Long non-coding RNA TUG1 is downregulated in Friedreich’s ataxia

Abstract

Friedreich's ataxia is a neurodegenerative disorder caused by reduced frataxin levels. It leads to motor and sensory impairments and has a median life expectancy of around 35 years. As the most common inherited form of ataxia, Friedreich’s ataxia lacks reliable, non-invasive biomarkers, prolonging and inflating the cost of clinical trials. This study proposes TUG1, a long non-coding RNA, as a promising blood-based biomarker for Friedreich’s ataxia, which is known to regulate various cellular processes. In a previous study using a frataxin knockdown mouse model, we observed several hallmark Friedreich’s ataxia symptoms. Building on this, we hypothesized that a dual-source approach—comparing the data from peripheral blood samples from Friedreich’s ataxia patients with tissue samples from affected areas in Friedreich’s ataxia knockdown mice, tissues usually unattainable from patients—would effectively identify robust biomarkers. A comprehensive reanalysis was conducted on gene expression data from 183 age- and sex-matched peripheral blood samples of Friedreich’s ataxia patients, carriers and controls and 192 tissue data sets from Friedreich’s ataxia knockdown mice. Blood and tissue samples underwent RNA isolation and quantitative reverse transcription polymerase chain reaction, and frataxin knockdown was confirmed through enzyme-linked immunosorbent assays. Tug1 RNA interaction was explored via RNA pull-down assays. Validation was performed in serum samples on an independent set of 45 controls and 45 Friedreich’s ataxia patients and in blood samples from 66 heterozygous carriers and 72 Friedreich’s ataxia patients. Tug1 and Slc40a1 emerged as potential blood-based biomarkers, confirmed in the Friedreich’s ataxia knockdown mouse model (one-way ANOVA, P ≤ 0.05). Tug1 was consistently downregulated after Fxn knockdown and correlated strongly with Fxn levels (R² = 0.71 during depletion, R² = 0.74 during rescue). Slc40a1 showed a similar but tissue-specific pattern. Further validation of Tug1's downstream targets strengthened its biomarker candidacy. In additional human samples, TUG1 levels were significantly downregulated in both whole blood and serum of Friedreich’s ataxia patients compared with controls (Wilcoxon signed-rank test, P < 0.05). Regression analyses revealed a negative correlation between TUG1 fold-change and disease onset (P < 0.0037) and positive correlations with disease duration and functional disability stage score (P < 0.04). This suggests that elevated TUG1 levels correlate with earlier onset and more severe cases. This study identifies TUG1 as a potential blood-based biomarker for Friedreich’s ataxia, showing consistent expression variance in human and mouse tissues related to disease severity and key Friedreich’s ataxia pathways. It correlates with frataxin levels, indicating its promise as an early, non-invasive marker. TUG1 holds potential for Friedreich’s ataxia monitoring and therapeutic development, meriting additional research.

Graphical Abstract

Open in new tabDownload slide

Introduction

Friedreich's ataxia is a debilitating neurodegenerative disorder characterized by the progressive degeneration of the nervous system, resulting in significant motor and sensory impairments.^1,2 The disease ultimately leads to severe physical disability and reduced life expectancy, with the median age of death being 35 years.^3,4 Friedreich's ataxia, being the most common inherited ataxia, is caused by a guanine–adenine–adenine (GAA) trinucleotide repeat expansion in the first intron of the frataxin (FXN) gene.^1,2 This genetic change has a strong correlation with disease severity and its onset.⁵ Notably, while heterozygous carriers of the GAA expansion remain asymptomatic, their prevalence ranges from 1:60 to 1:110 in European populations.^6,7 The need for an effective treatment for Friedreich's ataxia is challenged by the extensive and expensive clinical trials necessary for drug validation.⁸ Therefore, there is an increased emphasis on identifying molecular biomarkers that can quickly monitor disease progression. These biomarkers may expedite evaluations of possible treatments, enhancing patient care and prognosis.⁸

Recent advancements in Friedreich's ataxia research have enabled the assessment of over 20 potential therapeutic interventions in clinical trials.⁸ The Food and Drug Administration (FDA) has recently approved omaveloxolone, the first drug for the treatment of Friedreich's ataxia.⁹ However, its moderate efficacy and associated side effects underscore the continuing need for more effective treatments.⁹ A major obstacle in developing these therapies is the time-consuming and costly nature of clinical trials.⁸ This highlights the urgent need for reliable molecular biomarkers for quicker, evidence-based evaluation of potential treatments. Frataxin, central to Friedreich's ataxia pathology, exists in two primary forms: mitochondrial frataxin (frataxin-M) and an erythrocyte-specific variant (frataxin-E). The full-length precursor, a 210-amino acid protein, undergoes mitochondrial processing to yield frataxin-M, while frataxin-E, lacking a mitochondrial targeting sequence, is found predominantly in erythrocytes.^10-12 This distribution challenges the direct measurement of frataxin in serum or plasma, as it is not commonly secreted into systemic circulation. Moreover, the presence of these proteoforms in different cell types adds variability to frataxin measurements in whole blood.^10-12 Given the challenges in detecting FXN directly, and the intricacies of frataxin's distribution, identifying alternative biomarkers that correlate with FXN levels or disease progression is paramount. For Friedreich's ataxia therapies targeting increased frataxin levels, monitoring frataxin itself is the most direct approach. However, for symptomatic treatments of Friedreich's ataxia, additional biomarkers measurable in peripheral tissues, such as blood samples, are essential. Discovering such biomarkers in Friedreich's ataxia has been challenging, primarily because of difficulties accessing affected tissues at various disease stages and the lack of animal models that faithfully mimic human disease manifestations.

In our prior research, we presented the Friedreich's ataxia knockdown (FRDAkd) mouse model, which closely replicates the symptoms seen in Friedreich's ataxia patients, offering insights into disease progression and the potential for recovery when frataxin expression is restored.¹³ In this model, where FXN expression is reduced, key Friedreich's ataxia symptoms are evident, including ataxia, early mortality, muscle atrophy, degeneration of the dorsal root ganglia (DRG), impaired structural integrity of spinal cord axons and myelin, cardiomyopathy and iron overload.¹³ Utilizing this model, this study aims to address the absence of a high confidence molecular biomarker for Friedreich's ataxia. We hypothesize that through the integrative analysis of genomic data sets from (i) Friedreich's ataxia patient samples and (ii) samples from the tissues of FRDAkd mice, which are primarily affected in Friedreich's ataxia patients, we can identify robust biomarkers intricately associated with the severity and progression of the disease.

In this study, we focus on the long non-coding RNA (lncRNA) taurine-upregulated gene 1 (TUG1) as a prospective biomarker for Friedreich's ataxia. By analysing genomics data from both human and mouse models, we detected reduced expression of TUG1 in Friedreich's ataxia. Our results reveal a significant correlation between Tug1 and Fxn levels, highlighting the association of Tug1 downregulation with the disease. We suggest that evaluating TUG1 levels could provide a non-invasive metric for disease onset, progression and severity. Such a biomarker could significantly enrich the therapeutic landscape of Friedreich's ataxia, enabling prompt therapeutic interventions and improving long-term evaluations. In this paper, we report these findings and discuss their implications for the prognosis and management of Friedreich's ataxia. These observations, based on rigorous genomics analysis and a deep understanding of Friedreich's ataxia biology, underscore the potential of lncRNA TUG1 as a promising molecular biomarker in Friedreich's ataxia.

Materials and methods

Analysis of human and mouse gene expression data

A total of 733 peripheral blood samples from Friedreich's ataxia patients (411), carriers (228) and controls (94) were reanalysed to identify blood-based biomarker for Friedreich's ataxia [Gene Expression Omnibus (GEO) data set: GSE102008].¹⁴ To avoid any confounding effect in this data series, we conducted differential gene expression analyses on age- and sex-matched samples (Fig. 1A). We established a subcohort of 183 samples from the GEO data set, with stringent matching for age and sex. The subcohort included 72 Friedreich's ataxia patients (53% male, 47% female, average age 41), 68 carriers (56% male, 44% female, average age 46) and 43 controls (58% male, 42% female, average age 44). This careful selection resulted in a subcohort that mirrors the larger data set with a gender balance of 101 males and 82 females and relevant age averages across the groups, permitting accurate differential gene expression analyses. For the mouse data, a total of 192 microarray data sets from 64 RNA samples derived from FRDAkd mice were analysed to compare and examine the overlap of the differentially expressed genes obtained from Friedreich's ataxia patient data and mouse data (GEO data set: GSE98790).¹³ Both raw data were log transformed and checked for outliers. Interarray Pearson correlation and clustering based on variance were used as quality control measures. Quantile normalization was used and contrast analysis of differential expression was performed by using the LIMMA package (RRID:SCR_010943). Briefly, a linear model was fitted across the data set, contrasts of interest were extracted and differentially expressed genes for each contrast were selected using an empirical Bayes test statistic.¹⁵

Animal study design and ethics

All animal experiments were carried out in accordance with relevant guidelines and regulations and upon the approval of an Institutional Animal Care and Use Committee (IACUC) at the University of Florida (protocol number # 201909663) and in compliance with the ARRIVE guidelines. Wild-type mice were C57BL/6J from the Jackson Laboratory, and transgenic mice were FRDAkd mice in C57BL/6J background. Mice were sorted into four different groups: wild type with doxycycline (Wt + dox), transgenic with doxycycline (Tg + dox), without doxycycline (Tg − dox) and transgenic with dox removal–rescue (Tg + res). The number of mice (N) was four for each time point. The male to female ratio was 1:1 for all groups. Animals were to be euthanized at Weeks 0, 2, 3, 4, 6, 8, R1, R2, R3, R4, R6 and R8 (R = rescue − dox removal). Transgenic with dox and wild type with dox animals were given 2000 mg/kg dox hyclate (∼87% dox) diet, TD.09633 from Envigo. Transgenic without dox animals were given normal diet. Transgenic with rescue were switched from dox hyclate diet to normal diet.

Mice tissue and blood sample collection protocol

Animals were deeply anaesthetized with isoflurane, and cardiac puncture was performed to collect the blood in ethylenediamine tetraacetic acid (EDTA)-covered RNAse-free tube. The blood was then centrifuged at 3000 × g for 15 min. The supernatant (the plasma) was collected and snap frozen with liquid nitrogen. Mice were placed on petri dish on ice, and dissection of the heart, liver, brain and muscle from the femur was performed. The spinal column was isolated, and hydraulic extrusion of the spinal cord was performed with ice-cold 1× phosphate-buffered saline (PBS). All tissues were cut and rinsed in ice-cold 1× PBS and quickly transferred to 1.7 mL sterile, RNAse-free tubes to be snap frozen with liquid nitrogen. All samples were stored in −80°C.

RNA isolation, cDNA preparation and quantitative reverse transcription polymerase chain reaction protocols

Tissue samples of the blood, muscle, heart, spinal cord, brain and liver from FRDAkd mice were retrieved from a −80°C freezer for RNA isolation. The PAXgene blood RNA kit (cat #762164) was used for blood samples, while the miRNeasy mini kit (cat #2170040) was utilized for other tissues. SuperScript VILO master mix was employed to synthesize cDNA from the extracted RNA. Depending on the tissue type, total RNA concentrations ranging from 0.5 to 1.0 µg were used to prepare 10 µL cDNA aliquots as per the manufacturer's guidelines. For quantitative reverse transcription polymerase chain reaction (qRT-PCR), iTaq Universal SYBR Green Supermix was used. Each well contained 0.2 µL of the cDNA prep and 500 nM each of forward and reverse primers. The qRT-PCR was conducted using the Bio-Rad CFX96 real-time PCR system. Melt curve analysis was performed to confirm the presence of a single amplicon for each of the primer sets (Table 1).

Validation of frataxin knockdown using enzyme-linked immunosorbent assays in heart tissue

To validate the knockdown of frataxin, heart tissue samples were selected for enzyme-linked immunosorbent assay (ELISA) analysis. Snap-frozen heart tissue aliquots from all 40 mice were retrieved from a −80°C storage and weighed in new tubes. For protein extraction, 10 µL of 1× ELISA lysis buffer (Abcam ab176112 ELISA kit) per milligram of tissue was used in conjunction with a pestle tissue homogenizer. Protein concentration was quantified using the BCA assay kit (cat #23225). Subsequently, 30 µg of total protein from each heart sample was loaded into ELISA wells, and the protocol outlined in the ELISA kit was followed. Frataxin concentrations in the samples were interpolated using a second-order polynomial regression model based on a standard curve generated from a human lyophilized recombinant frataxin protein provided in the kit. Data were plotted using Prism 6.

RNA pull-down assay to investigate Tug1 RNA interactome

To explore the RNA interactome of Tug1, a previously established RNA pull-down protocol for lncRNAs was employed.¹⁶ The secondary structure of lncRNA Tug1 was predicted using RNAstructure Webserver, and an antisense biotinylated DNA oligonucleotide probe with minimal probability of internal base pairing was designed via the Fold algorithm.¹⁷ A non-specific DNA oligonucleotide probe served as a negative control. Heart tissues were initially crosslinked with paraformaldehyde, lysed and sonicated. Lysates were mixed with a hybridization buffer, and an input sample was frozen until use. Subsequent hybridization involved adding biotinylated DNA probes to the lysates for 4 h, followed by overnight incubation with streptavidin beads. The beads were magnetically separated and washed several times. Proteinase K treatment was performed before RNA isolation using the miRNeasy mini kit (cat #2170040). Reverse transcription was carried out with SuperScript IV VILO master mix with ezDNase (cat #11766050). The assay was analysed using reverse transcription quantitative PCR (RT-qPCR) to determine relative enrichment of target molecules in comparison with input samples.

Clinical sample acquisition and ethical approvals

Acquisition protocols for human clinical samples were approved by the Institutional Review Boards (IRBs) at the University of California, Los Angeles (UCLA) and the Children's Hospital of Philadelphia (CHOP). Informed written consent was obtained from all participants or from legal guardians in the case of subjects under 18 years. The sample set comprised 45 healthy controls and 45 Friedreich's ataxia patients for serum samples. Additionally, blood samples were collected from 66 heterozygous carriers of the FXN gene with GAA repeats (unaffected) and 72 Friedreich's ataxia patients.

Venous blood and serum sample collection and storage

Blood samples were collected from each participant using 8.5 mL purple-cap Vacutainer EDTA tubes (Becton, Dickinson and Company, NJ, USA, cat #367861). The tubes were gently inverted 8–10 times postcollection for anticoagulant mixing. The samples were transferred into 2.4 mL Eppendorf tubes (Eppendorf AG, Hamburg, Germany) with sterile pipettes and immediately stored at −80°C in an ultralow freezer. For serum, blood samples were allowed to clot at room temperature for 30 min before undergoing centrifugation at 1500 × g for 10 min at 4°C using a refrigerated centrifuge (Eppendorf AG, Hamburg, Germany). The supernatant serum was pipetted into sterile 1.5 mL Eppendorf tubes and stored at −80°C until further analysis.

Statistical analysis and data interpretation

Data normality was assessed using Kolmogorov–Smirnov, Shapiro–Wilk, or D'Agostino–Pearson tests based on the data set size. For normally distributed data, a one-way or two-way ANOVA was employed for statistical analysis. In cases where the data did not conform to a normal distribution, alternative statistical tests were applied as indicated in the respective figure legends. Grouped data were analysed using a one- or two-way ANOVA with appropriate post hoc multiple comparison tests. All statistical analyses were conducted using the GraphPad Prism 9 software (GraphPad Prism, RRID: SCR_002798). A P-value of ≤0.05 was considered statistically significant. Results were calculated based on N = 4 or more per time point and are presented as mean ± SEM.

Results

Blood-based biomarker discovery in Friedreich's ataxia

In this study, we reanalysed an extensive series of previously published unbiased gene expression profiles comprising 733 individuals, including 411 Friedreich's ataxia patients, 228 carriers and 94 controls, with the aim of identifying blood-based biomarkers specific to Friedreich's ataxia.¹⁴ Differential gene expression analyses were conducted on age- and sex-matched 183 samples, made up of 72 patients, 68 carriers and 43 controls, to eliminate any confounding effects in these data series (Fig. 1A). This analysis identified 293 genes that were differentially expressed with a false discovery rate (FDR) of <10% (Fig. 1B). Subsequent functional annotation analysis of these transcripts revealed predominant Gene Ontology categories, such as adaptive and innate immune response. This finding aligns with previous research, where immune system activation was one of the earliest regulated pathways post-Fxn knockdown.^13,14 Thus, utilizing the differentially expressed 293-gene list derived from Friedreich's ataxia patient whole blood data, we postulate that it is feasible to rank and validate these genes for the identification of potential biomarkers for Friedreich's ataxia.

Biomarker validation of Tug1 and Slc40a1 in FRDAkd mice

Utilizing gene expression data from the FRDAkd mouse model,¹¹ we examined the overlap of the 293 differentially expressed genes obtained from Friedreich's ataxia patient data and mouse data. This analysis focused on three tissues primarily affected in Friedreich's ataxia: the heart, DRG neurons and the cerebellum. We investigated these tissues after Fxn knockdown, achieved through dox treatment, and subsequent rescue following dox removal.¹¹ We found 49 genes that were differentially expressed in both Friedreich's ataxia patient data and FRDAkd mouse gene expression data, with an FDR of <5%. These genes were then ranked based on consistency across samples and a non-parametric Kruskal–Wallis P < 1.2 × 10⁻⁴, and the top nine genes (Slc40A1, Rab32, Tug1, Pabpc4, Gzmm, Hmgb2, Camk2N1, Ccnd2 and Pik3Ip1) were selected for further validation (Fig. 1C). To validate these nine genes and determine if these candidate biomarkers are direct targets due to Fxn knockdown, we performed a focused examination. Specifically, we assessed whether they were differentially expressed in the whole blood as early as 2 weeks after Fxn knockdown in FRDAkd mice. This process aimed to discern if the observed alterations in gene expression were a direct consequence of changes in Fxn levels, thus strengthening the hypothesis that these genes could serve as potential biomarkers for Friedreich's ataxia. Their expression levels were assessed by qRT-PCR at various intervals (0, 2, 3 and 6 weeks) post dox treatment. Among the nine genes, taurine-upregulated gene 1 (Tug1) and ferroportin-1 (Slc40a1) were found to be significantly differentially expressed (one-way ANOVA, P ≤ 0.05) as early as 2 weeks after Fxn knockdown in whole blood (Fig. 1D). The differential expression of Tug1 and Slc40a1 in intracardiac blood positions these genes as exemplary candidate biomarkers for Friedreich's ataxia, particularly appealing due to the minimal invasiveness required for patient sample collection.

Assessment of frataxin levels and correlation with candidate biomarkers in FRDAkd mouse model

Carriers and Friedreich's ataxia patients exhibit reduced frataxin levels when compared with controls. The lateral flow immunoassays conducted on buccal cells revealed frataxin protein levels at 50.5% in carriers and 21.1% in Friedreich's ataxia patients.¹⁸ The FRDAkd mouse model facilitates an exploration of various frataxin-level profiles across different tissues, mirroring those observed in controls, carriers and Friedreich's ataxia patients. These profiles vary based on the dose, duration and rescue of dox treatment. This model afforded the opportunity to investigate the correlation between the expressions of the top two candidate biomarkers (Tug1 and Slc40a1) at different time points and the varying levels of FXN. Utilizing ELISA on heart samples, we demonstrated the FXN-level profiles within the FRDAkd mouse model. After dox treatment, the FXN levels in the heart samples of FRDAkd mice were observed to decrease by 47% by Week 2, and 93% by Week 6 (Supplementary Fig. 1A). Extended exposure to dox treatment yielded lower FXN levels in the heart, while Tg − dox and Wt + dox remained unchanged. At the mRNA level, a significant knockdown of Fxn was detected at Week 6 in the heart, muscle, spinal cord, brain and liver, with evidence of partial rescue of Fxn expression by 8 weeks postdox removal (R8) in the heart, muscle, brain and liver (Fig. 2A).

We next delved into the examination of Tug1 and Slc40a1 gene expression across various tissues in the FRDAkd mouse model to explore the consistency of expression changes across different tissues. Specifically, we analysed the gene expression levels of Tug1 and Slc40a1 in five diverse tissues (heart, muscle, spinal cord, brain and liver) at three time points: immediately after Fxn knockdown at 0 and 6 weeks postdox treatment, followed by rescue at 8 weeks postdox removal (Fig. 2B and C). After reducing Fxn levels, we noted a marked decrease in Slc40a1 expression in both spinal cord and brain tissues. Notably, Slc40a1 mRNA levels did not return to normal immediately upon Fxn restoration, suggesting a delayed recovery process. Conversely, in the heart tissue, Fxn knockdown led to an upregulation of Slc40a1, with subsequent Fxn rescue resulting in a reversal of this expression pattern. In the liver and muscle tissues, there were no evident alterations in Slc40a1 expression attributed to Fxn knockdown (Fig. 2B). It is essential to note that ferroportin-1 is a critical transmembrane protein involved in iron export.¹⁹ Previous findings have highlighted iron metabolism dysregulation in heart autopsies of Friedreich's ataxia patients.^20,21 Moreover, iron concentrations in plasma samples were found to be significantly lower in Friedreich's ataxia patients compared with healthy controls.²² These observations align cohesively with our results, further emphasizing the potential role of Slc40a1 in the pathological mechanisms underlying Friedreich's ataxia.

Analysis of the potential of Tug1 as a specific biomarker for Friedreich's ataxia

Our next candidate, lncRNA Tug1, is known to interact with the polycomb repressor complex and functions in the epigenetic regulation of transcription. Tug1 has been shown as a regulatory factor, involved in cellular processes such as cell proliferation,^23-25 apoptosis,^26-30 cell cycle^24,26,31,32 and mitochondrial bioenergetics.³³ Notably, these processes are also primarily impacted in Friedreich's ataxia. Through our analysis, we discovered that Tug1 was markedly downregulated in all examined tissues following Fxn knockdown, with the exception of the liver. During the Fxn restoration phase, the expression level of Tug1 exhibited complete recovery in the heart and muscle tissues. However, there was only a limited rescue of Tug1 expression in the spinal cord and brain, as depicted in Fig. 2C. This limited rescue can be attributed to residual levels of dox remaining in the CNS after its withdrawal. In summary, these experiments reveal a consistent expression pattern of Tug1 in response to Fxn knockdown and restoration, highlighting its significant downregulation across various tissues and its potential role as a specific biomarker for Friedreich's ataxia.

Analysis of Tug1 and Slc40a1 expression in whole blood during Fxn knockdown and rescue in FRDAkd mice

Following the preliminary screening of candidate genes at early time points, we evaluated the expression of Tug1 and Slc40a1 genes in whole blood across 11 specific time points, consisting of five stages during Fxn knockdown and six stages during Fxn rescue. During the Fxn depletion phase, both Tug1 (R² = 0.71) and Slc40a1 (R² = 0.53) exhibited a strong linear decline. Conversely, during the Fxn rescue phase, Tug1 (R² = 0.74) and Slc40a1 (R² = 0.47) demonstrated a marked linear increase (Fig. 2D). To control for the potential confounding effects of dox treatment, we examined the expression of Tug1 and Slc40a1 in wild-type mice across 0, 6 and R8 weeks of dox treatment and subsequent removal (Supplementary Fig. 1B). No significant alterations in mRNA expression of Tug1 and Slc40a1 were observed in wild-type animals during either dox treatment or withdrawal, supporting the notion that the changes in transgenic mice were attributable to the changes in Fxn levels. In the context of these findings, Tug1 emerged as a more suitable biomarker for Friedreich's ataxia compared with Slc40a1. This is based on Tug1's consistent expression across various tissues and patient samples, making it a reliable biomarker for Friedreich's ataxia. While Slc40a1's role in iron metabolism is crucial for FRDA, Tug1's robust and consistent expression profile highlights its potential as a more effective peripheral biomarker.

Our previous gene expression data in FRDAkd mice revealed significant downregulation of Tug1 in the cerebellum and DRG throughout disease progression, paralleling the trend in Fxn expression (Supplementary Fig. 2A). In Friedreich's ataxia, significant neuronal loss in the dentate nuclei and extensive cerebellar damage have been reported.^4,34,35 Interestingly, we observed that TUG1 expression is highest in the cerebellum among all human brain regions (Supplementary Fig. 2B). Furthermore, in FRDAkd mice treated with dox (Fxn knockdown), we observed a downregulation of Tug1 as early as the third week of treatment (Supplementary Fig. 2C). In summary, Tug1's strong linear correlation with Fxn levels during knockdown and rescue phases, consistency in expression across various tissues and early detection of downregulation in specific regions such as the cerebellum present it as a suitable and compelling candidate biomarker for Friedreich's ataxia.

Tug1 downstream targets are altered in FRDAkd mice

Tug1 functions as a multifunctional regulatory factor and influences a range of cellular mechanisms that are disrupted in Friedreich's ataxia, including cell proliferation,^23-25 apoptosis,^26-30 cell cycle^24,26,31,32 and mitochondrial bioenergetics.³³ To further examine the role of Tug1 in Friedreich's ataxia and validate its potential to serve as a biomarker, we investigated the downstream target genes of Tug1. One study identified 630 genes as Tug1 targets using a modified RNA pull-down assay with promoter microarray analysis in BrU-labelled Tug1-transfected glioma cells.³⁶ Cross-comparison of these 630 targets of Tug1 combined with 33 experimentally verified Tug1 targets against the list of differentially expressed genes in the heart, cerebellum and DRG of FRDAkd mice led to the identification of 40 overlapping genes (Fig. 3A). To validate this, we manually selected the top 16 Tug1 targets based on differential expression P-value, expression in multiple tissues and association with cellular processes affected in Friedreich's ataxia as evidenced by the literature. The selected genes for validation were Vsig4, Nedd1, Mmp2, Casp1, Lcp1, Acads, Bdnf, Acsl4, Cd86, Ly9, Gtdc1, Omg, Clcn3, Apbb1ip, Crym and Ccnd2.

We conducted qRT-PCR experiments on blood, heart, muscle, brain and spinal cord tissues after Fxn knockdown and rescue in FRDAkd mice to validate Tug1 targets. Out of 16 genes tested, 9 in the heart, 5 in the brain, 3 in the muscle, 3 in the blood and 2 in the spinal cord were differentially expressed. Nedd1, for instance, was significantly downregulated in the blood, heart and muscle in FRDAkd mice treated with dox for 6 weeks (Fig. 3). Depletion of NEDD1, a centrosome-localized protein, is linked to senescence induction in mouse embryonic fibroblasts,³⁷ a phenomenon also observed in frataxin-deficient human neuroblastoma cells.³⁸ CCND2, a protein responsible for regulating cell cycle, was regulated by lncRNA TUG1 in bladder cancer, and the authors also reported that expression of CCND2 positively correlated with TUG1 expression.³⁹  Ccnd2 expression was downregulated in the spinal cord and heart. Clcn3, a gene thought to affect vesicle trafficking and exocytosis, was downregulated in the muscle and brain. Disrupted Clcn3 leads to impairment of acidification of synaptic vesicles, which causes severe neurodegeneration.⁴⁰ With frataxin knockdown and the corresponding Tug1 deficiency, Lcp1 was notably downregulated in the blood and brain but upregulated in the heart. Ly9 exhibited downregulation in the blood and upregulation in the spinal cord and heart. Both Lcp1 and Ly9 have been implicated in T-cell activation,^41,42 essential for immune system regulation. In FRDAkd mice, immune system activation was among the earliest pathways affected after frataxin knockdown.¹³ We also observed that several of these target genes including NEDD1, CASP1, CD86, LY9, OMG, CCND2 and CRYM showed differential expression in Friedreich's ataxia patients compared with age- and sex-matched controls. This was evidenced in a data set comprising 183 human blood samples, which included 72 individuals with Friedreich's ataxia, 68 carriers and 43 unaffected controls (Supplementary Fig. 3). These findings strengthen the case for Tug1 as a promising biomarker for Friedreich's ataxia, given the significant alterations in its downstream targets caused by Fxn knockdown in FRDAkd mice.

RNA pull-down analysis of Tug1 target gene associations in FRDAkd heart tissue

To further elucidate the interaction between Tug1 and its target genes in the heart tissue of FRDAkd mice (without frataxin knockdown), we conducted an RNA pull-down experiment. Utilizing a data-driven method with biotinylated probes, we aimed to isolate and specifically identify the target genes binding to Tug1 (Fig. 4A). A subsequent analysis using quantitative PCR on the Tug1 RNA pull-down samples obtained from the heart tissue of FRDAkd mice shed light on the enrichment of specific genes, reinforcing the interactions between Tug1 and its target genes. Through this meticulous approach, we successfully validated the significant association of Tug1 with several target genes, namely Nedd1, Ccnd2 and Lcp1 (Fig. 4B). These results not only suggest but also statistically associate Tug1's involvement in key molecular mechanisms of Friedreich's ataxia, warranting further investigation into its functional roles and potential as a therapeutic target.

Quantification of TUG1 levels in Friedreich's ataxia patients and correlation analysis with key clinical parameters

To investigate TUG1 expression levels in patients with Friedreich's ataxia, we analysed the publicly available microarray data set GSE102008. Our comparative evaluation focused on whole-blood TUG1 expression levels across age- and sex-matched individuals, including Friedreich's ataxia patients (n = 72), heterozygous carriers (n = 68) and healthy controls (n = 43). By employing a one-way ANOVA with Holm–Sidak's multiple comparisons test, we identified a significant downregulation of TUG1 expression in the FRDA cohort (Fig. 5A). Given the previous detection of TUG1 from human serum samples in patients with multiple myeloma,⁴³ we explored its expression in Friedreich's ataxia serum samples. Our analysis of serum samples from 45 patient serum and 45 healthy control serum samples showed a significant downregulation of serum TUG1 expression levels in Friedreich's ataxia patients compared with healthy controls, as confirmed through RT-qPCR, employing the Wilcoxon signed-rank test (Fig. 5B). We also examined and quantified the whole-blood TUG1 expression levels between age- and sex-matched Friedreich's ataxia patients (n = 72) and heterozygous carriers (n = 66). Through RT-qPCR, we detected a significant downregulation (P < 0.05) in comparison with the control group (Wilcoxon signed-rank test; Fig. 5C). The results collectively demonstrated a marked downregulation of TUG1 expression in both whole blood and serum from Friedreich's ataxia patients (Fig. 5A–C). Subsequently, to uncover the functional relevance of TUG1 expression, we constructed a correlation heat map that revealed the relationship between blood TUG1 expression levels (fold-change) and various demographic and clinical characteristics of Friedreich's ataxia patients. The correlation studies determined a substantial inverse correlation between TUG1 fold-change and disease onset (Fig. 5D). Notably, we detected positive correlations with disease duration and the functional disability stage (FDS) score, which are clinical parameters known to directly influence each other (Fig. 5D). In summary, our thorough examination of TUG1 expression levels across multiple cohorts and experimental frameworks has unveiled a distinct gene expression profile in Friedreich's ataxia patients. This novel understanding not only underscores the potential of TUG1 as a therapeutic target but also lays the groundwork for subsequent investigations into the molecular pathways modulated by TUG1 in Friedreich's ataxia.

Regression analyses on TUG1 expression levels and Friedreich's ataxia clinical variables

Given the evidence indicating significant correlations between TUG1 fold-change and various factors including disease onset, duration and the FDS score in Friedreich's ataxia, we further explored these relationships through linear regression analyses. As expected, a key observation emerging from these analyses was a marked association between the FDS and age, emphasizing the clinical significance of these parameters in predicting the progression of the disease (Fig. 6A). Further, the linear regression results validated a negative correlation between TUG1 fold-change and disease onset (P < 0.0037; Fig. 6B), reinforcing the relevance of TUG1 in understanding disease onset. Additionally, positive correlations were established between TUG1 fold-change and both disease duration (P < 0.04) and the FDS score (P < 0.04; Fig. 6C and D).

To further examine the impact of other variables, namely age, sex and GAA expansion (genetic marker of disease severity), on the correlation between TUG1 fold-change, disease onset and FDS score, we conducted a multivariate regression analysis. The FDS emerged with an estimated coefficient of 0.1721, hinting at a positive relationship with TUG1 fold-change. This infers that an increase in the FDS score is paralleled by a comparable increase in TUG1 fold-change. However, this association did not reach the threshold of statistical significance (P = 0.0751), calling for further investigation. In contrast, the variable ‘disease onset’ was characterized by an estimated coefficient of −0.02661, representing a statistically significant negative relationship with TUG1 fold-change (P = 0.0047). This compelling observation implies that elevated TUG1 fold-change is associated with earlier disease onset and hence, more genetically severe cases. Residual analysis further validated our regression model, with residuals showing no clear patterns of heteroscedasticity or non-linearity. This supports the robustness of our findings and underscores the significant association between TUG1 levels and disease onset (Supplementary Fig. 4). Taken together, our findings suggest that patients with increased levels of TUG1 typically experience an earlier disease onset and present with higher FDS scores. On the other hand, patients with lower TUG1 expression tend to have a delayed disease onset and exhibit less severe disease manifestations. These results underscore the potential of TUG1 as a crucial biomarker in the prognosis and management of Friedreich's ataxia.

Discussion

Friedreich's ataxia is a neurodegenerative disorder characterized by complex molecular mechanisms and limited therapeutic interventions.⁴⁴ A significant challenge in the management of Friedreich's ataxia is the rapid monitoring of disease progression and the expedited evaluation of the efficacy of potential treatments.⁸ This challenge motivated our study to explore potential blood-based molecular biomarkers for Friedreich's ataxia. Upon re-examining a comprehensive data set comprising 733 individuals, we identified 293 genes showing differential expression in Friedreich's ataxia patients compared with controls. These genes predominantly function in immune system activities, aligning with existing literature that identifies immune activation as an early pathway regulated following frataxin (Fxn) knockdown.^11,12 This finding not only lends credence to previous studies but also opens avenues for future research in identifying specific biomarkers for Friedreich's ataxia.

Extending the analysis to FRDAkd mice, we identified Tug1 and Slc40a1 as particularly promising biomarkers. These genes demonstrated consistent differential expression in both human patients and the FRDAkd mouse model. We validated the expression of Tug1 and Slc40a1 in both the mouse model and human blood samples, underscoring their potential as biomarkers and highlighting the benefits of minimally invasive sample collection. In FRDAkd mice, these genes were validated in different tissues primarily affected in Friedreich's ataxia, such as the heart, DRG neurons and the cerebellum, among others. Interestingly, Tug1 and Slc40a1 expression was significantly altered as early as 2 weeks following Fxn knockdown. This suggests the prospective value of these genes as early-stage biomarkers, whose expression is directly influenced by Fxn levels.

The FRDAkd mouse model was instrumental in assessing the correlation between frataxin levels and candidate biomarkers. Importantly, Slc40a1 (ferroportin), involved in iron metabolism linked to Friedreich's ataxia pathology, showed tissue-specific expression changes after Fxn knockdown and partial restoration upon Fxn rescue. This observation correlates with previous findings that have implicated iron metabolism dysregulation in Friedreich's ataxia pathology,^45,46 adding an extra layer of complexity and importance to our study. These collective findings contribute to a deeper understanding of the molecular mechanisms underpinning Friedreich's ataxia and underscore the potential utility of Slc40a1 as a therapeutic marker.

lncRNAs such as TUG1 have multifaceted roles in cellular biology, particularly in the epigenetic regulation of transcription.^47-49 Our findings establish Tug1 as a key molecular player that is downregulated in multiple tissues affected by Friedreich's ataxia pathology, specifically in response to Fxn knockdown. Given that TUG1’s involvement extends to cellular processes such as cell proliferation,^23-25 apoptosis^26-30 and mitochondrial bioenergetics³³—processes that are also disrupted in Friedreich's ataxia—the importance of TUG1 in Friedreich's ataxia pathogenesis becomes increasingly evident. We observed a stark downregulation of Tug1 in various tissues, with an exception in the liver, upon Fxn knockdown. Its high expression and unchanged levels in the liver may imply tissue-specific functions of Tug1. Additionally, it is plausible that the half-life of Tug1 RNA in the liver tissue is longer compared with other tissues. Strikingly, the levels of Tug1 were partially restored following Fxn recovery, particularly in heart and muscle tissues. This emphasizes Tug1's potential as a highly specific biomarker for Friedreich's ataxia.

Our analysis further extends to the exploration of Tug1 and Slc40a1 expression in whole blood across 11 time points during both Fxn depletion and recovery stages. Remarkably, Tug1’s strong linear correlation with Fxn levels, especially in comparison with Slc40a1, makes it a more suitable candidate as a peripheral biomarker. The trend of Tug1 expression was statistically significant, showing a linear decrease during the Fxn depletion phase and a linear increase during the Fxn rescue phase. It is important to note that our findings rule out the confounding influence of dox treatment in FRDAkd mice, strengthening Tug1’s candidacy as a Friedreich's ataxia-specific biomarker. Our study also sheds light on the specific downregulation of Tug1 in the cerebellum, which is crucial given the extensive cerebellar damage observed in Friedreich's ataxia.^32,33 The early detection of Tug1 downregulation in the cerebellum and its strong linear correlation with Fxn levels provide compelling evidence for Tug1’s importance in understanding the regional specificity of neuronal damage in Friedreich's ataxia.^50-53 Ultimately, our work delves into the downstream target genes of Tug1, which include genes implicated in cellular processes disrupted in Friedreich's ataxia. For instance, Nedd1, a gene involved in cellular senescence,³⁷ exhibited significant downregulation in multiple tissues in FRDAkd mice, reinforcing the vital role Tug1 may play in the cellular biology disrupted in Friedreich's ataxia. Another notable gene, Ccnd2, which is implicated in cell cycle regulation,⁵⁴ was also known to be regulated by Tug1 and showed tissue-specific alterations. We observed varying expressions of Tug1 targets across different tissues, indicative of Tug1's multifaceted roles in Friedreich's ataxia. This suggests its potential influence on disease progression in a tissue-specific manner. Such variation aligns with current literature on the diverse functions of lncRNAs and their context-dependent regulatory mechanisms. These findings solidify the case for Tug1 not just as a biomarker but as a critical molecular component in understanding the complex mechanisms underlying Friedreich's ataxia pathogenesis.

Our investigation into TUG1 expression in Friedreich's ataxia patients, heterozygous carriers and healthy controls, utilizing public microarray data sets¹⁴ and RT-qPCR, reinforces TUG1's potential as a biomarker. The results show a notable downregulation of TUG1 in both whole blood and serum of Friedreich's ataxia patients. The linear and multivariate regression analyses highlight a significant association between TUG1 levels and clinical variables such as disease onset, duration and the FDS score. Notably, we observed that elevated TUG1 fold-change correlates with earlier disease onset, suggesting its utility in disease monitoring and therapeutic development. Furthermore, TUG1 exhibits a complex relationship with both age of onset and disease duration in Friedreich's ataxia, indicating its diverse roles across different disease stages. Elevated levels of TUG1 in the early stages may indicate early onset, while increases during disease progression could contribute to greater severity and longer duration. This intricate involvement of TUG1 in Friedreich's ataxia's neurodegenerative pathology underscores its multifaceted role, emphasizing the need for further exploration of its potential as a biomarker across various stages of the disease. Periodic assessments of TUG1 levels, using easily accessible samples such as blood and serum, could serve as tools for tracking disease progression and severity. The implications for therapeutic development are profound, potentially transforming drug development processes and guiding targeted interventions. Insights into the molecular pathways of TUG1 could lead to targeted therapies, transforming Friedreich's ataxia drug development.

Our study employs a comprehensive approach and diverse sample analysis, incorporating both human and mouse model data, to substantiate the validity of our findings. The use of robust validation methods and the FRDAkd mouse model¹³ further enhances the reliability of our study, positioning non-coding RNA TUG1 as a promising biomarker for Friedreich's ataxia. Nevertheless, our study has limitations that warrant attention. Despite our efforts to eliminate confounding effects, the potential impact of unrecognized confounding variables on our results cannot be completely ruled out. Therefore, clinical validation through trials is essential for further substantiation. Additionally, our findings require replication in larger cohorts to establish broader validity, and longitudinal studies are crucial for understanding the long-term association of TUG1 with disease progression. Furthermore, the complex interplay between FXN, TUG1 and other cellular processes necessitates more detailed investigation.

In conclusion, our rigorous study underscores TUG1's critical role as a prospective blood-based biomarker for Friedreich's ataxia. Utilizing a robust methodology and in-depth analyses, we have confirmed the functional importance of Tug1 and its downstream targets in tissues specific to Friedreich's ataxia. Importantly, our data reveals a strong correlation between TUG1 expression and key clinical indicators, such as disease onset and FDS scores, further establishing its clinical significance. TUG1 stands out as an early-stage, blood-based marker, thereby emphasizing its potential for minimally invasive diagnostic applications with substantial clinical and therapeutic implications. As such, TUG1 offers a promising path for both monitoring the disease and guiding therapeutic development, contributing to improved patient care and deeper understanding of this complex neurodegenerative disorder. Future research should aim to validate these findings through larger and more diverse patient studies while also focusing on the translational potential of these scientific insights into effective clinical applications.

Supplementary material

Supplementary material is available at Brain Communications online.

Funding

The research presented in this manuscript was financially supported by grants from the Friedreich’s Ataxia Research Alliance (FARA), specifically grant numbers AWD07433 and AWD03468 awarded to V.C., as well as a grant from the Muscular Dystrophy Association (MDA), grant number MDA479997 also awarded to V.C.

Competing interests

The authors report no competing interests.

Data availability

The raw gene expression data for Friedreich's ataxia patients can be accessed via the NCBI Gene Expression Omnibus (GEO) database, under the accession number GSE102008. Similarly, gene expression data sets for FRDAkd mice are publicly available on the GEO database under the accession number GSE98790. All additional data pertinent to this study can be obtained upon reasonable request directed to the corresponding author.

References