INTRODUCTION
Alzheimer’s disease (AD) is a neurodegenerative disease characterized by amyloid beta deposition and pathologic tau. It manifests as a cognitive decline that worsens daily behavior [1-3]. Unlike familial AD, where genetic factors are relatively well understood, sporadic AD involves diverse etiologies such as senile plaques, neurofibrillary tangles, endoplasmic reticulum (ER) stress, cholesterol dysregulation, and genetic factors affected by ethnic diversity and sex. These diverse etiologies present additional challenges for drug and biomarker discovery [4-6]. Recently, the approval of the monoclonal antibody drugs aducanumab and lecanemab by the United States Food and Drug Administration represented a significant advancement in anti-amyloid therapy for AD treatment. While these drugs demonstrated efficacy, some patients do not respond, and some experience disease progression despite amyloid clearance [7,8]. Another study found that a single-nucleotide polymorphism in the APOE promoter, whose minor allele frequency varies among populations, has a higher prevalence in East Asians than in other groups [9]. Given these persistent limitations in drug development and the intricate nature of AD across ethnicity and etiology, there is a clear need for various pathological approaches and alternative therapeutic targets.
The ER is involved in essential cellular functions, including protein folding, Ca2+ balance maintenance, and cholesterol synthesis. Under certain pathological states, the accumulation of unfolded and misfolded proteins exceeds the protein-folding capacity of the ER lumen, leading to ER stress and activation of the unfolded protein response (UPR). Numerous studies have demonstrated that the ER stress response contributes to the onset and progression of AD. However, the mechanism underlying ER stress in AD pathophysiology remains poorly understood, hindering the development of effective biomarkers for its diagnosis and treatment [10-12]. However, the mechanism underlying ER stress remains poorly understood, hindering the development of effective biomarkers for its diagnosis and treatment. Although some animal models and autopsy studies can mimic certain aspects of the disease mechanism and have significantly advanced the comprehension of AD, their findings often face challenges when applied directly to patients [13,14]. Therefore, we emphasize the need for alternative methodological approaches.
Skin fibroblasts, which are easily obtained from the peripheral tissues of individual patients, are valuable models for studying disease. Patient-derived human dermal fibroblast (HDF) cultures exhibit aging-related changes based on donor age, reflecting the phenotype and brain pathology associated with neurodegenerative diseases [15]. It is known that HDF from AD patients shows transcriptomic response and elevations in markers to metabolic stress [16,17]. Their ex vivo expansion reveals molecular anomalies that disrupt cellular homeostasis. Neuropathological studies have suggested that early pathological changes in the AD brain can also be detected in peripheral tissues such as skin fibroblasts [6,18,19]. Additionally, fibroblasts provide a convenient source of primary cells for various research purposes. While fibroblasts are frequently reprogrammed into induced pluripotent stem cells (iPSCs) for brain research, this process can result in the loss of epigenetic information and aging signatures [20]. Therefore, leveraging the preserved aging signatures and individual heterogeneity of ex vivo fibroblast models can enable the identification of differential responses to neurodegenerative risk factors. This approach allows for more accurate identification of potential targets [15,21,22].
The aim of this study was to characterize the transcriptional profile of fibroblast derived from patients with AD, utilizing an ex vivo platform. We hypothesized that HDFs from healthy controls and patients with AD would exhibit differential responses to ER stress.
METHODS Study participants
Study participants
All participants were enrolled in the BICWALZS. It was launched in October 2016 by the Korea Disease Control and Prevention Agency and is part of the National Korea Biobank Project conducted at Ajou University Hospital. This initiative is part of a national effort to improve the infrastructure for biomedical and healthcare research. Individuals reporting cognitive changes, either self-reported or observed by others, were enrolled in the BICWALZS from five tertiary referral hospital outpatient clinics and a community mental health center. Dermal fibroblasts were collected from participants at Ajou University Hospital. We cultured and developed cells from patient-derived dermal fibroblasts for our ex vivo studies that reflect the participants’ genetic profiles. While not all participants provided consent, a subset agreed to undergo skin biopsies for the collection of dermal fibroblasts. All participants underwent comprehensive clinical assessments, including neuroimaging, cognitive tests, genetic evaluations, and blood-based biomarkers.
Participants were classified into healthy control and AD groups based on specific criteria. These included in vivo amyloid pathology assessed via brain positron emission tomography (PET) imaging and clinical assessment through global deterioration scale (GDS) scores. Scores of 4 or higher indicated AD, while scores of 3 or lower, along with the absence of in vivo amyloid pathology, indicated healthy controls. Among the participants who provided cell samples, 20 met the AD criteria (positive amyloid PET and a GDS score of 4 or higher). For healthy controls, the following additional exclusion criteria were applied: medial temporal lobe neurodegeneration on brain magnetic resonance imaging (MRI), notable disabilities reported by caregivers, and white matter hyperintensities on MRI. Participants who met the following criteria were also excluded from the study: notable hearing or visual impairments hindering interview participation; suspected behavioral-variant frontotemporal lobar degeneration or Lewy body dementia; a history of neurological disorders, including brain tumors, subarachnoid hemorrhage, epilepsy, encephalitis, or metabolic encephalopathy; diagnosis of psychiatric disorders such as intellectual disability, schizophrenia, bipolar disorder, or other psychiatric conditions; use of psychoactive substances other than alcohol; and physical illnesses that could impede study participation, including cancer, renal or hepatic failure, severe asthma, or chronic obstructive pulmonary disease. Additionally, blood-based biomarkers were analyzed, including phosphorylated tau-217 (pTau217), indicative of tau pathology; glial fibrillary acidic protein (GFAP), a marker of nonspecific inflammatory responses. To measure these biomarkers, we used an ultrasensitive immunoassay technology called the single-molecule array (Simoa) assay, performed on the Simoa HD-X instrument (Quanterix). Specifically, Aβ42 and Aβ40 were measured using the Neurology 4-Plex E (#103670), and p-tau217 levels were measured with the ALZpath Simoa® p-Tau 217 V2 Assay Kit (#104371). Ultimately, 22 participants were included as healthy controls in this study.
The BICWALZS is registered in the Korean National Clinical Trial Registry and can be accessed through the Clinical Research Information Service under the identifier KCT0003391. The research protocol was approved by the Institutional Review Board of Ajou University Hospital (AJIRB-BMR-SUR-16-362) prior to the study’s commencement. Written informed consent was obtained from all the participants and their caregivers. Additionally, approval for this study was obtained from the Institutional Review Board (AJOUIRB-EXP-2021-690). Detailed information on the BICWALZS can be found in other sources [23].
Cell cultures
To perform ex vivo studies, we cultured and established patient-derived dermal fibroblast cells. A subset of participants agreed to donate skin biopsy samples for patient-derived dermal fibroblast sampling. A skin biopsy sample was obtained within 4 weeks of the baseline assessment. We planned and modified our biopsy and culture protocols based on previous studies and our preliminary experience [23]. Briefly, a 3-mm3 skin biopsy sample was taken from the upper inner arm using a disposable punch under local anesthesia and complete aseptic conditions. The tissue was immediately immersed in 5 mL of warm culture medium and transferred to the laboratory at the same institution. The biopsy samples were then placed in a 100-mm dish and cut into small fragments. After 1 minute of attachment, 10 mL of Dulbecco’s Modified Eagle’s Medium (DMEM) with high glucose (Hyclone) supplemented with 20% fetal bovine serum (FBS) and 1% penicillin/streptomycin was added. After 1 week, the culture medium was changed every 2-3 days. Once enough fibroblasts were identified around the biopsy samples, we trypsinized and transferred the cells into another 100-mm dish for further processing (passage 1). More than 95% of participant’s skin biopsy samples yielded viable fibroblast cultures, following a rigorously maintained standard protocol. Fibroblasts were grown in DMEM, supplemented with 10% FBS and GibcoTM 1X Antibiotic-Antimycotic (Thermo Fisher Scientific). The cells were maintained in an incubator at 37°C and 5% CO2. The culture medium was changed every 2 days until the cells reached 90% confluency.
Immortalized human astrocytes (IM-HAs) have been developed by immortalizing primary human cortical astrocytes using the SV40 Large T antigen (Innoprot). IM-HAs were cultivated on poly-L-lysine (Sigma-Aldrich)-coated cultureware and grown in an appropriate medium purchased from Innoprot. All cell lines were maintained in an incubator at 37°C and 5% CO2. Immortalized human microglia (IM-HM) have been developed by immortalizing primary human cortical astrocytes using the SV40 Large T antigen (Innoprot). IM-HM were cultivated on GibcoTM collagen I (Thermo Fisher Scientific)-coated cultureware and grown in an appropriate medium purchased from Innoprot. The culture medium was changed every third day until the cells reached 70% confluence and changed every day until 90% confluence. All cell lines were maintained in an incubator at 37°C and 5% CO2.
Experimental scheme and thapsigargin treatments
In this study, we mimicked the ER stress environment in vitro by treating cells with thapsigargin at concentrations that allowed for cell survival. Each cell line was seeded in six-well plates with 70% confluency after 24 hours. The fibroblasts were treated with 10 nM of thapsigargin in 10% FBS media, IM-HAs with 1 µM in 2% FBS media, and the IM-HM cells with 100 nM in 5% FBS media. Control groups were treated with culture medium without thapsigargin. The treated and control cells were incubated at 37°C and 5% CO2. After 24 hours, RNA was collected from each well for sequencing.
RNA sequencing and data normalization
Total RNA was extracted using the RNwasy Plus Mini Kit (Qiagen) following the manufacturer’s guidelines. Bulk RNA sequencing was performed by Macrogen. Fibroblasts were selected from 22 healthy controls and 20 patients with AD. To induce ER stress, thapsigargin was administered as described in Section 2.3. Sequencing was conducted on an Illumina platform to generate paired end reads of 101 bp. Gene expression levels were quantified using StringTie and normalized to fragments per kilobase of transcript per million mapped reads or transcripts per million. Quality control was performed in two steps. First, the Limma package was used to remove batch effects from raw data. Then, normalization and principal component analysis were performed using Deseq2, employing GLM modeling and Wald’s tests. Genes that achieved 50 read counts in at least five samples were retained, whereas the low-expressed genes were filtered out. All RNA-Seq data were deposited in the GEO database (accession number GSE 283128).
Differentially expressed genes analysis
Considering the conditions, we performed differential gene expression analysis. We compared the effects of thapsigargin treatment by calculating fold changes relative to vehicle treatment in both patients with AD and healthy controls. We then recalculated these fold changes to identify genes that had a more sensitive response to thapsigargin treatment in AD. The selected genes were analyzed using the Shapiro-Wilk test to assess normality. Genes with equal variance were analyzed using Student’s t-test, whereas genes with unequal variance were analyzed using Welch’s t-test. Statistical significance was set at p<0.05. The Gene Ontology (GO), Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment, and protein-protein interaction analysis were performed using Enricher (http://amp.pharm.mssm.edu/Enrichr) and Metascape (https://metascape.org). The cholesterol biosynthesis pathway (WP197) was obtained using NDEx [24] and edited using BioRender. To select significant biomarkers, we assessed the coefficient of variation (CV, standard deviation/mean) and applied a cutoff value of <50%. Additionally, we filtered the genes using an open-source AD-related gene list. We complied this list from four sources: Alzdata (http://www.alzdata.org/index.html): we obtained genes with p<0.05 from data on the frontal cortex, temporal cortex, hippocampus, and entorhinal cortex, identifying 5,398 genes with at least two overlaps. GSEA (https://www.gsea-msigdb.org/gsea/index.jsp): we included genes from the following datasets: “blalock_alzheimers_disease_dn,” “blalock_alzheimers_disease_incipient_dn,” “blalock_alzheimers_disease_incipient_up,” “blalock_alzheimers_disease_up,” and “kegg_alzheimers_disease,” totaling 3,278 genes. PheGenI (https://www.ncbi.nlm.nih.gov/gap/phegeni): we included 1,509 AD-related genes from the Homo sapiens category. ALZGENE: we included 680 genes from ALZGENE. Genes present in at least two of these databases were considered AD-related.
Knockdown of DCTN2
All siRNAs were synthesized and chemically modified by Dharmacon, Inc. Cells were transfected with siRNA using the Lipofectamine RNAiMAX transfection reagent (Invitrogen) in GibcoTM Opti-MEM Reduced Serum Medium (Thermo Fisher Scientific) containing the ON-TARGETplus DCTN2 siRNA and ON-TARGETplus non-targeting control siRNAs (Dharmacon). During siRNA treatment, the culture medium was kept serum-free. After 6 hours, the culture medium was replaced with a complete medium. DCTN2 expression was analyzed using reverse transcription-quantitative polymerase chain reaction (RT-qPCR) and western blotting, following standard protocols.
Western blot analysis
Protein lysates were collected at 4°C using the radio-immunoprecipitation assay buffer supplemented with a protease inhibitor cocktail (Bio Rad). Western blotting was performed using Mini-Protean TGX Precast Gels (Bio Rad) following standard protocols. The antibodies used included rabbit β-actin (#4970, CST) and rabbit p50 dynamitin (ab248521, abcam). All the secondary antibodies (anti-mouse and anti-rabbit) were purchased from Cell Signaling Technology. Chemiluminescent signals were detected using the Invitrogen iBright imaging system, and quantification was performed using the ImageJ software (NIH).
Cell viability assay
To directly test whether DCTN2 inhibition affects cell viability and proliferation under ER stress, an EZ-cytox assay was conducted according to the manufacturer’s protocol (DoGenBio). Briefly, IM-HAs and IM-HM were seeded in a 12-well plate at concentrations of 8×104 and 4×104 cells, respectively. After transfection, control cells and those treated with 1 µM thapsigargin (IM-HAs) or 100 nM (IM-HM) were incubated for 24 hours at 37°C under a 5% CO2. Following incubation, 100 µL EZ-cytox agent was added into each well and incubated for 1.5 hours under the same conditions. Then, 100 µL of the culture medium from each well was transferred to a 96-well plate. The absorbance of each sample was measured using a microplate reader. Cell viability was assessed by measuring the absorbance at 450 nm using a microplate absorbance reader (BioteK, Synergy HTX). Each experiment was repeated three times.
Flow cytometry assay
After DCTN2 transfection and thapsigargin treatment, an Annexin V-FITC Apoptosis Detection Kit (#6592, CST) was used to identify apoptotic cells via flow cytometry. Briefly, after transfection and treatment, the cells were collected through centrifugation, washed with ice-cold phosphate-buffered saline, and resuspended at 4×105 cells/mL with 1X Annexin V Binding Buffer. Then, 1 μL of the Annexin V-FITC conjugate and 12.5 μL of the propidium iodide solution were added to each cell suspension (96 μL) and incubated on ice in the dark for 10 minutes. The suspension was then diluted to a final volume of 250 μL/assay with ice-cold 1X Annexin V Binding Buffer and analyzed immediately using BD FACS CantoTM II (BD Biosciences). Data were acquired and analyzed using the BD FACSDiva software. A total of 10,000 events were collected per sample, and the percentages of early apoptotic cells (Annexin V+/PI-) and late apoptotic cells (Annexin V+/PI+) were determined by gating based on fluorescence intensity. Three independent experiments were conducted.
Statistical analysis
Statistical analyses were performed using GraphPad Prism 9.0 (GraphPad Software). Data are presented as mean±standard error of the mean from at least three independent experiments. Statistical significance was determined using an unpaired two-tailed Student’s t-test or Mann-Whitney U test, depending on the normality of the data distribution, as assessed through the Shapiro-Wilk test. Pearson’s or Spearman’s correlation coefficients were calculated to assess the relationships between variables depending on data normality. Statistical significance was set at p<0.05. All statistical tests were two-sided.
RESULTS
Demographic characteristics of study participants
Table 1 describes the demographic characteristics of patients with AD based on their clinical diagnosis. Compared with healthy controls, patients with AD were well classified based on demographic characteristics. Patients with AD exhibited more advanced disease features than did healthy controls, as evidenced by a lower Mini-Mental State Examination (MMSE) score, impaired memory function on the Seoul Verbal Learning Test (SVLT) and Rey Complex Figure Test (RCFT), and higher disability scores on the Seoul Instrumental Activities of Daily Living (S-IADL). The average Scheltens scale score indicated more severe hippocampal neurodegeneration in patients with AD than in controls. All scores had p-value<0.001. Moreover, we performed detailed group comparisons of MMSE scores and amyloid deposition. Most healthy controls achieved a perfect score on the MMSE, whereas the disease group demonstrated poorer scores. Additionally, the disease group exhibited greater amyloid deposition than did the healthy control group, whose levels were nearly zero (Figure 1).
Dermal fibroblasts derived from patients with AD exhibit a cholesterol biosynthetic gene expression profile in response to ER stress
To understand the regulatory mechanisms related to ER stress, transcriptome analysis was conducted on ex vivo dermal fibroblasts obtained from patients with AD and healthy controls. These fibroblasts were exposed to thapsigargin at appropriate concentrations to induce ER stress, with a vehicle-treated group serving as the control. Bulk RNA sequencing revealed differential gene expression between patients with AD and healthy controls (Figure 2A). Batch-effect removal and normalization procedures were applied. In total, 13,309 genes were identified, and 468 of them showed significantly different responses between healthy controls and patients with AD (p<0.05). We first assessed the fold changes between thapsigargin treatment and vehicle in both the control and AD group samples. Then, we identified genes in patients with AD that showed an altered response to thapsigargin compared to those in controls by recalculating the fold changes between the two groups (Figure 2A). For functional annotation, GO and KEGG pathway enrichment analyses were performed. GO analysis of the biological process in highly responsive genes included terms such as “Negative regulation of NIK/NF-kappaB signaling (GO:1901223),” “Neuron projection extension (GO: 1990138),” and “Late endosome to lysosome transport (GO: 1902774)” (Figure 2B). In the KEGG pathway analysis, we observed “Herpes simplex virus 1 infection (hsa05168).” No significant protein-protein interactions were observed (data not shown). Additionally, GO analysis of the biological process in low responsive genes includes terms such as “Cholesterol biosynthetic process (GO:0006695),” “Sterol biosynthetic process (GO:0016126),” and “Cholesterol metabolic process (GO: 0008203)” (Figure 2C). In the KEGG pathway analysis, we also observed a significant enrichment of lipid-related terms such as “Steroid biosynthesis (hsa00100)” and “Biosynthesis of unsaturated fatty acids (hsa01040)” (Figure 2C). Consistent with this finding, the protein-protein interaction analysis was performed on the low responsive genes, revealing interactions between genes related to steroid and cholesterol biosynthesis, such as DHCR7, SC5D, NSDHL, MSMO1, CYP51A1, and FDFT1. These genes are highlighted in the cholesterol synthesis pathway (Figure 3). Their gene expression increased in response to ER stress; however, these increases were less pronounced in AD. Given that cholesterol metabolism dysfunction is a major risk factor for AD, these altered gene expression profiles prompt further investigation into its protective or pathological role in AD.
Identification of potential biomarkers of dermal fibroblasts derived from patients with AD under ER stress
To achieve disease-targeted approaches, we identified 49 differentially expressed genes (DEGs) from the interaction with the following three filtering conditions (Figure 4A): p<0.05, a CV <50%, and inclusion in open-source AD-related genes (Supplementary Material). The volcano plot in Figure 4B displays the 49 selected genes. DCTN2 emerged as the most significantly responsive gene when sorted by p-value, along with other notable genes such as HMGCR and LDLR (Table 2). DCTN2 read counts and fold changes for each patient are represented as points on the dot plot, with the controls marked in blue and patients with AD marked in red (Figure 4C). Similar data for HMGCR and LDLR are shown in Supplementary Figure 1. The read count results showed that, in the vehicle-treated state, DCTN2 expression was similar between controls and patients with AD. However, after thapsigargin treatment, DCTN2 expression decreased significantly more in patients with AD. These identifications suggest that the downregulation of DCTN2, which is more pronounced in AD, may be associated with AD pathophysiology.
Association of DCTN2 fold change induced by ER stress with clinical and biomarker characteristics
To validate whether the response of DCTN2 to thapsigargin in patient-derived HDFs was pathologically significant, we examined its association with various clinical, blood-based, and neuroimaging biomarkers. Biomarkers for cognitive function showed a positive correlation with DCTN2 fold change. General cognition was evaluated using the MMSE (Figure 5A), verbal memory function was assessed using the SVLT delayed recall (Figure 5B), and visuospatial memory function was evaluated using the RCFT (Figure 5C). The strongest correlation was observed with verbal memory function (r=0.6, p<0.001). Additionally, a weak positive correlation was observed with frontal function (Figure 5D). Conversely, neuroimaging results revealed a negative correlation between neurodegeneration in the hippocampus (Figure 5E) and the blood-based marker pTau217 (Figure 5G). Another blood-based marker, the Aβ42/Aβ40 ratio, demonstrated a positive correlation (Figure 5F). Supplementary Figure 2 indicates the distribution of normal controls and AD patients using different colors. These results highlight the importance of DCTN2 as a potential biomarker. Therefore, we conducted a validation experiment to investigate the effects of DCTN2 expression in vitro.
DCTN2 knockdown modulates cell viability in astrocyte and microglia cell lines
To investigate the biological role of DCTN2 during thapsigargin-induced ER stress in the central nervous system (CNS), we transfected IM-HA and IM-HM cell lines with either DCTN2-specific siRNA (Si-DCTN2) or control siRNA (Si-control). Western blot analysis revealed that transfection with si-DCTN2 significantly downregulated DCTN2 expression in both cell lines (Figure 6A), and qPCR analysis confirmed a corresponding decrease at the RNA level (Figure 6B). Cell viability was assessed after 24 hours of thapsigargin exposure using the EZ-cytox assay. DCTN2 knockdown decreased cell viability in both IM-HA and IM-HM cells (Figure 6C and D). These findings showed that DCTN2 knockdown in CNS glial cell lines may contribute to cellular damage under ER stress, as indicated by the observed decrease in cell viability.
DCTN2 knockdown modulates early apoptosis in astrocyte cell lines
To determine whether DCTN2 affects thapsigargin-induced ER stress cytotoxicity, we performed flow cytometric analysis using Annexin V/FITC staining. DCTN2 inhibition modestly enhanced thapsigargin-induced apoptosis in IM-HA (Figure 7A and B). In contrast, neither thapsigargin treatment nor DCTN2 knockdown significantly affected apoptosis in IM-HM (Figure 7C and D). This result indicates that knockdown of DCTN2 not only decreases cell viability (Figure 6C) but also slightly increases ER-stress-induced apoptosis in IM-HA.
DISCUSSION
In this study, we utilized an ex vivo platform to characterize the transcriptional profile of fibroblasts derived from patients with AD in response to ER stress and propose potential biomarkers. Following thapsigargin treatment, we observed distinct expression patterns of DEGs between healthy controls and patients with AD. These genes were categorized into two groups based on their responses to thapsigargin: those with either greater or lesser responses in AD, involving both upregulation and downregulation.
The GO analysis revealed that genes with a greater response in AD were significantly enriched in the “negative regulation of NF-kappa-B signaling” pathway. The genes associated with this pathway, such as CPNE1, MKRN2, and RELA, were significantly downregulated in AD. MKRN2, a known ubiquitin E3 ligase for the p65 subunit of NF-kB, negatively regulates inflammatory responses [25]. MKRN2 downregulation in AD suggests increased NF-kB signaling. Additionally, ER stress leads to the upregulation of toll-like receptor 4 (TLR4), which activates the NF-kB signaling pathway [10,26,27]. This indicates that ER stress may exacerbate inflammatory response in AD. The KEGG pathway analysis revealed significant enrichment of genes in the “Herpes simplex virus 1 infection” pathway. Most genes corresponding to this pathway belong to the Krueppel C2H2-type zinc-finger protein family. Ongoing research suggests a link between herpes simplex virus 1 and AD [28]. Moreover, we found that genes with lesser responses in AD were significantly enriched in cholesterol-related terms and pathways, including “Cholesterol biosynthetic process,” “Sterol biosynthetic process,” “Steroid biosynthesis,” and “Biosynthesis of unsaturated fatty acids.” Genes involved in cholesterol biosynthesis, such as DHCR7, SC5D, NSDHL, MSMO1, CYP51A1, and FDFT1, demonstrated increased expression in AD, although not as much as that in controls. These findings align with previous research linking ER stress to cholesterol biosynthesis upregulation [29]. and are further supported by the result of a study showing reduced cholesterol levels in the cerebrospinal fluid (CSF) of patients with AD [30]. Cholesterol metabolism dysfunction is a known risk factor for the development of AD [31]. Although the impact of ER stress on cholesterol metabolism is not fully understood, our findings from ex vivo patient-derived fibroblasts provide further insights.
Patient-derived HDFs can be easily obtained and stabilized. HDF cultures reflect the phenotype and pathology of neurodegenerative disease. Their ex vivo expansion reveals molecular anomalies that disrupt cellular homeostasis. Our ex vivo studies using HDFs replicated the responses observed in patients’ CSF, highlighting their potential to accurately reflect the disease environment. This suggests that HDF disease models could be a beneficial approach for identifying early biomarkers of AD. Given their ability to capture individual differences, ex vivo expansion is effective for developing personalized therapeutic strategies.
Among the genes showing significant responsiveness, we selected those related to AD for a more disease-specific approach (Figure 4A and Supplementary Material 1). The most significant DEG identified was DCTN2 (p<0.001). Thapsigargin treatment reduced DCTN2 expression, showing a greater downregulation in AD. To assess its pathological significance, we analyzed the correlation between DCTN2 and other clinical biomarkers in patients with AD. Cognitive function biomarkers (MMSE, SVLT delayed recall, RCFT delayed recall, and Controlled Oral Word Association Test) showed significant positive correlations with DCTN2, suggesting its potential as a biomarker for cognitive function. In contrast, neuroimaging results revealed a negative correlation between the DCTN2 fold change and Schelten’s scale score, suggesting that DCTN2 may be associated with hippocampal neurodegeneration. Furthermore, blood-based biomarkers demonstrated a positive correlation with plasma amyloid β (Aβ) protein and a negative correlation with plasma phospho-tau. These findings emphasize the potential of DCTN2 as a biomarker for AD.
DCTN2, the second dynactin subunit, mediates retrograde vesicle transport by interacting with dynein. Previous autopsy studies have shown that dynactin levels are significantly reduced in AD. This reduction is even more pronounced in patients with the APOE e4/4 genotype, highlighting a genotype-specific vulnerability to protein dysregulation and neuronal damage [32]. Our bulk RNA sequencing results reflect these findings, as DCTN2 expression was significantly decreased in AD under ER stress. Given that DCTN2 is an essential subunit of dynactin, its downregulation may destabilize dynactin and indirectly impair dynein function by attenuating the interaction between dynactin-dynein complexes. Dysfunction of dynein has been shown to cause intracellular accumulation of endosomal β-amyloid precursor protein and tau [33,34]. DCTN2 is known to affect dynein-dependent transport [35]. Therefore, we propose that the downregulation of DCTN2 under ER stress may ultimately exacerbate Alzheimer’s disease pathology by contributing to abnormal accumulation of AD-related proteins and intracellular trafficking defects. Subsequently, immortalized CNS cells were used to evaluate the role of DCTN2 in vitro. We found that DCTN2 knockdown in astrocytes under ER stress led to decreased cell viability and modestly increased apoptosis. Previous studies have demonstrated that dynein impairment interferes with Aβ clearance in astrocytes [36]. Thus, it is possible that DCTN2 downregulation, by potentially affecting dynein function, impairs Aβ clearance, which in turn may contribute to the reduced viability and increased apoptosis. Collectively, these findings suggest that inhibition of DCTN2 contributes to ER stress-induced cytotoxicity in astrocyte. Although the mechanism underlying ER stress-induced downregulation of DCTN2 remains unclear, and cell lines have inherent limitations, these results indicate that DCTN2 performs essential biological functions in cell survival.
Despite numerous studies on AD pathology over the past few decades, disease modeling has faced significant challenges, primarily because the symptoms manifest predominantly in the human brain. In this study, we employed an improved method using ex vivo patient-derived fibroblasts. This approach offers a more accessible methodology and is consistent with autopsy and animal study results. Notably, DCTN2, a potential target identified through our ex vivo platform, exhibited an effective role in the viability of CNS cell lines, further supporting its relevance to AD pathology. While the exact functions of DCTN2, including its role in misfolded protein accumulation, require further investigation, the present study demonstrates that an ex vivo approach could help identify potential biomarkers. Furthermore, identifying potential target genes using patient-derived fibroblasts contributes to establishing new therapeutic targets, addressing the limitations of conventional drug development, and facilitating drug repositioning strategies. We anticipate that this identification will provide insight into patient-specific targeting, diagnostic possibilities, and potential AD biomarkers for future studies. With these strengths in mind, we acknowledge potential limitations to the current study. The small sample size may limit the reliability of the results. However, given that the patients were diagnosed with AD based on multiple criteria, we believe and have observed that they exhibit significantly distinct pathological responses compared to healthy controls. Additionally, since the study was conducted exclusively with an East Asian South Korean population, further research is needed to determine whether the results can be validated in other populations.
In conclusion, this study revealed differential transcriptome expressions induced by ER stress in fibroblasts derived from patients with AD. In total, 468 genes were differentially expressed. Genes that showed a less pronounced response in patients with AD than in healthy controls were enriched in lipid-related GO terms, including cholesterol biosynthesis. The gene with the most significant differential expression, DCTN2, was downregulated in response to thapsigargin treatment, with a more pronounced reduction in patients with AD. DCTN2 positively correlated with several clinical biomarkers, and knockdown experiments indicated that DCTN2 dysfunction reduces cell viability in glial cell lines and increases apoptosis in astrocytes. These findings suggest that DCTN2 is a potential biomarker associated with AD pathology and indicates the possibility of patient-specific targeting through an ex-vivo approach.
PDF Links
PubReader
ePub Link
XML Download
Full text via DOI
Download Citation
Supplement1
Print
