Connecting HTT intermediate alleles and microRNA dysregulation to enhanced tauopathy in Late-Onset Alzheimer's Disease

doi:10.21203/rs.3.rs-7621820/v1

Download PDF

Research Article

Connecting HTT intermediate alleles and microRNA dysregulation to enhanced tauopathy in Late-Onset Alzheimer's Disease

https://doi.org/10.21203/rs.3.rs-7621820/v1

This work is licensed under a CC BY 4.0 License

You are reading this latest preprint version

Background

Late-onset Alzheimer´s disease (LOAD) is a heterogenous disorder influenced by genetic factors. In fact, we have previously described intermediate alleles (IAs) in the huntingtin (HTT) gene as potential modifiers in around 6% of AD population. The caudate nucleus, the most affected region in Huntington's disease, is highly sensitive to these HTT CAG expansions, as they can induce epigenetic changes, including altered microRNA profiles. All this implies a potential source of gene expression deregulation, affecting disease onset and/or progression in LOAD patients with HTT IAs.

Methods

We investigated the impact of HTT IAs on LOAD progression and neuropathology using a comprehensive approach, genotyping HTT CAG repeats in 323 LOAD patients and 335 healthy controls and further performing histopathological and molecular analyses on caudate nucleus samples in a small subcohort (6 healthy controls, 14 LOAD non-HTT IA carriers, and 13 LOAD HTT IA carriers).

Results

HTT IAs carriers patients exhibited decreased survival after disease onset, suggesting accelerated progression. Histopathologically, while LOAD patients showed increased soluble HTT levels and altered tau pathology compared to controls, these changes were consistently and markedly exacerbated in HTT IA carriers, characterized by heightened diffuse HTT immunoreactivity, pronounced tau 3R isoform imbalance, and increased 3R tau-enriched ghost tangles. Interestingly, this pathological exacerbation was supported by alterations in key splicing factors, including decreased SRSF6 and increased FUS-SFPQ complex formation. Analysis of microRNA (miRNA) profiling in the caudate nucleus revealed not only a LOAD-associated miRNA dysregulation that was significantly amplified in HTT IA carriers, but also five HTT IA signature miRNAs (miR-100-5p, miR-218-5p, miR-27b-3p, miR-487-3p, and miR-9-3p). In silico analyses, supported by network modeling and direct target validation, demonstrated that altered miRNAs target components of the nuclear spliceosome machinery, such as SRSF family, along with MAPT and HTT genes, suggesting a direct link to the observed tau 3R/4R imbalance.

Conclusions

Our findings underscore that HTT IAs as critical players in LOAD progression through an intricate network involving miRNA-mediated dysregulation of splicing. Thus, identifying HTT IAs through routine blood genetic screening offers a practical, non-invasive biomarker for patient stratification, taking a step closer to personalized therapeutic strategies in LOAD.

proteinopathies

alternative splicing

prognosis

post-transcriptional regulation

clinical heterogeneity

Alzheimer’s disease (AD), which accounts for 60–80% of dementia cases, is primarily characterized by the presence of extracellular beta-amyloid plaques and intracellular accumulation of tau protein in the form of neurofibrillary tangles (1). However, up to 50% of affected individuals present with mixed dementia (1) associated with the existence of co-pathologies involving the accumulation of proteins, such as alpha-synuclein or TDP-43 (2), especially in the more advanced stages of the disease (1, 2). While the precise cause of this phenomenon is still unknown, genetic factors like the APOE*ε4 allele have been associated with an increased risk of developing co-pathologies (1). In fact, this polymorphic variant introduces significant variability in AD development between carriers and non-carriers (3, 4), causing diverse clinical and phenotypic manifestations that hinder the application of optimal therapeutic and care interventions (1). Thus, identifying specific genetic risk factors for AD is essential for developing more appropriate, individualized treatment (5, 6).

Previous results from our laboratory suggest that there is a higher prevalence of intermediate alleles in the huntingtin gene (HTT) in AD patients with respect to control subjects (6.03% vs. 2.9%) (7), extending this fact to other tauopathies, such as frontotemporal dementia (FTD) (8, 9), and synucleinopathies (10). Exon 1 of the HTT gene presents CAG repeats in variable number, being considered normal below 27 repeats. When the number of repeats is above 35, it determines the age of onset of Huntington's disease (HD) (11), with incomplete penetrance when the range is between 36 and 39 repeats and complete when this number is higher (12). Between 27 and 35 repeats are referred to as intermediate alleles (HTT IAs), which are genetically unstable at the germline level, although they are not considered pathogenic (13).

However, in the case of AD, it has not yet been possible to determine the effect of the presence of HTT IAs. At the neuropathological level, an increase in diffuse HTT accumulation has been found in hippocampal regions associated with memory and in layer III of the frontal cortex in AD patients compared to healthy subjects (14), although in this case the number of CAG repeats in the HTT gene has not been evaluated. On the other hand, there are studies showing that elderly HD subjects show co-pathology with AD in up to 82% of cases, with prominent dementia in the clinic (15), and increased aggregation of other proteins, such as phosphorylated tau, alpha-synuclein, and TDP-43, can also be observed in the later stages of HD (16). It has even been proposed that HD is a 4R tauopathy, as alterations in the splicing of the MAPT gene have been described, modifying tau 3R/4R balance, and the absence of this gene in murine models of HD diminishes the characteristic motor alterations (17). Therefore, it is reasonable to hypothesize that HTT IAs in AD could also influence MAPT gene splicing, affecting the progression of AD pathology.

Studies developed in murine models describe changes in gene expression as a function of CAG repeat length (including intermediate ranges), being the striatum the most sensitive region (18), and such alteration modifies the expression of microRNAs (miRNAs) in this brain area (19). miRNAs are small non-coding RNAs (≈ 18–22 nucleotides) that are an integral part of the epigenetic regulatory network, since they can act as modulators that influence gene expression at the post-transcriptional level, including that of other genetic modifiers, such as DNA methyltransferases and histone acetylases (20). In addition, they themselves are subject to epigenetic regulation, as their expression profile can be altered under pathological conditions, as shown for both HD (21) and AD (22) or by lifestyle (23). Thus, it is possible that both AD and the presence of HTT IAs modify further the profile of miRNAs in carrier individuals, thereby affecting gene expression of key genes for disease progression.

Here, we investigated the impact of HTT IAs on the neuropathological features of late-onset AD (LOAD). At the clinical level, we found that the presence of HTT IAs is associated with accelerated disease progression. Building on this observation, we first hypothesized that HTT IAs influence LOAD pathology by altering MAPT gene splicing and, consequently, the tau 3R/4R balance. To investigate this, we focused on the caudate nucleus, demonstrating that its neurons present increased levels of soluble HTT and that HTT IAs further exacerbate this increase. Furthermore, LOAD patients carrying HTT IAs exhibited a pronounced imbalance of tau isoforms, characterized by an excess of 3R tau and a greater abundance of 3R tau-enriched ghost tangles, compared to non-carrier LOAD subjects. These findings were supported by lower levels of the splicing factor SRSF6 and increased formation of FUS–SFPQ nuclear complexes in the caudate nucleus. To elucidate the possible underlying molecular mechanisms driving these changes, we further hypothesized that the presence of HTT IAs modifies LOAD-associated miRNA profile in caudate nucleus. This was confirmed by the first detailed miRNA profiling study in this region. Importantly, the altered miRNAs target key components of the nuclear spliceosome machinery, suggesting a role in the observed tau 3R/4R imbalance. Moreover, the identified miRNA dysregulation represents a promising avenue for novel therapeutic targets in LOAD. Collectively, our findings underscore that HTT IAs serve as critical risk/progression biomarkers in AD, offering a practical, non-invasive genetic biomarker for clinical implementation, thereby facilitating patient classification and personalized therapeutic interventions.

Subjects

This study included LOAD patients (N = 323) and healthy controls (N = 335) from a previously described cohort (8). Detailed clinical data (diagnosis, gender, onset age, age at death, HTT gene genotyping) are summarized in Table 1. Based on caudate postmortem sample availability, a sub-cohort was obtained for LOAD non-carrier patients (N = 14), LOAD HTT IAs carrier patients (N = 13) and healthy controls (N = 6) for histological and molecular approaches, matched by demographic and neuropathological variables, which are summarized in Supplementary Table S1. Additionally, details of the subject samples used in each experiment are listed in Supplementary Table S2.

Table 1
Demographic data of cohorts studied.
Demographics	Control (N = 335)	LOAD (N = 323)	p-value
Gender (female %) ^a	188 (57.5%)	231 (71.51%)	< 0.001^c
Onset age^b	-	75.22 ± 6.02	N. A
Death age^b	-	83.25 ± 6.33	N. A
Final follow-up^b	70.84 ± 8.47	-	N. A
Disease duration^b	-	9.35 ± 4.86	N. A
HTT IAs carriers	14 (4.16%)	23 (7.12%)	0.102 ^d
HTT short allele^c	17 [16–17]	17 [17–18]	0.045^e
HTT long allele^c	18 [17–21]	20 [18–22]	< 0.0001^e
Abbreviations: LOAD, late-onset Alzheimer’s disease; N.A: not available. ^aData are shown as n (%). ^bData are shown as mean ± SD. ^cData are shown as median [IQR]. ^d Statistical analysis: Fisher's test. ^e Statistical analysis: Mann Whitney.

Material and data supply for patients were provided by the HCB-IDIBAPS Biobank (B.0000575), integrated in the platform ISCIII Biobanks and Biomodels, and by the Principado de Asturias BioBank (PT23/0077), member of the Spanish National Biobanks and Biomodels Network. Both materials and data were processed following standard operating procedures with the appropriate approval of the Ethics and Scientific Committees. The neuropathological LOAD diagnosis was performed according to standard international consensus criteria (24). All the participants or legal representatives gave written informed consent to participate in the study. All procedures have been approved by the Research Ethics Committee of the Principality of Asturias (CEImPA nº 2022.266).

Immunohistochemistry and immunofluorescence

For immunohistochemistry, formalin-fixed paraffin-embedded tissue sections (5 µm) were processed using the EnVision FLEX + kit (K8002, Agilent-Dako, Agilent, USA) and the Dako Autostainer. Epitope retrieval was performed at 95 ºC for 20 min in PT-LINK (Dako) with pH 6.0 buffer (K8005) for HTT and tau 3R, or pH 9.0 (58004) for tau 4R. Endogenous peroxidase was blocked and sections were incubated with primary antibodies in FLEX antibody diluent (K8006), for 20 min at room temperature and overnight at 4°C. Signal was developed with diaminobenzidine (DAB; DM827 + SM803) and counterstained with hematoxylin reagent (K8008). Negative controls were processed, omitting the primary antibody. All the primary antibodies used are summarized in Supplementary Table S3.

Images were acquired using the NanoZoomer-SQ Digital Slide Scanner (C13140-01, Hamamatsu Photonics, Germany) and analyzed with QuPath (version 0.4.2) and Fiji (ImageJ, version 1.6). All acquisition settings, including magnification (40×), were kept constant across experimental groups. For 3R-tau (05-803), 4R-tau (05-804), and HTT clone EM-48 (MAB5374; Merck Millipore, USA) quantification, the number of immunopositive neurons was manually counted in caudate nucleus sections from each patient. Five randomly selected regions of interest (ROI) per case were analyzed, each measuring 1 mm². Results were expressed as the mean number of positive neurons/mm². For HTT clone EPR5526 (ab109115; Abcam, UK), a fixed intensity threshold was applied using Fiji to quantify positive staining in five randomly selected 1 mm² ROIs per section. The signal intensity was then averaged across the five ROIs and expressed as mean signal intensity/mm².

For double immunofluorescence, after antigen retrieval, sections were washed with distilled water and subsequently rinsed three times with 0.1M phosphate buffer, followed by a rinse in PBS for 10 min. Tissue sections were then blocked for 45 min at room temperature in a blocking solution consisting of 1% of serum albumin and 1% Triton X-100 in PBS. Primary antibodies (see Supplementary Table S3) were diluted in blocking solution supplemented with normal serum overnight at 4°C. The following day, sections were rinsed with PBS and distilled water. Secondary antibodies (donkey anti-mouse Alexa Fluor 488 and donkey anti-rabbit Alexa Fluor 594, 1:1000; Invitrogen, Thermo Fisher Scientific, USA) were incubated for 2 hours at room temperature. Finally, nuclei were counterstained with DAPI, included in the mounting medium (Fluoroshield™, Sigma-Aldrich), and coverslipped for imaging.

Images were acquired on a Leica TSC-SP8X spectral confocal microscope (Leica DMI8 microscope) with excitation lines between 470–670 nm and PLA APO 20X/0.75 IMM CORR CS2 or PLA APO 40X /1.30 CS2 oil-immersion objective (Leica Microsystems, Germany), using Leica Application Suite X software (version 1.8.1, Copyright 1997–2015; Leica Microsystems), and processed with LAS_X_SMALL (version 1.0.0) and Uniovi Fiji Confocal/ImageJ (version 1.6) software.

Nuclear colocalization analysis

To quantitatively assess the colocalization of FUS (A300-293A; Bethyl Laboratories, USA) and SFPQ (WH0006421M2, Sigma Aldrich, USA) within neuronal nuclei, all sections were imaged under identical confocal microscope settings and laser intensities. Random images were acquired from two distinct regions of the caudate nucleus, capturing six Z-planes (1 µm step size) to cover the full nuclear volume at 63× magnification. Neuronal nuclei were identified based on size (~ 10–12 µm in diameter), and those smaller than 8 µm were excluded from analysis. Colocalization analysis was conducted on 30 randomly selected nuclei per section using the Coloc 2 plug-in in FIJI. Thresholded Manders’ coefficients (tM1 and tM2) [56], along with Pearson’s correlation coefficient (R) (25), were computed to quantify colocalization and assess the correlation between fluorescence signals.

Proximity ligation assay

The proximity ligation assay (PLA) for protein complexes detection was carried out as previously described (26). Briefly, paraffin-embedded brain sections were rehydrated and subjected to antigen retrieval in citrate buffer (pH 6.0) at 95°C for 20 min. Following permeabilization with 0.1% Triton X-100 in PBS for 15 min, sections were blocked for 45 min using the Duolink blocking solution (DUO92009, Sigma-Aldrich). Samples were then incubated overnight at 4°C with primary antibodies diluted in antibody diluent (see Supplementary Table S3).

After washing twice in 1× buffer A (DUO82049), sections were incubated for 1 h at 37°C with PLA probes PLUS and MINUS (DUO92002 and DUO92004, respectively; Sigma-Aldrich), each diluted 1:5 in antibody diluent. Ligation was carried out using ligase diluted 1:40 in ligation buffer (DUO92014) at 37°C for 30 min. Signal amplification was performed with polymerase diluted 1:80 in amplification buffer, incubated at 37°C for 100 min. Sections were then washed sequentially in 1× and 0.01× buffer B (DUO82049), mounted with DAPI-containing medium (DUO82040), and stored until imaging.

To assess PLA signal intensity, all sections were imaged under identical conditions using a confocal microscope (Leica SP8), with fixed laser power and acquisition settings. For each section, five random Z-stacks were acquired at 1 µm intervals, covering the entire nuclear volume at 63× magnification. Nuclei were selected based on size (10–12 µm diameter); nuclei smaller than 8 µm were excluded. Seventeen nuclei were randomly selected per section for quantification.

For each Z-stack, a maximum intensity projection was generated, and mean fluorescence intensity values (range: 0–255 grayscale units) were measured using the stack profile tool in LAS X software (Leica Microsystems), by manually outlining each nucleus. Fluorescence values were normalized to the nuclear area, and the average intensity per nucleus was calculated to yield a representative value per subject.

Western blotting

Human caudate tissue (30 mg) was homogenized in lysis buffer (20 mM HEPES pH 7.4, 100 mM NaCl, 50 mM NaF, 5 mM EDTA, 1% Triton X-100) with protease inhibitors (cOmplete™, Mini Protease Inhibitor Cocktail; Roche, Switzerland) in a glass homogenizer on ice. Subsequently, the samples were centrifuged for 10 min at 12 000 × g at 4°C and the supernatant was collected for analysis. Protein concentration was determined with the NanoDrop One 2000c spectrophotometer (Thermo Fisher Scientific, USA).

Fifty micrograms of total protein were loaded onto 4–12% SDS-polyacrylamide gels (M00653; GeneScript, USA) and transferred to PVDF membranes (0.2 µm, Amersham™ Hybond; Cytiva, USA), which were then blocked in TBS-T (Tris-buffered saline, 1% Tween-20) supplemented with 5% bovine serum albumin (BSA). Membranes were incubated with primary antibodies (see Supplementary Table S3), overnight at 4°C, and, after TBS-T washes, with HRP-conjugated goat anti-mouse IgG (1:20000, ab97040, Abcam, UK) or anti-rabbit IgG (1:20000, ab7090, Abcam) for 1h, at room temperature. Protein detection was performed with Amersham™ ECL™ Prime Western Blotting Detection Reagent (Cytiva). Anti-β-Actin (HRP) conjugated antibody (1:4000, sc-47778, Santa Cruz Biotechnologies, USA) was used as loading control. Western blot images were acquired with a ChemiDoc-It® BioChemi HR Camera (UVP) and quantified using ImageJ (Uniovi Fiji Confocal).

mRNA expression analysis by RT-qPCR

Frozen postmortem caudate nucleus samples (N = 33, approximately 5 mg each) were homogenized under cold conditions using Qiazol lysis reagent (Cat. No. 79306, QIAGEN, Germany). Total RNA was extracted using the miRNeasy Micro Kit (Cat. No 74004, QIAGEN) following the manufacturer’ protocol. RNA concentration and purity were determined on a NanoDrop One 2000c spectrophotometer (Thermo Fisher Scientific). RNA integrity was evaluated for each sample using the Agilent 2200 TapeStation system (Agilent Technologies) and the RNA integrity number (RIN) are provided in Supplementary Table S4. Samples were stored at -80°C until further processing.

For cDNA synthesis, 500 ng of total RNA were used for cDNA synthesis (StaRT reverse transcription kit, AnyGenes, France). The conditions used were: 10 min at 25°C, 120 min at 37°C, and 5 min at 85°C. Quantification was performed from cDNA using the Perfect Master Mix SYBR Green Kit (AnyGenes) on the 7900HT rapid real-time PCR system (Applied Biosystem, USA). The primers for the MAPT, MAPT 4R, and MAPT 3R genes analyzed were specifically designed by Eurogentec (Belgium). GAPDH and β-ACTIN were used to perform the internal normalization of the results, and the one with the most stable values for our dataset was chosen a posteriori, using the RefFinder tool (27). Primer information and amplification conditions are listed in Supplementary Tables S5 and S6. The mRNA levels are represented as relative quantification (2^−ΔΔCt), as described in (28).

small RNA-sequencing

Small RNA-Seq was performed by Seqplexing (Sequencing Multiplex S.L., Spain) using 200 ng of caudate nucleus total RNA per sample for miRNA library preparation. Library quality was assessed using the QIAxcel Advanced System (QIAGEN). Sequencing was performed on an Illumina NovaSeq X platform (Illumina, USA), generating paired-end reads (2 x 150bp). The sequencing depth ranged from 8.64 to 144 million reads per sample (median = 25.89 million).

Raw sequencing data were obtained in FASTQ format and quality control was performed using FastQC. Adaptor sequences were trimmed and miRTrace software (29) was used to classify the types of RNA present in each sample, distinguish between miRNAs, rRNA, tRNA and artifacts. miRNA counts were generated using miRDeep2 (30) and miRNA sequences were mapped to the human genome reference version hg38/GRCh38. Human miRNAs identification was performed using miRBase (v. 22) (31). All data generated and used in this study are publicly available in ZENODO (DOI: 10.5281/zenodo.15230070) and NCBI GEO DataSets (ID: GSE300433; The following secure access code has been created to allow review of record GSE300433 while it remains in private status: mtwfggcozpalrwn).

miRPM custom R package

A custom R package called miRPM was developed and used, integrating the entire bioinformatics workflow. In this package, the count matrices were normalized using the Reads Per Million (RPM) approach, based on the total number of reads per sample. To further assess the RPM normalization performance, the DANA approach was employed (32). DANA is an R-based method that evaluates the effectiveness of different normalization strategies in miRNA-Seq data by computing two metrics: concordance correlation coefficient (cc), which measures the preservation of biological signals, and mscr, which quantifies the reduction of handling effects. In our analysis, we incorporated the RPM normalization method into the DANA framework and made several modifications to improve the computation of the concordance correlation coefficient. These changes addressed issues like excessive miRNA filtering and strict cluster size constraints in the original DANA implementation, ensuring a more robust and interpretable evaluation of normalization methods while maintaining the biological integrity of the data. R code is available at https://github.com/sergio30po/miRNA-RPM-DE-Analysis.git.

Differential expression of miRNA sequencing data

For the differential expression analysis of miRNAs, using the miRPM package, an expression filter with two inclusion criteria, applied consecutively, were considered: 1) more than 50% of valid miRNA data in at least one group; 2) an expression level of > 1000 RPM in all subjects of at least one group. Before conducting the statistical analysis, the necessary assumptions were evaluated to determine the suitability of using non-parametric tests. Data normality was assessed using the Shapiro-Wilk test, and homoscedasticity between groups was evaluated using Levene’s test. Additionally, the coefficient of variation was calculated to evaluate the relative dispersion of the data. Since the assumptions of normality and homogeneity of variances were not met in most cases, non-parametric statistical tests were chosen. Thus, Kruskal-Walli’s test was employed to assess overall differences in the expression levels across the three groups. To control the error arising from multiple comparisons, FDR correction was applied. Only those miRNAs with a significant adjusted p-value were subsequently subjected to Dunn’s post-hoc test, which inherently accounts for the multiple pairwise comparisons, to determine which specific group comparisons were significant. Graphical representation of miRNAs levels in heatmap was performed by normalization through Z-score.

In silico prediction of target genes and pathway analysis of miRNAs

To explore the biological relevance of the differentially expressed miRNAs, enrichment analysis of signaling pathways was performed using DIANA-miRPath v.4, incorporating experimentally validated miRNA-gene interactions from miRTarBase v.8.0 (33). The analysis was conducted using both Kyoto Encyclopedia of Genes and Genomes (KEGG) and REACTOME repositories, and statistical significance was set at p-value < 0.05 with FDR correction.

Among the enriched pathways, particular attention was paid to the spliceosome, given its relevance to tau isoform regulation. For this pathway, validated gene targets associated with the miRNAs of interest were retrieved from miRTarBase v.8.0. The subset of genes involved in the spliceosome was then subjected to Gene Ontology (GO) analysis using PantherDB v.19.0 tool, to characterize enriched biological processes, molecular functions, and cellular components in which they are predominantly represented. GO terms meeting the significance threshold (p-value < 0.05) were retained. In silico graphs were created using the SRPlot module (34).

Finally, gene interaction networks were generated to visualize validated miRNA–target connections. Interaction data were imported into Cytoscape v3.10.3 for visualization and network analysis.

General statistical analysis

Statistical analyses were performed using IBM SPSS Statistic 27.0 software (IBM, USA). Categorical variables were described as frequencies and percentages. For between-group comparisons of the frequencies of HTT IAs, APOE alleles or gender frequency, Chi-square and Fisher tests were performed. To determine the relationship between the presence of HTT IAs and the probability of survival in the LOAD groups, the Kaplan-Meier estimation method and the log-rank test were used. Survival rate was calculated as the time elapsed from pathology diagnosis to age at death (event = 1 per subject). Additionally, multivariate Cox-PH regression models were run to identify predictors of survival risk, such as age or the presence of HTT IAs. Kolmogorov-Smirnov and Shapiro-Wilks normality tests (for N < 50) were performed and the corresponding parametric or nonparametric tests were then applied. Mean and standard deviation or median and interquartile range, as appropriate, were used to describe quantitative variables. To compare age at onset, age at death, disease duration, immunohistochemistry and immunofluorescence results between controls vs. LOAD group or between LOAD subgroups, Student’s t-test or Mann-Whitney U-test was performed, as appropriate. Also, the parametric analysis of variance (ANOVA) test, followed by Tukey's comparison, or the nonparametric Kruskal-Wallis’ test, followed by Dunn's test, was applied when comparisons were made considering the three groups. To determine whether miRNAs levels correlate with the recording of different clinical variables, a Spearman correlation test was applied. The threshold for statistical significance was set at p < 0.05. The exact levels of the p-values are indicated in each figure or additional table, if applicable. Graphs were created using GraphPad Prism 10.2.0 (GraphPad, USA).

The presence of HTT IAs modifies LOAD progression

We have previously analyzed the possible influence of the HTT IAs on the development of tauopathies, including AD, and our results suggested that LOAD patients had a higher frequency of HTT IAs than patients with early-onset AD (8). However, we have not established whether this could be affecting disease progression. Thus, to further explore the effect of HTT IAs on LOAD, we considered for this study only LOAD patients from our previous cohort and a group of healthy subjects (8). The age of LOAD onset was 75.22 ± 6.02 years and disease duration until death was 9.35 ± 4.86 years, with a higher percentage of women in the LOAD group versus the control (Table 1). The distribution frequency of the different APOE gene isoforms was consistent with previous reports (Table S7) (8). The presence of subjects carrying HTT IAs was higher in the case of LOAD patients versus healthy controls, with a frequency similar to that observed in previous studies (7.12% vs. 4.16%; p-value = 0.102; Fisher's test, Table 1) (7, 8). Moreover, the number of CAG repeats was significantly higher in the case of long HTT allele of LOAD subjects with respect to healthy controls (20 [18–22] vs. 18 [17–21], respectively; p-value < 0.0001; Mann Whitney; Table 1 and Figure S1B). Importantly, within the range of IAs in the LOAD group, the most frequent repeat was 27 CAGs, with subjects presenting even 35 CAGs (Figure S2).

To study the effect of the presence of HTT IAs specifically in LOAD donors and clinical variables, we divided this group into non-carrier (N = 300) and carrier (N = 23) patients (Table 2). The number of CAG repeats in the long HTT allele was higher in carriers versus non-carriers (28 [27–30] vs. 19 [18–22]; p-value = 0.001, Mann Whitney; Table 2). No significant differences were found in the percentage of females, onset age and death age between the two groups.

Table 2
Demographics data of LOAD cohort studied.
Demographics	LOAD non-HTT IAs (N = 300)	LOAD HTT IAs (N = 23)	p-value
Gender (female %)^a	216 (66.87%)	15 (65.21%)	0.488 ^d
Onset age^b	75.24 ± 5.98	75.75 ± 6.64	0.855 ^e
Death age^b	84.33 ± 6.31	83.43 ± 6.79	0.492^e
Disease duration^b	9.50 ± 4.96	7.40 ± 2.89	0.053^e
HTT short allele^c	17 [17–17]	17 [13–18]	0.065^f
HTT long allele^c	19 [18–22]	28 [27–30]	< 0.0001^f
Abbreviations: LOAD, late-onset Alzheimer’s disease ^aData are shown as n (%). ^bData are shown as mean ± SD. ^cData are shown as median [IQR]. ^dStatistical analysis: Fisher's test. ^eStatistical analysis: Student’s t-test. ^fStatistical analysis: Mann-Whitney.

However, the disease duration, after diagnosis, was shorter in carrier subjects with respect to non-carrier (7.40 ± 2.89 years vs. 9.50 ± 4.96 years, respectively; p-value = 0.053, Student’s t-test; Table 2). In fact, the survival rate after diagnosis showed a clear reduction of this rate in LOAD donors with HTT IAs (p-value = 0.0017; Kaplan-Meier estimator; Fig. 1). The disease duration was estimated based on the available clinical records of age at disease onset and age at death, allowing us to calculate survival time from onset to death. To verify this result, logistic regression models were used. The model proposed revealed a significant fit (Cox-Snell R² = 0.01) and a significant association between disease duration and the presence of HTT IAs (p-value = 0.012). These results suggest that the HTT IAs modifies LOAD progression, decreasing patient survival after disease onset.

Increased diffuse HTT protein levels in caudate neurons of LOAD patients are further exacerbated by HTT IAs

It has been previously described that HTT levels are increased in the neurons of hippocampus and frontal cortex in LOAD (14). To determine whether HTT IAs affect HTT total protein burden and its distribution in LOAD patients, a histopathological analysis was performed on caudate nucleus (Fig. 2). First, we explored an antibody that recognizes both wild-type and mutant HTT (Fig. 2A, EPR5526 clone). The results show significantly higher HTT intensity signal in the cytoplasm of caudate neurons in both LOAD HTT IAs non-carriers and carriers compared to healthy controls (p-value = 0.0002 and p-value < 0.0001 respectively, ANOVA followed by Tuckey´s test; Fig. 2B), as reported by Axenhus et al. (14). Notably, LOAD group carrying HTT IAs had an even higher intensity signal in HTT-positive neurons, compared to non-carriers (p-value < 0.0001, ANOVA followed by Tuckey´s test; Fig. 2B), which has not been described before. Given this, we next sought to ascertain the presence of intranuclear inclusions, a typical hallmark of HD pathology (Fig. 2B, EM-48 clone). However, HTT inclusions were only detected in HD samples used as positive controls. Interestingly, we observed cytoplasmic HTT EM-48 labeling in a subset of neurons, exhibiting a cytoplasmic pattern consistent with our previous analysis with HTT EPR5226. Quantitatively, both LOAD HTT IAs non-carrier and carrier groups showed a significant increase in these HTT-positive neurons compared to control subjects (p-value < 0.0001, ANOVA followed by Tuckey´s test; Fig. 2C). Furthermore, LOAD HTT IAs carriers displayed a higher number of HTT-positive neurons than non-carriers (p-value < 0.0001, ANOVA followed by Tuckey’s test; Fig. 2C) and HD subjects revealed a significantly greater increase compared to all three other groups (p-value < 0.0001, ANOVA followed by Tuckey´s test; Fig. 2C).

These results were consistent with previously published data (14), extending the regions in which there is an increase of HTT-positive neurons in the brains of LOAD patients. Moreover, this finding was further exacerbated in HTT IAs carriers, although no sign of HTT aggregation or inclusion formation was observed. All in all, the increased HTT levels observed in LOAD patients represent another component in the co-occurrence of proteinopathies (16) at the brain level. Furthermore, the presence of HTT IAs appears to amplify this phenomenon, potentially promoting a more disturbed neuronal environment that detrimentally modifies disease progression in these subjects.

HTT IAs increase tau 3R and 4R isoforms in the caudate nucleus of LOAD patients

An imbalance in tau 3R and 4R isoforms has been previously reported in the striatum and cerebral cortex of HD patients (17). Thus, we explored whether a disruption of the 3R/4R balance of tau isoforms could also be present in LOAD patients with HTT IAs.

We first analyzed mRNA expression levels of total MAPT and its MAPT 3R and MAPT 4R transcripts in the caudate nucleus of healthy controls and LOAD patients (Figs. 3A-B). No significant differences in total MAPT and MAPT 3R mRNA levels were observed among the three groups. However, MATP 4R mRNA levels were significantly higher in the non-carrier LOAD group compared to controls (p-value = 0.02, Kruskal-Wallis followed by Dunn´s test; Fig. 3B). Next, total tau protein and its 3R and 4R isoforms were analyzed by Western blot (Fig. 3C and Figure S3). Due to comprised integrity, control samples could not be reliable included in this protein analysis, thus comparisons were restricted to the LOAD groups. Total tau levels were significantly higher in LOAD carrier patients (p-value = 0.0006, Student´s t-test; Fig. 3D), with a trend toward higher levels of the 3R isoform in the HTT IAs carrier group (p-value = 0.06, Student´s t-test; Fig. 3E). No changes in tau 4R levels were detected between the LOAD groups (Fig. 3F).

To further evaluate the distribution and abundance of tau 3R and 4R isoforms, we performed immunohistochemical analysis of tau 3R- and 4R-positive neurons across different regions of the caudate nucleus in healthy subjects and in both LOAD groups (Fig. 3G). Tau 4R quantification revealed a higher number of tau 4R-positive neurons in LOAD patients compared to controls, although this increase was only statistically significant in the HTT IAs carriers (p-value = 0.0017, Kruskal-Wallis followed by Dunn´s test; Fig. 3I). Interestingly, both LOAD non-carrier and carrier groups presented a higher number of tau 3R-positive neurons compared to controls (Fig. 3H). Furthermore, LOAD subjects carrying HTT IAs exhibited a significantly more pronounced increase in tau 3R-positive neurons compared to LOAD non-carriers (p-value < 0.0001, ANOVA followed by Tuckey´s test; Fig. 3H). Collectively, our immunohistochemical findings demonstrate a complex modulation of tau 3R and 4R isoform levels and distribution within the caudate nucleus of LOAD patients. While both isoforms show elevated neuronal counts in LOAD patient groups, the significantly more pronounced increase of the 3R isoform in LOAD patients carrying HTT IAs strongly suggest that the presence of these alleles distinctively influences the tau 3R/4R balance.

Final stage neurofibrillary tangles predominate in LOAD patients with HTT IAs

In AD, both 3R and 4R tau isoforms are present in neurofibrillary tangles (NFTs), key elements in AD pathophysiology. This contrasts with other tauopathies like Pick’s disease, which predominantly features 3R tau, or Progressive Supranuclear Palsy (PSP), which mainly involves tau 4R (35–38).

NFTs begin as fibrillar bundles in neurons, evolve into mature tangles, and are externalized after neuronal death (39, 40). This progression appears to be unidirectional and correlates with the sequential predominance of tau isoforms. Ghost tangles, representing the final stage of NFTs degeneration (Fig. 4A), are primarily composed of the 3R isoform, suggesting a temporal shift from 4R to 3R tau during the course of AD pathology (40). Accordingly, we hypothesized that the increased number of 3R-positive neurons observed in LOAD patients carrying HTT IAs may reflect a more advanced stage of NFT maturation in these individuals.

We first analyzed the proportion of each stage of NFT maturation intra-group. Within non-carrier subjects, we observed a significantly higher number of pretangle-positive cells and mature tangles compared to ghost tangles (p-value = 0.024 and p-value = 0.004, respectively; Kruskal-Wallis followed by Dunn’s test; Fig. 4B). Conversely, within HTT IAs carrier patients, there was a significantly less pretangles regarding mature tangles (p-value = 0.0005; Kruskal-Wallis followed by Dunn’s test), with no significant differences observed in ghost tangles compared to other stages (Fig. 4B).

When comparing NFTs structures between LOAD groups, the primary differences emerged at the early and late stages of maturation. Specifically, LOAD patients with HTT IAs displayed significantly fewer pretangles (p-value = 0.002; Mann Whitney U-test) and a greater number of ghost tangles than the non-carrier group (p-value = 0.003; Student’s t-test; Fig. 4B). These findings suggest that the presence of HTT IAs is associated with faster neuropathological progression, affecting tau aggregation and indicating a more accelerated neurodegenerative process in LOAD patients carrying HTT IAs.

SRSF6 splicing factor is decreased in caudate nucleus of LOAD patients with HTT IAs

Given the observed increase in tau 3R isoform levels in LOAD HTT IAs patients, we explored the possibility of disturbances in the spliceosome pathway, as it has been well documented in AD and other neurodegenerative pathologies (41, 42). Within the factors involved in MAPT splicing, the serine/arginine splicing factor family (SRSF) stands out (43). In fact, increased levels of SRSF6 protein, which is involved in the inclusion of exon 10 of the MAPT gene, and potentially related to increased tau 4R isoform, have been previously described in HD patients (17).

In our study, we found no changes in SRSF6 mRNA expression between groups (Fig. 5A). However, SRSF6 immunoblot analysis in LOAD patients showed that this protein is decreased in HTT IAs carriers compared to non-carriers (p-value = 0.017; Student’s t-test; Figs. 5B-C; Figure S4A, B). These results suggest that downregulation of SRSF6 protein could be one of the factors contributing to the observed increase in tau 3R in LOAD HTT IAs patients.

We also conducted an exploratory analysis of mRNA expression levels for other members of the SRSF family known to be involved in MAPT splicing (43). The results showed that SRSF1 and SRSF9, both implicated in the inclusion of MAPT exon 10, had lower expression levels in LOAD patients with HTT IAs compared to controls (p-value = 0.01 and p-value = 0.02, respectively; Kruskal-Wallis followed by Dunn’s test; Figure S4B). Lower mRNA expression levels were also observed in two other family members involved in MAPT exon 10 exclusion, such as SRSF3 (although not significantly) and SRSF4 (p-value = 0.003; Kruskal-Wallis followed by Dunn’s test) in carrier patients. However, due to the low RIN levels of the RNA samples and the absence of the corresponding protein levels analyses for the specific splicing factors, these mRNA findings should be interpreted with caution.

In summary, our findings demonstrate a significant decrease in SRSF6 protein levels in LOAD patients with HTT IAs, which potentially contributes to the observed tau 3R isoform imbalance. While exploratory mRNA analyses suggest broader alterations in other SRSF family members, these observations require further validation, particularly at the protein level, to definitively ascertain whether the presence of HTT IAs induces a widespread alteration within the SRSF family.

Increased formation of nuclear FUS-SFPQ complexes in caudate neurons of HTT IAs carrier LOAD patients

Previous studies have demonstrated that the formation of a nuclear complex between fused in sarcoma (FUS) and the proline/glutamine splicing factor (SFPQ) plays a critical role in the regulation of MAPT pre-mRNA splicing, facilitating exon 10 (E10) exclusion through the assembly of an intranuclear dimer (44). This interaction is disrupted in several neurodegenerative diseases, including tauopathies, where it contributes to splicing defects and the aberrant expression of tau isoforms (25).

To elucidate the role of the FUS-SFPQ complex, and its possible modulation by the presence of HTT IAs, we assessed the levels of FUS and SFPQ in the caudate nucleus. While no differences were detected at the mRNA expression level (Figs. 6A–B), protein levels of both FUS and SFPQ were significantly elevated in LOAD patients carrying HTT IAs compared to non-carriers (p-value = 0.03 and p-value = 0.005, respectively; Student’s t-test; Figs. 6D–E; Figure S5). This upregulation of FUS and SFPQ protein levels in HTT IAs carriers suggested enhanced assembly and/or stability of the FUS–SFPQ complex, which, given its crucial role in MAPT splicing and E10 inclusion/exclusion (25), could profoundly influence LOAD pathogenesis. Therefore, we explored whether this elevation involved a change in FUS-SFPQ complex formation.

To assess complex formation, we investigated nuclear localization and colocalization of FUS and SFPQ (Fig. 6F). Quantitative analysis using Pearson’s correlation coefficient (R) revealed increased nuclear colocalization in LOAD HTT IA carriers, with no differences observed between controls and non-carrier patients (p-value < 0.0001 carriers vs. other groups; Kruskal-Wallis followed by Dunn’s test; Fig. 6G). These findings were further supported by Mander’s threshold colocalization coefficients (tM1 and tM2; Fig. 6H). For more validation at protein–protein interaction level, we employed proximity ligation assays (PLA; Figs. 6I–J). Quantification of the PLA signal demonstrated significantly higher interaction levels in LOAD HTT IA carriers than in the other groups (p-value < 0.0001 carriers vs. other groups; Kruskal-Wallis followed by Dunn’s test; Fig. 6J), consistent with enhanced FUS–SFPQ complex formation.

Collectively, these results provide compelling evidence that the presence of HTT IAs in LOAD patients promotes aberrant assembly of splicing regulatory complexes in the caudate nucleus. This mechanism may underlie the observed shift in tau isoform expression, reinforcing the emerging role of RNA-binding proteins in modulating tau pathology through alternative splicing dysregulation.

HTT CAG repeat size modulates caudate nucleus miRNA profiles in LOAD patients

Our previous findings have revealed that LOAD patients with HTT IAs presented a shift in tau isoform balance towards tau 3R and an increased burden of ghost tangles in the caudate nucleus, consistent with the lower survival rate observed in these patients. These histopathological changes could be due to alterations in splicing factors dynamics, as evidenced by reduced SRSF6 levels and increased FUS/SFPQ complex formation, providing a potential mechanistic link to the observed increase of tau 3R. Beyond alterations in splicing factor activity, post-transcriptional gene regulation by small non-coding RNAs, such as miRNAs, plays a critical role in the intricate molecular landscape of neurodegenerative diseases, including MAPT alternative splicing (45). Interestingly, impaired miRNA expression as a function of the number of CAG repeats in the HTT gene was observed in different brain regions in HD mouse models, with the striatum the most vulnerable area (19). Therefore, we hypothesized that the caudate nucleus of LOAD patients carrying HTT IAs might exhibit an altered miRNA profile.

We performed small RNA-Seq on postmortem caudate nucleus samples from a subset of individuals within the study cohort, carefully selected based on demographic factors such as sex and age at death to minimize potential confounding effects, and Braak stage in the case of LOAD patients to account for disease progression (see Table S2 for details). Initial analysis identified a total of 1953 miRNAs with at least 1 RPM in at least one subject. To focus on consistently expressed miRNAs, we applied a first inclusion criterion requiring expression in at least 50% of the subjects within at least one of the groups. This reduced the database to 1187 miRNAs. Subsequently, to prioritize highly abundant miRNAs likely to have a greater biological impact, we applied a second inclusion criterion, preserving those miRNAs that exhibited an expression level > 1000 RPM in all subjects within at least one of the groups, retaining 39 miRNAs (Figure S6A).

To further assess the robustness of our normalization strategy and the biological integrity of the data, we employed the DANA approach (32), incorporating RPM normalization method into its framework. This analysis demonstrated that our RPM normalization effectively preserved biological signals and mitigated handling effects, providing robust and interpretable evaluation of the miRNA-Seq data (Figure S6B).

Based on this filtered set, differential expression analysis was performed using our custom R package, miRPM, which integrates the entire bioinformatics workflow for miRNA-Seq data analysis. From the retained 39 miRNAs, we performed a principal component analysis (PCA; Figure S6C), from which PC1 and PC2 components explained 58.9% of the variance observed in the subjects, exhibiting a consistent distribution pattern. In the PCA plot shown in Figure S6C, the control group appeared separated from the LOAD groups along PC1 and PC2, with an observable gradient within the LOAD group, in which HTT IAs carriers tended to position further along these components compared to non-carriers. However, despite these discernible trends, the overall separation was not sufficient to establish a clear distinction between the three groups solely based on PCA.

The analysis among the three groups revealed that 26 of the 39 miRNAs that passed the screening process were significantly altered (Fig. 7A and Table 3). Of these 26 miRNAs, all of them exhibited higher expression levels in LOAD HTT IA carriers compared to control group, while 21 showed significant differences between controls and LOAD non-carriers (Table 3). Thus, specifically, miR-100-5p, miR-218-5p, miR-27b-3p, miR-487b-3p, and miR-9-3p were differentially expressed regarding controls in LOAD HTT IAs patients. On the other hand, comparison between the two LOAD groups showed that 14 miRNAs were more overexpressed in HTT IA carriers than in non-carriers. Collectively, these findings indicate that the miRNA profile is significantly affected by the AD pathology itself, and that HTT IAs further amplifies these alterations.

Table 3
Differentially expressed miRNAs in healthy controls vs. LOAD groups.
miRNAs	Control	LOAD non-HTT IAs [Adjusted p-value vs. Control]	LOAD HTT IAs [Adjusted p-value vs. Control \| LOAD non-HTT IAs]
miR-99b-5p	706.99 ± 284.88	1213.29 ± 232.88 [0.0170]	1682.73 ± 332.49 [< 0.0001 \| 0.0024]
miR-9-5p	16386.6 ± 7417.38	32643.36 ± 5013.56 [0.0096]	39014.65 ± 5207.62 [< 0.0001 \| 0.0076]
miR-30d-5p	2819.36 ± 1657.27	5793.21 ± 1134.76 [0.0084]	7545.85 ± 1484.60 [< 0.0001 \| 0.0102]
miR-128-3p	4182.43 ± 2453.05	7897.71 ± 1813.42 [0.0228]	11928.67 ± 3091.31 [< 0.0001 \| 0.0026]
miR-23b-3p	626.34 ± 269.65	1428.61 ± 284.35 [0.0068]	1691.89 ± 213.97 [< 0.0001 \| 0.0174]
miR-191-5p	746.22 ± 424.02	1596.30 ± 297.25 [0.0083]	1932.93 ± 390.83 [< 0.0001 \| 0.0193]
miR-151a-5p	918.83 ± 460.58	1704.72 ± 375.72 [0.0127]	2079.86 ± 455.23 [< 0.0001 \| 0.0168]
miR-125a-5p	837.72 ± 400.65	2889.29 ± 859.72 [0.0021]	4390.30 ± 3066.51 [< 0.0001 \| n. s.]
miR-30a-5p	3941.43 ± 2223.77	7719.60 ± 1488.43 [0.0089]	9224.76 ± 1615.98 [< 0.001 \| 0.0282]
miR-487b-3p	889.57 ± 571.90	1477.67 ± 402.52 [n.s]	2171.85 ± 648.72 [0.0002 \| 0.0044]
miR-125b-5p	4169.52 ± 2147.22	11502.07 ± 3279.17 [< 0.0029]	13890.95 ± 4082.34 [0.0002 \| n. s.]
miR-221-3p	946.90 ± 466.80	1827.74 ± 641.08 [0.0052]	2250.74 ± 693.33 [< 0.0004 \| n. s.]
miR-139-5p	669.33 ± 319.52	1345.48 ± 420.42 [0.0043]	1698.56 ± 453.94 [0.0006 \| n. s.]
miR-103a-3p	3531.15 ± 2064.04	7974.13 ± 2293.12 [0.0034]	9017.74 ± 2049.48 [0.0008 \| n. s.]
miR-99a-5p	3093.67 ± 1188.20	5094.84 ± 1597.55 [0.0225]	6288.39 ± 1970.75 [0.0011\| n. s.]
miR-29a-3p	5212.68 ± 2435.92	9829.22 ± 2232.98 [0.0043]	10515.86 ± 3491.47 [0.0016 \| n. s.]
miR-218-5p	1495.30 ± 904.06	2005.07 ± 656.07 [n. s.]	2685.15 ± 433.87 [0.0017 \| 0.0024]
miR-126-3p	1127.99 ± 647.37	1832.21 ± 638.10 [0.0421]	2354.87 ± 704.51 [0.0019 \| 0.0398]
miR-100-5p	1946.82 ± 799.29	2979.13 ± 803.40 [n.s]	3582.45 ± 951.83 [0.0033 \| 0.0289]
miR-124-3p	3016.37 ± 1393.79	6776.95 ± 3470.01 [0.0053]	6340.28 ± 2815.46 [0.0033 \| n. s.]
miR-24-3p	1305.32 ± 611.40	2751.71 ± 557.81 [0.0012]	2583.71 ± 471.96 [0.0035 \| n. s.]
let-7a-5p	5551.12 ± 2778.08	9428.42 ± 1717.47 [0.0084]	10713.78 ± 3368.41 [0.0043 \| n. s.]
miR-181a-5p	3125.08 ± 1920.87	5420.40 ± 1409.52 [0.0428]	7066.36 ± 2796.03 [0.0049]
miR-30c-5p	1855.95 ± 964.69	3017.19 ± 536.48 [0.0148]	3246.25 ± 798.72 [0.001 \| n. s.]
miR-9-3p	8417.09 ± 6358.31	9191.86 ± 1803.44 [n.s]	12182.08 ± 1913.93 [0.0084 \| 0.0042]
miR-27b-3p	2717.82 ± 1559.54	3526.19 ± 808.69 [n.s]	4626.31 ± 1250.03 [0.0095 \| 0.0265]
Data are shown as mean ± SD. Statistical analysis: Kruskal-Wallis (miRNAs selected based on FDR-adjusted p-value < 0.05). Pairwise comparisons were performed using Dunn’s post-hoc test, with p-values adjusted for multiple comparisons. Ordered from highest to lowest significance according to LOAD HTT IAs.

To investigate whether altered miRNA expression is associated with clinical features in this genetic context of LOAD, we examined the relationship between the 14 differentially expressed miRNAs between the LOAD groups (Table 3) and key clinical variables. No significant correlations were observed with age at onset, age at death, disease duration, or Braak stage (Table S8). In contrast, CAG repeat size showed a positive, and significant, correlation with the expression of six miRNAs (Fig. 7B and Table S8): miR-128-3p (R = 0.48, p-value = 0.011), miR-99b-5p (R = 0.59, p-value = 0.001), miR-9-5p (R = 0.45, p-value = 0.017), miR-9-3p (R = 0.58, p-value = 0.001), miR-218-5p (R = 0.52, p-value = 0.005), and miR-27b-3p (R = 0.46, p-value = 0.015). Of these, the last three are among the five miRNAs that we have previously found to be specific to the presence of HTT IAs. These results suggest that the altered miRNA profile in the caudate nucleus is significantly influenced by HTT CAG repeat length, potentially contributing to the accelerated disease progression observed in LOAD individuals carrying HTT IAs.

In silico analysis reveals spliceosome pathway as a key target of dysregulated miRNAs in LOAD patients with HTT IAs

Since the LOAD-associated miRNA profile in caudate is even more impaired by the presence of HTT IAs, we next explored whether this miRNA pattern could be affecting genes specifically involved in the splicing of the MAPT gene. As a first approach to our in silico study, we performed an initial enrichment analysis using validated gene targets. This analysis revealed that the miRNAs differentially expressed between LOAD donors and controls, as well as those distinguishing the LOAD subgroups, were significantly enriched in 151 and 147 biological pathways, respectively. Notably, 13 of these KEGG pathways were directly associated with neurodegenerative processes (Figure S6D-E, Table S9 and S10). To further explore the molecular mechanisms potentially regulated by these miRNAs, we performed an additional enrichment analysis using the REACTOME encyclopedia. This analysis identified over 470 potentially modulated processes, among which ten stood out with highly significant relevance and were all related to RNA biology (Fig. 7C, Table S11), including mRNA splicing and metabolism of non-coding RNAs (such as miRNAs themselves). These findings reinforce our hypothesis of a functional connection between the identified miRNA profile and the spliceosome pathway in this subgroup of patients.

To explore this pathway more thoroughly, we retrieved the genes included in the spliceosome-related pathway from REACTOME and subjected them to a Gene Ontology (GO) enrichment analysis using PantherDB v.19.0. A total of 246 genes were extracted and categorized according to cellular component, molecular function, and biological process. In terms of cellular component, the nucleus was the most enriched compartment (47%), followed by the cytosol (38%) (Fig. 7D, Table S12). Regarding molecular function, the majority of genes were associated with nucleic acid-related activities, with DNA binding (31%) and RNA binding (26.14%) being the most prominent (Fig. 7E, Table S13). When analyzing enrichment in biological processes, a large proportion of genes were involved in RNA processing (25.4%), RNA splicing (19.8%), and, more specifically, mRNA splicing (16.07%) (Fig. 7F, Table S14). These data indicate that the spliceosome machinery is among the most enriched pathways targeted by the miRNAs identified in our analysis.

Finally, experimentally validated miRNA-mRNA interactions were retrieved from miRTarBase, focusing on the 246 genes linked to the spliceosome pathway. Subsequent network analysis using Cytoscape revealed that ten out of the 14 miRNAs differentially expressed between LOAD subgroups shared common target genes with ten members of the SRSF family (Fig. 7G). Taken together, these findings provide compelling evidence for a significant functional interplay between the miRNA expression profile identified in the caudate nucleus and the regulation of the spliceosome pathway. This suggests that dysregulation of specific miRNAs may critically influence alternative splicing mechanisms, potentially contributing to the molecular pathology underlying LOAD. These findings underscore the importance of spliceosome as a key regulatory center targeted by miRNAs in this neurodegenerative context, warranting further experimental validation.

Inter-relationship between dysregulated miRNAs, HTT CAG repeat size, and key neuropathological hallmarks

To better integrate our findings within a neuropathological framework, we next investigated the associations between miRNA expression profiles, CAG repeat length, and the severity of tau pathology, including both pretangles and ghost tangles, as well as the soluble HTT signal in caudate neurons (Fig. 8A, Table S15). We observed a consistent and statistically significant positive correlation between the intensity of soluble HTT immunoreactivity and the expression levels of eleven miRNAs, with miR-218-5p and miR-27b-3p, included in the group of five miRNAs specific for LOAD HTT IAs patients, showing the strongest associations. This suggests a transcriptional footprint associated with the presence of an exacerbated HTT protein profile in the caudate nucleus. Only miR-30d-5p, miR-100-5p, and miR-126-3p failed to show significant associations with increased HTT soluble levels.

Conversely, the burden of pretangles exhibited a predominantly negative correlation with miRNA levels, reaching statistical significance for 6 out of the 14 miRNAs analyzed (Fig. 8A, Table S15). In contrast, ghost tangles, indicative of more advanced tau pathology, displayed a positive correlation with miR-218-5p and miR-30a-5p (Fig. 8A, Table S15). Interestingly, the number of CAG repeats correlated positively with both HTT soluble signal and the abundance of ghost tangles, while showing an inverse correlation with pretangle burden. This finding reinforces the notion that CAG repeat length correlates with HTT protein burden and is associated with a shift toward more mature tau aggregates.

Building upon our in silico target analysis, we examined experimentally validated miRNA-mRNA interactions involving the MAPT and HTT genes (Fig. 8B). Network modeling identified three miRNAs (miR-23b-5p, miR-99-5p, and miR-128-3p) that are validated regulators of both MAPT and HTT. Additionally, other miRNAs showed gene-specific interactions: miR-191-5p targets MAPT exclusively, while miR-27b-3p is selectively linked to HTT. These direct validated interactions provide a molecular basis for the correlations observed, strengthening the evidence that the post-transcriptional effects of HTT extend beyond its canonical role in HD, reshaping the miRNA scenario in a way that promotes tau pathology in the context of LOAD. Such miRNA-mediated cross-talk between HTT and MAPT may contribute to the acceleration of tau-driven neurodegeneration.

Collectively, these data underscore a dual role for HTT IAs: both as histopathological hallmarks, by increased HTT and ghost tangles, and as modulators of miRNA-mediated gene regulation, synergistically accelerating tau-mediated neurodegeneration. However, further experimental studies are necessary to elucidate the precise molecular mechanisms and the biological significance of these miRNAs in the context of this accelerated neurodegeneration.

Late-onset Alzheimer´s disease (LOAD) patients represent a heterogeneous population. In fact, there is a complex influence of genetic, environmental, and other factors on disease development and progression. Our previous studies showed a higher frequency of intermediate alleles in the HTT gene among AD patients (7, 8). Therefore, in this work, we aimed to investigate whether this genetic characteristic influences LOAD progression and/or its neuropathological hallmarks, with a special focus on MAPT gene splicing.

A key finding from our study is that LOAD patients carrying HTT IAs exhibited a lower survival rate after disease onset compared to non-carrier patients, a clinical outcome not previously described. We have analyzed the caudate nucleus, a region particularly sensitive to alterations in HTT CAG repeat number (18). Histopathological and molecular analysis of this brain region revealed that while LOAD non-carriers subjects already display alterations (e.g., increased soluble HTT levels or altered miRNA profiles) compared to healthy controls, these changes were consistently and further exacerbated in HTT IAs carrier patients. This observed severity in neuropathology included a heightened diffuse HTT immunoreactivity, in a non-aggregated state, distinguishing it from the profound pathology seen with fully mutated HTT in HD, and a more advanced tau pathology. Both factors likely contribute to the compromised survival in HTT IAs carrier patients by favoring a gain of toxic function of the non-aggregated HTT. The tau 3R predominance observed here, characterized by a tau 3R isoform imbalance and a significant increase in 3R tau-enriched ghost tangles, aligns with its association with axonal cytoskeleton destabilization and reduced neuronal survival (46). This also positions HTT IA-associated LOAD in later stages of disease, where tau 3R becomes more prevalent (47, 48) and may promote the progression of NFTs toward their final, neurotoxic stage (40, 49, 50).

Our study elucidates molecular mechanisms underlying this acceleration in neuropathology. It represents the first detailed miRNA profiling in the caudate nucleus of LOAD patients, showing that this profile is already altered in LOAD, and profoundly amplified in HTT IA carriers. We employed a novel RPM-based analysis method (miRPM) validated by the DANA approach (32). This stringent analysis revealed that while 26 miRNAs were generally altered in LOAD versus controls, HTT IAs specifically exacerbated these changes, with 14 miRNAs being significantly more overexpressed in HTT IA carriers compared to non-carriers. Five miRNAs (miR-100-5p, miR-218-5p, miR-27b-3p, miR-487-3p, and miR-9-3p) were uniquely altered in LOAD HTT IAs patients regarding healthy controls, highlighting their specificity, and emerging as a HTT IA signature miRNAs. Notably, miR-100-5p has also been linked to disease progression in HD (21), suggesting a shared molecular pathway across HTT-related conditions. Furthermore, a subset of miRNAs showed a positive correlation with CAG repeat length, indicating a CAG-dependent modulation. Importantly, miR-218-5p was the only miRNA analyzed that exhibited positive correlations with CAG repeat length, HTT protein levels, and ghost tangles, and negative with pretangles. This suggests miR-218-5p may play an essential role in the alterations leading to the accelerated disease progression observed in these patients.

Our in silico studies and network analysis provide mechanistic insights, suggesting that these altered miRNAs directly target key components of the splicing machinery and both MAPT and HTT genes. This strongly supports our observations at the histopathological levels regarding tau isoform imbalance and splicing factors dysregulation, such as SRSF6 and FUS-SFPQ, in HTT IAs carriers. This miRNA-mediated effect suggests that the presence of HTT IAs significantly modifies the regulatory environment of gene expression to aggravate splicing machinery alterations and accelerate tau-mediated neurodegeneration. This scenario of HTT-tau co-occurrence, where HTT IAs favor increased HTT and tau 3R levels, driving accelerating NFTs maturation, is in line with studies suggesting HTT pathology as a trigger for multiple proteinopathies, including TDP-43, alpha-synuclein, beta-amyloid, and tau alterations in advanced HD (16). While tau isoform shifts differ between HD (4R) (17) and LOAD (3R), the positive correlation of tau alterations with CAG repeat length in both conditions is striking. This suggests that the final HTT protein state (mutated versus intermediate) may dictate the precise tau imbalance. Our proposed mechanism integrating these findings is schematically represented in Fig. 9.

These findings collectively underscore the profound impact of HTT IAs on LOAD pathogenesis, offering another clue to understand the disease’s clinical heterogeneity. This comprehensive neuropathological and molecular characterization reveals HTT IAs as critical risk biomarkers. Crucially, as HTT genotyping is a widely used and easily accessible technique, our results can contribute to improved clinical practice by enabling a more precise stratification of LOAD patients for clinical trials and facilitating the development of more focused, personalized, therapeutic interventions. The identified miRNA dysregulation, particularly the HTT IA signature miRNAs, represents a promising avenue for novel therapeutic targets, as their dysregulation offers a precise avenue for future intervention.

Limitations

Working with postmortem human brain tissue, while invaluable for understanding neuropathology, presents unique limitations that can influence study design and sample inclusion. We acknowledge the variability in sample size (N) across our different experimental assays (e.g., RT-qPCR, Western blot, immunohistochemistry), which is primarily due to the finite nature of our well-characterized donor cohort and the differing quality requirements of each analytical method. For RNA-based analyses, such as RT-qPCR or mRNA expression and miRNA-Seq, we faced challenges with RNA integrity. As indicated by the low RIN values in Table Supplementary 4, RNA is highly susceptible to postmortem degradation. While this significantly affects mRNA expression levels, miRNA sequencing data is less compromised, as several studies indicate that miRNAs, being small nucleotide chains with high resistance to degradation, allow for reliable data acquisition despite low RIN levels (51–53). Similarly, the integrity of proteins in postmortem brain tissue poses significant challenges for Western blot analysis. We encountered difficulties due to the degradation of protein samples, especially for control subjects, which limited their inclusion. However, it is important to note that the primary objective of our Western blot analyses was to elucidate differences between LOAD HTT IAs carriers and LOAD non-carrier patients, a critical comparison that remained fully addressed by the available samples. For immunohistochemistry and immunofluorescence, optimal tissue morphology and antigen preservation are critical for accurate high-resolution analysis and cell counting. Despite differences in N across the different techniques, consistent trends and complementary information observed at various molecular levels (e.g., miRNA, mRNA, protein, and IHC profiles) derived from these human samples significantly enhance the overall robustness and interpretability of our conclusions.

AD: Alzheimer’s disease

APOE: Apolipoprotein E

E10: exon 10 of MAPT gene

FDR: false discovery rate

FTD: frontotemporal dementia

FUS: fused in sarcoma protein

GO: gene ontology

HD: Huntington’s disease

HTT IAS: intermediate alleles in HTT gene

HTT: huntingtin protein

KEGG: Kyoto Encyclopedia of Genes and Genomes

LOAD: Late- onset Alzheimer’s disease

MAPT: microtubule associated tau gene

miRNAs: microRNAs

NDs: neurodegenerative diseases

NFTs: neurofibrillary tangles

PCA: principal component analysis

PLA: proximity ligation assay

PMI: postmortem interval

PSP: progressive supranuclear palsy

RBP: RNA binding protein

RPM: reads per million

RIN: RNA integrity number

ROI: region of interest

RPM: read per million

SFPQ: Proline/Glutamine rich splicing factor

SRSF: Serine/Arginine rich splicing factor

Ethical approval

All procedures were performed after obtained the approval of Ethical Committees of Neurological Tissue Bank (NTB) of the Hospital Clinic-FRCB-IDIBAPS (Barcelona, Spain), the Principado de Asturias BioBank and Research Ethics Committee of the Principality of Asturias (CEImPA nº 2022.266).

Consent for publication

All authors have approved the content of this manuscript and provided consent for publication.

Availability of data and materials

All data generated and used in this study are publicly available in ZENODO (DOI: 10.5281/zenodo.15230070) and NCBI GEO DataSets (ID: GSE300433; The following secure access code has been created to allow review of record GSE300433 while it remains in private status: mtwfggcozpalrwn). R code used to analyze miRNA database is available at https://github.com/sergio30po/miRNA-RPM-DE-Analysis.git.

Competing interests

The authors declare no conflicts of interest.

Author's contribution

J.C-S performed all histological and molecular experiments, conducted data analysis, created the figures and wrote the manuscript. S.P-G performed genotyping the initial samples to establish the studied cohort, developed the miRPM R package, and carried out bionformatic analysis. P.P-H, M.F-S and E.I-G provided guidance on the analysis and interpretation of RNAseq and miRNAs experiments. MD.C-T developed the technical tasks for the preparation of the samples provided by the Biobank of the Principality of Asturias. V.A and M.M-G conceptualized, designed and obtained funding for the study, interpreted the epidemiological data, executed the genetic objective of the study, and supervised the final versions of the manuscript. C.T-Z designed and supervised the molecular and histopathological analysis experiments, interpreted the data, and supervised the different versions of the manuscript.

Acknowledgments

We extend our gratitude to the research staff for their valuable support, including: J. Bermejo-Pampliega (Physiology Area of University of Oviedo; AYUD/2021/5134); M. Alonso-Guervos (Microscopy and Image Processing Unit, Scientific and Technical Services, University of Oviedo); the Molecular Histopathology Service in Animal Cancer Models at Instituto Universitario de Oncología del Principado de Asturias (IUOPA), and the Molecular Genetics Laboratory of Hospital Universitario Central de Asturias (HUCA). We also acknowledge the collaboration of the Principado de Asturias BioBank (PT23/0077), financed by Servicio de Salud del Principado de Asturias and ISCIII/FEDER, and HCB-IDIBAPS Biobank for sample and data procurement. Finally, we are indebted to Asociación Parkinson Asturias-Obra Social Cajastur and to all the patients and their families for their invaluable contributions.

Funding

This study was funded by Instituto de Salud Carlos III (ISCIII) and co-funded by the European Union (FEDER/FSE) through PI21/00467 (V.A., M. M-M). J.C-S was supported by ISCIII grant AC20/00017, co-founded by EuroNanoMed III (20-0084, M. M-M) and Asociación Parkinson Asturias-Obra Social Cajastur. S.P-O is supported by Fundación para la Investigación e Innovación Biosanitaria del Principado de Asturias (FINBA). P.P-H was supported by grant AYUD/2021/5134 from Fundación para el Fomento en Asturias de la Investigación Científica y la Tecnología (FICYT), co-funded by FEDER.

Alzheimer's disease facts and figures. Alzheimers Dement. 2024;20(5):3708-821.
Robinson JL, Lee EB, Xie SX, Rennert L, Suh E, Bredenberg C, et al. Neurodegenerative disease concomitant proteinopathies are prevalent, age-related and APOE4-associated. Brain. 2018;141(7):2181-93.
Fortea J, Pegueroles J, Alcolea D, Belbin O, Dols-Icardo O, Vaque-Alcazar L, et al. APOE4 homozygozity represents a distinct genetic form of Alzheimer's disease. Nat Med. 2024;30(5):1284-91.
Yamazaki Y, Zhao N, Caulfield TR, Liu CC, Bu G. Apolipoprotein E and Alzheimer disease: pathobiology and targeting strategies. Nat Rev Neurol. 2019;15(9):501-18.
Andrews SJ, Renton AE, Fulton-Howard B, Podlesny-Drabiniok A, Marcora E, Goate AM. The complex genetic architecture of Alzheimer's disease: novel insights and future directions. EBioMedicine. 2023;90:104511.
Identification of Genetic Factors that Modify Clinical Onset of Huntington's Disease. Cell. 2015;162(3):516-26.
Menendez-Gonzalez M, Clarimon J, Rosas-Allende I, Blazquez M, San Martin ES, Garcia-Fernandez C, et al. HTT gene intermediate alleles in neurodegeneration: evidence for association with Alzheimer's disease. Neurobiology of aging. 2018.
Perez-Oliveira S, Castilla-Silgado J, Painous C, Aldecoa I, Menendez-Gonzalez M, Blazquez-Estrada M, et al. Huntingtin CAG repeats in neuropathologically confirmed tauopathies: Novel insights. Brain Pathol. 2024;34(4):e13250.
Rosas I, Martinez C, Clarimon J, Lleo A, Illan-Gala I, Dols-Icardo O, et al. Role for ATXN1, ATXN2, and HTT intermediate repeats in frontotemporal dementia and Alzheimer's disease. Neurobiology of aging. 2020;87:139 e1- e7.
Perez-Oliveira S, Alvarez I, Rosas I, Menendez-Gonzalez M, Blazquez-Estrada M, Aguilar M, et al. Intermediate and Expanded HTT Alleles and the Risk for alpha-Synucleinopathies. Mov Disord. 2022;37(9):1841-9.
Handsaker RE, Kashin S, Reed NM, Tan S, Lee WS, McDonald TM, et al. Long somatic DNA-repeat expansion drives neurodegeneration in Huntington's disease. Cell. 2025;188(3):623-39 e19.
Faquih TO, Aziz NA, Gardiner SL, Li-Gao R, de Mutsert R, Milaneschi Y, et al. Normal range CAG repeat size variations in the HTT gene are associated with an adverse lipoprotein profile partially mediated by body mass index. Hum Mol Genet. 2023;32(10):1741-52.
Apolinario TA, Paiva CL, Agostinho LA. REVIEW-ARTICLE Intermediate alleles of Huntington's disease HTT gene in different populations worldwide: a systematic review. Genet Mol Res. 2017;16(2).
Axenhus M, Winblad B, Tjernberg LO, Schedin-Weiss S. Huntingtin Levels are Elevated in Hippocampal Post-Mortem Samples of Alzheimer's Disease Brain. Curr Alzheimer Res. 2020;17(9):858-67.
Davis MY, Keene CD, Jayadev S, Bird T. The co-occurrence of Alzheimer's disease and Huntington's disease: a neuropathological study of 15 elderly Huntington's disease subjects. J Huntingtons Dis. 2014;3(2):209-17.
St-Amour I, Turgeon A, Goupil C, Planel E, Hebert SS. Co-occurrence of mixed proteinopathies in late-stage Huntington's disease. Acta neuropathologica. 2018;135(2):249-65.
Fernandez-Nogales M, Cabrera JR, Santos-Galindo M, Hoozemans JJ, Ferrer I, Rozemuller AJ, et al. Huntington's disease is a four-repeat tauopathy with tau nuclear rods. Nat Med. 2014.
Langfelder P, Cantle JP, Chatzopoulou D, Wang N, Gao F, Al-Ramahi I, et al. Integrated genomics and proteomics define huntingtin CAG length-dependent networks in mice. Nat Neurosci. 2016;19(4):623-33.
Langfelder P, Gao F, Wang N, Howland D, Kwak S, Vogt TF, et al. MicroRNA signatures of endogenous Huntingtin CAG repeat expansion in mice. PLoS One. 2018;13(1):e0190550.
Yao Q, Chen Y, Zhou X. The roles of microRNAs in epigenetic regulation. Curr Opin Chem Biol. 2019;51:11-7.
Diez-Planelles C, Sanchez-Lozano P, Crespo-Lorenzo MC, Gil-Zamorano J, Ribacoba R, Gonzalez N, et al. Circulating microRNAs in Huntingtons Disease: emerging mediators in metabolic impairment. Pharmacol Res. 2016.
Takousis P, Sadlon A, Schulz J, Wohlers I, Dobricic V, Middleton L, et al. Differential expression of microRNAs in Alzheimer's disease brain, blood, and cerebrospinal fluid. Alzheimers Dement. 2019;15(11):1468-77.
Pinto-Hernandez P, Castilla-Silgado J, Coto-Vilcapoma A, Fernandez-Sanjurjo M, Fernandez-Garcia B, Tomas-Zapico C, et al. Modulation of microRNAs through Lifestyle Changes in Alzheimer's Disease. Nutrients. 2023;15(17).
Montine TJ, Phelps CH, Beach TG, Bigio EH, Cairns NJ, Dickson DW, et al. National Institute on Aging-Alzheimer's Association guidelines for the neuropathologic assessment of Alzheimer's disease: a practical approach. Acta neuropathologica. 2012;123(1):1-11.
Ishigaki S, Riku Y, Fujioka Y, Endo K, Iwade N, Kawai K, et al. Aberrant interaction between FUS and SFPQ in neurons in a wide range of FTLD spectrum diseases. Brain. 2020;143(8):2398-405.
Hegazy M, Cohen-Barak E, Koetsier JL, Najor NA, Arvanitis C, Sprecher E, et al. Proximity Ligation Assay for Detecting Protein-Protein Interactions and Protein Modifications in Cells and Tissues in Situ. Curr Protoc Cell Biol. 2020;89(1):e115.
Xie F, Wang J, Zhang B. RefFinder: a web-based tool for comprehensively analyzing and identifying reference genes. Funct Integr Genomics. 2023;23(2):125.
Livak KJ, Schmittgen TD. Analysis of relative gene expression data using real-time quantitative PCR and the 2(-Delta Delta C(T)) Method. Methods. 2001;25(4):402-8.
Kang W, Eldfjell Y, Fromm B, Estivill X, Biryukova I, Friedlander MR. miRTrace reveals the organismal origins of microRNA sequencing data. Genome Biol. 2018;19(1):213.
Friedlander MR, Mackowiak SD, Li N, Chen W, Rajewsky N. miRDeep2 accurately identifies known and hundreds of novel microRNA genes in seven animal clades. Nucleic Acids Res. 2012;40(1):37-52.
Kozomara A, Birgaoanu M, Griffiths-Jones S. miRBase: from microRNA sequences to function. Nucleic Acids Res. 2019;47(D1):D155-D62.
Düren Y, Lederer J, Qin L-X. Depth normalization of small RNA sequencing: using data and biology to select a suitable method. Nucleic Acids Res. 2022;50(10):e56-e.
Vlachos IS, Zagganas K, Paraskevopoulou MD, Georgakilas G, Karagkouni D, Vergoulis T, et al. DIANA-miRPath v3.0: deciphering microRNA function with experimental support. Nucleic Acids Res. 2015;43(W1):W460-6.
Tang D, Chen M, Huang X, Zhang G, Zeng L, Zhang G, et al. SRplot: A free online platform for data visualization and graphing. PLoS One. 2023;18(11):e0294236.
Chain DG, Margolin RA, Pollack S, Goncalves M. AlphaFold 2 predicts a markedly different MTBR conformation for 3-repeat and 4-repeat tau isoforms. Alzheimer´s & Dementia. 2025;20.
Espinoza M, de Silva R, Dickson DW, Davies P. Differential incorporation of tau isoforms in Alzheimer's disease. J Alzheimers Dis. 2008;14(1):1-16.
Kovacs GG, Lukic MJ, Irwin DJ, Arzberger T, Respondek G, Lee EB, et al. Distribution patterns of tau pathology in progressive supranuclear palsy. Acta Neuropathol. 2020;140(2):99-119.
Kraus A, Saijo E, Metrick MA, 2nd, Newell K, Sigurdson CJ, Zanusso G, et al. Seeding selectivity and ultrasensitive detection of tau aggregate conformers of Alzheimer disease. Acta Neuropathol. 2019;137(4):585-98.
Moloney CM, Lowe VJ, Murray ME. Visualization of neurofibrillary tangle maturity in Alzheimer's disease: A clinicopathologic perspective for biomarker research. Alzheimers Dement. 2021;17(9):1554-74.
Uchihara T. Neurofibrillary changes undergoing morphological and biochemical changes - How does tau with the profile shift of from four repeat to three repeat spread in Alzheimer brain? Neuropathology. 2020;40(5):450-9.
Petry S, Nateghi B, Keraudren R, Sergeant N, Planel E, Hebert SS, et al. Differential Regulation of Tau Exon 2 and 10 Isoforms in Huntington's Disease Brain. Neuroscience. 2023;518:54-63.
Schilling J, Broemer M, Atanassov I, Duernberger Y, Vorberg I, Dieterich C, et al. Deregulated Splicing Is a Major Mechanism of RNA-Induced Toxicity in Huntington's Disease. Journal of molecular biology. 2019;431(9):1869-77.
Park SA, Ahn SI, Gallo JM. Tau mis-splicing in the pathogenesis of neurodegenerative disorders. BMB Rep. 2016;49(8):405-13.
Ishigaki S, Fujioka Y, Okada Y, Riku Y, Udagawa T, Honda D, et al. Altered Tau Isoform Ratio Caused by Loss of FUS and SFPQ Function Leads to FTLD-like Phenotypes. Cell Rep. 2017;18(5):1118-31.
Buchholz S, Zempel H. The six brain-specific TAU isoforms and their role in Alzheimer's disease and related neurodegenerative dementia syndromes. Alzheimers Dement. 2024;20(5):3606-28.
Fernandez-Nogales M, Santos-Galindo M, Hernandez IH, Cabrera JR, Lucas JJ. Faulty splicing and cytoskeleton abnormalities in Huntington's disease. Brain Pathol. 2016;26(6):772-8.
Cherry JD, Esnault CD, Baucom ZH, Tripodis Y, Huber BR, Alvarez VE, et al. Tau isoforms are differentially expressed across the hippocampus in chronic traumatic encephalopathy and Alzheimer's disease. Acta Neuropathol Commun. 2021;9(1):86.
Uematsu M, Nakamura A, Ebashi M, Hirokawa K, Takahashi R, Uchihara T. Brainstem tau pathology in Alzheimer's disease is characterized by increase of three repeat tau and independent of amyloid beta. Acta Neuropathol Commun. 2018;6(1):1.
Saroja SR, Sharma A, Hof PR, Pereira AC. Differential expression of tau species and the association with cognitive decline and synaptic loss in Alzheimer's disease. Alzheimers Dement. 2022;18(9):1602-15.
Uchihara T. Pretangles and neurofibrillary changes: similarities and differences between AD and CBD based on molecular and morphological evolution. Neuropathology. 2014;34(6):571-7.
Gallego Romero I, Pai AA, Tung J, Gilad Y. RNA-seq: impact of RNA degradation on transcript quantification. BMC Biol. 2014;12:42.
Potla P, Ali SA, Kapoor M. A bioinformatics approach to microRNA-sequencing analysis. Osteoarthr Cartil Open. 2021;3(1):100131.
Sonntag KC, Woo TW. Laser microdissection and gene expression profiling in the human postmortem brain. Handb Clin Neurol. 2018;150:263-72.

No competing interests reported.

Download PDF

Editorial decision: Revision requested
30 Nov, 2025
Reviews received at journal
25 Nov, 2025
Reviews received at journal
23 Nov, 2025
Reviewers agreed at journal
09 Nov, 2025
Reviewers agreed at journal
07 Nov, 2025
Reviewers invited by journal
28 Sep, 2025
Editor assigned by journal
17 Sep, 2025
Submission checks completed at journal
17 Sep, 2025
First submitted to journal
15 Sep, 2025

You are reading this latest preprint version

Connecting HTT intermediate alleles and microRNA dysregulation to enhanced tauopathy in Late-Onset Alzheimer's Disease

Status:

Version 1

Abstract

Background

Methods

Results

Conclusions

Figures

Introduction

Methods

Subjects

Immunohistochemistry and immunofluorescence

Nuclear colocalization analysis

Proximity ligation assay

Western blotting

mRNA expression analysis by RT-qPCR

small RNA-sequencing

miRPM custom R package

Differential expression of miRNA sequencing data

General statistical analysis

Results

Discussion

Conclusions

Limitations

Abbreviations

Declarations

References

Additional Declarations

Supplementary Files

Status:

Version 1