Dual Targeting of Tau Kinases and Autophagy by Abemaciclib Independent of CDK4/6 Inhibition
Jihui Han, June‐Hyun Jeong, Dongjoon Lee, Yujin Jung, Yujin Jin, Eun Sun Jung, Haeng Jun Kim, Woo Youn Kim, Chang‐Han Lee, Inhee Mook‐Jung
TL;DR
Abemaciclib, a cancer drug, shows promise for treating Alzheimer's by targeting tau proteins and boosting brain cell cleanup, unrelated to its original use.
Contribution
Abemaciclib's novel dual action on tau kinases and autophagy in Alzheimer's is revealed, independent of its CDK4/6 inhibition mechanism.
Findings
Abemaciclib improves cognition and reduces neurodegeneration in Alzheimer's models without affecting amyloid or glial activation.
The drug inhibits tau kinases CaMKII and GSK3β and enhances autophagy to clear pathological tau proteins.
Findings are validated in both mouse models and human-derived brain organoids, supporting drug repurposing for Alzheimer's.
Abstract
Alzheimer's disease (AD) is marked by progressive cognitive decline driven largely by tau pathology, yet disease‐modifying therapies targeting tau remain limited. In this study, we re‐evaluated abemaciclib, a clinically approved CDK4/6 inhibitor for breast cancer and uncovered its previously unrecognized therapeutic potential in AD via CDK4/6‐independent mechanisms. Using the APPNL−F/MAPT double knock‐in mouse model (dKI) and AD patient‐derived brain organoids, we found that abemaciclib robustly ameliorates cognitive deficits and reduces neurodegeneration without altering amyloid burden or glial activation. Mechanistically, abemaciclib selectively inhibited key tau kinases, particularly Ca2⁺/calmodulin‐dependent protein kinase II (CaMKII) and glycogen synthase kinase 3β (GSK3β), independent of CDK4/6 inhibition, as confirmed by lentiviral knockdown experiments. Furthermore, abemaciclib…
Genes, proteins, chemicals, diseases, species, mutations and cell lines named across the full text — each resolved to its canonical identifier and authoritative record.
Click any figure to enlarge with its caption.
FIGURE 1
FIGURE 2
FIGURE 3
FIGURE 4
FIGURE 5
FIGURE 6
FIGURE 7- —Ministry of Education10.13039/501100002701
- —Korea Dementia Research Center10.13039/100020206
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsAdvanced Breast Cancer Therapies · Cancer-related cognitive impairment studies · Cancer-related Molecular Pathways
Introduction
1
Alzheimer's disease (AD) is one of the most prevalent neurodegenerative disorders, characterized by progressive memory loss and cognitive decline. The hallmark pathological features of AD include the accumulation of amyloid‐beta (Aβ) plaques, neurofibrillary tangles composed of phosphorylated tau proteins, and neuronal loss [1]. Despite intensive research efforts, effective disease‐modifying treatments for AD remain elusive, underscoring the urgent need for novel therapeutic approaches. Recent advances in amyloid‐β (Aβ)‐targeted therapies have led to the U.S. Food and Drug Administration (FDA) approval of several agents that reduce plaque burden and modestly slow cognitive decline, representing important progress in AD treatment [2, 3]. Nevertheless, these advances have also highlighted the multifactorial nature of AD pathogenesis, where selective targeting of amyloid may not fully address the complex mechanisms of neurodegeneration. Consequently, expanding the therapeutic strategies to target tau pathology remains crucial, particularly given the robust correlation between tau burden and cognitive decline in AD patients [4]. Notably, tau pathology exhibits a hierarchical spatiotemporal progression that closely parallels the sequence of neurodegeneration and cognitive impairment, further validating tau as a critical therapeutic target.
In the search for effective AD treatments, drug repurposing has emerged as a promising strategy. This approach, which involves the application of approved drugs to new indications, offers several advantages over traditional drug discovery methods. Repurposed drugs have established safety profiles and well‐characterized pharmacokinetic properties, which significantly reduce time and costs for drug development [5]. Moreover, existing clinical data can facilitate more rapid translation to AD trials, addressing the urgent need for therapeutic interventions in the rapidly growing patient population. To harness the potential of drug repurposing for AD, we previously developed a network‐based drug screening platform that leverages AD patient‐derived brain organoids to evaluate drug efficacy against key AD pathological features [6]. This in silico mathematical model predicts intricate neural dynamics and assesses molecular regulatory interactions, computing probable impacts of prospective medications on five prominent AD phenotypes: Aβ, phosphorylated tau, synaptic loss, apoptosis, and autophagy. These phenotypes are weighed differently for optimal candidate selection, specifically for AD. Through this systematic approach, we identified abemaciclib, an FDA‐approved drug, as a promising candidate for AD treatment.
Abemaciclib is currently approved for the treatment of hormone receptor‐positive, human epidermal growth factor receptor 2‐negative advanced or metastatic breast cancer. Its primary mechanism of action involves inhibiting cyclin‐dependent kinases 4 and 6 (CDK4/6), which play crucial roles in cell cycle progression. Specifically, abemaciclib prevents the phosphorylation of the retinoblastoma (Rb) protein, a key regulator of the G1‐S phase transition in the cell cycle [7]. Importantly, abemaciclib demonstrates superior blood‐brain barrier penetration compared to other CDK4/6 inhibitors, a crucial property for central nervous system (CNS)‐targeted drugs [8, 9].
Recently, a study reported the therapeutic effects of abemaciclib using multiple mouse models, each addressing distinct aspects of AD pathology [10]. In 5xFAD transgenic mice, abemaciclib improved cognitive function and reduced Aβ plaque burden through modulation of neprilysin, ADAM17, and presenilin‐1 (PSEN1). Using PS19 mice, a tauopathy model, they demonstrated decreased phosphorylated tau levels via inhibition of DYRK1A and GSK3β. Additionally, in lipopolysaccharide‐treated wild‐type mice, abemaciclib attenuated neuroinflammation through inhibition of AKT/STAT3 signaling. While these results provide compelling evidence for abemaciclib's potential as an AD therapeutic, the study relied primarily on transgenic mouse models with non‐physiological overexpression of pathogenic proteins, limiting its translational relevance. Furthermore, the study lacked a comprehensive model exhibiting both amyloid and tau pathologies simultaneously, which is critical for understanding therapeutic efficacy in the complex pathological environment of AD. The mechanisms underlying abemaciclib's effects on tau pathology also remained incompletely characterized, particularly in human‐relevant model systems.
To address these limitations and further elucidate the therapeutic potential of abemaciclib in AD, we employed a multi‐model approach centered on physiologically relevant models. We utilized the APP^NL−‐F^/MAPT (amyloid precursor protein/microtubule‐associated protein tau) double knock‐in (dKI) mouse model, which expresses human amyloid precursor protein (APP) and tau at physiological levels and exhibits spontaneous, age‐dependent development of both amyloid and tau pathologies without overexpression artifacts [11]. Additionally, we employed brain organoids derived from sporadic AD patient‐induced pluripotent stem cells (iPSCs), which recapitulate key aspects of human brain development and disease in 3D environment models [12, 13, 14, 15]. This human‐relevant model system offers distinct advantages for investigating drug effects in a physiologically relevant context that closely mimics human pathology, potentially bridging the gap between preclinical studies and clinical outcomes.
In this study, we demonstrate that abemaciclib not only ameliorates cognitive deficits and attenuates neurodegeneration in a physiologically relevant AD animal model, but also exerts its therapeutic effects through previously unrecognized, CDK4/6‐independent mechanisms. Specifically, abemaciclib selectively reduces tau phosphorylation without altering amyloid burden or glial activation, acting through a dual mechanism of CaMKII and GSK3β inhibition and enhanced autophagy‐mediated clearance of pathological tau proteins. These findings redefine abemaciclib's pharmacological profile beyond its canonical role as a CDK4/6 inhibitor, uncovering a novel therapeutic mode of action that positions it as a multifunctional and disease‐modifying candidate for AD and related tauopathies.
Results
2
Abemaciclib Ameliorates Cognitive Deficits and Attenuates Neurodegeneration in AppNL−F/MAPT Double Knock‐in Mice
2.1
To evaluate the therapeutic potential of abemaciclib in AD, we utilized the App^NL−F^/MAPT double knock‐in (dKI) mouse model, which expresses human APP and human tau at physiological levels. Unlike traditional transgenic models that rely on overexpression of pathogenic proteins, this model exhibits spontaneous, age‐dependent development of both amyloid and tau pathologies, thus providing a more clinically relevant model of AD pathogenesis. Twelve‐month‐old dKI mice received daily oral administration of abemaciclib (50 mg/kg) or vehicle for four weeks (Figure 1A). This dose and duration were selected based on previous studies demonstrating effective brain penetration at 30 mg/kg and therapeutic efficacy across dose ranges of 25–100 mg/kg administered for 10 days to 4 weeks [8, 9]. Although previous studies have confirmed the blood‐brain barrier (BBB) penetration of abemaciclib, we first performed western blot analysis of retinoblastoma protein (Rb), the primary target of abemaciclib, and its phosphorylated form to ensure penetration and efficacy of abemaciclib in our experimental model. As expected, abemaciclib treatment substantially reduced phosphorylated Rb (pRb) levels in the brain, confirming BBB penetration of abemaciclib (Figure S1A).
Abemaciclib treatment ameliorates cognitive deficits and prevents neuronal cell death in AppNL−F / MAPT knock‐In mouse (A) Experimental design showing PBS or abemaciclib (50 mg/kg) daily oral administration to 12‐month‐old AppNL−F/MAPT knock‐in mice for 1 month. (B) Schematic illustration of Y‐maze test used to assess working memory. (C) Spontaneous alternation percentage in Y‐maze test from WT and AppNL−F/MAPT double knock‐in (dKI) mice treated with saline (Sal) or abemaciclib (Abe, 50 mg/kg daily oral administration) for 4 weeks. ** p<0.01, * p<0.05. (D) Total entry number during Y‐maze test. n.s., not significant. (E) Representative swim path traces in the Morris water maze test. (F) Average swim velocity. n.s., not significant. (G) Escape latency to platform during 5‐day training in the Morris water maze test. ** p<0.01. (H) Number of times the platform was reached during learning sessions. *** p<0.001, ** p<0.01. (I) Percentage of time spent in the target quadrant during probe trial. * p<0.05. (J) Representative images of NeuN immunostaining in CA1 region of hippocampus. Scale bar: 50 µm. (K) Quantification of NeuN‐positive area as a percentage of control. *** p<0.001, **** p<0.0001. (L) Representative images of cleaved caspase‐3 (red) with DAPI (blue) immunostaining in CA1 region. Scale bar: 50 µm. (M) Quantification of cleaved caspase‐3 positive area. * p<0.05, ** p<0.01. All analyses were performed with 9–10 mice per group. Data are presented as mean ± s.e.m. Statistical analysis was performed using one‐way ANOVA followed by Tukey's test.
Working memory, assessed by the Y‐maze spontaneous alternation test, was significantly reduced in vehicle‐treated dKI mice compared to wild‐type (WT) littermates. Abemaciclib treatment significantly improved alternation performance in dKI mice to levels comparable with WT controls (Figure 1B,C). The total number of arm entries remained consistent across all experimental groups, indicating that the observed cognitive effect was not confounded by locomotor activity (Figure 1D). We further assessed spatial learning and memory using the Morris water maze test. Vehicle‐treated dKI mice exhibited persistent deficits across the five‐day training period, with elevated escape latencies compared with WT controls (Figure 1G). In contrast, abemaciclib treatment significantly improved learning performance, as evidenced by reduced escape latencies, increased platform crossing frequency, and greater time spent in the target quadrant (Figure 1E–I). Swimming velocity remained comparable across all experimental groups (Figure 1F).
Given that cognitive deficits in AD correlate with synaptic dysfunction and neuronal loss, we examined whether abemaciclib affected neurodegeneration in the hippocampus, a region critically involved in learning and memory. Immunohistochemical analysis revealed a significant reduction in NeuN‐positive neurons in the CA1 region of vehicle‐treated dKI mice compared to WT controls (Figure 1J). Abemaciclib treatment significantly preserved neuronal density in dKI mice (Figure 1K). Consistent with this neuroprotective effect, abemaciclib reduced levels of cleaved caspase‐3, a canonical marker of apoptotic cell death, in the hippocampus of dKI mice (Figure 1L,M). Together, these findings demonstrate that abemaciclib effectively ameliorates cognitive deficits and attenuates neurodegeneration in a physiologically relevant AD mouse model that spontaneously develops age‐dependent pathology. This represents a significant advance over previous studies that utilized transgenic models with non‐physiological protein overexpression, thereby strengthening the translational potential of abemaciclib as a disease‐modifying therapeutic for AD.
Abemaciclib Selectively Reduces Tau Pathology but Not Amyloid Burden or Glial Activation
2.2
To elucidate the molecular mechanisms underlying the cognitive improvements and neuroprotection observed with abemaciclib treatment, we investigated its effects on the primary pathological features of AD, tau phosphorylation, and amyloid‐β (Aβ) deposition. Immunohistochemical analysis of hippocampal sections revealed abundant AT8‐positive (phospho‐Ser202/Thr205) tau in vehicle‐treated dKI mice, whereas abemaciclib administration significantly reduced AT8 intensity (Figure 2A,B). This effect was also observed in cortical regions, indicating a global reduction in tau phosphorylation throughout the brain (Figure S1B). Given that the dKI model does not develop robust neurofibrillary tangles but rather exhibits progressive age‐dependent tau accumulation, we employed biochemical fractionation to assess distinct pathological tau species. We isolated sarkosyl‐insoluble and sarkosyl‐soluble tau fractions, which represent aggregation‐prone and soluble tau pools, respectively. Western blot analysis demonstrated that abemaciclib significantly reduced both AT8 and phospho‐Thr231 (pT231) in the sarkosyl‐insoluble fraction (Figure 2C), indicating a reduction in the pathologically relevant tau species. This finding is particularly significant as sarkosyl‐insoluble phosphorylated tau represents an early marker of pathological tau accumulation that precedes the formation of mature neurofibrillary tangles. Similarly, abemaciclib decreased AT8 and pT231 levels in the sarkosyl‐soluble fraction (Figure 2D). Notably, total tau levels detected with Tau13 remained unchanged in both fractions (Figure S1C,D), indicating that abemaciclib specifically affected tau phosphorylation rather than tau expression.
Tau pathology is reduced by abemaciclib in AppNL−F / MAPT knock‐In mouse (A) Representative immunofluorescence images of AT8 (green) and DAPI (blue) staining in the hippocampus of vehicle (Veh) and abemaciclib (Abe) treated mice. Scale bar: 20 µm. (B) Quantification of AT8 intensity in hippocampu. ** p<0.01. (C) Western blot analysis of sarkosyl‐insoluble tau fractions from hippocampal homogenates probed with AT8 (phospho‐Ser202/Thr205), pT231 (phospho‐Thr231), and Tau13 (total tau) antibodies, with corresponding quantification. * p<0.05. (D) Western blot analysis of sarkosyl‐soluble tau fractions with corresponding quantification. ** p<0.01, * p<0.05. (E) Pearson correlation analysis between sarkosyl‐insoluble AT8 levels and time spent in the target quadrant during Morris water maze probe trial. R2 = 0.3469, * p<0.0101. (F) Pearson correlation analysis between sarkosyl‐insoluble pT231 levels and time spent in the target quadrant. R2 = 0.3251, * p<0.0135. All analyses were performed with 9–10 mice per group. Data are presented as mean ± s.e.m. Statistical analysis was performed using an unpaired t‐test.
To establish the functional relevance of these biochemical changes, we performed correlation analyses between sarkosyl‐insoluble phosphorylated tau levels and cognitive performance. Notably, both sarkosyl‐insoluble AT8 and pT231 showed significant inverse correlations with the time spent in the target quadrant during the Morris water maze probe trial (Figure 2E,F). These correlations demonstrate that the reduction in pathological tau species directly corresponds to improved spatial memory performance, establishing a mechanistic link between abemaciclib's effects on tau pathology and its therapeutic benefits on cognitive function.
In contrast to previous reports in transgenic mouse models [10]. Abemaciclib treatment did not significantly alter Aβ plaque burden in dKI mice, as assessed by 6E10 immunostaining (Figure S2A–C). To determine whether this lack of effect on amyloid pathology was model‐specific, we examined the effects of abemaciclib in the ADLP^APT^ mouse model, which expresses both amyloid and tau pathologies through overexpression of mutant APP/PS1 and P301L tau [16]. Consistent with our findings in dKI mice, abemaciclib did not reduce Aβ plaque burden in this alternative AD model (Figure S2D). This divergence from prior findings in 5xFAD mice suggests that the cognitive and neuroprotective benefits of abemaciclib observed in our physiologically relevant models are primarily mediated through modulation of tau pathology rather than amyloid processing.
Given the established link between neuroinflammation and neurodegeneration in AD, we also examined whether abemaciclib affected glial activation. Immunohistochemical analysis revealed comparable levels of GFAP and Iba1, markers of astrocytes and microglia, respectively, in vehicle and abemaciclib‐treated dKI mice, indicating that the effects of abemaciclib were not mediated through modulation of glial activation in this model (Figure S3). These results demonstrate that abemaciclib selectively reduces tau phosphorylation without affecting Aβ deposition or glial activation in the dKI mouse model, and that this selective effect on tau pathology directly correlates with cognitive improvements.
Abemaciclib Attenuates Tau Phosphorylation in AD Patient‐Derived Brain Organoids
2.3
While our in vivo data provided compelling evidence for abemaciclib's efficacy in mitigating tau pathology, we next sought to validate these findings in a human‐relevant model. To this end, we utilized brain organoids derived from sporadic AD patient‐induced pluripotent stem cells (iPSCs), which recapitulate key aspects of human brain development and AD pathology in a 3D environment. Following established protocols, iPSCs from sporadic AD patients were differentiated into brain organoids and maintained for 70 days to allow for maturation and development of AD‐related pathologies (Figure 3A). To determine the optimal concentration for investigating abemaciclib's effects on tau phosphorylation in our brain organoid system, we first performed a dose‐response experiment. Abemaciclib demonstrated a dose‐dependent reduction in phosphorylated tau levels, with significant effects observed at concentrations as low as 500 nm (Figure S4A). Cell viability remained unaffected at concentrations up to 10 µm, with modest reductions in viability observed only at the highest tested concentration of 20 µm (Figure S4B). Based on these findings, we selected 1 µm as our working concentration for subsequent experiments, as it demonstrated robust efficacy while maintaining excellent cell viability. At day 70, organoids were treated with abemaciclib or vehicle for 24 h, and the effects on tau phosphorylation were assessed. To confirm target engagement, we evaluated Rb phosphorylation, which was markedly reduced in abemaciclib‐treated organoids, validating effective target inhibition in the organoid model (Figure S4C).
Tau pathology is diminished by abemaciclib in AD patient‐derived brain organoids (A) Schematic illustration of the generation of brain organoids from sporadic AD patient iPSCs, showing the timeline and media conditions for each developmental stage. (B) Representative immunofluorescence images of pT181 (phospho‐Thr181, green) and DAPI (blue) staining in brain organoids treated with vehicle (Veh) or abemaciclib (Abe, 1 µm) for 24 h. Scale bar: 50 µm. (C) Quantification of pT181 intensity as percentage of area. * p<0.05. n = 5 for each group. (D) Levels of pT181 detected by ELISA. **** p<0.0001. n = 12 for each group. (E) Representative immunoblots of pT181, AT8 (phospho‐Ser202/Thr205), pS396 (phospho‐Ser396), T22 (oligomeric tau), Tau13 (total tau), and β‐actin in brain organoids. (F) Quantification of relative pT181/Tau13 ratio. ** p<0.01. n = 10 for each group. (G) Quantification of relative AT8/Tau13 ratio. ** p<0.01. n = 10 for each group. (H) Quantification of relative pS396/Tau13 ratio. ** p<0.01. n = 10 for each group. (I) Quantification of relative T22/Tau13 ratio. ** p<0.01. Data are presented as mean ± s.e.m. Statistical analysis was performed using a two‐tailed unpaired t‐test.
Immunofluorescence analysis revealed abundant phosphorylated tau in vehicle‐treated organoids, predominantly localized in neuronal processes (Figure 3B). Consistent with the dose‐response data, treatment with abemaciclib significantly reduced pT181 intensity (Figure 3C). Quantitative ELISA further confirmed this observation, demonstrating a significant decrease in pT181 levels in abemaciclib‐treated organoids compared to vehicle‐treated controls (Figure 3D). To further characterize the effect of abemaciclib on tau pathology, we examined multiple phosphorylation sites associated with different stages of AD progression and distinct pathogenic tau species. Phosphorylation at Thr181 (pT181) is considered an early marker of tau pathology, while phosphorylation at Ser202/Thr205 (AT8) is associated with more advanced stages of tau pathology [17]. Abemaciclib treatment significantly reduced tau phosphorylation at both sites, suggesting its potential efficacy across different stages of tau pathology (Figure 3E–G). Moreover, western blot analysis demonstrated reduced phosphorylation at Ser396 (pS396), a site associated with paired helical filaments (PHF), and T22‐positive tau oligomers (Figure 3H,I). The reduction of these pathogenic tau species is particularly significant, as these species are increasingly recognized as key pathogenic drivers of neurodegeneration. Notably, total tau levels measured by Tau13 remained unchanged, confirming that abemaciclib specifically affects tau phosphorylation rather than overall tau expression.
To evaluate potential toxicity and delayed cytotoxicity associated with long‐term use of abemaciclib, we extended the treatment period to four weeks. Cell viability assays demonstrated that abemaciclib treatment maintained organoid viability at both 2 weeks and 4 weeks (Figure S4E,F). Moreover, western blot analysis of caspase‐3 revealed insignificant changes in apoptotic cell death after 2 weeks of treatment (Figure S4D). These long‐term treatment data establish that abemaciclib's therapeutic effects on tau pathology can persist without compromising organoid viability or inducing cellular stress responses, even with prolonged exposure. Collectively, these findings in human AD patient‐derived brain organoids are consistent with the observations in the dKI mouse model, indicating that abemaciclib's potential to reduce tau phosphorylation is conserved across species and models. Moreover, the validation in patient‐derived organoids represents a significant advance over previous studies that relied solely on mouse models, further highlighting the robust translational relevance of our findings.
Selective Inhibition of CaMKII and GSK3β by Abemaciclib Reduces Pathological Tau Phosphorylation
2.4
To elucidate the molecular mechanisms underlying the effects of abemaciclib on tau phosphorylation, we investigated its impact on key tau kinases. Given that tau phosphorylation in AD is mediated by several kinases, including GSK3β, CaMKII, cyclin‐dependent kinase 5 (CDK5), and extracellular signal‐regulated kinases (ERK1/2), we examined whether abemaciclib modulated the activity of these kinases [18]. In AD patient‐derived brain organoids, abemaciclib treatment induced a dose‐dependent reduction in the phosphorylation of CaMKII, with significant inhibition observed at concentrations as low as 100 nm and almost complete suppression at 10 µm (Figure S5A,B). Similarly, abemaciclib dose‐dependently reduced GSK3β phosphorylation, with significant effects observed from 500 nm (Figure S5C). In contrast, CDK5 and ERK expression levels remained unchanged across all tested concentrations in organoids, indicating a selective effect on specific tau kinases (Figure S5D,E).
To investigate whether abemaciclib directly interacts with CaMKII, we performed molecular docking analysis. Molecular docking analysis revealed that abemaciclib binds to CaMKII with a binding affinity of −10.0 kcal/mol (Figure 4A). The interaction was stabilized through π–π stacking with Phe89, hydrogen bonds with Val92 and Lys42, and ionic/hydrogen bonding with Glu99 (Figure 4B). Surface plasmon resonance (SPR) analysis further validated this prediction, demonstrating concentration‐dependent binding of abemaciclib to CaMKII, with a dissociation constant (KD) of 38.8 ± 14.8 µm (Figure 4C,D).
Abemaciclib directly binds to and inhibits CaMKII and GSK3β (A) Molecular docking image of abemaciclib and CaMKII. (B) Schematic illustration of key binding interactions between abemaciclib and CaMKII, showing π–π stacking, hydrogen bonds, and ionic interactions with specific residues. (C) Surface plasmon resonance (SPR) sensorgrams showing concentration‐dependent binding of abemaciclib to recombinant CaMKII protein at indicated concentrations. (D) Quantification of SPR binding response showing dissociation constant (KD) of 38.8 ± 14.8 µm, Rmax of 20.4 RU, and Chi2 of 0.748. RU, response units. (E) Molecular docking image of abemaciclib and GSK3β. (F) Schematic illustration of key binding interactions between abemaciclib and GSK3β. (G) SPR sensorgrams showing concentration‐dependent binding of abemaciclib to recombinant GSK3β protein. (H) Quantification of SPR binding response showing KD of 1.706 ± 0.005 m, Rmax of 30.6 RU, and Chi2 of 0.345. (I) Schematic illustration of the proposed mechanism where abemaciclib inhibits CaMKII phosphorylation, which can be activated by ionomycin treatment. (J) Representative immunoblots of AT100 (phospho‐Thr212/Ser214 tau), Tau13 (total tau), and β‐actin in vehicle‐treated, abemaciclib‐treated, or abemaciclib plus ionomycin‐treated organoids. (K) Quantification of relative AT100/Tau13 ratio. ** p<0.01, *** p<0.001. n = 6 for each group. All data are presented as mean ± s.e.m. Statistical analysis was performed using one‐way ANOVA followed by Tukey's test.
To confirm whether direct binding of abemaciclib to CaMKII results in functional kinase inhibition, we examined phosphorylation of GluA1 at Ser831, an established CaMKII substrate. Along with the dose‐dependent decrease in phosphorylation of CaMKII, abemaciclib significantly reduced pGluA1 levels (Figure S5A,B,F,G), demonstrating effective inhibition of CaMKII activity toward its endogenous substrate. To validate these findings in vivo, we examined the effects of abemaciclib on CaMKII activity in the dKI mouse model. Western blot analysis of brain homogenates revealed a significant reduction in pCaMKII levels, while other tau kinases remained unchanged (Figure S6A–E). Consistently, immunohistochemical analysis showed markedly decreased pCaMKII expression in the hippocampus of abemaciclib‐treated dKI mice compared to vehicle‐treated controls (Figure S6F,G).
To further establish the role of CaMKII in mediating abemaciclib's effects on tau phosphorylation, we performed a rescue experiment using ionomycin, a calcium ionophore that activates CaMKII. Co‐treatment of abemaciclib with ionomycin significantly reversed abemaciclib's inhibitory effect on tau phosphorylation, as assessed by western blot using the AT100 antibody, which detects tau phosphorylated at Thr212/Ser214, a phosphorylation site directly regulated by CaMKII (Figure 4I–K). Conversely, treatment with KN93, a selective CaMKII inhibitor, recapitulated abemaciclib's effects by reducing phosphorylation at pT181, AT8, and AT100, with no additive effects upon co‐treatment (Figure S5H–J).
In addition to CaMKII, abemaciclib dose‐dependently reduced GSK3β phosphorylation in AD patient‐derived brain organoids, with significant effects observed from 500 nm (Figure S5C). Notably, molecular docking analysis revealed that abemaciclib can also directly binds to GSK3β with a binding affinity of −7.6 kcal/mol through π–π stacking interactions with Tyr134 and a hydrogen bond with Cys199 (Figure 4E–F). SPR analysis further demonstrated concentration‐dependent binding of abemaciclib to GSK3β, with a KD of 1.706 ± 0.005 m, establishing that abemaciclib can directly engage multiple tau kinases simultaneously (Figure 4G,H). To verify whether the direct binding of abemaciclib to GSK3β translates to functional kinase inhibition, we examined the phosphorylation of GSK3β and its well‐established substrate, β‐catenin. Western blot analysis revealed increased total β‐catenin levels, confirming effective GSK3β inhibition (Figure S5F,G). Collectively, these findings establish that abemaciclib directly binds and inhibits CaMKII and GSK3β, two major tau kinases implicated in AD pathogenesis, reducing pathological tau phosphorylation in both mouse models and human brain organoids.
Abemaciclib Promotes Autophagic Degradation of Pathological Tau Proteins
2.5
While our previous results demonstrated that abemaciclib effectively reduces tau phosphorylation via kinase inhibition, we further hypothesized that abemaciclib might enhance the clearance of pathological tau through autophagy, a primary degradation pathway for misfolded proteins that is frequently compromised in neurodegenerative diseases. Notably, previous studies have suggested that abemaciclib can modulate autophagic processes in cancer cells [19]. Therefore, we investigated whether abemaciclib could activate autophagy to facilitate abnormal tau degradation under AD‐relevant pathological conditions.
In AD patient‐derived brain organoids, abemaciclib treatment reduced levels of p62/SQSTM1, a selective autophagy cargo receptor that is typically degraded during successful autophagic clearance, while concurrently increasing LC3‐II expression, suggesting enhanced autophagosome formation (Figure 5A–C). However, the accumulation of autophagosomes alone could indicate either enhanced autophagy initiation or impaired autophagosome‐lysosome fusion and degradation, the latter representing a dysfunctional state. To distinguish between these possibilities and establish whether abemaciclib induces functional autophagic flux, we conducted an autophagic flux assay utilizing bafilomycin A1 (Baf), a vacuolar H⁺‐ATPase inhibitor that prevents lysosomal acidification and autophagosome‐lysosome fusion. In this assay, if autophagic flux is enhanced, bafilomycin treatment should result in greater LC3‐II accumulation in abemaciclib‐treated samples compared to controls, reflecting increased autophagosome formation that is subsequently blocked from degradation. Indeed, western blot analysis demonstrated that bafilomycin treatment led to significantly greater LC3‐II accumulation in abemaciclib‐treated organoids compared to vehicle‐treated controls (Figure 5D,E). Quantification of the LC3‐II ratio between the bafilomycin‐treated to untreated groups further confirmed that abemaciclib significantly enhanced autophagic flux, indicating that the observed increase in autophagosome formation represents activation of the autophagic pathway rather than a blockade in autophagosome clearance (Figure 5F).
Abemaciclib enhances autophagic flux for pathological tau clearance (A) Representative immunoblots of p62, LC3, and β‐actin in vehicle or abemaciclib‐treated brain organoids. (B) Quantification of p62/β‐actin ratio. * p<0.05. n = 10 for each group. (C) Quantification of LC3‐II/LC3‐I ratio. ** p<0.01. n = 10 for each group. (D) Representative immunoblots of LC3 and β‐actin in brain organoids treated with vehicle (Veh), bafilomycin A1 (Baf, 5 nm), abemaciclib (Abe, 1 µm), or both bafilomycin A1 and abemaciclib for 24 h. (E) Quantification of LC3‐II/LC3‐I ratio in the indicated treatment groups. ** p<0.01, *** p<0.001, **** p<0.0001. n = 3 for each group. (F) Quantification of autophagic flux (calculated as the difference between LC3‐II levels with and without bafilomycin A1). * p<0.05. n = 4 for each group. (G) Representative immunoblots of AT8, LC3, and β‐actin in brain organoids treated with vehicle, abemaciclib, or abemaciclib plus bafilomycin A1. (H) Quantification of AT8/β‐actin ratio. * p<0.05. n = 4 for each group. (I) Representative fluorescence images of MAP2, lysotracker, and their 3D rendering with enlarged regions in iPSC‐derived neurons treated with vehicle or abemaciclib (1 µm) for 24 h. Scale bar: 10 µm. (J) Quantification of lysotracker intensity in MAP2‐positive neurons in arbitrary units. **** p<0.0001. n = 8 for each group. All data are presented as mean ± s.e.m. Statistical analysis was performed using a two‐tailed unpaired t‐test or one‐way ANOVA followed by Tukey's test where appropriate.
To elucidate whether enhanced autophagy is functionally necessary for abemaciclib‐mediated tau clearance, we blocked autophagy with bafilomycin before abemacicilb treatment and assessed changes in tau levels. While abemaciclib treatment alone significantly reduced AT8‐positive tau levels, bafilomycin abolished this effect, restoring tau phosphorylation to levels comparable to vehicle‐treated controls (Figure 5I,J). To further validate the effects of abemaciclib on autophagy, we examined lysosomal functio n using LysoTracker staining, which selectively labels acidic organelles. 3D rendering of confocal images stained for neurons with MAP2 and Lysotracker revealed significantly increased LysoTracker intensity in abemaciclib‐treated neurons compared to controls (Figure 5G,H). This enhancement of acidified, functionally active lysosomes provides additional evidence that abemaciclib treatment promotes the complete autophagic process, including the degradative phase that is critical for effective clearance of pathological tau proteins. On the other hand, expression of autophagy‐related genes in organoids remained consistent over abemaciclib treatment (Figure S7). Collectively, these findings demonstrate that abemaciclib enhances functional autophagic flux in AD model brain organoids, establishing a dual mechanism of action where abemaciclib not only prevents the formation of newly phosphorylated tau via kinase inhibition but also facilitates the clearance of existing pathological tau through enhanced autophagy‐lysosomal degradation pathways.
Knockdown of CDK4 or CDK6 Does Not Affect Tau Pathology
2.6
Given that abemaciclib is primarily characterized as a CDK4/6 inhibitor, we sought to determine whether its observed effects on tau phosphorylation were directly mediated through CDK4 or CDK6 inhibition. As CDK4/6 knockdown during organoid development could compromise organoid formation due to their essential roles in cell cycle regulation, we employed mature iPSC‐derived neurons to assess whether the effects of abemaciclib on tau pathology are mediated through CDK4/6. Lentiviral knockdown of CDK4 or CDK6 was validated by mRNA and protein expression analysis, showing successful reductions in CDK4 and CDK6 levels (Figure 6A–C).
Abemaciclib modulates tau phosphorylation through CDK4/6‐independent mechanisms (A) Relative CDK4 mRNA expression level in negative control (NC) or CDK4 knockdown (KD) induced pluripotent stem cell (iPSC)‐derived neurons. *** p<0.001. n = 4 for each group. (B) Relative CDK6 mRNA expression level in NC or CDK6 KD iPSC‐derived neurons. *** p<0.001. n = 4 for each group. (C) Protein levels of CDK4 and CDK6 in NC or knockdown iPSC‐derived neurons, with β‐actin as loading control. (D) Representative immunoblots of AT8 (phospho‐Ser202/Thr205), Tau13 (total tau), and β‐actin in lentivirus‐treated iPSC‐derived neurons treated with vehicle (Veh) or abemaciclib (Abe, 1 µm). (E) Quantification of AT8/Tau13 ratio. * p<0.05, ** p<0.01, *** p<0.001. n = 6 for each group. (F) Representative immunoblots of pCaMKII, CaMKII, β‐actin, pGSK3β, and GSK3β in iPSC‐derived neurons subjected to the indicated treatments. (G) Quantification of pCaMKII/CaMKII ratio. *** p<0.001, **** p<0.0001. n = 5 for each group. (H) Quantification of pGSK3β/GSK3β ratio. * p<0.05, ** p<0.01. n = 5 for each group. (I) Representative immunoblots of LC3 and β‐actin. (J) Quantification of LC3/β‐actin ratio. ** p<0.01. n = 5 for each group. All data are presented as mean ± s.e.m. Statistical analysis was performed using a two‐tailed unpaired t‐test for A and B; one‐way ANOVA followed by Tukey's test for E, G, H, and J.
Despite the substantial decrease in CDK4 or CDK6 levels, we observed no significant alteration in phosphorylated tau levels in either knockdown condition, suggesting that CDK4/6 inhibition alone is insufficient to recapitulate the effects of abemaciclib on tau pathology (Figure 6D,E). To further elucidate the molecular mechanisms underlying abemaciclib's tau‐modulation, we examined whether CDK4 or CDK6 knockdown could mimic abemaciclib's previously observed effects on key tau kinases that we had identified as targets of abemaciclib in our earlier experiments. Notably, CDK4/6 knockdown failed to alter the phosphorylation states of CaMKII or GSK3β (Figure 6F–H). In contrast, abemaciclib treatment significantly reduced CaMKII and GSK3β phosphorylation, consistent with our previous findings in AD mouse models and patient‐derived brain organoids. Moreover, as our earlier observation showed that abemaciclib enhances functional autophagic flux, we next assessed autophagy markers following CDK4 or CDK6 knockdown. While abemaciclib treatment significantly upregulated the autophagosome marker LC3‐II, neither of the knockdowns showed effects on the autophagic marker (Figure 6I,J). These findings indicate that abemaciclib's effects on tau phosphorylation are mediated through CDK4 and CDK6‐independent mechanisms, revealing novel off‐target benefits that fundamentally distinguish its neuroprotective mode of action from its established anti‐cancer mechanism, which primarily functions through inhibition of CDK4/6‐mediated cell cycle arrest.
Collectively, these data elucidate a dual mechanism of action whereby abemaciclib ameliorates tau pathology through inhibition of key tau kinases, primarily CaMKII and GSK3β, and enhancement of autophagic clearance of pathological tau proteins (Figure 7). The inhibition of tau kinases prevents the formation of newly phosphorylated tau, while the concurrent enhancement of autophagy‐lysosomal degradation pathways facilitates the clearance of existing pathological tau proteins. Importantly, these neuroprotective effects occur independently of CDK4 and CDK6 inhibition, suggesting an additional therapeutic application that extends beyond its canonical role as an anti‐cancer agent. This mechanistic profile differentiates abemaciclib from both traditional tau kinase inhibitors and other investigational AD therapeutics that target single pathogenic mechanisms, underscoring its unique potential as a multifunctional disease‐modifying agent for AD and related tauopathies.
Graphic summary Schematic illustration comparing the established mechanism of abemaciclib in breast cancer with its repurposed actions in Alzheimer's disease, highlighting modulation of autophagy and tau‐related kinase pathways.
Discussion and Conclusion
3
In this study, we demonstrated the novel disease‐modifying potential of abemaciclib in AD through a comprehensive multi‐model approach utilizing physiologically relevant models. Our findings revealed that abemaciclib selectively attenuates tau pathology via dual mechanisms, inhibition of key tau kinases, primarily CaMKII and GSK3β, and enhancement of autophagic clearance of pathological tau proteins. Importantly, these effects occured independently of CDK4 and CDK6 inhibition, establishing a unique therapeutic profile that distinguishes abemaciclib's neuroprotective mode of action from its established anti‐cancer mechanism.
We employed the App^NL−F^/MAPT double knock‐in (dKI) mouse model, which represents a significant advancement over previous studies using transgenic models with non‐physiological overexpression of pathogenic proteins. Unlike traditional transgenic models, the dKI mice express human APP and tau at physiological levels and exhibit spontaneous, age‐dependent development of both amyloid and tau pathologies, thus providing a more human‐relevant model of the progressive nature of AD pathogenesis [11]. Our findings in this physiologically relevant model differ from previous research utilizing separate transgenic models for each pathological aspect of AD [10]. While the previous study reported that abemaciclib reduced Aβ plaque burden in 5xFAD mice via modulation of neprilysin and ADAM17, we observed no significant effect on amyloid deposition in either dKI or the ADLP^APT^ double transgenic model. This discrepancy likely reflects model‐specific differences and the distinct administration protocols. Our study administered abemaciclib orally at 50 mg/kg for four weeks, whereas the previous work used intraperitoneal administration at 30 mg/kg. Similarly, while they demonstrated anti‐inflammatory effects in lipopolysaccharide‐treated wild‐type mice, we observed no significant alteration in glial activation in our model.
Importantly, both studies consistently confirmed abemaciclib's robust ability to reduce tau phosphorylation across all models tested, establishing tau as the primary and most reliable therapeutic target of abemaciclib. This consistency is particularly significant given the strong correlation between tau pathology and cognitive decline in AD patients, as supported by our correlation analyses demonstrating that sarkosyl‐insoluble phosphorylated tau levels (both AT8 and pT231) inversely correlate with spatial memory performance. This analysis is consistent with the growing recognition that tau pathology correlates more strongly with cognitive decline in AD than amyloid burden [4]. The absence of reduced glial activation in our model despite cognitive improvements further supports the hypothesis that abemaciclib exerts its beneficial effects through direct neuronal protection rather than modulation of neuroinflammatory processes. Glial cells, particularly microglia, play a crucial role in regulating Aβ levels through phagocytosis [20, 21]. Therefore, it is plausible that the dysfunctional microglia observed in dKI mice could not effectively phagocytose Aβ plaques regardless of drug treatment. Nonetheless, our findings demonstrate that abemaciclib effectively prevented neuronal damage through its robust effects on tau phosphorylation.
A significant advancement of our study is the validation of abemaciclib's effects in human AD patient‐derived brain organoids, which recapitulate key aspects of human brain development and AD pathology in a 3D environment. This human‐relevant model system addresses a critical gap in translational research, providing insights that may accurately predict clinical outcomes compared to animal models alone. Our organoid experiments not only confirmed abemaciclib's ability to reduce tau phosphorylation at multiple epitopes but also demonstrated dramatic reduction of T22‐positive tau oligomers. These soluble tau oligomers are increasingly recognized as highly toxic species that may be more pathogenic than mature neurofibrillary tangles, as they actively disrupt synaptic function and propagate between neurons. While the dKI model does not develop robust neurofibrillary tangles, our biochemical fractionation revealed significant reductions in sarkosyl‐insoluble phosphorylated tau, representing aggregation‐prone species that precede mature tangle formation. Together, these findings demonstrate that abemaciclib effectively targets multiple pathogenic tau species across the aggregation continuum, from early phosphorylation events to toxic oligomeric assemblies. The consistent effects observed across murine and human models reinforce the robustness of our findings and strengthen the case for abemaciclib as a promising therapeutic candidate for AD.
Our mechanistic investigation revealed that abemaciclib directly binds to and selectively inhibits both CaMKII and GSK3β, two major tau kinases implicated in AD pathogenesis. CaMKII is well‐established as a key tau kinase that phosphorylates tau at multiple pathologically relevant sites, including serines 131, 214, 262, 356, and 416, as well as threonines 135 and 212, many of which are phosphorylated in AD brains [18]. The critical role of CaMKII in tau pathology is substantiated by several studies demonstrating that pharmacological inhibition of CaMKII diminishes phosphorylated tau and prevents neurodegeneration [22, 23]. Importantly, Aβ is known to activate CaMKII, which subsequently destabilizes synaptic functions, leading to functional deficits in neurons [24]. Therefore, abemaciclib's inhibition of CaMKII activity provides a mechanistic explanation for its neuroprotective effects despite the persistence of amyloid plaques. By suppressing CaMKII activation, abemaciclib could prevent the downstream neuronal dysfunction and cell death that would otherwise be induced by Aβ, thus effectively uncoupling amyloid pathology from its neurotoxic consequences.
Notably, our docking and SPR analyses also revealed that abemaciclib directly binds to GSK3β, another well‐characterized tau kinase. GSK3β knockdown has been shown to reduce neurofibrillary tangles, tau phosphorylation, and mitigate conformational changes in AD model mice [25]. The inhibitory effects of abemaciclib on multiple tau kinases are consistent with previous multi‐omics profiling data demonstrating that abemaciclib can reduce the activities of several kinases beyond its canonical CDK4/6 targets [26, 27]. Phosphoproteomics data have revealed that CaMKII activity is downregulated in cells treated with abemaciclib, and multiple in vitro assays (KINOMEscan profiling, multiplex inhibitor bead assay, and kinase activity assay) have identified CaMKII, GSK3, Adaptor‐associated kinase 1 (AAK1), and Cdc‐like kinase (CLK) as secondary targets of abemaciclib.
The dual mechanism of action we have elucidated, simultaneous inhibition of tau kinases and enhancement of autophagy, positions abemaciclib as a multifunctional therapeutic agent with unique advantages over compounds targeting single mechanisms. By both preventing the formation of newly phosphorylated tau and facilitating the clearance of existing pathological tau proteins, abemaciclib addresses tau pathology through complementary approaches, offering more comprehensive therapeutic benefits. This mechanistic profile differentiates abemaciclib from both traditional tau kinase inhibitors and other investigational AD therapeutics.
Several limitations of this study warrant consideration. First, our in vivo design employed group‐level comparisons rather than within‐animal pre/post‐treatment assessments. While our experimental design demonstrates that abemaciclib‐treated animals performed significantly better than vehicle controls, longitudinal within‐subject designs in future studies would provide stronger evidence for drug‐induced cognitive recovery and account for baseline pathology variability among individual animals. Second, the four‐week treatment in dKI mice does not address long‐term safety over extended periods relevant to chronic AD treatment. Future studies should evaluate prolonged administration to assess potential adverse effects and determine whether therapeutic benefits are maintained or enhanced with longer treatment durations. Third, while we demonstrated dose‐dependent effects in organoids, the animal model was tested on a single dose. Systematic dose‐response studies in vivo would help establish optimal therapeutic windows and determine whether lower doses might achieve similar efficacy with improved safety profiles. Lastly, although our study establishes the importance of CaMKII and GSK3β inhibition and autophagy enhancement, the relative contributions of these mechanisms to overall therapeutic efficacy remain to be quantified.
As an FDA‐approved drug with established safety profiles in cancer patients, abemaciclib can be evaluated for clinical efficacy in AD more rapidly than novel compounds, significantly reducing the time and costs associated with traditional drug development [5, 28]. Importantly, abemaciclib demonstrates superior blood‐brain barrier penetration [9], a critical property for CNS‐targeting therapeutics. The real‐world data of abemaciclib in cancer patients provides valuable insights into its long‐term safety profile, potential adverse effects, and optimal dosing strategies, further facilitating its potential translation to AD clinical trials.
Methods
4
Animal
4.1
All animal experiments were approved by the Animal Care and Use Guidelines of Seoul National University, Korea (approval number: SNU‐220620‐4). Mice were maintained in a pathogen‐free facility, caged in groups of five, on a 12‐h light/dark cycle with ad libitum access to food and water. Only female mice were used in all experiments. App^NL−F^/MAPT double knock‐in (dKI) mice were bred by crossing App^NL−F^ knock‐in mice that carry APP KM670/671NL (Swedish), APP I716F (Iberian) [29], and MAPT knock‐in mouse that replaces murine tau with human tau [11]. For drug administration, 12‐month‐old wild‐type (WT) or dKI mice were orally administered with phosphate‐buffered Saline (PBS) as vehicle or abemaciclib methanesulfonate (MedChemExpress, trade name Verzenio) at 50 mg/kg daily for four weeks. After four weeks, mice were sacrificed, perfused with PBS, and brain samples were collected. One hemisphere was used for protein extraction, and the other for immunohistochemistry.
Behavioral Tests
4.2
Y‐maze Spontaneous Alternation Test
4.2.1
Mice were introduced to the center of a Y‐shaped maze comprised of three white and opaque plastic arms at a 120° angle from each other. The mice were allowed to explore the arms for 8 min while the video was recorded. The total number of entries and the number of alternations were counted.
Morris Water Maze
4.2.2
The Morris water maze test was conducted in a circular tank (120 cm diameter) filled with water (23°C) rendered opaque with non‐toxic white paint. A hidden platform was submerged 1 cm below the water surface in one quadrant, positioned 30 cm from the tank wall. During the five‐day training phase, mice performed four consecutive trials daily, with 60‐s maximum search time per trial and 15‐s platform retention followed by 1‐min inter‐trial intervals. Starting positions (north, south, east, west) varied between trials. If mice failed to locate the platform within the allotted time, it was guided to the platform and marked as 60 s. On the probe trial day, the platform was removed, and mice were placed in the center of the tank for a 60‐s free swim. Performance was recorded and analyzed using EthoVisionXT software (Noldus Information Technology), measuring escape latency, swimming velocity, platform crossings, and time spent in the target quadrant.
Generation of Human Brain Organoids Derived from hiPSCs
4.3
Brain organoids were derived from Alzheimer's disease patient‐induced pluripotent stem cells (hiPSCs, Camden CW50039) following established protocols [6, 15]. Briefly, hiPSCs were dissociated using ReLeSR (STEMCELL Technologies) and centrifuged to form embryoid bodies in EB formation medium containing Y‐27632 (STEMCELL Technologies) using AggreWell800 plates (STEMCELL Technologies). On the next day, day 1, the medium was replaced with EB formation medium without Y‐27632. From day 2 to day 5, the medium was replaced daily with DMEM/F‐12 supplemented with GlutaMAX (Gibco), 20% KnockOut Serum Replacement (Gibco), 1% MEM Non‐Essential Amino Acids Solution (Gibco), 0.1 mm 2‐mercaptoethanol (Gibco), 100 U/ml penicillin and 100 µg/ml streptomycin (Merck), and SMAD inhibitors, 10 µm dorsomorphin (Merck) and 10 µm SB‐431542 (TOCRIS). On day 7, individual embryoid bodies were transferred to 96‐well ultra‐low‐attachment microplates (Corning) and maintained in neural differentiation medium (Neurobasal‐A Medium; Gibco) with B‐27 supplement (minus vitamin A; Gibco), P/S, GlutaMAX, and 0.5% Matrigel (Corning) supplemented with 20 ng/ml epidermal growth factor (EGF; Merck) and 20 ng/ml basic fibroblast growth factor (bFGF; R&D Systems) until day 24, with daily medium changes until day 15 and every other day thereafter. From day 25 to 42, organoids were cultured in neurobasal medium supplemented with 20 ng/ml brain‐derived neurotrophic factor (BDNF; Peprotech) and 20 ng/ml neurotrophin‐3 (NT‐3; Peprotech) with medium changes every two days. After day 43, organoids were maintained in neurobasal medium without growth factors until day 70, with medium changes every four days. For drug treatment, organoids were treated with abemaciclib for 24 h.
Generation of iPSC‐Derived Neurons (iN)
4.4
iPSC‐derived neurons (iN) were generated following the previosly published protocol [30]. Briefly, hiPSCs were transduced with TetO‐Ngn2‐GFP‐Puro and rtTA lentiviruses and plated on Geltrex‐coated plates (A1413302, Thermo) at 75,000 cells/cm^2^ in mTeSR Plus supplemented with Y‐27632. On day 0, the medium was replaced with DMEM/F12 + GlutaMAX containing 1.5% (v/v) 20% glucose, 1% N2, SB431542 (10 µm), XAV939 (2 µm), LDN‐193189 (200 nm), and doxycycline (2 µg/mL). On day 1, cells were maintained in the same base medium with 1% N2, SB431542 (5 µm), XAV939 (1 µm), LDN‐193189 (100 nm), and puromycin (5 µg/mL). On day 2, cells were dissociated with Accutase in the presence of Y‐27632, replated at 125,000 cells/cm^2^ on Geltrex, and cultured in NPC maintenance medium (DMEM/F12 + GlutaMAX with penicillin/streptomycin, 1% NEAA, B‐27 minus vitamin A, N‐2, EGF 10 ng/mL, bFGF 10 ng/mL, puromycin, and Y‐27632; 3 mL per well of a 6‐well plate). On day 3, puromycin and Y‐27632 were omitted, and NPCs were passaged every 5–6 days. For neuronal differentiation, NPCs were dissociated, replated on Matrigel hESC‐qualified matrix–coated dishes, and maintained in Neurobasal‐A with B‐27 minus vitamin A and penicillin/streptomycin, supplemented with BDNF (10 ng/mL) and NT‐3 (10 ng/mL) for 14 days.
Lentivirus Production
4.5
To produce lentiviral particles, 293FT cells were seeded in 10 cm dishes at a density of 3.5 × 10⁶ cells per dish one day prior to transfection. The cells were cultured overnight at 37°C and 5% CO_2_ in a growth medium consisting of Dulbecco's Modified Eagle Medium (DMEM) supplemented with 10% fetal bovine serum (FBS), 1% penicillin‐streptomycin, 1% GlutaMAX, 1% sodium pyruvate, and 1% MEM non‐essential amino acids (MEM‐NEAA). Transient transfection was performed using Lipofectamine 2000 with a total of 20 µg of plasmid DNA per dish. The plasmid mixture contained 10 µg of the transfer plasmid encoding the sgRNA of interest, 8.9 µg of the psPAX2 packaging plasmid (Addgene #12260), and 1.1 µg of the pMD2.G envelope plasmid (Addgene #12259). The DNA‐lipid complexes were formed in Opti‐MEM I medium and added to the cells. 6 h post‐transfection, the medium was replaced with fresh growth medium. At 72 h post‐transfection, the viral supernatant was harvested and subsequently concentrated using the Lenti‐X Concentrator (Takara Bio) following the manufacturer's protocol.
Tau Fractionation
4.6
To measure pathological tau levels, the hippocampus from one hemisphere was homogenized with Tris‐buffered saline (TBS; 25 mm Tris‐HCl pH 7.4, 150 mm NaCl, 1 mm EDTA, and 1 mm EGTA with protease and phosphatase inhibitors) with tissue grinder. The homogenate was centrifuged at 14,000 g for 5 min at 4°C. Next, the pellet and supernatant were sonicated to fully lyse the cells and centrifuged again for 15 min. The supernatant was collected, and its concentration was measured with a BCA protein assay (Thermo Fisher Scientific). Next, 200 µg proteins were incubated with 1% N‐ lauroylsarcosine sodium salt solution (Sigma‐Aldrich) on a rotator at 37°C for 1 h. Then, the samples were fractionated with ultracentrifugation at 150,000 g for 1 h. The supernatant, sarkosyl‐soluble fraction, was collected, measured with BCA protein assay, and immunoblotted for quantitative analysis. The pellet, sarkosyl‐insoluble fraction, was washed with TBS containing 1% sarkosyl solution and ultra‐centrifuged at 150,000 g for 1 h. The supernatant was removed, and the remaining pellet was resuspended in 5X sample buffer (Serva blue G) at 70°C for 10 min.
Western Blot
4.7
Brain tissues and cells were lysed in RIPA buffer containing protease inhibitors, phosphatase inhibitor cocktails, and PMSF, then sonicated and centrifuged at 13,000 rpm at 4°C for 15 min. Protein concentration was determined by BCA assay. Samples were denatured at 95°C for 3 min, separated on 4–12% Bis‐Tris gels by SDS‐PAGE, and transferred to PVDF membranes. Membranes were blocked with 5% skim milk in TBST for 1 h at room temperature, then incubated with primary antibodies at 1:1000 dilution at 4°C overnight. The primary antibodies used were pRb (Cell Signaling Technology, #8180), Rb (Cell Signaling Technology, #9313), AT8 (Ser202, Thr205; Invitrogen, #MN1020), pT181 (Thr181; Invitrogen, #701530), pS396(Invitrogen, #44752G), T22 (Milipore, #ABN454), Tau13 (Abcam, #ab13090), AT180 (Thr231; Invitrogen, MN1040), β‐actin (Cell Signaling Technology, #3700), APP (Biolegend, #803001), pCaMKII (Cell Signaling Technology, #12716), CaMKII (Abcam, #52476), pGSK3β (BD science #612313), GSK‐3β (Cell Signaling Technology, #12456), pERK1/2 (Cell Signaling Technology, #9101), ERK1/2 (Cell Signaling Technology, #9102), p35/25 (Cell Signaling Technology, #2680), p62 (Sigma, #P0067), LC3 (Cell Signaling Technology, #2775), CDK4 (Santa Cruz, #sc‐260). All the primary antibodies were diluted in 1:1000. On the next day, the membranes were washed with TBST before labeling with the corresponding horseradish peroxidase (HRP)‐conjugated secondary antibody (Invitrogen), incubation for 1 h. Proteins were visualized using ECL reagents (Abfrontier) on an Amersham Imager 600 (GE Healthcare Life Sciences). Densitometric analysis was performed using Multigauge software (Fujifilm Corporation) with normalization to β‐actin.
Immunohistochemistry
4.8
Mouse brain samples and brain organoids were fixed in 4% PFA overnight, cryoprotected in 30% sucrose for three days, and sectioned at 25–30 µm using a Leica CM3050s cryostat. Mouse brain sections underwent antigen retrieval with 70% formic acid for 20 min prior to staining. All samples were permeabilized with 0.3% Triton X‐100 and blocked with 5% horse serum (Vector Laboratories). Primary antibody incubation was performed overnight at 4°C using: NeuN (Millipore, 1:500), Caspase3 (Cell Signaling, 1:250), AT8 (Invitrogen, 1:200), pT181 (Invitrogen, 1:500), GFAP (Thermo Fisher Scientific, 1:1000), Iba1 (Wako, 1:1000), 6E10 (Biolegend, 1:1000), pCaMKII (Cell Signaling, 1:500), MAP2 (Abcam, 1:5000), LC3 (Cell Signaling, 1:200), and pS262 (Invitrogen, 1:200). Following PBS washes, samples were incubated with appropriate secondary antibodies and counterstained with DAPI (Sigma, 1:5000). Images were acquired using a Zeiss LSM700 confocal microscope and analyzed with ImageJ software.
Molecular Docking
4.9
Molecular docking was performed using the HyperLab platform (HITS Inc.) to predict the binding mode of abemaciclib to CaMKIIα and GSK3β and to estimate relative binding affinities [31]. Abemaciclib was obtained from PubChem as a SMILES string and converted to 3D using HyperLab. Receptor structures were obtained from the Protein Data Bank (PDB) as CaMKIIα (PDB ID: 6VZK) and GSK3β (PDB ID: 3SAY).
Surface Plasmon Resonance (SPR)
4.10
SPR experiments were performed on a Biacore T200 instrument (Cytiva) at 25°C using CM5 sensor chips. Proteins were immobilized via amine coupling in 10 mm sodium acetate buffer at a flow rate of 10 µL/min. GSK3β was immobilized at pH 5.5 to a surface density of ∼ 6400 response units (RU), and CaMKIIα was immobilized at pH 4.0 to ∼ 8000 RU. The running buffer consisted of HBS‐EP (10 mm HEPES, 150 mm NaCl, 3 mm EDTA, 0.005% (v/v) Tween‐20) supplemented with 2% (v/v) DMSO. DMSO solvent correction was performed using eight concentrations according to the manufacturer's recommendations to minimize refractive‐index effects. Serially diluted analytes were injected at concentrations of 4–32 µm for GSK3β‐abemaciclib, 10–40 µm for GSK3β‐CHIR99021, 2–16 µm for CaMKIIα‐abemaciclib, and 5–40 µm for CaMKIIα‐KN93 in the running buffer containing 2% (v/v) DMSO. Injections were performed at a flow rate of 30 µL/min with a contact time of 100 s and a dissociation time of 60 s. Between injection cycles, the surface was regenerated twice with 1 M NaCl for 30 s at 30 µL/min. Equilibrium response values across various analyte concentrations were plotted to generate binding isotherms. The equilibrium dissociation constant (KD) was subsequently derived by fitting these isotherms using a steady‐state affinity model in the Biacore Evaluation Software (Cytiva).
RNA Isolation and Real‐Time Quantitative RT‐PCR
4.11
Total RNA was extracted using RNeasy kit (QIAGEN) according to the manufacturer's protocols. cDNA was synthesized by reverse transcription using Maxime RT Premix (Intron Bio). Target genes were amplified using SYBR FAST reagents (Kapa Biosystems). Primers for human CDK4 and CDK6 were purchased from Bioneer and used following the manufacturer's instructions (CDK4: P268249 V, Human; CDK6: P273199 V, Human), and the sequences for all other primers are provided in Supplementary Figure S8. Relative mRNA levels were calculated by the ΔΔCt method.
Enzyme‐Linked Immunosorbent Assay (ELISA)
4.12
Sandwich ELISAs for pTau was performed with Tau (Phospho) [pT181] Human ELISA Kit (Invitrogen) according to the manufacturer's protocols.
Autophagic Flux Assay
4.13
To assess autophagic flux, cells were treated with 1 µm abemaciclib for 24 h in the presence or absence of 5 nm bafilomycin A1 (Sigma‐Aldrich). Autophagic flux was quantified by calculating the ratio of LC3II levels in bafilomycin‐treated versus untreated conditions for each experimental group.
Lysotracker Staining
4.14
The cells were stained with LysoTracker Red DND‐99 (Invitrogen) according to the manufacturer's instructions. Briefly, after 24 h of abemaciclib treatment, the cells were washed with prewarmed PBS and incubated with 100 nm Lysotracker at 37°C for 30 min. The images were acquired using a BC43 Microscope (Oxford Andor Technology) and analyzed with IMARIS software (Oxford Instruments).
Statistical Analysis
4.15
Statistical analyses were performed using GraphPad Prism 8.0 (GraphPad Software). A two‐tailed unpaired t‐test was used for comparisons between two groups, and one‐way analysis of variance (ANOVA) with Tukey's post hoc test was used for comparisons for groups of four. All data are shown as the mean, and error bars were drawn with the standard error of the mean (SEM). Outliers were detected with Grubbs’ test with alpha levels of 0.05, and p < 0.05 was considered statistically significant.
Author Contributions
J.H. and I.M.‐J. conceived the project. J.H. and J.H.J. performed animal experiments. J.H., D.L., Y.J., Y.J., W.Y.K., and C.‐H. L. performed and analyzed the results. H.J.K generated hiPSC‐derived brain organoids. J.H. wrote the manuscript. I.M.‐J. and E.S.J. revised the manuscript. I.M.‐J. supervised the study.
Funding
This research was supported by a grant of the Korea Dementia Research Project through the Korea Dementia Research Center (KDRC), funded by the Ministry of Health & Welfare and Ministry of Science and ICT, Republic of Korea (grant number: RS‐2020‐KH106747 & RS‐2020‐KH106773). This research was supported by the “Regional Innovation System & Education (RISE)” through the Seoul RISE Center, funded by the Ministry of Education (MOE) and the Seoul Metropolitan Government, Republic of Korea. (grant number: 2025‐RISE‐01‐016‐01).
Conflicts of Interest
The authors declare no conflicts of interest.
Supporting information
Supporting file: advs73588‐sup‐0001‐SuppMat.pdf.
The reference list from the paper itself. Each links out to its DOI / PubMed record.
- 1C. L. Masters , R. Bateman , K. Blennow , C. C. Rowe , R. A. Sperling , and J. L. Cummings , “Alzheimer's disease,” Nature Reviews Disease Primers 1 (2015): 15056, 10.1038/nrdp.2015.56.27188934 · doi ↗ · pubmed ↗
- 2S. Salloway , S. Chalkias , F. Barkhof , et al., “Amyloid‐Related Imaging Abnormalities in 2 Phase 3 Studies Evaluating Aducanumab in Patients with Early Alzheimer Disease,” JAMA Neurology 79, no. 1 (2022): 13–21, 10.1001/jamaneurol.2021.4161.34807243 PMC 8609465 · doi ↗ · pubmed ↗
- 3C. J. Swanson , Y. Zhang , S. Dhadda , et al., “A Randomized, Double‐blind, Phase 2b Proof‐of‐concept Clinical Trial in Early Alzheimer's Disease With Lecanemab, an anti‐Aβ Protofibril Antibody,” Alzheimer's Research & Therapy 13, no. 1 (2021): 80, 10.1186/s 13195-021-00813-8.PMC 805328033865446 · doi ↗ · pubmed ↗
- 4A. Bejanin , D. R. Schonhaut , R. La Joie , et al., “Tau Pathology and Neurodegeneration Contribute to Cognitive Impairment in Alzheimer's disease,” Brain 140, no. 12 (2017): 3286–3300, 10.1093/brain/awx 243.29053874 PMC 5841139 · doi ↗ · pubmed ↗
- 5S. Pushpakom , F. Iorio , P. A. Eyers , et al., “Drug Repurposing: Progress, Challenges and Recommendations,” Nature Reviews Drug Discovery 18, no. 1 (2019): 41–58, 10.1038/nrd.2018.168.30310233 · doi ↗ · pubmed ↗
- 6J.‐C. Park , S.‐Y. Jang , D. Lee , et al., “A Logical Network‐based Drug‐screening Platform for Alzheimer's Disease Representing Pathological Features of human Brain Organoids,” Nature Communications 12, no. 1 (2021): 280, 10.1038/s 41467-020-20440-5.PMC 780413233436582 · doi ↗ · pubmed ↗
- 7G. W. Sledge , “MONARCH 2: Abemaciclib in Combination With Fulvestrant in Women With HR+/HER 2− Advanced Breast Cancer Who Had Progressed While Receiving Endocrine Therapy,” Journal of Clinical Oncology 35, no. 25 (2017): 2875–2884, 10.1200/JCO.2017.73.7585.28580882 · doi ↗ · pubmed ↗
- 8T. J. Raub , G. N. Wishart , P. Kulanthaivel , et al., “Brain Exposure of Two Selective Dual CDK 4 and CDK 6 Inhibitors and the Antitumor Activity of CDK 4 and CDK 6 Inhibition in Combination With Temozolomide in an Intracranial Glioblastoma Xenograft,” Drug Metabolism and Disposition 43, no. 9 (2015): 1360–1371, 10.1124/dmd.114.062745.26149830 · doi ↗ · pubmed ↗