Alternative splicing in Alzheimer's disease: Mechanisms,therapeutic implications,and 3D modeling approaches

Abstract

Alzheimer's disease (AD) is characterized by progressive cognitive decline, memory loss, and behavioral changes. AD is pathologically marked by the accumulation of extracellular amyloid-β (Aβ) oligomers, amyloidogenic plaques, and intracellular neurofibrillary tangles. Amyloid-β protein precursor (AβPP) plays a central role in AD pathology, as it is cleaved by β-secretase and γ-secretase enzymes to generate Aβ peptides and oligomers which aggregate to form neurotoxic fibrils and plaques in the brain. Increased AβPP expression has been correlated to Aβ suggesting a larger role for AβPP function potentially through AβPP isoforms. Alternative splicing (AS) of APP pre-mRNA has emerged as a key regulatory mechanism influencing AβPP function through the generation of distinct isoforms. Similarly, the microtubule-associated protein tau (MAPT) is also subject to alternative splicing, producing isoforms that can contribute to hyperphosphorylation and neurodegeneration. In this review, we explore the role of APP alternative splicing and the regulation of its isoform expression in AD and other neurodegenerative disorders, with a focus on its potential impact on Aβ peptide production. We also discuss recent advances in therapies targeting dysregulated splicing in neurodegenerative diseases and their potential relevance to AD. Finally, we highlight the use of three-dimensional culture models as a platform to study AS regulation in AD and other neurodegeneration-related disorders.

Keywords

alternative splicing Alzheimer's disease amyloid beta-peptides amyloid beta-protein precursor antisense oligonucleotides ApoE organoids presenilin-1 tau protein

AD pathogenesis and AβPP interactome

Alzheimer's disease (AD) is often characterized by the accumulation of soluble and insoluble oligomeric species of amyloid-β (Aβ) peptides, primarily Aβ₄₀ and Aβ₄₂, which are prone to aggressive aggregation and are neurotoxic, as well as the formation of neurofibrillary tangles primarily composed of misfolded and phosphorylated tau protein.¹ These oligomers generated from amyloid peptides are a byproduct of differential cleavage of amyloid-β protein precursor (AβPP), which is integral to the amyloid cascade hypothesis.² Processing of AβPP is categorized into two canonical pathways: non-amyloidogenic and amyloidogenic pathways. As depicted in Figure 1, the non-amyloidogenic pathway involves the initial cleavage of AβPP by α-secretase (ADAM10) in exon 16 within the Aβ region for assembly of Aβ. The cleaved byproduct known as the soluble fragment of AβPP (sAβPP-α) is released from the cell membrane and the C-terminal α-fragment (CTF-α), also known as C83, is then cleaved by γ-secretase complex which is composed of the presenilin proteins family (PSEN1 and PSEN2).³ One of the cleaved C83 byproduct, termed p3 fragment, is then processed and is not involved in further amyloid synthesis, whereas the CTF-α is internalized to produce the AβPP intracellular C-terminal domain (AICD) fragment which is involved in regulating transcription of APP and is pro-apoptotic.³ On the other hand, the amyloidogenic process begins with internalization of AβPP at the endosomal membrane to be cleaved by β-secretase (BACE1) to the amyloidogenic soluble β-fragment (sAβPP-β) and an endosomal C-terminal β-fragment (CTF-β).⁴ The CTF-β, also known as the C99 fragment, is subsequently cleaved by the γ-secretase complex into Aβ peptides. The peptides Aβ₄₀ and Aβ₄₂ form monomers, oligomers, and Aβ₄₂ oligomers aggregate into protofibrils and senile plaques from exosomes, where they impact cell survival, synaptic signaling, impair mitochondria, and promote neurodegeneration.⁵ The functionality of AβPP to undergo these canonical and other non-canonical processing pathways is crucial in the development of amyloid pathology seen in early to late onset of AD. Gene mutations at several checkpoints of AβPP processing including β- and γ-secretases, such as beta-site AβPP cleaving enzyme 1 (BACE1), presenilin (PSEN) 1, and PSEN2 respectively, have been shown to promote the preferential production of Aβ₄₂ peptide in the early-onset AD.^6,7 Alongside these findings, alternatively spliced isoforms of APP may play a crucial role in regulating AβPP processing and may modulate the effects of Aβ accumulation and aggregation.⁸

Figure 1.

Amyloid precursor protein (APP) processing. Canonical APP processing and proteolytic products following secretase activity in extracellular and cytoplasmic context are shown. Briefly, the non-amyloidogenic pathway involves the initial cleavage of APP by α-secretase (ADAM10). The cleaved byproduct known as the soluble fragment of APP (sAPP-α) is released from the cell membrane and the C-terminal fragment (CTF-α), also known as C83, is then cleaved by γ-secretase. The cleaved C83 byproducts are p3 and APP intracellular C-terminal domain (AICD) fragments. AICD is involved in regulating transcription of APP and is pro-apoptotic. On the other hand, the amyloidogenic pathway begins with the cleavage of APP by β-secretase (BACE1) into the amyloidogenic soluble fragment (sAPP-β) and an endosomal C-terminal fragment (CTF-β). The CTF-β also known as the C99 fragment is subsequently processed by the γ-secretase complex into amyloid β peptides (Aβ). The peptides Aβ₄₀ and Aβ₄₂ form monomers, oligomers, and aggregates into protofibrils and senile plaques. α-sec: α-secretase; sAPP-α: soluble α-fragment of APP; CTF-α or C83: C-terminal α-fragment; γ-sec: γ-secretase; p3: CD83 byproduct; AICD: APP intracellular C-terminal domain; β-sec: β-secretase; sAPP-β: soluble β-fragment of APP; CTF-β or C99: C-terminal β-fragment; Aβ: amyloid-beta; ADAM10: A disintegrin and metalloproteinase domain-containing protein 10; BACE1: beta-site APP cleaving enzyme 1; PSEN: presenilin. (Created in BioRender. Bellizzi, A. (2024) https://BioRender.com/n65l926).

APP pre-mRNA alternative splicing and APP isoforms

Specific splice variants of APP may preferentially undergo cleavage by β-secretase and γ-secretase enzymes, leading to differential production of Aβ isoforms such as Aβ₄₀ and Aβ₄₂.⁹ The APP gene is located on chromosome 21 and undergoes complex alternative splicing (AS), resulting in the production of several different isoforms such as APP770, APP751, and APP695 as depicted in Figure 2.^10,11 AS events within the APP transcript can occur in multiple regions, including the acidic extracellular domain, the transmembrane domain, and the cytoplasmic domain.¹² These splicing events give rise to AβPP isoforms with varying lengths and domain compositions after inclusion and/or exclusion of exons 7 and 8 prevalently, which may influence their processing and function in the context of AD. Alongside AβPP, there is a family of proteins known as the amyloid precursor like-proteins (APLP) that share several conserved domains such as the E1 and E2 dimerization domains of AβPP but lack the transmembrane domain encoding for Aβ.¹² Knockout models also independently show that APP, APLP1, or APLP2 produce viable mice but double knockout of APP with APLP2 are embryonic lethal.^13,14 AβPP is a type 1 single-pass transmembrane glycoprotein composed of a large acidic extracellular domain and several subdomains which are conserved through each unique isoform. The E1 and E2 dimerization domains aid with synaptic adhesion and contain a copper ion-binding structural domain (CuBD) and a heparin-binding domain (HBD), which aid in developing the AβPP interactome at the synapse and neuronal development.^1,15 Alongside these dimerization domains, each of the secretase cleavage sites are conserved within each major isoforms of AβPP.¹⁵ However, the Kunitz protease inhibitor (KPI) domain and the OX-2 antigen recognition and binding domain are only present in APP770 whereas the OX-2 domain is absent in both APP751 and APP695 isoforms.¹⁶ Unlike the isoforms containing an KPI domain, APP695 is devoid of both the OX-2 and KPI domains and is predominantly expressed in neurons.¹⁷ Studies have shown that the KPI domain may play several roles including inhibiting serine proteases, disrupting cell metabolism, and dysregulating homodimerization of AβPP which are all associated with Aβ production and regulation.¹⁸ Increased homodimerization of AβPP has been shown to increase amyloid production through apposition of Glycine-XXX-Glycine motifs (GxxxG) in the extracellular transmembrane sequences of AβPP.¹⁹ Notably, the study examined the role of GxxxG motifs in facilitating dimerization without interfering with the dimerization of full-length AβPP or its large ectodomain, which includes the Kunitz protease inhibitor (KPI) domain. Isoforms containing the KPI domain, such as APP751, have also been shown to exhibit a marked increase in AβPP homodimerization.¹⁹ However, homodimerization of APP751 via the KPI domain has shown to produce fewer amyloidogenic metabolites than APP695, but impairing KPI folding through cysteine modifications increased Aβ₄₂ production in CHO cells and showed increased retention of APP751 dimers in the endoplasmic reticulum (ER) and trans-Golgi network (TGN).¹⁸ Studies utilizing neuronal SHSY5Y cells overexpressing the major AβPP isoforms have shown higher formation of amyloidogenic cleavage products in APP695 isoform overexpression versus KPI domain-containing isoforms which include APP751 and APP770.^20,21 These findings suggest that APP695 isoform promotes AβPP C-terminal fragment internalization and amyloid clearance through downstream regulation of amyloid clearance enzyme neprilysin transcription versus that of APP751 and APP770 isoforms. Additionally, a study targeting the spliceosome factor U1 showed that efficient knockdown of this splicing regulatory protein shifted APP expression of primarily APP695 isoforms to KPI containing isoforms APP751 and APP770, and also showed increase in expression of Aβ₄₀ in both HEK and SHSY-5Y cultures.²² Alongside these findings, Chua et al.²³ have shown that APP deficient cells transfected with APP751 isoform impair mitochondrial function by lowering citrate synthase expression, lowering membrane potential, and increasing NAD+/NADH ratio which are all crucial in maintaining mitochondrial functionality in neurons.²³ Furthermore, APP751 dimerization has been shown to form more complexes through interactions with KPI-E1 and KPI-KPI domains versus that of APP695 isoform as confirmed through pull-down assays and in silico modeling.²⁴ Taking this information together demonstrates that APP AS and isoform expression correlatively affect Aβ production and AβPP function by tampering with AβPP dimerization mediated through the KPI domain. KPI containing isoforms such as APP770 and APP751 may aid in initial homodimerization to promote non-amyloidogenic processing but later contrast this preference because of KPI misfolding, thus promoting amyloidogenic processing through increased retention in the endoplasmic reticulum (ER) and trans-Golgi network (TGN).^18,24 These findings also suggest that more information is needed to characterize the OX-2 and KPI domains to further reveal the AβPP isoform functions.

Figure 2.

Structure of the amyloid precursor protein (APP). (A) The APP gene has 18 exons. APP has an N-terminal signal peptide (SP) in exon 1. The E1 dimer domain has a heparin-binding domain (HBD1), and a copper-binding domain (CuBD). Exon 6 encodes for an acidic domain containing glutamic acid residues, whereas exon 7 and exon 8 encode for a Kunitz protease inhibitor (KPI) domain and an OX-2 antigen domain, respectively. The E2 dimer domain contains a second heparin-binding domain (HBD2). Between the E2 and amyloid-β (Aβ) region are two potential N-linked glycosylation sites. The exon 16 and 17 encode mostly for Aβ peptides along with the secretase cleavage sites. The APP intracellular C-terminal domain (AICD) encoded by exon 18 is also shown. (B) Differential mRNA splicing of exons 7 (red) and 8 (clear blue) can lead to the expression of APP-695, −751 and −770 amino-acid isoforms. Only the isoforms APP751 and APP770 contain intact KPI and OX-2 antigen domains. hAPP: human amyloid precursor protein; NH₂: N-terminal; COOH: C-terminal; SP: N-terminal signal peptide; E1: dimer domain 1; E2: dimer domain 2; HBD1: heparin-binding domain 1; CuBD: copper-binding domain; KPI: Kunitz protease inhibitor domain; OX-2: OX-2 antigen domain; HBD2: second heparin-binding domain; Aβ: amyloid-β; AICD: APP intracellular C-terminal domain; aa: amino acids. (Created in BioRender. Bellizzi, A. (2024) https://BioRender.com/n65l926).

Impacts of alternative splicing of MAPT and other genes on AD

AS allows for the generation of multiple mRNA transcripts from a single gene, leading to the production of protein isoforms with distinct structural and functional properties. Changes in AS patterns of genes in the brain can affect gene's function and regulate the onset of several diseases, including AD and many other nervous system diseases.²⁵ One of the main factors affecting the AS in the brain seems to be hypoxia, which has shown a higher risk for AD onset and aging.^26–29 In a study conducted in mouse hippocampal neurons, it has been shown that hypoxia regulated the expression of several splicing factors, including Rbm15. In turn, Rbm15 seemed to affect changes in the ratio of Dicer1 transcripts, participating in the neuronal hypoxic response by regulating miR-29b.³⁰ Rbm15 was also found to increase significantly in AD patients compared to controls, was 60% higher in expression in later stage AD patients versus age-matched controls, and may regulate hypoxia-mediated AS.^28–30 Additionally, in a comprehensive splicing transcriptome-wide association study, 18 causal splicing intron of 15 novel genes were identified to be related to AD susceptibility, even if the role of AS g in etiology of AD remain largely elusive.³¹

Familial and sporadic AD progression involves mutations in multiple aging-associated genes including APP, BACE, microtubule associated protein tau (MAPT), and the presenilin family genes (PSEN1, PSEN2).^32–34 Tau functions as a regulator of axonal transport, neuronal communication, and development through microtubule regulation.³⁵ In pathophysiological conditions, Tau contributes to neurodegeneration and self-aggregation by forming neurofibrillary tangles and propagation of tau seeding that disrupts axonal transport and subsequent neuronal function.³⁶ Tau seeding utilizes misfolded tau proteins to induce the aggregation of normal tau, thereby propagating tau pathology. Oligomeric Aβ has been shown to potentiate the process of tau seeding. Evidence suggests that Aβ oligomers may facilitate tau aggregation by enhancing the cellular uptake of tau seeds.³⁷

The interaction between tau and Aβ is mediated through specific binding regions in tau.³⁸ Molecular dynamics simulations have revealed that the microtubule-binding repeats (R1–R4) in tau interact with Aβ fibrils, with the R2 and R4 regions exhibiting a higher affinity for the fibril elongation ends and promoting β-sheet formation and promoting tau aggregation.³⁹ The enhanced uptake of tau seeds by Aβ oligomers suggests a mechanistic link where Aβ facilitates the intracellular aggregation of tau which can exacerbate AD pathology.

Alongside Tau, the PSEN family proteins serve to establish the γ-secretase enzyme complex which is crucial in severing the C terminal β-fragment of AβPP for further processing of Aβ.⁴⁰ These associated AD-aging genes mentioned previously also have aberrant splicing patterns seen in AD. APP isoforms APP770 and APP751 which both share the KPI domain are upregulated in cases of AD versus age matched controls.⁴¹ Furthermore, alternatively spliced variants forming truncated PSEN2 are upregulated in both sporadic and familial cases of early onset AD versus PSEN1.⁴² Tau is also alternatively spliced into six major isoforms and promotes aggregation of tau following 3R/4R isoform imbalance mediated by AS of exon 10.^43,44 A novel truncated intron 12 including isoform of Tau has been shown to be less aggregation prone than other tau isoforms and is downregulated in late Braak-stage AD patients compared to age-matched controls.⁴⁵ Long-range sequencing of rTg4510 tau transgenic mice have shown novel AβPP and Tau isoforms associated with tauopathy progression which also showed a similar trend in AD brain samples.⁴⁶ These genes work cooperatively to regulate AβPP processing, promote amyloid deposition, and disrupt amyloid clearance but the function of the splice variants of these associated genes remains elusive. In sporadic cases of AD, apolipoprotein E4 (APOE4) also confers a higher risk for AD progression.⁴⁷ Overall, there are several genes associated with APP that are alternatively spliced in AD which affect their function and may play a larger role in AD development and other neurodegenerative disease progression than what has been previously studied.

Furthermore, recent studies have highlighted the role of AβPP isoforms in modulating tau pathology suggesting broader implications for disease progression and therapeutic targeting.⁴⁸ Understanding the specific roles of different AβPP isoforms in AD pathogenesis is crucial for developing targeted therapies aimed at mitigating Aβ toxicity and preserving neuronal function as well as uncovering specific roles of AβPP isoforms in overall AβPP physiology.

APP isoforms and their role in AD pathology and other neurodegenerative diseases

Perturbations in APP AS have been implicated in AD pathogenesis and potentially could affect Aβ production, plaque formation, and neuronal toxicity (Table 1).⁴⁹ Specifically within the hippocampus and dorsolateral prefrontal cortex of those with AD, there is more exon skipping events and increased AS susceptibility compared to that of age-matched controls.^50,51

Table 1.

Expression of APP isoforms in various disease paradigms.

Disease/condition	APP Isoforms
Alzheimer's disease^41,49,52	*↑APP770; ↑APP751; †↓APP695
Huntington's disease⁵³	↑APP770; ↑APP751; ↓ APP695
Cerebrovascular disease (including cerebral amyloid angiopathy)⁵⁴	↑APP770
Coronary artery disease⁵⁵	↓APP770
Neurodegeneration.⁵⁶	↑APP770; ↑APP751; ↓APP695
Down's syndrome^53,57,58	↑APP770; ↑APP751
Parkinson's disease⁵³	↑APP770; ↑APP751
Inflamed endothelial cells and activated platelets^55,59	↑APP770

*↑: Upregulated.

†↓: Downregulated.

In AD, APP770 and APP751 isoforms are increased within most regions of the brain over APP695 (Table 1).⁴⁸ Within the temporal and frontal lobes of the brain, significant changes of splicing events within AD brains demonstrated marked differences in APOE and MAPT isoform expression over healthy age-matched controls.⁵² Alongside these differences, the cerebral cortex, hippocampus, and cerebellum also shows the highest distribution of AβPP expression compared to other brain regions as well as compared to other organs.^60,61 Studies have also shown that in brains of people with Huntington's disease, Down's syndrome (DS), frontal lobe atrophy, and Parkinson's disease all show a higher transcriptional count of APP751 isoform over that of APP695 in both the temporal and frontal lobe of post-mortem brains (Table 1).^53,56

Specifically, Huntington's and Parkinson's diseases show a marked increase of APP751/APP695 ratio. Amyloid pathology is also typically seen in people with either Parkinson's disease or DS.⁵³ In people living with DS, extracellular amyloid plaque pathology typically arises around 40 years of age and specifically shows a marked increase of KPI-containing AβPP expression.^57,58 Recently identified in brains of post-mortem people with DS, dual-specificity tyrosine-phosphorylation regulated kinase 1 (Dyrk1a) has been shown to promote exon 7 inclusion of AβPP which primarily affects KPI domain containing isoforms of AβPP and humanized Dyrk1a expressing mice also exhibited profound amyloid production.⁵⁸ Also within this study, N2a cells transfected with APP770, APP751, and APP695 showed a marked increase in sAβPP-β and C99 fragment in APP770 and APP751 isoforms versus that of APP695 isoform.⁵⁸ Overall, KPI containing isoforms of AβPP have been shown to increase amyloidogenic cleavage products in AD but how KPI containing isoform expression of AβPP changes over time in other neurodegenerative diseases as well as how KPI containing isoforms affect AβPP dimerization and localization in other neurodegenerative diseases remain elusive and deserve further exploration. Uncovering the complexities of AβPP isoform expression and the effect of KPI dimerization on AβPP processing may provide novel therapeutic avenues to understand and prevent KPI misfolding and generation of amyloidogenic peptides. Moreover, beyond Aβ production, AS of APP may influence other aspects of AD pathophysiology. AβPP isoforms can differ in their interactions with cellular receptors and signaling pathways, affecting synaptic function, neuronal survival, and inflammation.^55,62,63 Dysregulated splicing events that alter the balance between neuroprotective and neurotoxic AβPP isoforms could contribute to synaptic dysfunction, neuronal and glial mitochondrial impairment, and oxidative stress observed in AD.^64,65

Alongside neuronal and glial cells within the brain, dysregulation of AβPP expression also affects other cell types such as vascular endothelial cells and platelets. Platelets express high levels of soluble APP770 specifically over other blood cell types and are positively correlated with CD40 ligand which is a classical marker of platelet activation.^55,59 Moreover, patients undergoing dual anti-platelet therapy to treat coronary artery disease (CAD) showed a marked decrease in soluble APP770, whereas vascular endothelial cells express high levels of soluble APP770 (Table 1).⁵⁹ These findings demonstrate that APP770 may be used as an effective marker for vascular and platelet associated diseases such as CAD or other cerebrovascular diseases.⁵⁵ Mice expressing human APP770 in endothelial cells demonstrated a mounted amyloid deposition response specifically in blood vessels which was further exacerbated after crossing with APP knock-in mice (Table 1).⁵⁴ This amyloid deposition response presented as a classical sign of cerebral amyloid angiopathy (CAA) which is typically seen in people living with AD. These findings suggest that neuronal and endothelial cells work cooperatively to regulate APP770 expression to control amyloid deposition and plaque formation. This leaves multiple areas to study APP770 and APP751 isoform expression in the context of endothelial and neuronal coordination of Aβ deposition in CAA, inflammatory responses, and other neurodegenerative diseases. The coordination of APP770 expression also demonstrates how splicing regulation or therapeutic targeting of these AβPP isoforms may play a protective or preventive role in early onset of these disease conditions.^54,55,59

Therapies for treating splicing dysregulation and recent advances in three-dimensional culture systems to study AD pathology

The treatment of dysregulated splicing has emerged as a promising therapeutic strategy for various neurodegenerative diseases.⁶⁶ Currently, the U.S. Food and Drug Administration (FDA) has approved three splice-switching oligonucleotides (SSOs) for clinical use: eteplirsen, golodirsen, and nusinersen.⁶⁷ Eteplirsen and golodirsen are specifically designed for Duchenne muscular dystrophy (DMD), where they promote exon skipping to restore the reading frame of the DMD gene in patients with frameshift mutations.^68,69 Nusinersen, approved for spinal muscular atrophy (SMA), works by enhancing the inclusion of exon 7 in the SMN2 gene through the steric blockade of an intronic splicing silencer, thereby improving the production of functional survival motor neuron (SMN) protein.⁷⁰

Antisense oligonucleotides (ASOs) have also been employed in gene therapy to address alternative splicing events in neuromuscular diseases like Huntington's disease and DMD.⁶⁷ In AD, an ASO targeting trisomy 21 of the APP gene has been shown to reduce both intracellular and extracellular soluble amyloid and mitigated endosomal enlargement.⁷¹ Despite the progress with ASOs in neuromuscular diseases, ASOs have yet to reach clinical trials for AD targeting amyloid production or γ-secretase site exclusion, although they have advanced in trials aimed at reducing tau hyperphosphorylation.^72–74

In addition to ASOs, CRISPR/Cas9-based approaches have shown promise, particularly in correcting the apolipoprotein E (APOE4) genotype to the APOE3 allele in human induced pluripotent stem cells (hiPSCs) derived from AD patients, potentially alleviating the harmful effects of the APOE4 genotype.⁷⁵ The FDA's approval of SSOs, along with the growing application of ASOs and CRISPR technologies, highlights the increasing potential of RNA-based therapies to address genetic dysfunction in AD and other neurodegenerative diseases. The success of SSOs underscores the precision of exon-skipping strategies in partially restoring gene function.^66–74 However, challenges such as patient-specific variability in responses, difficulties with intracellular delivery, and concerns about long-term efficacy remain, emphasizing the need for continued refinement of SSO and RNA-based strategies for therapeutic use.⁷⁶

The main challenges presented by the new SSO and RNA-based therapeutic strategies can be addressed through the utilization of three-dimensional (3D) culture systems which have introduced novel models for studying neurodegenerative diseases, including AD.^77–79 Human-induced pluripotent stem cell (hiPSC)-derived organoids have become valuable tools for investigating the mechanisms of AD in an environment that closely mimics the native human brain, offering insights into the complex pathology of the disease.^79,80 Organoids unique to each AD patient can be generated from hiPSCs, which can recapitulate key pathological features, such as hyperphosphorylated tau protein, Aβ plaques, and endosomal abnormalities.^81,82 For instance, Raja et al.⁸² used neural organoids derived from familial AD (FAD) patients to establish a 3D culture system, revealing increased production and aggregation of Aβ. Similarly, Choi et al.⁸⁰ conducted a study using 3D culture systems and observed the induction of amyloid plaques and fibrillary p-tau aggregates resulting from the overexpression of APP and PSEN1, along with multiple FAD mutations. Notably, organoids derived from AD patient-specific hiPSCs have also been utilized to examine the impact of the APOE4 allele on splicing and disease pathology.⁸³ These studies found that APOE4-expressing organoids exhibit increased lipid droplet formation and cholesterol accumulation, both of which are associated with tau pathology and Aβ deposition.⁸³

However, the effects of neurovascular amyloid deposits and the complexities of brain vasculature on AβPP expression require more comprehensive models, such as 3D vessel organoids or cerebral organoid cultures. These models could improve understanding of splicing regulation, AβPP isoform expression, and the downstream effects of these isoforms in translationally relevant models.⁷⁷ Moreover, the inherent variability of organoids and the absence of a cohesive neurovascular 3D culture system pose challenges in accurately examining the intricate roles of various brain regions in splicing regulation and expression within patient-derived AD models, as well as models for other neurodegenerative diseases.⁸⁴ Cerebral organoids also face challenges such as inconsistent microglial integration and the lack of self-organizing vascular networks—both essential for proper neurodevelopment and the establishment of a functional blood-brain barrier.^85,86

While efforts have been made to vascularize brain organoids through assembloid cultures that integrate vascular organoids, these models have yet to achieve full integration or establish a perfused vascular network.⁸⁷ Emerging technologies, such as microfluidic systems and 3D-printed vascular meshes, offer promising strategies for enhancing vascularization and better replicating the brain microenvironment.⁸⁸ Additionally, co-culturing with vascular progenitors has shown potential in promoting more complex and physiologically relevant brain models within 3D culture systems.⁸⁹ These systems offer a more relevant cellular microenvironment for studying splicing regulation in real-time and could allow for the exploration of novel RNA-binding partners, such as RBFOX1, which plays a regulatory and protective role in maintaining transcriptional stability in AD brains.^90,91

Targeting both familial and sporadic AD cases in 3D cell cultures will help identify unique splicing events associated with early versus late-onset AD, as well as provide insights into overall splicing regulation as the brain ages. In conclusion, 3D cultures models have emerged as a powerful strategy for studying AD mechanisms and investigating the roles of AβPP isoforms and their genetic risk factors, opening new avenues for anti-neurodegeneration and anti-aging therapeutics. These systems offer a comprehensive cellular environment that is scalable, easily generated, and suitable for studying genotypic changes. They can also be applied in microfluidic systems to simulate blood flow, further enhancing their potential for disease modeling and drug testing.⁸⁴

Conclusions

This review gathers evidence highlighting the need for further exploration of the physiological roles of AβPP isoforms to better understand AβPP function and domain characteristics. Specifically, it underscores the cooperative roles of the KPI and OX-2 domains, and how cells regulate AS of APP751 and APP770 in relation to their potential preventative or contributory roles in disease. The expression of APP770 and APP751 isoforms has been shown to be elevated in various neurodegenerative diseases, including AD, Parkinson's disease, Huntington's disease, frontotemporal atrophy, and DS. However, the definitive roles of APP770 and APP751 in these conditions remain unclear, offering abundant opportunities for research into how splicing regulation influences disease progression.

In contrast, the potential protective role of APP695 has been understudied. Specifically, exons 7 and 8 of APP695 may be targeted by RNA-binding proteins, such as spliceosome regulatory factors, which could contribute to inflammation and amyloid production.⁶⁴ Additionally, APP770 expression has been implicated in other diseases, such CAA, suggesting that AβPP may have broader functions at the neurovascular interface that warrant further investigation.

Finally, the use of advanced translational models of AD could provide valuable insights into splicing regulation at a more relevant biological level. These models would enable the detection and examination of pathogenic isoform expressions, which could be crucial for reducing disease progression.

Footnotes

Acknowledgments

The authors wish to thank past and present members of the Department of Microbiology, Immunology and Inflammation and Center for Neurovirology and Gene Editing for sharing of ideas, reagents, and equipment.

Author contributions

Matthew Pasteris: Conceptualization; Methodology; Writing – original draft.

Senem Cakir: Conceptualization; Methodology; Writing – original draft.

Anna Bellizzi: Conceptualization; Methodology; Writing - review & editing.

Ilker K Sariyer: Conceptualization; Funding acquisition; Resources; Supervision; Writing – review & editing.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported in part by NIH grants R01 DA052284 (NIDA) and R01 MH110360-6 (NIMH) and used services offered by core facilities of the Comprehensive NeuroHIV Center (CNHC) P30MH092177 (NIMH).

Declaration of conflicting interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

ORCID iDs

Senem Cakir

Anna Bellizzi

Ilker K Sariyer

References

Guo

Gaddam

SSK

, et al. Amyloid precursor protein revisited. J Biol Chem 2012; 287: 2437–2445.

Esch

Keim

Beattie

, et al. Cleavage of amyloid β peptide during constitutive processing of its precursor. Science 1990; 248: 1122–1124.

Das

Wang

Ganguly

, et al. Visualizing APP and BACE-1 approximation in neurons yields insight into the amyloidogenic pathway. Nat Neurosci 2016; 19: 55–64.

Gervais

Robertson

, et al. Involvement of caspases in proteolytic cleavage of Alzheimer’s amyloid-β precursor protein and amyloidogenic Aβ peptide formation. Cell 1999; 97: 395–406.

Soriano

Chandra

, et al. The amyloidogenic pathway of amyloid precursor protein (APP) is independent of its cleavage by caspases. J Biol Chem 2001; 276: 29045–29050.

Mullan

Crawford

Axelman

, et al. A pathogenic mutation for probable Alzheimer’s disease in the APP gene at the N–terminus of β–amyloid. Nat Genet 1992; 1: 345–347.

Kelleher

Shen

. Presenilin-1 mutations and Alzheimer’s disease. Proc Natl Acad Sci U S A 2017; 114: 629–631.

Fukuchi

Younkin

. The alternatively spliced Kunitz protease inhibitor domain alters amyloid β protein precursor processing and amyloid β protein production in cultured cells. J Biol Chem 1996; 271: 30929–30934.

Chang

Hinrich

Roman

, et al. Targeting amyloid-β precursor protein, APP, splicing with antisense oligonucleotides reduces toxic amyloid-β production. Mol Ther 2018; 26: 1539–1551.

10.

Dawkins

Small

. Insights into the physiological function of the β-amyloid precursor protein: beyond Alzheimer’s disease. J Neurochem 2014; 129: 756–769.

11.

Kim

Han

Cho

, et al. Brain region-dependent alternative splicing of Alzheimer disease (AD)-risk genes is associated with neuropathological features in AD. Int Neurourol J 2022; 26: S126–S136.

12.

van der Kant

Goldstein

LSB

. Cellular functions of the amyloid precursor protein from development to dementia. Dev Cell 2015; 32: 502–515.

13.

Rassoulzadegan

. APLP2, A member of the Alzheimer precursor protein family, is required for correct genomic segregation in dividing mouse cells. EMBO J 1998; 17: 4647–4656.

14.

Mallm

Tschäpe

Hick

, et al. Generation of conditional null alleles for APP and APLP2. Genesis 2010; 48: 200–206.

15.

Müller

Deller

Korte

. Not just amyloid: physiological functions of the amyloid precursor protein family. Nat Rev Neurosci 2017; 18: 281–298.

16.

Clark

Gagnon

Williams

, et al. MRC OX-2 antigen: a lymphoid/neuronal membrane glycoprotein with a structure like a single immunoglobulin light chain. EMBO J 1985; 4: 113–118.

17.

McDonald

Kwong

. Does Noggin head a new class of Kunitz domain? Trends Biochem Sci 1993; 18: 208–209.

18.

Khalifa

Tyteca

Marinangeli

, et al. Structural features of the KPI domain control APP dimerization, trafficking, and processing. FASEB J 2012; 26: 855–867.

19.

Kienlen-Campard

Tasiaux

Van Hees

, et al. Amyloidogenic processing but not amyloid precursor protein (APP) intracellular c-terminal domain production requires a precisely oriented APP dimer assembled by transmembrane GXXXG motifs. J Biol Chem 2008; 283: 7733–7744.

20.

Kulatunga

DCM

Ranaraja

Kim

, et al. A novel APP splice variant-dependent marker system to precisely demarcate maturity in SH-SY5Y cell-derived neurons. Sci Rep 2024; 14: 12113.

21.

Kulatunga

DCM

Ranaraja

Kim

, et al. Precise characterization of differentiated SH-SY5Y cells using a novel APP splice variant dependent method: For intended use in AD research. Res Sq 2023. https://doi.org/10.21203/rs.3.rs-3479761/v1 [Preprint]. Posted 29 October 2023.

22.

Bai

Hales

Chen

P-C

, et al. U1 small nuclear ribonucleoprotein complex and RNA splicing alterations in Alzheimer’s disease. Proc Natl Acad Sci U S A 2013; 110: 16562–16567.

23.

Chua

L-M

Lim

M-L

Wong

B-S

. The Kunitz-protease inhibitor domain in amyloid precursor protein reduces cellular mitochondrial enzymes expression and function. Biochem Biophys Res Commun 2013; 437: 642–647.

24.

Byun

Vellampatti

Chatterjee

, et al. Characterization of the role of Kunitz-type protease inhibitor domain in dimerization of amyloid precursor protein. J Comput Chem 2023; 44: 1437–1445.

25.

Nik

Bowman

. Splicing and neurodegeneration: insights and mechanisms. WIRES RNA 2019; 10: e1532.

26.

Tao

Gong

, et al. The relationship between hypoxia and Alzheimer’s disease: an updated review. Front Aging Neurosci 2024; 16: 1402774.

27.

Sriram

Mehkri

Quintin

, et al. Shared pathophysiology: understanding stroke and Alzheimer’s disease. Clin Neurol Neurosurg 2022; 218: 107306.

28.

Vijayan

Reddy

. Stroke, vascular dementia, and Alzheimer’s disease: molecular links. J Alzheimers Dis 2016; 54: 427–443.

29.

Adeyemi

Awakan

Afolabi

, et al. Hypoxia and the kynurenine pathway: implications and therapeutic prospects in Alzheimer’s disease. Oxid Med Cell Longev 2021; 2021: 5522981.

30.

Liu

Sun

, et al. The changes in the ratio of Dicer1 transcripts can participate in the neuronal hypoxic response by regulating miR-29b. Cereb Cortex 2025; 35: bhae490.

31.

Sun

Bae

Zhu

, et al. A splicing transcriptome-wide association study identifies novel altered splicing for Alzheimer’s disease susceptibility. Neurobiol Dis 2023; 184: 106209.

32.

Xiang

Meng

Wang

, et al. Two novel DNA motifs are essential for BACE1 gene transcription. Sci Rep 2014; 4: 6864.

33.

Deng

Gao

. Analysis of genetic mechanisms, clinical characteristics and pathological changes in familial Alzheimer’s disease and sporadic Alzheimer’s disease. BIO Web Conf 2024; 111: 03001.

34.

Van Cauwenberghe

Van Broeckhoven

Sleegers

. The genetic landscape of Alzheimer disease: clinical implications and perspectives. Genet Med 2016; 18: 421–430.

35.

Kidd

. Alzheimer’s disease—an electron microscopical study. Brain 1964; 87: 307–320.

36.

Endoh

Ogawara

Iwatsubo

, et al. Lack of the car□yl terminal sequence of tau in ghost tangles of Alzheimer’s disease. Brain Res 1993; 601: 164–172.

37.

Shin

Cao

, et al. Amyloid β-protein oligomers promote the uptake of tau fibril seeds potentiating intracellular tau aggregation. Alzheimers Res Ther 2019; 11: 86.

38.

Darricau

Dou

Kinet

, et al. Tau seeds from Alzheimer’s disease brains trigger tau spread in macaques while oligomeric-Aβ mediates pathology maturation. Alzheimers Dement 2024; 20: 1894–1912.

39.

Song

Gatch

Sun

, et al. Differential binding and conformational dynamics of tau microtubule-binding repeats with a preformed amyloid-β fibril seed. ACS Chem Neurosci 2023; 14: 1321–1330.

40.

Arber

Belder

CRS

Tomczuk

, et al. The presenilin 1 mutation P436S causes familial Alzheimer’s disease with elevated Aβ43 and atypical clinical manifestations. Alzheimers Dement 2024; 20: 4717–4726.

41.

Menéndez-González

Pérez-Pinera

Martínez-Rivera

, et al. APP Processing and the APP-KPI domain involvement in the amyloid cascade. Neurodegener Dis 2005; 2: 277–283.

42.

Course

Gudsnuk

Keene

, et al. Aberrant splicing of PSEN2, but not PSEN1, in individuals with sporadic Alzheimer’s disease. Brain 2023; 146: 507–518.

43.

Corsi

Bombieri

Valenti

, et al. Tau isoforms: gaining insight into MAPT alternative splicing. Int J Mol Sci 2022; 23: 15383.

44.

Niblock

Gallo

J-M

. Tau alternative splicing in familial and sporadic tauopathies. Biochem Soc Trans 2012; 40: 677–680.

45.

García-Escudero

Ruiz-Gabarre

Gargini

, et al. A new non-aggregative splicing isoform of human tau is decreased in Alzheimer’s disease. Acta Neuropathol (Berl) 2021; 142: 159–177.

46.

Leung

Bamford

Jeffries

, et al. Long-read transcript sequencing identifies differential isoform expression in the entorhinal cortex in a transgenic model of tau pathology. Nat Commun 2024; 15: 6458.

47.

Chartier-Hariln

M-C

Parfitt

Legrain

, , et al. ε4 allele as a major risk factor for sporadic early and late-onset forms of Alzheimer’s disease: analysis of the 19q13.2 chromosomal region. Hum Mol Genet 1994; 3: 569–574.

48.

Chen

Fan

, et al. APP Mediates tau uptake and its overexpression leads to the exacerbated tau pathology. Cell Mol Life Sci 2023; 80: 123.

49.

Matsui

Ingelsson

Fukumoto

, et al. Expression of APP pathway mRNAs and proteins in Alzheimer’s disease. Brain Res 2007; 1161: 116–123.

50.

Raj

Wong

, et al. Integrative transcriptome analyses of the aging brain implicate altered splicing in Alzheimer’s disease susceptibility. Nat Genet 2018; 50: 1584–1592.

51.

Han

Miller

Byun

, et al. Identification of exon skipping events associated with Alzheimer’s disease in the human hippocampus. BMC Med Genomics 2019; 12(Suppl 1): 13.

52.

Singh

Tosakulwong

Graff-Radford

, et al. APOE Ε4 influences medial temporal atrophy and tau deposition in atypical Alzheimer’s disease. Alzheimers Dement 2023; 19: 784–796.

53.

Panegyres

Zafiris-Toufexis

Kakulas

. Amyloid precursor protein gene isoforms in Alzheimer’s disease and other neurodegenerative disorders. J Neurol Sci 2000; 173: 81–92.

54.

Tachida

Miura

Muto

, et al. Endothelial expression of human amyloid precursor protein leads to amyloid β in the blood and induces cerebral amyloid angiopathy in knock-in mice. J Biol Chem 2022; 298: 101880.

55.

Kitazume

Yoshihisa

Yamaki

, et al. Soluble amyloid precursor protein 770 is released from inflamed endothelial cells and activated platelets. J Biol Chem 2012; 287: 40817–40825.

56.

Iverfeldt

Walaas

Greengard

. Altered processing of Alzheimer amyloid precursor protein in response to neuronal degeneration. Proc Natl Acad Sci U S A 1993; 90: 4146–4150.

57.

Oyama

Cairns

Shimada

, et al. Down’s syndrome: up-regulation of β-amyloid protein precursor and τ mRNAs and their defective coordination. J Neurochem 1994; 62: 1062–1066.

58.

Chu

Lei

, et al. Dual-specificity tyrosine phosphorylation-regulated kinase 1A promotes the inclusion of amyloid precursor protein exon 7. Biochem Pharmacol 2024; 224: 116233.

59.

Miura

Yoshihisa

Misaka

, et al. Amyloid precursor protein 770 is specifically expressed and released from platelets. J Biol Chem 2020; 295: 13194–13201.

60.

Cha

Shen

Kang

. Regulation of gene expression by the APP family in the adult cerebral cortex. Sci Rep 2022; 12: 66.

61.

Aydin

Filippov

Tschäpe

J-A

, et al. Comparative transcriptome profiling of amyloid precursor protein family members in the adult cortex. BMC Genomics 2011; 12: 160.

62.

Fogel

Frere

Segev

, et al. APP Homodimers transduce an amyloid-β-mediated increase in release probability at excitatory synapses. Cell Rep 2014; 7: 1560–1576.

63.

Brugg

Dubreuil

Huber

, et al. Inflammatory processes induce beta-amyloid precursor protein changes in mouse brain. Proc Natl Acad Sci U S A 1995; 92: 3032–3035.

64.

Yue

Xie

, et al. Ontology specific alternative splicing changes in Alzheimer’s disease. Front Genet 2022; 13: 926049.

65.

Vaillant-Beuchot

Eysert

Duval

, et al. The amyloid precursor protein and its derived fragments concomitantly contribute to the alterations of mitochondrial transport machinery in Alzheimer’s disease. Cell Death Dis 2024; 15: 367.

66.

Schmok

Yeo

. Strategies for programmable manipulation of alternative splicing. Curr Opin Genet Dev 2024; 89: 102272.

67.

Aoki

Wood

MJA

. Emerging oligonucleotide therapeutics for rare neuromuscular diseases. J Neuromuscul Dis 2021; 8: 869–884.

68.

Heo

Y-A

. Golodirsen: first approval. Drugs 2020; 80: 329–333.

69.

Lim

Maruyama

Yokota

. Eteplirsen in the treatment of Duchenne muscular dystrophy. Drug Des Devel Ther 2017; 11: 533–545.

70.

Arechavala-Gomeza

Graham

Popplewell

, et al. Comparative analysis of antisense oligonucleotide sequences for targeted skipping of exon 51 during dystrophin pre-mRNA splicing in human muscle. Hum Gene Ther 2007; 18: 798–810.

71.

Hung

Fertan

Livesey

, et al. APP Antisense oligonucleotides reduce amyloid-β aggregation and rescue endolysosomal dysfunction in Alzheimer’s disease. Brain 2024; 147: 2325–2333.

72.

Farr

Erickson

Niehoff

, et al. Central and peripheral administration of antisense oligonucleotide targeting amyloid-β protein precursor improves learning and memory and reduces neuroinflammatory cytokines in Tg2576 (AβPPswe) mice. J Alzheimers Dis 2014; 40: 1005–1016.

73.

Grabowska-Pyrzewicz

Want

Leszek

, et al. Antisense oligonucleotides for Alzheimer’s disease therapy: from the mRNA to miRNA paradigm. eBioMedicine 2021; 74: 103691.

74.

Mignon

Kordasiewicz

Lane

, et al. Design of the first-in-human study of IONIS-MAPTRx, a tau-lowering antisense oligonucleotide, in patients with Alzheimer disease (S2.006). Neurology 2018; 90: S2.006.

75.

Lin

Y-T

Seo

Gao

, et al. APOE4 Causes widespread molecular and cellular alterations associated with Alzheimer’s disease phenotypes in human iPSC-derived brain cell types. Neuron 2018; 98: 1141–1154.e7.

76.

Sparmann

Vogel

. RNA-Based medicine: from molecular mechanisms to therapy. EMBO J 2023; 42: e114760.

77.

Sharma

Karan

Lee

, et al. Advances in modeling Alzheimer’s disease in vitro. Adv NanoBiomed Res 2021; 1: 2100097.

78.

Choe

Yeo

Kim

, et al. Simple modeling of familial Alzheimer’s disease using human pluripotent stem cell-derived cerebral organoid technology. Stem Cell Res Ther 2024; 15: 118.

79.

Heinzelmann

Piraino

Costa

, et al. iPSC-derived and patient-derived organoids: applications and challenges in scalability and reproducibility as pre-clinical models. Curr Res Toxicol 2024; 7: 100197.

80.

Choi

Kim

Quinti

, et al. 3D Culture models of Alzheimer’s disease: a road map to a “cure-in-a-dish”. Mol Neurodegener 2016; 11: 75.

81.

Papaspyropoulos

Tsolaki

Foroglou

, et al. Modeling and targeting Alzheimer’s disease with organoids. Front Pharmacol 2020; 11: 396.

82.

Raja

Mungenast

Lin

Y-T

, et al. Self-organizing 3D human neural tissue derived from induced pluripotent stem cells recapitulate Alzheimer’s disease phenotypes. PLoS One 2016; 11: e0161969.

83.

Huang

Zhang

Cao

, et al. Chimeric cerebral organoids reveal the essentials of neuronal and astrocytic APOE4 for Alzheimer’s tau pathology. Signal Transduct Target Ther 2022; 7: 176.

84.

Sreenivasamurthy

Laul

Zhao

, et al. Current progress of cerebral organoids for modeling Alzheimer’s disease origins and mechanisms. Bioeng Transl Med 2023; 8: e10378.

85.

Matsui

Tsuru

Hasegawa

, et al. Vascularization of human brain organoids. Stem Cells 2021; 39: 1017–1024.

86.

Sun

X-Y

X-C

, et al. Generation of vascularized brain organoids to study neurovascular interactions. eLife 2022; 11: e76707.

87.

Dao

You

, et al. Modeling blood-brain barrier formation and cerebral cavernous malformations in human PSC-derived organoids. Cell Stem Cell 2024; 31: 818–833.e11.

88.

Ding

, et al. Artificial meshed vessel-induced dimensional breaking growth of human brain organoids and multiregional assembloids. ACS Nano 2024; 18: 26201–26214.

89.

Kistemaker

Van Bodegraven

De Vries

, et al. Vascularized human brain organoids: current possibilities and prospects. Trends Biotechnol 2025; 43: 1275–1285.

90.

Alkallas

Fish

Goodarzi

, et al. Inference of RNA decay rate from transcriptional profiling highlights the regulatory programs of Alzheimer’s disease. Nat Commun 2017; 8: 909.

91.

Alam

Suzuki

Tsukahara

. Alternative splicing regulation of APP exon 7 by RBFox proteins. Neurochem Int 2014; 78: 7–17.