TDP-43-related cryptic splicing reflects pathological heterogeneity in amyotrophic lateral sclerosis

Abstract

Amyotrophic lateral sclerosis (ALS) involves TAR DNA-binding protein 43 (TDP-43) dysfunction, leading to cryptic splicing, including UNC13A cryptic exon inclusion and truncated STMN2 production. Pathologically, ALS is classified into three phosphorylated TDP-43 subtypes (Types 1, 2a, and 2b) based on the distribution of neuronal and glial cytoplasmic inclusions and the density of dystrophic neurites. However, the relationship between these pathological subtypes and TDP-43-associated cryptic splicing remains unclear. This retrospective postmortem study aimed to analyze motor cortex tissue from sporadic ALS cases (Type 1, n = 12; Type 2a, n = 12; Type 2b, n = 11) with previously described subtype-specific morphological features. UNC13A cryptic RNA and truncated STMN2 RNA levels were quantified using quantitative real-time PCR (qRT-PCR). Compared with Types 1 and 2a, Type 2b, characterized by abundant thread-, dot-, and grain-like dystrophic neurites, exhibited significantly higher levels of UNC13A cryptic RNA and truncated STMN2 RNA. Droplet digital PCR revealed a similar subtype-dependent trend and showed strong correlations with qRT-PCR measurements, providing partial support for the primary findings. UNC13A cryptic RNA and truncated STMN2 RNA levels were positively correlated (Spearman’s ρ = 0.58), indicating coordinated dysregulation of multiple TDP-43-dependent splicing targets. Our results demonstrate that the burden of TDP-43-associated cryptic splicing differs across ALS pathological subtypes. Additional analyses indicated that these subtype-dependent differences were not readily accounted for by disease duration, RNA quality, normalization strategy, or semiquantitative neuronal loss scores. These findings suggest that molecular heterogeneity parallels pathological heterogeneity in ALS and provide additional insight into the biological diversity of TDP-43 proteinopathy.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1186/s40478-026-02400-0.

Introduction

Approximately 97% of sporadic amyotrophic lateral sclerosis (ALS) cases show nuclear loss of TAR DNA-binding protein 43 (TDP-43), resulting in disrupted RNA processing and pathological overlap with frontotemporal lobar degeneration (FTLD) [10, 16]. In the nucleus, TDP-43 regulates pre-messenger RNA splicing and suppresses the inclusion of exon-like sequences within introns, a process known as cryptic splicing [1, 9]. Notably, cryptic splicing abnormalities have been reported even in the absence of overt cytoplasmic TDP-43 inclusions, suggesting that nuclear TDP-43 dysfunction may precede detectable aggregate formation [19]. Accordingly, cryptic exon inclusion in UNC13A (UNC13A cryptic exon) [3, 11] and cryptic splicing-mediated truncation of stathmin-2 (truncated STMN2) [6, 7, 12, 17] have emerged as established molecular readouts of TDP-43 loss-of-function.

Pathologically, ALS can be classified into three subtypes (Types 1, 2a, and 2b) based on the presence of phosphorylated TDP-43-positive neuronal cytoplasmic inclusions in the hippocampal dentate gyrus and the density of phosphorylated TDP-43-positive dystrophic neurites in the temporal cortex [14, 20, 21]. Among these subtypes, ALS Type 2b is characterized by abundant thread-, dot-, and grain-like dystrophic neurites in addition to numerous neuronal and glial cytoplasmic inclusions, and is associated with shorter disease duration [20]. Our previous neuropathological study demonstrated that neuronal loss in the anterior horn was relatively less pronounced in this subtype [20]. Notably, this morphological pattern of phosphorylated TDP-43 aggregates closely resembles that of FTLD-TDP Type E [8, 13], and has therefore also been referred to as ALS-TDP Type E [21]. Because phosphorylated TDP-43 aggregation is accompanied by nuclear depletion of TDP-43, pathological subtypes may differ in the burden of TDP-43-associated cryptic splicing. However, how this pathological heterogeneity relates to molecular abnormalities reflecting TDP-43 loss-of-function remains unclear.

In this study, we aimed to investigate whether UNC13A cryptic exon inclusion and truncated STMN2 expression differ across phosphorylated TDP-43 pathological subtypes using motor cortex tissue from pathologically confirmed ALS cases. We further sought to determine whether the burden of cryptic splicing reflects pathological heterogeneity in ALS.

Methods

Subjects

This was a retrospective postmortem observational study. Postmortem motor cortex samples were obtained from sporadic Japanese patients with ALS without ALS-related gene mutations (n = 35) and from control individuals (n = 5) without neuropathological evidence of neurodegenerative diseases (Additional file 1, Supplementary Table S1). ALS cases with dementia preceding motor symptoms by more than one year were excluded. All cases were pathologically confirmed as having ALS or no neurodegenerative diseases at Niigata University, including cases from the previously reported cohort. The Institutional Review Board of Niigata University approved the study (G2020-0031), and written informed consent was obtained from the families.

ALS cases were classified into three pathological subtypes (Types 1, 2a, and 2b) as described by Takeuchi et al [20]. Type 1 lacked phosphorylated TDP-43-positive neuronal cytoplasmic inclusions in the dentate gyrus. Type 2 cases exhibited neuronal cytoplasmic inclusions and were further subdivided based on semi-quantitative assessment of phosphorylated TDP-43-positive thread-, dot-, and grain-like dystrophic neurites in the temporal cortex: cases with no or few dystrophic neurites were classified as Type 2a, whereas those with abundant dystrophic neurites were classified as Type 2b (Additional file 2, Fig. S1a, b). In the previous report, the cohort included Type 1 (n = 63), Type 2a (n = 22), and Type 2b (n = 11) cases [20].

In the present study, cases were selected to achieve comparable age distributions across ALS pathological subtypes and between ALS and control groups. Initially, 12 cases were selected for each ALS pathological subtype (Types 1, 2a, and 2b). During a subsequent review of clinical records, one Type 2b case was excluded because dementia had preceded the onset of motor symptoms by several years. The final cohort therefore consisted of 35 ALS cases: Type 1 (n = 12), Type 2a (n = 12), and Type 2b (n = 11). Of these cases, 23 cases (Type 1, n = 5; Type 2a, n = 10; Type 2b, n = 8) overlapped with cases included in our previous study (indicated in (Additional file 1, Supplementary Table S1), while the remaining cases were newly included in this study. The TDP-43 pathological subtypes in the motor cortex were assessed to confirm previously established morphological features. Immunohistochemical analysis of phosphorylated TDP-43 was performed using a mouse monoclonal antibody to phosphorylated TDP-43 [pTDP-43; Ser409/410; clone 11–9; Cosmo Bio] as previously described [20]. In ALS Type 2 subtypes (Types 2a and 2b), dystrophic neurites exhibited subtype-specific patterns, appearing either sparse or abundant. In contrast, Type 1 cases demonstrated no or only sparse dystrophic neurites, similar to Type 2a (representative images are presented in Fig. 1). Because previous studies have reported shorter disease duration in ALS pathological subtype Type 2b, disease duration was compared among pathological subtypes in the present cohort using the Kruskal–Wallis test followed by Dunn’s multiple-comparisons test (Additional file 2, Fig. S1c). Using 4-μm-thick hematoxylin–eosin stained sections, neuronal loss in the motor cortex was independently assessed semiquantitatively by two neuropathologists (SH and MT) employing a four-point scale (0 = absent, 1 = mild, 2 = moderate, 3 = severe) according to the criteria described by Takeuchi et al. [20] (Additional file 1, Supplementary Table S2). Neuronal loss scores were compared among pathological subtypes using the Kruskal–Wallis test followed by Dunn’s multiple-comparisons test.

Fig. 1.

Representative phosphorylated TDP-43 immunohistochemistry in the motor cortex. Representative motor cortex pathology from ALS cases with TDP-43 Type 1 (ALS_9), Type 2a (ALS_13), and Type 2b (ALS_28). Phosphorylated TDP-43-positive neuronal and glial cytoplasmic inclusions, and dystrophic neurites are shown. Scale bars: 50 μm

RNA analysis

Total RNA was extracted from approximately 100 mg of postmortem frozen motor cortex tissue using the mirVana miRNA Isolation Kit (Life Technologies), followed by purification with the Direct-zol RNA Miniprep Plus kit (Zymo Research) according to the manufacturers’ protocols. RNA integrity was assessed using the RNA Integrity Number (RIN), and samples with RIN ≥ 5.0 were included in the analyses. The analyzed cohort showed generally good RNA quality, with a median RIN of 7.7 (interquartile range, 6.8–8.5; range, 5.2–9.3) (Additional file 1, Supplementary Table S1). Complementary DNA (cDNA) was synthesized from 2,000 ng of total RNA using the Superscript IV VILO Master Mix (Thermo Fisher Scientific). Quantitative real-time polymerase chain reaction (qRT-PCR) was performed on a Thermal Cycler Dice Real-Time System III (Takara Bio) using SYBR Green Premix Ex Taq II (Takara Bio), according to the manufacturer’s instructions, with 40 cycles of amplification. Each sample was analyzed in technical triplicate. Amplification specificity was confirmed by melt curve analysis. The assay quantified UNC13A cryptic exon-containing RNA (UNC13A cryptic RNA) and truncated STMN2 RNA. For the primary analyses, cryptic RNA levels were normalized to the corresponding full-length UNC13A or STMN2 transcripts using the ΔCt method. To determine whether the observed subtype-dependent differences were influenced by the normalization method, cryptic RNA levels were additionally normalized to the housekeeping gene RPLP0 using the ΔCt method. Ct values ≥ 38 or undetermined signals were considered below the detection limit. For ΔCt-based analyses, these values were assigned a Ct value of 38 for computational purposes. Comparisons among pathological subtypes included only samples within the detectable range, whereas values below the detection limit were retained in analyses involving control samples. Droplet digital PCR (ddPCR) was performed in 26 ALS cases (Type 1: n = 9; Type 2a: n = 10; Type 2b: n = 7) that met predefined RNA quality criteria, including RNA storage duration (< 2 years) and robust amplification of endogenous controls (RPLP0, TBP), to ensure reliable quantification. Assays were performed using the QX200 Droplet Digital PCR System (Bio-Rad) with ddPCR Multiplex Supermix according to the manufacturer’s instructions. A fluorescence amplitude threshold was set based on no-template controls to distinguish positive from negative droplets. Data were analyzed using QX Manager Standard Edition software (Bio-Rad). UNC13A cryptic RNA and truncated STMN2 RNA levels were quantified as absolute copy numbers and normalized to the corresponding total UNC13A or STMN2 transcript levels within each sample to account for gene expression variability. Samples with no detectable positive droplets were assigned a value of 0, reflecting the absence of detectable transcripts. This differs from qRT-PCR, where undetermined values indicate measurements below the detection limit rather than true absence. Primer and probe sequences are listed in Additional file 1, Supplementary Table S3. UNC13A cryptic RNA and truncated STMN2 RNA were selected as primary readouts based on their well-established roles as markers of nuclear TDP-43 loss-of-function [3, 6, 7, 11, 12, 17] and prior validation of the full-length transcript normalization strategy in our laboratory [7]. During assay development, additional cryptic splicing candidates were explored using reference gene-based normalization (RPLP0, GAPDH); as these data were not directly comparable to the final protocol, they were not included in the subtype comparisons. Data acquisition was conducted independently of pathological subtype classification.

Statistical analysis

Statistical analyses were performed using GraphPad Prism version 10.5.0 (GraphPad Software). Group comparisons were performed using the Mann–Whitney U test or the Kruskal–Wallis test with Dunn’s multiple comparisons correction to account for multiple comparisons. Associations between cryptic splicing markers and disease duration, postmortem interval (PMI), and RIN were assessed using Spearman’s rank correlation coefficient. Statistical significance was set at p < 0.05.

Results

To determine whether TDP-43-associated cryptic splicing differed among pathological subtypes, we quantified UNC13A cryptic RNA and truncated STMN2 RNA expression levels in motor cortex samples using qRT-PCR. Cryptic RNA levels were normalized to their respective full-length transcripts using the ΔCt method to estimate the relative burden of cryptic splicing events. As a preliminary analysis, we compared age-matched control individuals and ALS Type 2b cases (n = 5 each), as Type 2b is characterized by abundant phosphorylated TDP-43-positive dystrophic neurites in the motor cortex. UNC13A cryptic RNA and truncated STMN2 RNA levels were either undetectable or present at very low levels in control samples (Additional file 2, Fig. S2). The five control subjects were comparable in age at death to the five Type 2b ALS cases shown in Additional file 2, Fig. S2 (median 66 vs 75 years, p = 0.159), although PMI was longer in controls (median 8.1 vs 2.3 h, p = 0.008). Similar findings were observed when controls were compared with the entire ALS cohort (Additional file 1, Supplementary Tables S1). We next compared cryptic splicing burden across all ALS pathological subtypes. Statistical comparisons were performed using the Kruskal–Wallis test followed by Dunn’s multiple-comparisons test. All samples included in the subtype analyses exhibited Ct values below 38 for both cryptic and corresponding full-length transcripts, indicating that all ΔCt values were derived from detectable signals and were not influenced by assigned Ct values. Significant subtype-dependent differences were observed for both UNC13A cryptic RNA and truncated STMN2 RNA (Fig. 2a, b). Type 2b cases exhibited higher levels of cryptic splicing than Type 1 and Type 2a cases (UNC13A cryptic RNA: p < 0.05; truncated STMN2 RNA: p < 0.01), whereas no significant difference was detected between Type 1 and Type 2a cases. Furthermore, UNC13A cryptic RNA and truncated STMN2 RNA levels were positively correlated across ALS cases (Spearman’s ρ = 0.574, 95% confidence interval [CI]: 0.288 to 0.766, p < 0.001; Fig. 2c), indicating coordinated dysregulation of multiple TDP-43-dependent splicing targets. Although the pattern was not identical to that reported previously, disease duration differed among pathological subtypes in the present cohort (Additional file 2, Fig. S1c). Median disease duration was 32 months in Type 1, 19 months in Type 2a, and 11 months in Type 2b. Type 2a and Type 2b cases exhibited shorter disease duration than Type 1 cases (p < 0.05 and p < 0.001, respectively), whereas no significant difference was observed between Type 2a and Type 2b cases. To determine whether disease duration contributed to the observed subtype-dependent differences in cryptic splicing, we performed correlation analyses across the entire ALS cohort. Neither UNC13A cryptic RNA nor truncated STMN2 RNA expression was significantly associated with disease duration (Additional file 2, Fig. S3). These findings suggest that disease duration alone is unlikely to explain the observed subtype-dependent differences in cryptic splicing.

Fig. 2.

TDP-43-associated cryptic splicing across phosphorylated TDP-43 pathological subtypes of ALS. a–b Expression levels of UNC13A cryptic RNA a and truncated STMN2 RNA b measured by quantitative real-time PCR across ALS pathological subtypes (Type 1, n = 12; Type 2a, n = 12; Type 2b, n = 11; total n = 35). UNC13A cryptic RNA and truncated STMN2 RNA levels were normalized to their respective full-length transcripts using the ΔCt method. Statistical comparisons were performed using ΔCt values with the Kruskal–Wallis test followed by Dunn’s multiple-comparisons test. Data are shown as individual values; horizontal bold lines indicate the median, and error bars represent the interquartile range (25th–75th percentiles). Asterisks indicate significant differences (*p < 0.05, **p < 0.01, ***p < 0.001); ns indicates not significant. c Correlation between UNC13A cryptic RNA and truncated STMN2 RNA expression levels across ALS cases. Expression levels were normalized to the corresponding full-length transcripts using the ΔCt method. Correlation was assessed using Spearman’s rank correlation coefficient. Each point represents an individual case. The Spearman correlation coefficient (ρ) and 95% confidence interval (CI) are shown in the figure. Asterisks indicate significant correlations (***p < 0.001). An inset provides a zoomed view of the low-expression range (UNC13A cryptic RNA: 0–0.005; truncated STMN2 RNA: 0–0.02)

To evaluate potential technical confounders, we examined the relationships of cryptic splicing levels with PMI and RNA quality. Neither UNC13A cryptic RNA (Spearman’s ρ =  − 0.156, p = 0.372, 95% CI: − 0.473 to 0.197) nor truncated STMN2 RNA (Spearman’s ρ =  − 0.241, p = 0.163, 95% CI: − 0.539 to 0.110) was significantly associated with PMI. Likewise, RIN did not correlate with UNC13A cryptic RNA (Spearman’s ρ =  − 0.135, p = 0.439, 95% CI: − 0.456 to 0.217) or truncated STMN2 RNA (Spearman’s ρ = 0.083, p = 0.634, 95% CI: − 0.267 to 0.414), suggesting that the observed subtype-dependent differences were unlikely to be explained by postmortem degradation or RNA quality. Semiquantitative neuronal loss scores in the motor cortex did not significantly differ among pathological subtypes (Additional file 1, Supplementary Table S2; Kruskal–Wallis test, p = 0.250).

To determine whether the observed subtype-dependent differences were influenced by the normalization method, qRT-PCR data were additionally normalized to the housekeeping gene RPLP0. Similar results were obtained using this alternative normalization strategy, with Type 2b cases showing the highest cryptic splicing burden (Additional file 2, Fig. S4).

To further evaluate the qRT-PCR findings using an independent quantification method, ddPCR analyses were performed in a subset of ALS cases (Type 1, n = 9; Type 2a, n = 10; Type 2b, n = 7) and control cases (n = 5). Cryptic transcript copy numbers were first normalized to the corresponding total UNC13A or STMN2 transcript levels. No detectable copies of either cryptic transcript were observed in controls. Type 2b cases exhibited higher levels of UNC13A cryptic RNA and truncated STMN2 RNA than Type 2a cases (p < 0.05), whereas differences between Type 1 and Type 2b did not reach statistical significance (Additional file 2, Fig. S5a, b). To evaluate the effect of an alternative normalization strategy, ddPCR data normalized to the housekeeping gene RPLP0 were also analyzed. The overall subtype-dependent pattern was comparable to that observed using total transcript normalization, although the significance of individual pairwise comparisons differed between the two normalization methods (Additional file 2, Fig. S5c, d). Nevertheless, qRT-PCR and ddPCR measurements were strongly correlated for both UNC13A cryptic RNA (Spearman’s ρ = 0.895, 95% CI: 0.701 to 0.937, p < 0.001) and truncated STMN2 RNA (Spearman’s ρ = 0.870, 95% CI: 0.722 to 0.942, p < 0.001) (Additional file 2, Fig. S5e, f), indicating overall concordance between the two quantification methods.

Discussion

In this study, we demonstrate that among pathological TDP-43 subtypes of ALS, Type 2b—associated with shorter disease duration—exhibited increased levels of UNC13A cryptic RNA and truncated STMN2 RNA, indicating a greater burden of TDP-43-associated cryptic splicing. In contrast, Types 1 and 2a showed overlapping molecular profiles. Notably, these findings were partially supported by ddPCR analyses using both total transcript and RPLP0 normalization, which revealed similar subtype-dependent patterns for UNC13A cryptic RNA and truncated STMN2 RNA, although the significance of individual pairwise comparisons differed between the two normalization strategies. ALS Type 2b has been reported to share morphological similarities with FTLD-TDP Type E [8, 13, 20]. In line with this, the pronounced elevation of cryptic splicing is consistent with previous observations suggesting a clinicopathological continuum characterized by shared TDP-43-associated molecular abnormalities between ALS and FTLD [10, 16]. Together, the concordant elevation of UNC13A cryptic RNA and truncated STMN2 RNA supports the presence of a greater burden of TDP-43-associated cryptic splicing in ALS Type 2b.

Cryptic splicing has recently emerged as a robust molecular indicator of TDP-43 loss-of-function-associated pathology across neurodegenerative diseases, including Perry Syndrome, limbic-predominant age-related TDP-43 encephalopathy, and Alzheimer’s disease [15, 22]. Cryptic splicing events correlate with phosphorylated TDP-43 pathology and define molecular subtypes that are largely independent of amyloid and tau pathology. These data highlight the potential of cryptic splicing as a cross-disease marker of TDP-43-related molecular dysfunction [22]. Notably, our findings extend these observations to TDP-43 pathological subtypes in ALS. To our knowledge, this study is among the first to demonstrate that the burden of TDP-43-associated cryptic splicing differs across ALS pathological subtypes.

Pathological patterns of phosphorylated TDP-43 in ALS have been interpreted either as distinct clinicopathological subtypes or as stages along a neuropathological continuum, in which involvement of the temporal cortex and hippocampus indicates more advanced disease [2]. An intriguing finding of the present study is that cryptic splicing burden differed substantially among ALS pathological subtypes despite the presence of TDP-43 pathology across all groups. In particular, Type 2b cases exhibited significantly higher levels of UNC13A cryptic RNA and truncated STMN2 RNA than Type 1 or Type 2a cases. Although disease duration differed among pathological subtypes in the present cohort, neither UNC13A cryptic RNA nor truncated STMN2 RNA levels were significantly associated with disease duration across all ALS cases. These findings suggest that the observed subtype-dependent differences in cryptic splicing are unlikely to be explained solely by disease duration. Consistent with previous neuropathological studies [20], Type 2b is characterized by abundant phosphorylated TDP-43-positive dystrophic neurites and widespread extranuclear TDP-43 pathology. These observations raise the possibility that the pathological features characteristic of Type 2b are associated with a greater burden of TDP-43-related cryptic splicing. However, the bulk-tissue nature of our analyses precludes cell-level assessment of the underlying biological basis of these differences. Importantly, these subtype classifications do not depend solely on the overall burden of phosphorylated TDP-43 deposition, but also on its regional distribution and morphological characteristics. The observed differences in cryptic splicing burden may therefore be related to qualitative features of TDP-43 pathology rather than simple quantitative pathological load.

This study has several limitations. First, cryptic splicing was quantified using bulk RNA extracted from motor cortex tissue, precluding cell-level assessment of the biological basis of the observed differences. Therefore, we could not determine whether the increased cryptic splicing burden observed in Type 2b reflects a greater number of affected neurons, differences in the extent of TDP-43 dysfunction within individual neurons, or other subtype-specific pathological processes. In addition, neuronal density was not directly quantified in the exact motor cortex punches used for RNA extraction. Although semiquantitative neuronal loss scores did not significantly differ among pathological subtypes in the present study, and a previous neuropathological study reported no significant subtype-associated differences in motor cortical neuronal loss [20], a contribution of subtype-specific differences in cellular composition cannot be completely excluded. Second, the study had a relatively small sample size, potentially reducing the statistical power of our analyses. In particular, ddPCR analyses were restricted to a subset of cases meeting strict RNA quality criteria, which may have introduced selection bias toward samples with enhanced RNA preservation. Although qRT-PCR and ddPCR measurements were strongly correlated for both UNC13A cryptic RNA and truncated STMN2 RNA, some subtype comparisons did not reach statistical significance in the ddPCR subset. This discrepancy may reflect the reduced sample size and the restriction of ddPCR analyses to cases meeting stringent RNA quality criteria. Therefore, the ddPCR findings should be interpreted as providing partial support for the primary qRT-PCR results. Third, our analyses were restricted to the motor cortex. Given the well-established regional heterogeneity of TDP-43 pathology in ALS, it remains unclear whether the subtype-specific differences in cryptic splicing observed here are similarly present in other affected brain regions. Future studies incorporating multi-regional sampling will be important to address this question. Finally, the present study focused on cryptic RNA accumulation and did not systematically assess cryptic peptide expression at the protein level. Integrating cell-level analyses, quantitative neuropathological assessments, and cryptic peptide biomarkers may provide further insight into the biological basis and clinical significance of subtype-specific differences in TDP-43-related cryptic splicing.

Recent studies have demonstrated that TDP-43-associated cryptic splicing can generate de novo proteins containing cryptic peptide sequences and that some of these proteins, including cryptic HDGFL2 peptides, may serve as sensitive molecular markers of TDP-43-associated pathology [4, 5, 18]. Furthermore, cryptic HDGFL2 peptides have been detected in cerebrospinal fluid and blood, and have emerged as promising fluid biomarkers of TDP-43 proteinopathy [5]. Although the present study focused on cryptic RNA accumulation in postmortem motor cortex tissue, integrating transcript-based measurements with cryptic peptide biomarkers may provide a more comprehensive assessment of TDP-43-associated molecular pathology.

Taken together, our findings suggest that integrating pathological subtypes with molecular readouts can reveal subtype-specific differences in TDP-43-associated cryptic splicing burden. These molecular measures complement conventional neuropathological classification and provide additional insight into the heterogeneity of TDP-43 proteinopathy in ALS. Future studies incorporating cell-level analyses and multi-regional sampling will be important for clarifying the biological basis of these subtype-specific differences.

Abbreviations

ALS

Amyotrophic lateral sclerosis

TDP-43

TAR DNA-binding protein 43

FTLD

Frontotemporal lobar degeneration

FTLD-TDP

Frontotemporal lobar degeneration with TDP-43 pathology

qRT-PCR

Quantitative real-time polymerase chain reaction

cDNA

Complementary DNA

RIN

RNA integrity number

ddPCR

Droplet digital PCR

PMI

Postmortem interval

Author contributions

MT, AKA, OO, and YMK contributed to the conception and design of the study. II, SH, AKO, YH-Y, MY, MT, and YMK contributed to the acquisition and analysis of data. II, SH, AKO, YH-Y, MY, AS, TI, MT, AKA, OO, and YMK contributed to drafting the manuscript and preparing the figures. All authors read and approved the final manuscript.

Funding

This work was supported by grants from the Japan Agency for Medical Research and Development (AMED) (23ek0109641h0001, MoonShot 7 Program JP24zf0127012, SICORP JP24jm0210097); the Japan Society for the Promotion of Science (JSPS) (Grants-in-Aid for Scientific Research (C) 24K10635 and 24K10506, Program for Forming Japan’s Peak Research Universities (J-PEAKS) JPJS00420240016); the Japan Science and Technology Agency (JST) (PRESTO, Sakigake, JPMJPR248A, Fundamental Understanding of Age-Related Organismal Transformations); and the Uehara Memorial Foundation.

Data availability

The datasets supporting the conclusions of this article are included within the article and its additional files. The datasets generated and/or analyzed during the current study are available from the corresponding author on reasonable request. Source data underlying all quantitative figures presented in the main text and supplementary materials are provided in Additional file 3 (Source Data).

Declarations

Ethics approval and consent to participate

The study was approved by the Institutional Review Board of Niigata University (G2020-0031), and written informed consent for autopsy and research use was obtained from the patients’ families.

Consent for publication

Not applicable.

Competing interests

The authors declare that they have no competing interests.