Abeta Clearance Removes Early Tau Pathology in Neuronal Processes

Figure 1. Examples of “the cerebral neocortex (a) and CA1 hippocampus (b) of Aβ42 (21F12 antibody) versus phospho-tau (AT8 antibody) staining of an unimmunised AD control (cAD) and an immunised AD case (iAD). From (2).

In probing the mechanism of clearance of early tau pathology in dystrophic neurites, the authors note the possibility that this could be due to the actual phagocytic removal of the phospho-tau-laden dystrophic neurites themselves. Their findings suggest otherwise, however, as

we have not observed any evidence of phospho-tau within microglia, whereas Aβ is readily detectable within microglia in many of these immunised AD cases. This observation may be interpreted as suggesting that the phospho-tau is not cleared by phagocytosis. … An alternative explanation is that the neurons are able to mobilise the aggregated tau in the dystrophic neurites and remodel the neurites. The mechanism of clearance of phospho-tau is difficult to address directly in a study of this type. …  However, experimental models with aggregation of both Aβ and tau do exist. In triple transgenic (3 × Tg-AD) mice early, but not the late, forms of phosphorylated human tau were removed by a single injection of anti-Aβ antibody into the brain [3]; reaccumulation of Aβ preceded reappearance of neuronal tau. In another model, active Aβ immunisation of APPSw/NOS2^−/− mice, which develop Aβ deposits as well as hyperphosphorylated mouse tau, produced a significant reduction of deposits of both Aβ and hyperphosphorylated tau[4]. Our findings in human AD following Aβ immunisation are consistent with these experimental observations linking Aβ and tau pathobiology.  … This is consistent with the amyloid cascade hypothesis which states that tau pathology is linked to and downstream from Aβ accumulation in AD and confirms the position of Aβ as a target for modifying both Aβ and tau aggregation in AD. (2)

Indeed, recent studies(notably (5-7)) would seem to highlight an even more optimistic reading of the specific clearance of phospho-tau from neuronal processes. In a variety of systems, these reports suggest that Aβ-associated neurotoxicity and synaptic dysfunction may in large part be mediated by aberrant redirection of tau from its normal axonal localization to the somatodendritic compartment, even in the absence of any change in the total levels of tau or of axonal transport proteins. Importantly, lowering tau protein levels affords significant protection against such toxic effects, despite a lack of effect on baseline axonal transport. This is consistent with many previous studies (eg, (8-10)) showing that depletion of native or transgenic tau protects rodent models of AD against Aβ neurotoxicity, even absent an effect on NFT. These findings therefore suggest that the ability of Aβ immunotherapy to clear early tau pathology from the dentrites, as found in the current report,(2) may have an unexpectedly powerful neuroprotective effect, if initiated at a relatively early stage in the “amyloid cascade,”  even if not accompanied by concomitant NFT clearance. Additionally, it seems reasonable to expect that the accumulation of NFT may be substantially prevented by such early immunotherapeutic clearance of Aβ, accompanied by clearance of its precursors in the dendrites. The new report does however seem consistent with a much more limited benefit of Aβ immunotherapy alone once the disease is well-esatablished, as in these patients and those in earlier autopsy reports:

One possible explanation for the continued decline in function [in AD patients after Aβ immunisation] may be the lack of protective effect of Aβ immunisation on accumulation of phospho-tau in the neuronal cell bodies, implying that additional therapy may need to be directed at neurofibrillary tangles. … Each of the iAD patients had cognitive function scores within the range MMSE score 15–25/30 at the time of their immunisation. This MMSE range is associated with Braak stages III–V whereas Braak stage VI is associated with MMSE scores in the range 0–10. The decline in MMSE score to 0 in all but one of the iAD cases prior to their death is therefore in keeping with progression in their Braak stage following Aβ immunisation. Correspondingly, the lack of difference in phospho-tau-positive neuron counts between the iAD and cAD cases suggests that phospho-tau continued to accumulate in neuronal cell bodies to reach, by the time they died, a level indistinguishable from that in unimmunised terminal AD. We cannot exclude an upstream role of smaller tau aggregates (not detected by AT8 staining) in the cognitive deterioration as observed in animal studies

In showing that Aβ immunisation can influence phospho-tau pathology, we confirm the position of Aβ as a target for modifying tau accumulation in AD and demonstrate a link between these proteins. However, the continuing progression of cognitive decline in AD patients after Aβ immunisation may be explained by its lack of apparent effect on tangles.(2)

As the authors note, and as animal studies reviewed in an earlier post suggest, immunotherapy directed at clearance of tau pathology would seem to offer a way to overcome this limitation.