Aβ and microglia analysis revealed unique regional patterns of pathology.
Characteristics of microglia and Aβ plaque pathology was quantified using Ilastik in the EC, HPC, SNr, RSc, and MSB (Fig. 1A and B). At one, three-, and five-months post seeding the immunohistochemical staining revealed a unique pattern of expression (Fig. 1C – H). Overall, the percent area covered by 82e1+ [F(2, 17) = 0.625, p = 0.547] and Iba1+ [F(2, 16) = 0.8688, p = 0.438] immunoreactivity was not significantly different between seeds (Fig. 1F.). Compared to earlier time points, the later time points showed a significant increase in the percent area covered by 82e1 + immunoreactivity [F(2, 17) = 4.415, p = 0.0286] but Iba1 + immunoreactivity [F(2, 16) = 3.435, p = 0.057].
Iba1 + immunoreactivity was found to be significantly different between regions of the brain sampled with immunoreactivity changing over time. One month post seeding the region sampled had a significant effect on Iba1 + immunoreactivity percent area [F(4, 12) = 7.39, p = 0.0030]. Overall, the SNr had significantly greater immunoreactivity compared to the EC, HPC, and MSB (p < 0.05). Three months post seeding, the region had significant effect on Iba1 + immunoreactivity [F(4, 20) = 21.08, p < 0.0001] and a significant region x seed interaction was found [F(8, 20) = 2.460, p = 0.049]. Overall, the SNr had significantly greater Iba1 + immunoreactivity compared to the EC, HPC, and MSB (p < 0.01). The HPC and RSc also had significantly greater Iba + immunoreactivity compared to the MSB (p < 0.05).
The interaction is due to the control and rpAD seeded mice showing significant differences between regions whereas the slAD seeded mice showed no regional differences. Consistently, the SNr had significantly greater Iba1 + immunoreactivity compared to all other regions in the control and rpAD seeded mice three months following seeding. In the rpAD seeded mice, the RSc, HPC, and SNr had significantly higher percent area covered compared to the MSB. Five months following seeding the effect of the region sampled had a significant effect on Ib1 + immunoreactivity [F(4, 32) = 4.83, p = 0.004]; the EC, SNr, and RSc all had significantly greater Iba1 + immunoreactivity compared to the MSB.
Analysis one month following seeding showed that the region sampled had a significant effect on 82e1+ immunoreactivity [F(4, 12) = 8.29, p = 0.002] but the seed did not [F(2, 3) = 0.0561, p = 0.946]. The EC had significantly greater immunoreactivity compared to the HPC, SNr, and MSB (p < 0.05). The RSc was also found to have significantly greater immunoreactivity compared to the MSB. Three months following seeding the region was also found to have a significant effect on 82e1+ immunoreactivity [F(4, 17) = 6.917, p = 0.002; mixed-effects analysis[1]]. The EC had significantly higher immunoreactivity compared to the SNr, and MSB (p < 0.05). Five months post seeding regional 82e1+ immunoreactivity differences were found [F(4, 31) = 17.74, p < 0.0001; mixed effects analysis for missing values]. The EC had significantly greater immunoreactivity compared to the HPC, SNr, and MSB (p < 0.001). The RSc had significantly greater immunoreactivity compared to the HPC, SNr, and MSB (p < 0.0005). The EC, the location targeted for seeding, had the highest overall 82e1+ immunoreactivity compared to all other regions. Consistently the cortical regions, EC and RSc had the highest levels of 82e1+ immunoreactivity – the MSB consistently had the least 82e1+ immunoreactivity.
Reduction in cholinergic tone in MSB following Aβ seeding.
The number of ChAT + neurons was reduced significantly in Aβ seeded compared to the control tissue seeded mice [t(9) = 3.355, p = 0.0042] and the non-seeded controls (non-seeded controls: 404.0 ± 134.4 cells vs. Aβ seeded: 152.6 ± 46.06; [t(8) = 2.319, p = 0.0245]. The number of ChAT + cells in the MSB was significantly reduced in the Aβ seeded mice compared to control seeded mice (152.6 ± 46.06 vs. 416.8 ± 66.4, respectively). In the control compared to the Aβ seeded mice, respectively; this represents a 273% reduction in cholinergic cells, with the Aβ seeded mice having only 36% of cholinergic cells compared to the control seeded mice (Fig. 2C)
Seeding did not impair spatial learning and memory.
The mice significantly reduced their average proximity to the target across training with no significant differences between groups, showing they were able to learn the location of target platform (one month (Fig. 3Ai): training effect [F(3, 44) = 0.556, p = 0.647] and no day x group interaction [F(27, 396) = 0.714, p = 0.855]; three months (Fig. 3Aiv): training effect [F(9, 324) = 17.82, p < 0.0001] between group [F(3, 36) = 2.03, p = 0.128], day x group interaction [F(27, 324) = 0.817, p = 0.730]; five months (Fig. 3Avii): training effect [F(9, 225) = 14.3, p < 0.0001], between groups [F(3, 25) = 0.579, p = 0.635], day x group interaction [F(27, 225) = 0.293, p = 0.999].
Training caused a significant increase in swim speed, but no group differences were found (one month (Fig. 3Aii): training effect [F(9, 396) = 6.37, p < 0.0001] between group [F(3, 44) = 1.49, p = 0.231], day x group interaction [F(27, 396) = 0.714, p = 0.855]; three month (Fig. 3Av): training effect [F(9, 279) = 4.59, p < 0.0001], group differences [F(2, 31) = 2.92, p = 0.069], day x group interaction [F(18, 279) = 0.996, p = 0.464]; five months (Fig. 3 viii): training effect [F(9, 225) = 0.93, p = 0.496], group differences [F(3, 25) = 0.546, p = 0.655] day x group interaction [F(27, 225) = 0.149, p > 0.9999]). In the probe trial, no significant difference was found between groups at each time point tested.
The mice all had similar swim speeds at each time point (one month (Fig. 3Aii): between group [F(3, 44) = 0.571, p = 0.637]; three months (Fig. 3Av): between group [F(3, 36) = 1.56, p = 0.216]; five months (Fig. 3Aviii): between group [F(3, 25) = 0.254, p = 0.858]). The mice also had similar proximity to the target (data not shown) (one month: between group [F(3, 44) = 0.207, p = 0.891]; three months: between group [F(3, 36) = 0.08, p = 0.971]; five months: between group [F(3.00, 19.59) = 0.497, p = 0.689; Brown-Forsythe ANOVA used due to significantly different standard deviations (p = 0.011)].
The mice all spent similar time in the target quadrant (one month (Fig. 3 Aiii): between group [F(3, 44) = 0.178, p = 0.911]; three months (Fig. 3Avi): between group [F(3, 36) = 0.095, p = 0.962]; five months (Fig. 3Avi) between group [F(3, 25) = 0.41, p = 0.7495]). When comparing the percentage of time spent in the target quadrant compared to chance all groups showed significant preference for the target quadrant. One month: non-seeded [t(5) = 5.56, p = 0.001], control [t(13) = 5.04, p = 0.0001], rpAD [t(13) = 5.39, p < 0.0001], and slAD [t(13) = 5.77, p < 0.0001]. Three months: non-seeded [t(5) = 3.59, p = 0.008], control [t(10) = 5.73, p < 0.0001], rpAD [t(11) = 3.86, p = 0.001], and slAD [t(10) = 4.36, p = 0.0007]. Five months: non-seeded [t(5) = 8.39, p = 0.0002], control [t(7) = 5.81, p = 0.0003], rpAD [t(6) = 5.20, p = 0.001], and slAD [t(7) = 7.94, p < 0.0001]
The measures of the MWT were compared across the three time points tested and the seed had no significant effect on proximity [F(2, 39) = 1.14, p = 0.332], swim speed [F(2, 39) = 1.74, p > 0.05], or time in target quadrant [F(2, 90) = 0.088, p > 0.05]. Proximity significantly decreased across time [F(2, 51) = 8.25, p = 0.0008]. Swim speed significantly increased over time [F(2, 51) = 7.75, p < 0.05], between one and three months (p = 0.0006). Time following seeding had no significant effect on the time spent in the target quadrant [F(2, 90) = 1.050, p > 0.05].
The mice in this study were able to sufficiently learn the location of target platform at each time point tested. Significant changes in swim speed occurred across training and between testing time points. The seed had no significant effect on spatial memory or swim speed.
NOR revealed unique behaviour towards objects used but discrimination between the novel and familiar object was not impaired.
At one and five months following seeding the mice had a significant preference for the novel object but spent equal time investigating the novel and familiar objects three months following seeding (one month (Fig. 3Bi): novel vs familiar [F(1, 35) = 40.2, p < 0.0001]; three month (Fig. 3Bi): [F(1, 29) = 0.69, p = 0.413]; five month (Fig. 3Bi): novel vs familiar [F(1, 25) = 57.2, p < 0.0001]). The mice spent equal time investigating the objects during the NOR at each time point tested (one month: [F(3, 35) = 1.36, p = 0.270]; three months: between groups [F(3, 29) = 0.742, p = 0.536]; five months: between groups [F(3, 25) = 0.71, p = 0.553]). The investigation ratio for the novel object was similar between groups at each time point (one month: between group [F(3, 35) = 1.26, p = 0.303]; three months: between groups [F(3, 29) = 0.124, p = 0.945]; five months: between groups [F(3, 25) = 0.71, p = 0.553]).
Freezing behaviour one month post seeding showed the mice froze significantly more during the tone test compared to baseline but only the rpAD and slAD mice showed significantly more freezing in the context condition (Fig. 3Ci). Three months post seeding the mice froze significantly more in the tone condition but only the control and slAD seeded mice froze significantly more in the context condition. Five months post seeding the mice froze significantly more in the tone and context condition compared to baseline. Over the course of the experiment, no significant changes in overall freezing during the tone test occurred, however, the slAD seeded mice showed a significant drop in freezing behaviour five months following seeding compared to their freezing behaviour three months post seeding. In the context condition, the control seeded, rpAD, and slAD mice showed significant increases in freezing behaviour. Freezing behaviour did not change significantly over time; however, the slAD seeded mice showed a significant drop in freezing behaviour five months following seeding compared to their freezing behaviour three months post seeding.
At each time point the mice froze more during the onset of a tone that was paired with a shock compared to baseline. The freezing percent one month after seeding in the context condition was significantly lower in the control, rpAD and slAD mice than at three months following seeding. By the third testing, the control seeded mice still had a significantly higher freezing percent than one month following seeding but the rpAD and slAD mice did not. There was no change in freezing behaviour in the non-seeded control mice across time.
The condition had a significant effect on freezing behaviour [F(2, 88) = 110.4, p < 0.0001]. No group differences in freezing behaviour were found for baseline, tone, or context conditions [F(3, 44) = 0.452, p = 0.717] and no group x condition interaction was found [F(6, 88) = 0.618, p = 0.716]. However, the non-seeded and control seeded mice did not show significantly greater freezing in the context condition, compared to baseline freezing (p < 0.05) whereas the rpAD and slAD seeded mice did. Three months following seeding only the slAD seeded mice froze significantly more in the context test than baseline. The conditions had significant effect on freezing behaviour [F(2, 70) = 57.5, p < 0.0001] but no group differences were found [F(3, 35) = 0.085, p = 0.968] and no group x condition interaction [F(6, 70) = 1.11, p = 0.367]. All groups spent significantly more time freezing during the onset of the tone compared to baseline but only the control seeded and slAD seeded mice spent significantly more time freezing during the context test compared to base line (Fig. 3Cii.). Five months following seeding all groups showed significant freezing in the context relative to baseline (Fig. 3Ciii.). Time post seeding had a significant effect on freezing during the tone [F(2, 60) = 5.92, p = 0.005] and context [F(2, 61) = 20.8, p < 0.0001] tests. No group differences were found in the tone [F(3, 44) = 0.046, p = 0.987] and context [F(3, 44) = 0.55, p = 0.650] tests. No group x time post seeding interaction was found for the tone [F(6, 60) = 0.868, p = 0.524] or context [F(6, 61) = 0.87, p = 0.525] tests.
No significant differences were found in tests of the balance beam at any time point tested (data not shown).
[1] Due to poor staining or damage of tissue, inconsistent sampling occurred for some regions resulting in missing values.