Amyloid Precursor Protein Can Cause Changes in Neuronal Firing Rates via Axon Initial Segment

Alzheimer's disease (AD) is the leading cause of dementia, affecting an estimated 55 million people worldwide. Most studies of AD pathogenesis focus on insoluble amyloid-β (Aβ) plaques that are found in the brains of postmortem AD patients. These plaques are produced by the proteolytic cleavage of amyloid precursor protein (APP). The link between APP and AD is reinforced by the fact that duplication and missense mutations in APP promote AD pathology (Tcw and Goate, 2017). While the role of Aβ in AD pathogenesis has received considerable attention, not much focus has been given to the possible contribution of APP. APP has been shown to localize to the axonal initial segment (AIS) (Fig. 1), the axonal region nearest to the soma, which is responsible for initiating action potentials (Huang and Rasband, 2018). This raises the possibility of a role for APP in regulating the propagation of action potentials along the axon and, hence, in patterns of synaptic release. In a recent article in The Journal of Neuroscience, Ma et al. (2023) reported that an increase in APP, as seen under AD-like conditions, shrinks and displaces the AIS from the neuronal soma. This change in the size and position of the AIS may change the pattern of action potential propagation and synaptic release, thus affecting cognition.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1.

Summary figure showing the change in the length and shift in the position of the AIS with an increase in APP, which alters propagation of action potentials to the synapses. Figure created using Biorender.com.

To assess possible effects of APP on AIS position and length, Ma et al. (2023) treated cortical neurons cultured from embryonic day 16.5 (E13.5) mice and exposed them to 1 μm glutamate. This concentration mimics the excitotoxic conditions that are commonly associated with the progression of AD (Wang and Reddy, 2017). Immunostaining revealed that APP levels increased in both dendrites and axons of neurons treated with glutamate compared with control conditions. This increase in APP was because of the activity of NMDARs, as shown by the absence of such changes in the presence of an NMDAR antagonist, dCPP. Importantly, increases in APP expression were accompanied by the shrinkage and displacement of the AIS, as shown by immunostaining of the scaffolding protein ankyrin G (AnkG), which localizes to the AIS. Thus, excess glutamate exposure can lead to an NMDA-dependent increase in APP expression, a shift in the AIS position away from the soma, and a reduction of AIS length.

To determine whether glutamate-induced changes in the AIS position and length depended on the glutamate-induced increase in APP expression, Ma et al. (2023) knocked down APP using RNA silencing. After APP knockdown, cultured cortical neurons were treated with 1 μm glutamate. The length and position of the AIS were not significantly different between APP-knockdown neurons and control neurons. Thus, this experiment conclusively showed that the glutamate-driven changes in the AIS length and position are driven by the APP protein.

To further examine the effects of APP on the AIS, Ma et al. (2023) transfected cultured neurons with plasmids encoding GFP-tagged APP. The resulting elevation in APP levels led to a ∼50% reduction in AIS length and a shift in the position of the AIS. Thus, APP overexpression alone was able to shift and reduce the length of the AIS without glutamate stimulation. To examine the relevance of this finding to AD, the authors transfected cortical cultured neurons with WT human APP (hAPP) or Swedish variant K95N/M596L (hAPP_Swe) which causes early-onset AD. Again, an increase in APP levels led to decreased length and shifted position of the AIS.

Notably, the concentration of APP in the above experiments was much greater than those found in AD patients. To address this, the authors examined the location of APP (via an N-terminal APP antibody) and the position of the AIS (via AnkG immunolabeling) in cultured neurons obtained from R1.40 mice, a model for AD. In general, they observed a reduction in the length of the AIS compared with control, but the reduction was substantially less than that in neurons transfected with hAPP_Swe. This suggested that the APP-driven shortening of the AIS is dependent on the dose and the genotype of APP. This genotype dependence illustrates the importance of considering genotype when drawing conclusions from mouse models of AD, as different models may lead to different results.

Since there was reasonable evidence that the change in length and position of the AIS was influenced by the level of APP expression in neurons, Ma et al. (2023) speculated that a physical interaction between APP and the protein components of the AIS underlay the effect. AnkG and βIV spectrin are proteins that are associated with the AIS region. Immunoprecipitation of APP from brain (Broadmann area 9) lysates obtained from control and AD patients pulled down AnkG and βIV spectrin, indicating that these proteins physically interact in vivo. Similar results were obtained using lysates from WT and R1.40 mice.

Because APP is the precursor protein of Aβ, which is thought to have toxic roles in AD, Ma et al. (2023) asked whether changes in AIS length associated with elevated APP result from increased production of Aβ. To test this hypothesis, the authors treated neuronal cultures with Aβ (fibrils and oligomers) and examined the AIS length and position. They observed no change in AIS length or position of the AIS. Similar results were obtained when they treated neurons derived from human iPSCs with Aβ. They therefore concluded that the changes in the AIS resulting from elevated APP levels are mediated by the precursor protein itself, not by increased accumulation of Aβ.

The authors speculated that, since the AIS is the site of action potential initiation, APP-dependent changes in AIS length and position would likely reduce the firing probability of neurons. To test this hypothesis, Ma et al. (2023) coinfected cultured neurons with lentivirus-expressing GCaMP8f (a calcium indicator) and APP-expressing virus. The calcium indicator is used to visualize neuronal activity. When monitored, APP-overexpressing cells showed reduced spontaneous GCaMP8f flashes, indicating reduced neuronal activity compared with control cells not overexpressing APP.

Overall, Ma et al. (2023) demonstrated that increased expression of APP, which can occur in response to excitotoxic stress associated with AD or as a result of genetic mutation, leads to shortening of the AIS and a shift in its position away from the soma. The change in the length and position of the AIS, in turn, reduces neuronal activity. The study provides evidence that increased expression of APP can influence AIS length and position independent of the accumulation of one of its key cleavage proteins, Aβ. This shifts the limelight from Aβ to APP as one of the pivotal molecules to be studied to better understand AD. The study might not only affect treatment strategies for AD, but also has implications in treating related diseases, such as Down syndrome (DS). DS patients make supranormal amounts of APP because of the trisomy of the 21st chromosome, which is incidentally the chromosome on which the APP gene is located (Ness et al., 2012). The increased accumulation of APP in DS patients has been shown to be associated with the early onset of Alzheimer's-like dementia. This study comes at a critical time in AD and DS research, where there is an increased drive to study additional mechanisms to understand and treat the progression of the diseases.