miR-361-5p contributes to the pathogenesis of Alzheimer’s disease

Abstract

Alzheimer’s disease (AD) is a progressive neurodegenerative disorder. This study investigated the roles of HOMER1, ATAD1, and miR-361 in AD pathogenesis using microarray (GSE106241, GSE157239; n = 60) and RT-PCR (n = 100; 50 AD patients, 50 controls from Northwest Iran) analyses. Decreased expression of HOMER1 and ATAD1, key regulators of glutamatergic synapses, and miR-361, a potential regulator of both, was observed in AD brain tissue (GSE106241, categorized into seven Braak stages), suggesting a link between their dysregulation, impaired synaptic function, and increased neuroinflammation. However, blood-based RT-PCR showed no significant difference in HOMER1 or ATAD1. miR-361 was significantly lower in AD patients (adjusted p < 0.043). These findings, limited by sample size and lacking a formal power analysis, require further investigation to validate their potential as peripheral biomarkers for AD. Future studies with larger sample sizes are warranted.

Introduction

Alzheimer’s disease (AD) is the most prevalent neurodegenerative cause of dementia, characterized by progressive memory loss, cognitive decline, and social dysfunction, significantly impacting global health^1,2. While both early-onset (EOAD) and late-onset (LOAD) forms exist, LOAD accounts for the vast majority of cases, exhibiting substantial heritability and a complex etiology involving genetic, environmental, and lifestyle factors^3–5. Key pathological hallmarks include the accumulation of amyloid-beta (Aβ) plaques and neurofibrillary tangles (NFTs), arising from the abnormal processing of amyloid precursor protein (APP) and the hyperphosphorylation of tau protein^6,7. The dysregulation of Aβ production, involving enzymes like BACE1 and γ-secretase, is a critical driver of AD pathogenesis⁶.

Although the amyloid-tau cascade is a key feature of AD, neuroinflammation plays a crucial role in disease progression. Alongside oxidative stress, mitochondrial dysfunction, and cholinergic deficits⁸. The excessive production of inflammatory molecules characterizes neuroinflammation. These molecules, often triggered by Aβ and tau, activate glial cells, releasing pro-inflammatory cytokines and chemokines, which in turn stimulate further Aβ production and deposition, creating a vicious cycle^9,10.

Glutamatergic synaptic transmission plays a critical role in cognitive function, and its dysregulation is increasingly recognized as a key contributor to AD pathogenesis. The precise balance of excitatory and inhibitory signaling is essential for maintaining healthy neuronal function, and disruptions in glutamatergic pathways lead to excitotoxicity, synaptic loss, and cognitive decline^8,11,12. Numerous studies have implicated alterations in glutamate receptor expression, glutamate transporter function, and synaptic plasticity in AD progression¹³.

HOMER1 is a scaffold protein that stands out in the HOMER family due to its unique feature of having both long and short variants resulting from alternative splicing. Both forms share the same N-terminal binding domain and target identical proteins^14,15. HOMER1 is essential to the postsynaptic density, impacting the regulation of metabotropic glutamate receptors (mGluR), inositol 1,4,5-triphosphate receptors (IP3R), and transient receptor potential channels (TRPC). It also influences the endocytosis of glutamatergic synapses, such as NMDAR and AMPAR, indirectly through linker proteins like SHANK and GRIP, thus affecting excitatory synapses^16–19. By regulating NMDAR and AMPAR endocytosis, HOMER1 plays a crucial role in synaptic plasticity, learning, and memory¹⁷.

ATAD1, also known as Thorase, is a neuroprotective protein that plays a crucial role in regulating the surface expression of glutamatergic receptors, particularly AMPAR and NMDAR. It achieves this by declustering the AMPAR/GluR-interacting protein 1 (GRIP1) complex, thereby contributing to the regulation of synaptic plasticity, learning, and memory^20–23. When ATAD1 levels drop, AMPAR internalization also decreases, leading to increased AMPA receptor activity. These heightened glutamate levels result in overactive receptors, triggering neuronal depolarization and subsequent influx of calcium and sodium ions. This process leads to glutamate excitotoxicity, as well as the production and release of inflammatory cytokines like TNF and IL-6. Ultimately, this neuronal death may contribute to the development of AD^24–26.

MicroRNAs (miRNAs) are small RNA molecules approximately 22 nucleotides long. They target genes by binding to specific sites in the 3′ untranslated region (3′UTR) of mRNA. Numerous studies have shown that dysregulation of miRNAs plays a significant role in various human diseases, such as AD^27–30.

miR-361 has a regulatory effect on the expression of both ATAD1 and HOMER1. The relationship between miR-361 and these genes underscores the complexity of gene regulation in the context of neurodegenerative diseases, indicating that alterations in miR-361 levels could have downstream effects on synaptic functioning and inflammation associated with AD.

Effective therapies for AD require a better understanding of the interplay between glutamatergic dysfunction and the expression of key genes and microRNAs. This study evaluates the potential of ATAD1, HOMER1, and miR-361-5p as peripheral biomarkers for AD by analyzing their expression levels in blood samples from AD patients and controls using bioinformatics and RT-PCR.

Materials and methods

Identification and differential analysis of ATAD1, HOMER1, and miR-361 expression levels as an inflammatory biomarker in peripheral blood

Participants and peripheral blood samples

This case–control study involved 100 participants, with 50 patients diagnosed with AD and 50 healthy controls matched for sex and age from northwest Iran. The study was approved by the clinical research ethics committee of Tabriz University of Medical Sciences (Ethical code: IR.TBZMED.REC.1398.1265). We recruited participants from Imam Reza Hospital at Tabriz University of Medical Sciences for this research study. A neurology expert used the DSM-V criteria³¹ to select the participants. We carefully reviewed the patients’ medical records to select the cases. This included examining the duration of the disease, medications prescribed, family history, education level, previous medical conditions and their duration, and the outcomes of the Mini-Mental State Examination. Only patients who met the inclusion criteria were chosen for the study. These criteria included being 65 years old or older and having no psychiatric or neurological diagnoses other than AD. The control group consisted of individuals in the same department who were 65 years old or older and did not have AD. We excluded individuals with the following conditions based on our criteria: active or chronic infectious diseases, diabetes, thyroid disorders, renal and liver failure, cancer, inflammatory diseases or those taking anti-inflammatory drugs, metabolic disease, severe ischemic heart disease, alcohol abuse, cerebrovascular accident, and those who had received corticosteroids within the eight weeks before assessment. We used the Mini-Mental State Examination (MMSE) to assess cognitive function in both groups. Before starting the research, all participants or their primary caregivers gave written informed consent. Peripheral blood (5 mL) was collected from each participant into BD Vacutainer EDTA tubes using standard venipuncture techniques. Samples were immediately placed on ice and processed within two hours to minimize RNA degradation. Blood samples were centrifuged at 10,000× g for 30 s at 4 °C using a Hettich ROTINA 46R centrifuge. Plasma was carefully separated from the buffy coat and red blood cells and transferred to Maxwell cryogenic tubes. Aliquots (1 mL) of plasma were stored at − 80 °C until RNA extraction.

Expression assays

To conduct the expression assay, total RNA was extracted from the samples using the Hybrid-RTM Blood RNA extraction kit (GeneAll, Seoul, South Korea). The RNA was then treated with DNase I to remove any DNA contamination. The quantity and quality of the isolated RNA were assessed using NanoDrop technology from Thermo Scientific (Wilmington, DE). Additionally, cDNA synthesis was performed according to the manufacturer’s instructions using the cDNA synthesis Kit from GeneAll. cDNA was stored at −20 °C. Primer sequences were initially designed using Primer3 and Primer-BLAST for ATAD1 and HOMER1, targeting exon-exon junctions to enhance specificity. These sequences were then manually adjusted to optimize parameters such as melting temperature and GC content to ensure high specificity and efficiency. Subsequent analysis using (Oligo 7 and GeneRunner) confirmed the suitability of the final primer sequences. sRNAPrimerDB was used for miR-361-5p and U6 (reference gene). All primer sequences are listed in Table 1.

Table 1.

Sequences of primers used in this research.

Gene name	Primer sequences	Product size	Tm
ATAD1	Forward primer: ACATGCTGAAGCCTTTTCTCGT Reverse primer: AGAGGGTCTACAAGATGAGCAG	235 nt	60
HOMER1	Forward primer: TGGAAAAGCAACTGCGAAATAGC Reverse primer: CTGAAGATAGGTTGTTCCCCCAT	234 nt	60.5
UBC	Forward primer: CAGCCGGGATTTGGGTCG Reverse primer: CACGAAGATCTGCATTGTCAAGT	72 nt	60
has-miR-361-5p	RT primer: GTCGTATCCAGTGCAGGGTCCGAGGTATTCGCACTGGATACGACGTACCC
	Forward primer: AAGCGCCTTTATCAGAATCTCC	–
	Reverse primer: GTCGTATCCAGTGCAGGGT		60
U6	RT primer: GTCGTATCCAGTGCAGGGTCCGAGGTATTCGCACTGGATACGACAAAAATAT
	Forward primer: GCTTCGGCAGCACATATACTAAAAT	–
	Reverse primer: CGCTTCACGAATTTGCGTGTCAT		60

We normalized miRNA and mRNA levels using U6 and ubiquitin C (UBC) as internal controls, respectively. qPCR was performed using the Step OnePlusTM Real-Time PCR and RealQ Plus2x Master Mix (Ampliqon, Odense, Denmark). All qPCRs were done in duplicate. Real-time PCR thermal cycling parameters are detailed in Table 2.

Table 2.

Real-time PCR thermal cycling parameters.

Parameter	miR-361 and U6	ATAD1, HOMER1 and UBC
Initial Denaturation	94 °C for 10 min	95 °C for 15 min
Number of cycles	40	40
Denaturation	94 °C for 15 sec	95 °C for 30 sec
Annealing	64 °C for 30 sec	60 °C for 30 sec
Extension	64 °C for 30 sec	72 °C for 30 sec

Statistical analysis for quantitative polymerase chain reaction (qPCR)

To analyze the data, we used R version 4.x, including the brms, stan, pROC, and GGally packages. Relative gene expression levels were calculated using the 2^− ΔΔCt method. Specifically, ΔΔCt values were calculated based on Ct values obtained from real-time PCR. Ct values for target genes (ATAD1, HOMER1, and miR-361) were normalized to the Ct values of housekeeping genes (UBC for mRNAs and U6 for miRNAs). Bayesian regression models were then used to evaluate the association between these relative expression levels and AD status, considering age and sex as covariates. Adjusted p values below 0.05 were deemed statistically significant. We also analyzed expression levels across different age groups and genders. Spearman correlation coefficients were used to examine the relationships between study variables. Furthermore, we evaluated the diagnostic capabilities of the genes through receiver operating characteristic (ROC) curve analysis.

Bioinformatics analysis using the brain microarray dataset

Gene expression profile data collection

In this research, we used a bioinformatics technique to analyze data from a microarray dataset of human temporal cortex tissue samples with differing levels of AD-related pathology (GSE106241). Our goal was to identify any changes in the expression of ATAD1 and HOMER1 in these tissue samples. The gene expression profile shown above was obtained from the NCBI Gene Expression Omnibus dataset (GEO)³². The dataset was generated using a chip-based platform called GPL24170, Agilent-044312, Human 8 × 60 K Custom Exon array (Probe Name version). GSE106241 contains samples from 60 human temporal cortical (TC) tissue samples, categorized into seven groups based on Braak staging (Braak 0: n = 6, Braak I: n = 11, Braak II: n = 11, Braak III: n = 6, Braak IV: n = 7, Braak V: n = 12, Braak VI: n = 7) to represent the disease severity³³. Braak staging categorizes AD progression into three stages: early, middle, and advanced, based on the spread of neurofibrillary tangles (NFTs) in the brain’s memory circuit. Stage 0 indicates no NFTs, stages I-II show NFTs in the entorhinal-perirhinal cortex, stages III-IV involve NFTs in the hippocampus, and stages V-VI exhibit NFTs in various neocortical areas³⁴.

Differentially expressed genes (DEGs) identification and data preprocessing

To correct the background for the data analysis, we applied the method described by Silver et al.³⁵, using an offset of 15 and the normexp technique. We also normalized the data between arrays using the quantile procedure from the linear models for microarray data (limma) R package³⁶. We only considered spots with a signal minus background that were flagged as “positive and significant” (field name gIsPosAndSignif) and not flagged as ControlType or IsManualFlag. We used the AgiMicroRNA Bioconductor package (version 2.40.0) to assess the quality. Bioconductor’s Differentially Expressed Gene Analysis (DEGA) compared AD samples to normal samples using the limma R package for linear models in microarray data analysis^36,37. To identify statistically significant genes, the Student’s t-test was utilized. The appendix established a cut-off value for aberrantly expressed mRNAs: a false discovery rate (adjusted P value) of < 0.01³⁸ and a log2 fold change (log2FC) of > 0.5. Visualization of DEGs was done through a volcano plot, while the expression patterns of ATAD1 and HOMER1 genes were shown in a heat map using Pheatmap (version 1.0.12) and Enhanced Volcano (version 1.8.0) R packages.

Identification of miR-361 expression levels associated with Alzheimer’s disease pathogenesis

To pinpoint the miRNAs linked to AD development, we employed the same bioinformatics approach as before. We obtained data on miRNA profiles from the NCBI GEO database for GSE157239. The miRNA-4Array platform (GPL21572) was used for this dataset. GSE157239 included samples from eight healthy controls and eight AD patients (Braak stage III or higher). The affy Bioconductor package includes a useful tool called the Robust Multichip Average (RMA), which is used to background correct and normalize raw data files³⁹. We used the interquartile range filter to narrow down the number of genes analyzed. This was followed by an intensity filter to remove probe sets with low expression signals (less than 100) across at least 25% of the arrays. This step aimed to eliminate insignificant genes that are either not expressed or not changing⁴⁰. We utilized the AgiMicroRNA Bioconductor package to assess the quality of the data. DEGA was conducted on both healthy and AD samples using the limma R package³⁶ in Bioconductor³⁷. The miRNA names were converted to miRbase v22 using the miRNAmeConverter Bioconductor package⁴¹. To identify miRNAs with statistical significance, we used Student’s t-test. We considered miRNAs as abnormally expressed if they met the following criteria: a false discovery rate (adjusted p value) < 0.01⁶ and |log2 fold change (log2FC)|> 0.1³⁸. We used the R package Pheatmap to create the DEmiRNA heatmap.

Correlation analysis between ATAD1 and HOMER1

We used Pearson’s correlation analysis to see if there were any positive correlations between ATAD1 and HOMER1. We used the Hmisc and psych packages for analyzing and visualizing the correlations.

Results

Reanalysis of brain microarray datasets

Differentially expressed genes identification

DEGA, background correction, batch modification, gene filtering, and normalization were all completed successfully. The quality was assessed using the AgiMicroRNA Bioconductor package. To analyze the data distribution post-normalization, box plots of the gene expression dataset were used. The box plots showed that the arrays had similar median expression levels, indicating that the adjustments were made correctly. Based on the findings, ATAD1 and HOMER1 genes were both significantly downregulated. The expression levels of these genes in Braak stages I-V are detailed in Table 3. Criteria for identifying differentially expressed genes (DEGs) included an adjusted p value of less than 0.01 and an absolute log2 fold change of greater than 0.5. In Fig. 1, a volcano plot of DEGs is displayed, along with a hierarchical clustering heatmap of ATAD1 and HOMER1 genes in Braak stage V.

Table 3.

Expression levels of ATAD1 and HOMER1 in TC tissue samples.

Gene symbol	Log2FC	AveExpr	t-test	P value	adj.P. val	B-statistic
ATAD1	− 0.790451482	10.29545762	− 12.00528071	8.36E − 15	1.56E − 13	23.50925357
HOMER1	− 1.688416312	6.267508418	− 12.99989488	6.60E − 16	1.86E − 14	26.05727925

Fig. 1.

Differentially expressed mRNAs in brain samples from patients with Alzheimer’s disease in Braak stage V compared to control samples. (A) See the gene-specific heatmaps for the ATAD1 and HOMER1 genes. Genes with high expression levels are in red, while those with low expression levels are in blue. (B) Look at the volcano plot for DEGs. We identified DEGs using a |(log2FC)|> 0.5 and an adjusted p value of 0.01.

Identification of miR-361 expression levels in AD

EGA analysis, gene filtering, normalization, batch adjustment, and background correction were conducted. The quality was monitored using the Bioconductor package. Box plots were generated post-normalization to evaluate the data distribution of gene expression profiles. Similar expression level medians in different arrays indicated successful adjustment. miR-361 was found to be down-regulated (log2FC = -0.10152, adj.P.Val = 0.000129) based on the criteria of adjusted p value < 0.001 and |log2 fold change (log2FC)|> 0.1 among DEmiRNAs. Figure 2 demonstrates a DEmiRNA volcano plot.

Fig. 2.

Volcano plot for the DEmiRs. The DEmiRs were screened based on a | (log2FC) |> 0.1 and an adjusted p value < 0.001.

Correlation analysis between ATAD1 and HOMER1

Pearson correlation analysis was performed to validate the correlation between ATAD1 and HOMER1 in regulating glutamatergic receptors. A significant positive correlation was discovered between the expression levels of these genes (r = 0.93, P < 0.001) (Fig. 3).

Fig. 3.

The distribution of each variable is shown on the diagonal. The lower portion of the diagonal shows bivariate scatter plots with a fitted line. On the upper part of the diagonal, the correlation coefficients plus the significance level as stars are displayed. *** is a significant correlation at P value < 0.001.

Identification and differential analysis of ATAD1/HOMER1/hsa-miR-361-5p expression level in peripheral blood

General demographic data

A total of 50 AD patients (31.4% male, 68.6% female) with an average age of 76.36 ± 6.26 and 50 healthy controls (30.6% male, 69.4% female) with an average age of 74.3 ± 6.22 were included in the study. The mean MMSE scores were 14.2 ± 5.94 for the patients and 27.5 ± 0.76 for the control group.

Expression assays

Figure 4 displays the relative expression levels of ATAD1, HOMER1, and hsa-miR-361-5p genes in AD patients and controls.

Fig. 4.

Expression of HOMER1, ATAD1, and hsa-miR-361-5p in cases and controls’ peripheral blood samples. Values are depicted as gray dots. Means of expression levels and interquartile range are displayed.

The expression levels of ATAD1 showed no significant differences in peripheral blood (PB) samples between AD patients and healthy controls (adjusted p value = 0.695) or within subgroups. Similarly, there were no significant differences in HOMER1 expression levels in PB samples between AD patients and healthy controls (adjusted p value = 0.810) or within subgroups. The expression of hsa-miR-361-5p was significantly lower in PB samples from patients with AD compared to controls (posterior beta = − 0.613, adjusted P value < 0.043). Both male and female subgroups showed a similar decrease in expression levels (posterior beta = − 0.605, adjusted p value = 0.047 for males and posterior beta = − 0.611, adjusted p value = 0.046 for females). Tables 4, 5 and 6 display the full data for the relative expressions of ATAD1, HOMER1, and hsa-miR-361-5p.

Table 4.

Relative levels of ATAD1 in AD cases and controls according to the Bayesian quantile regression model.

	ATAD1	Posterior beta of (2(−ddct))ʎ	SE	Adjusted P value*	95% Crl for beta
Total	Group, case vs. control	0.107	0.29	0.695	[− 0.48, 0.67]
	Sex, female vs. male	0.091	0.26	> 0.999	[− 0.42, 0.6]
	Age (years)	− 0.008	0.01	> 0.999	[− 0.02, 0.01]
	Group * sex	− 0.306	0.36	0.249	[− 1.03, 0.46]
Male	Case vs. control	− 0.077	0.17	0.55	[− 0.42, 0.25]
	Age	− 0.008	0.01	0.683	[− 0.02, 0.01]
Female	Case vs. control	− 0.075	0.17	0.669	[− 0.41, 0.26]
	Age	− 0.008	0.01	> 0.999	[− 0.02, 0.01]

*Estimated from frequentist methods; CrI: Credible interval, ʎ: Power transformation value estimated from Box-Cox or Yeo-Johnson methods.

Table 5.

Relative levels of HOMER1 in AD cases and controls according to the Bayesian quantile regression model.

	HOMER1	Posterior beta of (2(−ddct))ʎ	SE	Adjusted P value*	95% Crl for beta
Total	Group, case vs. control	− 0.2	0.27	0.810	[− 0.78, 0.34]
	Sex, female vs. male	− 0.061	0.25	0.895	[− 0.59, 0.42]
	Age (years)	− 0.007	0.01	0.871	[− 0.02, 0.01]
	Group * sex	0.245	0.33	0.848	[− 0.4, 0.91]
Male	Case vs. control	− 0.041	0.16	0.856	[− 0.35, 0.28]
	Age	− 0.006	0.01	0.812	[− 0.02, 0.01]
Female	Case vs. control	− 0.046	0.16	0.601	[− 0.36, 0.28]
	Age	− 0.006	0.01	0.848	[− 0.02, 0.01]

*Estimated from frequentist methods; CrI: Credible interval, ʎ: Power transformation value estimated from Box-Cox or Yeo-Johnson methods.

Table 6.

Analysis of miR-361 levels in AD cases compared to controls using a Bayesian quantile regression model.

	miR-361	Posterior beta of (2(−ddct))ʎ	SE	Adjusted P value*	95% Crl for beta
Total	Group, case vs. control	− 0.613	0.2	0.043	[− 1.002, − 0.22]
	Sex, female vs. male	− 0.293	0.28	0.12	[− 0.82, 0.25]
	Age (years)	− 0.007	0.01	0.439	[− 0.02, 0.01]
	Group * sex	− 0.176	0.4	0.807	[− 0.95, 0.6]
Male	Case vs. control	− 0.605	0.19	0.047	[− 0.98, − 0.24]
	Age	− 0.01	0.01	0.818	[− 0.03, 0.01]
Female	Case vs. control	− 0.611	0.19	0.046	[− 0.98, − 0.23]
	Age	− 0.01	0.01	0.409	[− 0.03, 0.01]

*Estimated from frequentist methods; CrI: Credible interval, ʎ: Power transformation value estimated from Box–cox or Yeo–Johnson methods.

Significant values are in bold

Correlation analysis

The expressions of ATAD1, HOMER1, and hsa-miR-361-5p were not found to be correlated with the age of patients. However, in patients with AD, the expression of ATAD1 was positively correlated with HOMER1 (r = 0.599, P < 0.001) and with miR-361 (r = 0.577, p < 0.001). Additionally, the correlation between HOMER1 and miR-361 was significant (r = 0.540, p < 0.001), with higher values observed in patients compared to control samples (Figure 5).

Fig. 5.

The distribution of variables is depicted on the diagonal. The correlation coefficients, plus the significance level as stars, are shown. *, **, and *** is significant correlation at P < 0.05, P < 0.01, and P < 0.001, respectively.

Receiver operating characteristic (roc) curve analysis

We evaluated the ATAD1, HOMER1, and has-miR-361-5p for their ability to differentiate between AD patients and controls. Our analysis showed a strong diagnostic value of 0.706 for has-miR-361-5p at the transcript level, as indicated by the area under the curve (AUC) (Fig. 6).

Fig. 6.

Receiver operating characteristic (ROC) curve analysis. (A) miR-361 transcript levels displayed a diagnostic power of 0.706.

Discussion

Recent studies have revealed a significant link between glutamatergic signaling and neurodegenerative diseases, such as AD^42,43. Research suggests that AD could result from the excessive activity of glutamatergic neurons, leading to excitotoxic damage in cortical afferent neurons^44,45. This indicates that focusing on this pathway could be a beneficial treatment choice for these conditions⁸.

Exploring the regulatory mechanisms of gene expression in AD would help to identify potential molecular targets and key factors influencing disease progression. This in-depth analysis allows us to pinpoint specific regulatory processes involved in driving the advancement of the disease. In this study, we investigated the expression levels of ATAD1, HOMER1, and miR-361-5p in the context of AD using blood samples from affected patients and bioinformatics analysis.

HOMER1 and ATAD1 are vital proteins that play essential roles in the brain’s synaptic architecture and function. HOMER1 is a scaffolding protein that contributes to the organization of the postsynaptic density by linking glutamate receptors, such as AMPARs and NMDARs, to intracellular signaling pathways^14,15. AMPARs and NMDARs are critical for synaptic transmission, synaptic plasticity, and overall neuronal communication^20,24. The proper functioning of these receptors is crucial for learning, memory, and cognitive processes, which are often compromised in neurodegenerative diseases such as AD^17,18.

Our bioinformatics analysis revealed that the expression levels of HOMER1 were significantly lower in TC samples from individuals with AD compared to the control group. This might reflect an underlying pathology in AD, as reduced HOMER1 levels could lead to over-internalization of receptors and subsequent synaptic loss. However, HOMER1 levels did not display significant differences in PB samples in our study.

The absence of significant differences in HOMER1 and ATAD1 expression in PB samples limits their utility as peripheral biomarkers. This finding suggests that these genes may be more relevant in brain tissue rather than blood. Further investigations in cerebrospinal fluid (CSF) or extracellular vesicles may provide better insights into their biomarker potential.

Previous research has not explored the expression of HOMER1 RNA variants in the human AD temporal cortex. A study showed that HOMER 2 and HOMER 3 specifically interact with APP, inhibiting its processing and reducing the secretion of Aβ peptide¹⁸. Also, other researchers indicated that the accumulation of Aβ in specific synapses can lead to the binding of glutamatergic receptors. This, in turn, triggers the internalization and disruption of the AMPA and NMDA signaling pathways²⁵.

ATAD1 is known to play a role in the endoplasmic reticulum and is involved in the regulation of various synaptic proteins, including glutamate receptors²³. By promoting the trafficking and stability of AMPARs and NMDARs at the synapse, ATAD1 contributes to the maintenance of effective synaptic signaling⁴⁶.

We showed that the expression of ATAD1 was decreased in TC samples from AD patients compared to controls. However, we could not find any difference in its expression in PB samples of patients and controls. The latter finding raises questions about its reliability as a peripheral biomarker for AD, yet its potential dysregulation in TC samples could indicate impaired receptor trafficking, leading to altered synaptic transmission and cognitive deficits.

Furthermore, our investigation highlighted the regulatory role of miR-361 as a miRNA that may influence the expression of both ATAD1 and HOMER1. The relationship between miR-361 and these genes underscores the complexity of gene regulation in the context of neurodegenerative diseases, indicating that alterations in miR-361 levels could have downstream effects on synaptic functioning and inflammation associated with AD.

The divergence between our real-time PCR findings in PB samples and the results of bioinformatics analyses in TC samples underscores the diverse patterns of gene regulation in PB versus CNS. Such discrepancies may highlight the compartmentalized nature of gene expression and the adaptation of genes to the specific biological environment. It also emphasizes that PB may not capture the intricate changes occurring in the brain relevant to neurodegeneration.

Continued research into the molecular interactions involving ATAD1, HOMER1, and miR-361 is warranted, as a deeper understanding of these relationships could lead to the identification of novel therapeutic targets and biomarkers essential for early diagnosis and intervention in AD.

Finally, the miR-361 gene has a significant ROC value, indicating its potential as a diagnostic marker for AD. Decreased expression of miR-361 in both the brain and PB is consistently found in AD patients. Further research in this area could help address this aspect of the disease.

Conclusion

This study elucidates the significance of ATAD1, HOMER1, and miR-361 in the context of AD. Our findings suggest that ATAD1 and HOMER1 play critical roles in regulating synaptic function and integrity, particularly in glutamatergic signaling pathways. The downregulation of these proteins in TC samples may indicate their involvement in the neuroinflammatory processes that characterize AD, emphasizing their potential as biomarkers for understanding the disease pathology.

Furthermore, our investigation suggested the regulatory role of miR-361 as a miRNA that may influence the expression of both ATAD1 and HOMER1. The relationship between miR-361 and these genes underscores the complexity of gene regulation in the context of neurodegenerative diseases, indicating that alterations in miR-361 levels could have downstream effects on synaptic functioning and inflammation associated with AD. Future studies are warranted to explore the function of these genes in different biological contexts, particularly in neural tissues, to better understand their contributions to AD and their potential as biomarkers for early diagnosis and therapeutic intervention.

Limitation

Our research has some limitations that need to be acknowledged. First, factors like different methods, data analysis techniques, platforms used, sample preparation methods, and patient characteristics can all influence gene expression patterns. Second, we acknowledge that our sample size is relatively limited, which may affect the generalizability of our conclusions. However, our study prioritized stringent inclusion/exclusion criteria to minimize confounding factors, and we employed rigorous statistical methods to maximize the reliability of our findings. Future studies with larger cohorts will be necessary to validate and extend our results. Additionally, our bioinformatics analysis requires confirmation through experimental techniques. In vitro or in vivo validation of miR-361’s regulatory effects on HOMER1 and ATAD1 (e.g., luciferase assays, miR-361 knockdown/overexpression) would strengthen our mechanistic conclusions. Due to resource constraints, we focused on bioinformatic predictions and initial RT-PCR validation. Besides, although we excluded major comorbidities, we did not perform multivariate adjustments for factors like education or family history. Finally, the MMSE alone is insufficient for definitive AD staging. While our study primarily used MMSE for cognitive assessment due to data availability, we acknowledge that incorporating additional tools (e.g., Clinical Dementia Rating, Montreal Cognitive Assessment) or clinical biomarkers (e.g., Aβ42, p tau) would enhance diagnostic precision.

Author contributions

AJ and MR wrote and revised the draft. AJ and MT conducted the experiments. SA-J and JG analyzed the data. SGF edited the manuscript and contributed to the study design. All authors reviewed and approved the final version.

Funding

The research protocol received approval and support from the Molecular Medicine Research Center at Tabriz University of Medical Sciences (grant no. 64670).

Data availability

The datasets generated and/or analysed during the current study are available in the Agilent repository (https://www.agilent.com/en/human-gene-expression-microarrays-details-specifications).

Declaration

Competing interests

The authors declare no competing interests.

Ethics approval and consent to participate

Human participants in the study received approval from the Tabriz University of Medical Sciences Clinical Research Ethics Committee (Ethical code: IR.TBZMED.REC.1398.1265). All patients/participants provided written informed consent to participate in the study. All methods were performed by relevant guidelines and regulations.

Informed consent

Informed consent has been obtained from all patients.