frontal cortex of wild-type rats. (F) CCK+ neurons and CCK+ plaques in the frontal cortex of transgenic rats. Black arrowheads indicate CCK+ neurons, whereas open arrowheads point to CCK+ plaques. CCK+ plaques were present mainly in the neocortex of the transgenic rats and these objects were most likely staining artefacts. CA, Cornu Ammonis; DG, dentate gyrus; gcl, granule cell layer. Scale bars represent 500 μm in (A,B), and 100 μm in (C–F).

Figure 9

TgF344-AD rats had reduced numbers of cholecystokinin-positive (CCK+) interneurons in the hippocampus. (A) Quantification of CCK+ cells revealed that TgF344-AD rats had a reduced number of CCK+ cells in the CA1 and CA3 areas, as well as in the entire hippocampus. Cell numbers indicate cell counts from both hemispheres. Statistical analysis: Two-way ANOVA (genotype × brain area) followed by Šídák’s multiple comparisons post-hoc test CA1: ** p = 0.0014; CA3 ** p = 0.0020; entire hippocampus: ****p < 0.0001. (B) Cell quantification data revealed no genotype effect on CCK+ cell densities in the mPFC. CA, Cornu Ammonis; DG, dentate gyrus; aCg, anterior cingulate; PrL, prelimbic; IL, infralimbic.

Figure 10

Adult hippocampal neurogenesis was not altered in the young TgF344-AD rats. (A,B) Representative images of doublecortin-positive (DCX+) immature neurons in the hippocampal dentate gyrus of wild-type and transgenic rats. (C) Systematic cell-count data. We found a small, but non-significant reduction of DCX+ cells numbers in the transgenic rats. Cell numbers indicate cell counts from both hemispheres. DG, dentate gyrus; gcl, granule cell layer. Scale bars represent 200 μm.

Figure 11

Learning and memory in the Barnes maze. (A,B) Line diagrams represent the learning curves in the Barnes maze task, where rats were tested in four subsequent trials on day 1 (A) and day 2 (B). Spatial learning of the wild-type rats was significantly faster in the trials of the first day. Repeated measures two-way ANOVA (time × genotype) detected significant difference between the learning curves of the wild-type and transgenic rats on the first training day (p < 0.01), and Šídák’s multiple comparisons post-hoc test detected a significant difference between the first trial of day 1 (* p < 0.05). (C,D) Results of the two probe trials on Day 3 (C) and day 10 (D) indicating spatial memory competences. On Day 3, no difference was present between the WT/WT and Tg/Tg animals, but on Day 10 transgenic rats spent significantly more time finding the escape box, indicating impaired spatial memory Statistics: unpaired Student’s t-test, ***p < 0.0001.

Figure 12

Correlations analysis between Aβ plaque numbers and cellular changes. (A) In the hippocampus, we found a negative correlation between Aβ plaque numbers and Iba-1+ cell numbers (p < 0.05). (B) Similarly, negative correlation was found between Aβ plaque numbers and CCK+ interneuron numbers (p < 0.05) of the hippocampus. (C) In the mPFC, there was a negative correlation between Aβ plaque density and astrocyte density (p < 0.05).