Jonny Daborg
33
Perhaps compromised Aβ clearance is an initial culprit in SAD, followed by
positive feedback loops, including RAGE, which will increase the production
of Aβ when the scales tip over.
5.1.2
Altered synaptic plasticity by A
β
Aβ inhibits LTP (Chang et al. 2006; Chapman et al. 1999; Kamenetz et al.
2003; Klyubin et al. 2008; Knobloch et al. 2007; Shankar et al. 2008;
Townsend et al. 2007; Walsh et al. 2002) and promotes LTD (Cheng et al.
2009; Hsieh et al. 2006; Li et al. 2009). Several mechanisms have been
proposed (Pozueta et al. 2012); of particular interest is a study showing that
Aβ blocks glutamate uptake, thereby altering the induction thresholds for
LTP and LTD, such that LTD is promoted on the expense of LTP (Li et al.
2009).
“What about tau?” some would ask. Tau is most certainly involved in AD
pathogenesis (Hardy et al. 1998). No FAD-mutations are found in the gene
encoding tau, however. Mutations do exist in tau, and causes dementing
disorders such as frontotemporal dementia with Parkinsonism, this is,
however, not associated with any Aβ pathology (Hardy and Selkoe 2002).
Thus, it seems that in AD, Aβ is the causative agent, and that tau follows.
Nevertheless, tau has been shown to be necessary for Aβ to inhibit LTP
(Shipton et al. 2011) suggesting that Aβ is sufficient to initiate the pathologic
cascade, but also that tau is necessary for its continuing rampage.
A possible reason for the shift towards LTD could be an increased tonic
NMDAR signalling since it has been shown that increasing the extracellular
concentration of glutamate by blocking the glutamate uptake promotes LTD
(Li et al. 2009; Yang et al. 2005). Additional findings pointing towards
dysregulation of NMDARs in AD are those that have implicated the prion
protein in Aβ pathophysiology (Stys et al. 2012; You et al. 2012; Parkin et al.
2007; Griffiths et al. 2011). The prion protein regulates NMDAR
desensitisation, and it has been shown that Aβ perturb this function of the
prion protein, thus increasing the tonic activity of these receptors. This could
potentially explain the findings of Aβ-induced excitotoxicity as well as the
shift towards LTD and eventually apoptosis.
5.1.3
Synapses are marked for destruction
A possible link between LTD and elimination of synapses is a family of
proteins called neuronal pentraxins (NP). These have been shown to be
necessary for LTD in the hippocampus (Cho et al. 2008) and serve as binding
sites for C1q (Perry and O'Connor 2008). Moreover, the levels of certain NPs
Synaptic elimination and the complement system in Alzheimer’s disease
34
are increased in FAD CSF (Ringman et al. 2012). NP1 and NP2 (a.k.a.
NARP) are secreted molecules that bind to, and cluster AMPARs. NP
receptor (NPR) is a transmembrane protein that binds to AMPAR and NP1
and NP2, thereby potentially anchoring AMPARs in the synaptic cleft.
Stimulation of LTD (which would be increased by Aβ) also stimulates the
enzyme tumor necrosis factor-alpha converting enzyme (TACE) to cleave the
transmembrane domain of NPR, thus releasing it from the membrane (Cho et
al. 2008). This event is associated with endocytosis of AMPAR and LTD
(Cho et al. 2008). One can speculate that the released AMPAR-NP complex,
act as an initiation point for the complement cascade when it moves outside
the synaptic cleft. The complement proteins that would then bind to the
AMPAR-NP complex could be of various origins. The main source of
complement proteins in the body is the liver. A compromised blood-brain
barrier can give peripheral complement access to the brain. This could be of
pathological relevance, and it is interesting to note that Aβ-RAGE interaction
can disrupt the blood-brain barrier (Kook et al. 2012). This pathway is not
necessary, however, since complement is also produced in the brain. In the
brain, complement components are secreted by both glia and neurons, mainly
during development, but also in the adult brain in response to disease or
trauma (Veerhuis et al. 2011; Stephan et al. 2012). Interestingly, Aβ can
modulate complement activation (Wang et al. 2009), and knockout of C1q in
a mouse model of AD reduced the decrease in synapse markers and
activation of glia, otherwise observed in this model (Fonseca et al. 2004).
5.1.4
Engulfment of synapses by microglia
Several scenarios could follow once complement has marked a synapse. The
complement cascade might reach its final stage in forming the membrane
attack complex with subsequent lysis of the synapse as a consequence.
Another possibility would be that the complement molecules act as a
phagocytosis signal, attracting microglia that would respond by engulfing the
synapse. Microglia have been implicated in the events of synaptic elimination
by impairing their motility, and the observation of synaptic proteins in
microglial endosomes (Paolicelli et al. 2011). Since hindering microglial
motility impairs synaptic elimination, their role is conceivably beyond merely
engulfing debris from cell-autonomous elimination. Therefore it seems more
likely that complement marks synapses for elimination, and that this process
is carried out in a most direct manner by phagocytosing microglia.
5.1.5
Neuronal death
An interesting consequence of fewer synapses is that it means less LTP,
which aside from impaired learning abilities, also would lead to a decrease of