Reactive astrogliosis is spread across frontal cortex in Tg2576 and AD human PM brain
Astrocytes were abundant and widely distributed in the grey matter of the frontal cortex and hippocampus in Tg2576 mice. Protoplasmic astrocytes are located throughout all grey matter, whereas fibrous astrocytes are located throughout all white matter (Kim et al., 2016) These studies and the finely branched and dense network of astrocytes observed, suggest that the grey matter astrocytes were protoplasmic.
The astrocytes appear to occupy non-overlapping domains. The sections have strong GFAP immunostaining and cellular hypertrophy which is associated with reactive astrogliosis (Jessen 2006). Sofroniew and Vinters 2010 demonstrated that in diseased human cortices, moderate astrogliosis can be present with the preservation of non-overlapping astrocytic domains. The Tg2576 mice sections may also present this. Nevertheless, mice and human brain tissue may differ in the presentation of astrocytic processes. Jin et al., 2013 shows similar findings in the hippocampus. However, Jin et al does not assess the distribution of astrocytes in the frontal cortex. Other differences between this present study and Jin et al include the IHC protocol and source of the anti-GFAP which may result in antibody specificity and sensitivity differences in resullts. Nevertheless, the similarities such as the type of AD mouse model, type of astrocytic marker, type of secondary antibody and DAB chromogen may have contributed towards the similarities in observed outcomes.
Protoplasmic astrocytes in the grey matter were abundant throughout the human PM frontal cortices. In 2 cases, astrocytes had overlapping domains and cellular hypertrophy which suggests moderate to severe astrogliosis. Perhaps these are AD cases. A previous study demonstrates findings of reactive astrogliosis in diseased human tissue (Sonfroniew and Vinters 2010). The details of the IHC protocols and the type of transgenic mouse used is not mentioned in the paper. Consequently, direct comparisons between the GFAP IHC assays and those in the present study cannot be made with certainty. 4 cases had evenly distributed astrocytes with little/no overlapping domains. Perhaps 3 cases were non-AD subjects. A limitation though is that anti-GFAP does not mark all non-reactive astrocytes and all astrocytes may not be detected in non-diseased tissue (Sonfroniew and Vinters 2010). This may be the case for the non-AD cases. Therefore, it is important to be careful when distinguishing the non-AD from AD cases based solely on GFAP immunostaining.
Neuronal distribution is widespread across frontal cortex in Tg2576 mice and human PM brain tissue
Widespread NeuN-positive immunostaining was observed in the Tg2576 mice frontal cortices. In the hippocampus, the dentate fascia was densely stained. Duffy et al., 2015 demonstrates similar findings in Tg2576 entorhinal and prefrontal cortices. Although, it is important to consider that the study does not show data for entire frontal cortex nor entire hippocampus. Having a WT mice would have helped to deduce differences in neuronal distributions through immunofluorescence assays. Quantification of this using confocal microscopy would be required for accurately deducing cell count and distribution.
Negative NeuN immunostaining was observed in the human PM frontal cortices. Although the retailer, Abcam presents that the antibody can bind to the human antigen, it is possible that the primary antibody was not specific to it. Alternatively, the antigens may be masked or degraded. This can occur when the tissue is fixed for a prolonged period of time. An alternative neuron-specific antibody, Anti-HuC/HuD produced positive immunostaining across the frontal cortices. Perhaps this outcome differs from the negative assays because the two antibodies bind different antigens (Thermofisher 2018; Abcam 2018). The AD and non-AD cases were indistinguishable. Perhaps a signal enhancing technique such as immunofluorescence and quantification of the neuron counts via. Confocal microscopy could aid this. Previous studies suggest that the hippocampal pyramidal layer has a reduced neuron number near Aβ deposits in APP23 mice (Schwab et al., 2004). Also, neuronal loss effects may be less than those found in post mortem AD brains (Schwab et al., 2004). It would be interesting to have seen if this was also the case for the Tg 2576 mouse and human tissue from this study.
NG2 cells distribution is widespread throughout frontal cortex in Tg2576 mice and human PM brain
NG2 cells were abundant and widely distributed across the frontal cortex and hippocampus in Tg2576 mice. Yao et al., 2010 finds similar evidence Tg 2576 mice brains. Additionally, immunostaining of NG2 cells decreased in all ages when compared with WT mice of the same age. Also, anti-NG2 positive cells had a small soma size and thinner processes than the WT mice of the same age. This led the researchers postulating that NG2 cells play a role in AD. Arguably, the results are not suggestive of this because the Tg2576 mice are partial models of AD (Elder et al., 2010; Kitazawa and LaFerla 2012). Perhaps, it can be assumed with more confidence that Yao et al’s findings show that NG2 cells play a role in AD pathology. Analysing the association between NG2 cells and Aβ plaques in the Tg 2576 mouse brain may provide evidence for a potential link of these cells to AD. It is possible that these findings could be present in the mice brains of this present study, however IHC assays using WT control mice brains were not carried out. This serves as a limitation to the present study.
NG2 cells appeared numerous and widely spread throughout 3/6 the human PM frontal cortices. Dawson et al 2000 finds of even distribution of multi-processed anti-NG2 positive cells in all post mortem non-diseased brains. It was difficult to visualise positivity in the remaining 3 cases. Differences in anti-NG2 expression was difficult to deduce due to the weak immunostaining. Perhaps this relates to findings that AD patients PM brains have reduced NG2 immunoreactivity compared to non-demented controls (Nielson et al.,2013). If there was more time, immunofluorescence labelling followed by quantification of anti-NG2 fluorescence would have illuminated expression differences.
RARα distribution is widespread throughout frontal cortex in Tg2576 mice and human PM brain but RARβ staining only positive in Tg2576 brain
Widespread RARα (Abcam 28767) and RARβ (Abcam 53161) positive nuclear staining was observed throughout the frontal cortex and hippocampus in Tg 2576 mice. These findings are partially supported by Krezel et al., 1999 where medium to strong RARα nuclear immunostaining was detected in the frontal and hippocampal regions. However, RARβ nuclear immunostaining was not detected. This finding may differ from the present study as 4-month-old C57/BL5 mice and Tg2576 mice may have different RARβ expression levels. Also, the brain slice thickness, anti-RARβ antibody and IHC methods differ between the two studies. Goodman et al., 2012 shows similar findings that RARα has widespread distribution with significantly high levels in the hippocampus and cortex. Although some researchers find that RARα distribution is conserved in rodents, it is important to consider that they may differ in RAR densities, and so one should be careful when comparing these studies.
Negative RARα (Abcam 28767) and RARβ (Abcam 53161) immunostaining was observed in all human PM frontal cortices. This differed from previous studies. Goodman et al., 2012 shows positive anti-RARα (ab28767) immunostaining to demonstrate the widespread distribution of nuclear and cytoplasmic RARα in the hippocampus and cortex. This infers that the RARα shuttles between the nucleus and the cytoplasm (Mackem et al., 2002). Widespread positive RARα (Santa Cruz 551) immunostaining was observed in all human post mortem frontal cortices. Corcoran et al., 2013 shows positive immunostaining using the same antibody. In the present study, a more sensitive DAB called DAB immpact produced the RARα (Santa Cruz 551) immunopositivity. However, standard DAB was used in RARα (Abcam 28767) and RARβ (Abcam 53161) immunostaining. Perhaps DAB chromogen sensitivity was the main factor influencing the detection of immunoreactivity.
Senescent astrocyte-staining was negative in control tissues and in human PM brain tissue
Negative p16INK4a immunostaining was observed in the rat liver, mouse embryo and mouse ovaries. There are not previous studies to directly compare with these. However, Wei et al 2018 shows that AD 5XFAD mice have increased neuronal p16 expression in the hippocampus compared with age matched
WT mice. Other studies support these findings (Bhat et al., 2012; Salminen et al., 2011).
Negative p16INK4a immunostaining was observed in all human PM frontal cortices. It is likely that in the present study, significant antigen masking or degradation may have occurred before or during performing IHC. This result differs from Bhat et al, 2012 who found that the frontal cortices of 15 AD subjects had elevated p16INK4a-positive astrocytes compared to 25 non-AD controls of similar ages. This suggested that senescent astrocytes may be associated with increased risk of sporadic AD. The two studies may differ as the source of the primary antibodies, the human tissues and IHC protocols differed.