A New Molecular Mechanism for the Late-Onset Alzheimer's Disease Risk Gene CD2AP

Alzheimer's disease (AD) is a progressive neurodegenerative disorder that impairs memory and cognitive functions; it is the leading cause of dementia worldwide. Familial Alzheimer's disease (FAD) is rare and typically appears earlier in life (first symptoms can appear before 40 years of age and fully developed pathology before 60–65 years of age). In contrast, late-onset Alzheimer's disease (LOAD), which accounts for the majority of cases, develops at older ages (after 60–65 years of age). Unlike autosomal-dominant FAD, in which mutations in specific genes (APP, PSEN1, and PSEN2) directly cause the pathology, LOAD is driven by a combination of environmental factors and genetic factors that increase the risk of developing the disease. LOAD-linked alleles such as APOE4 increase the risk of developing LOAD, but they do not guarantee its onset. How genetic and environmental factors contribute to the development of LOAD is not fully understood and there is an urgent need to address this gap in the field.

For decades, most studies on AD pathogenesis focused on the highly penetrant FAD mutations and on mechanisms leading to the deposition of amyloid-beta (Aβ) plaques and tau neurofibrillary tangles, the two main histopathological signs of AD. More recent studies have begun to investigate how the classical Aβ and tau hypotheses relate to LOAD initiation and progression, but other mechanisms have also been studied (Nasb et al., 2024). Such studies have shown that LOAD pathogenesis involves the disruption of several cellular pathways, with synaptic dysfunction and synapse loss as central pathological events (Dorostkar et al., 2015). How these processes are affected by specific gene variants in LOAD is a field of growing interest.

One significant genetic risk factor for LOAD is CD2AP (CD2-associated protein), in particular, the K633R mutation (Vardarajan et al., 2015). CD2AP is an actin-binding protein whose complete loss of function causes kidney glomeruli failure and premature death. But CD2AP is ubiquitously expressed, including in neurons, where it localizes to neuronal dendrites and participates in the regulation of neurite extension, endocytosis, endosomal maturation, and actin polymerization (Ubelmann et al., 2017). In the context of AD, CD2AP knockdown has been shown to increase tau toxicity and to promote intracellular accumulation of a particularly toxic isoform of Aβ, Aβ₄₂ (Shulman et al., 2014; Ubelmann et al., 2017). The role of CD2AP in the regulation of actin dynamics described in other cell types and its dendritic localization in neurons led Mirfakhar and colleagues to hypothesize that CD2AP affects spine and synaptic function by regulating actin remodeling. To test this hypothesis, they explored the contribution of CD2AP and its K633R mutant to LOAD-related synaptic dysfunction (Mirfakhar et al., 2024) in primary cultures of mouse cortical neurons.

Mirfakhar et al. (2024), first showed that a pool of CD2AP was present in dendritic spines and colocalized with PSD-95, indicating a partial postsynaptic localization. Silencing of CD2AP with two different shRNAs led to a decrease in spine volume, spine number, and synapse density, as indicated by immunostaining. The number of active synapses, assessed by FM4.64 uptake, was also decreased, probably as a consequence of the loss of functional and mature spines. Finally, multielectrode arrays revealed that CD2AP knockdown led to reduced neuronal network activity compared with control neurons.

The effects described above could be caused by an increase in intraneuronal Aβ₄₂ accumulation, which may happen upon CD2AP silencing and is known to cause spine reduction. To test this hypothesis, the authors used two different anti-Aβ treatments (BACE or γ-secretase inhibitor) to block Aβ₄₂ production. Although both treatments increased PSD-95 density in silenced neurons, the density did not return to control levels, suggesting the amyloid peptide is only partially responsible for spine loss. These findings indicate that CD2AP's influence on synapses is at least in part independent of Aβ-related pathways.

Mirfakhar et al. (2024) also overexpressed the AD-related CD2AP K633R mutant in neurons. In contrast to protein knockdown, CD2AP^K633R increased the volume and density of dendritic spines, particularly of those with more mature morphology (mushroom and stubby), while reducing their motility. Synaptic contacts were also morphologically increased but less active (decreased FM4.64 loading) than in control neurons.

Because actin dynamics are crucial for maintaining spine shape and stability and CD2AP binds actin, Mirfakhar et al. (2024) asked whether CD2AP controls F-actin polymerization/depolymerization in dendritic spines. They observed that a fraction of CD2AP colocalized with F-actin and with the actin-binding protein cortactin in dendritic spines. CD2AP was necessary for the correct distribution of F-actin between dendritic shafts and spines, with lower spine F-actin upon CD2AP silencing. On the other hand, CD2AP^K633R overexpression increased spine F-actin and decreased mCherry-actin recovery in fluorescence recovery after photobleaching (FRAP) experiments, indicating decreased F-actin turnover and overstabilization of the cytoskeleton. Finally, in depleted neurons, the reexpression of CD2AP^WT, but not CD2AP^K633R, rescued F-actin and PSD-95 densities, confirming a mutation-related functional weakening in the ability of CD2AP to regulate actin dynamics within spines, and thus spine formation and growth.

As mentioned above, most of the molecular mechanisms through which AD risk genes impair physiological functions and increase the chances of developing the disease are yet unknown and are an important research goal. This study revealed a novel molecular mechanism through which a LOAD risk gene affects synapses. Mirfakhar et al. (2024) showed that CD2AP is involved in dendritic spine morphogenesis through the modulation of F-actin dynamics, which is crucial for spine stability and plasticity. The overstabilization of dendritic spines caused by CD2AP^K633R may thus impair the ability of synapses to strengthen and weaken in synaptic plasticity events, which are at the basis of learning and memory and are defective in AD patients.

The identification of actin remodeling impairments caused by CD2AP^K633R provides a basis for targeted therapies to mitigate synaptic deficits in AD patients carrying this mutation. Even though challenging, preclinical and clinical actin-targeted therapies are already used for the treatment of intraocular hypertension/glaucoma and cerebral vasospasm. One actin-remodeling protein targeted for pharmacological intervention is the Rho kinase ROCK, which activates LIM-kinase 1, which in turn phosphorylates and inhibits the actin depolymerizing factor cofilin. ROCK inhibition might help to mitigate actin and spine overstabilization induced by CD2AP^K633R by promoting actin remodeling. Fasudil, a ROCK inhibitor, has already been shown to ameliorate cognitive deficits in animal models of alcohol-related brain damage, AD, and cerebral ischemia (Cai et al., 2024). Another target that may mitigate the effects of CD2AP^K633R is Rac1, a member of the Rho GTPase family implicated in the regulation of actin dynamics. Rac1 regulates the structural plasticity of dendritic spines and is upregulated in neurodevelopmental disorders like fragile X syndrome, in which an increased density of aberrant spines is present together with cognitive disabilities. The inhibition of Rac1 has been suggested as a tool to rescue aberrant spine phenotypes through its actin-remodeling activity (Bongmba et al., 2011).

Even if these therapies could be applied to CD2AP^K633R carriers with adequate adjustments, modifying actin dynamics could be risky given the critical role of these dynamics in many fundamental molecular processes throughout the body. Systemic delivery could lead to severe side effects (e.g., vasodilation, with low blood pressure and headaches, palpitations, chest pain, and muscle cramps for fasudil). Therefore, more targeted therapeutic approaches should be considered (e.g., nanoparticles as cell-specific drug delivery systems). Moreover, a careful risk/benefit assessment of the probability of developing AD and the side effects of actin-remodeling drugs in CD2AP^K633R carriers should be considered. An accurate stratification of LOAD patients based on genetics and maybe also on specific biomarkers would be vital in the future to select those with the greatest potential for beneficial therapeutic outcomes.

To select a rational therapeutic approach to decreasing AD risk in CD2AP^K633R carriers, the signaling pathways leading to actin remodeling downstream of CD2AP in spines need further exploration. In the postsynaptic compartment, CD2AP colocalized with cortactin, which promotes the formation of branched actin networks by stabilizing the Arp2/3 complex and inhibiting debranching. Whether CD2AP interacts with cortactin remains to be tested. Moreover, an interaction with capping proteins, which localize at the ends of the growing F-actin branches and block the attachment or loss of new actin monomers, may be involved. For example, CD2AP might inhibit capping proteins, preventing their action and allowing excessive growth of actin filaments. Direct interactions between CD2AP and F-actin might also be compromised, with stronger binding in the case of CD2AP^K633R. This interaction could be responsible for more stable ends of the actin cytoskeleton, with a decreased ability of actin monomers to detach from the branches.

After CD2AP-interacting proteins are identified, how these interactions intersect with other actin-related pathways involved in Alzheimer's disease should be considered. For example, activated cofilin interacts with tubulin, leading to tau protein displacement and microtubule destabilization, with implications for the development of tauopathy and synaptic dysfunction (Woo et al., 2019). Similarly, the activity of Rac1 has been found to have implications in APP processing and Aβ/tau accumulation in an AD mouse model (Borin et al., 2018).

AD is a complex and multifactorial disorder with very limited therapeutic options. By uncovering the role of CD2AP in regulating dendritic spine actin dynamics, Mirfakhar et al. help us understand synaptic dysfunction in at least some forms of LOAD. Although the results still need to be confirmed in vivo, the research underscores the potential for actin-targeting therapies to address synaptic deficits and offers a framework for future investigations into the molecular mechanisms underlying AD pathogenesis.