Blood Proteome Profiling Reveals Biomarkers and Pathway Alterations in Fragile X PM at Risk for Developing FXTAS

Fragile X-associated Tremor/Ataxia Syndrome (FXTAS) is a neurodegenerative disorder associated with the FMR1 premutation. Currently, it is not possible to determine when and if individual premutation carriers will develop FXTAS. Thus, with the aim to identify biomarkers for early diagnosis, development, and progression of FXTAS, along with associated dysregulated pathways, we performed blood proteomic profiling of premutation carriers (PM) who, as part of an ongoing longitudinal study, emerged into two distinct groups: those who developed symptoms of FXTAS (converters, CON) over time (at subsequent visits) and those who did not (non-converters, NCON). We compared these groups to age-matched healthy controls (HC). We assessed CGG repeat allele size by Southern blot and PCR analysis. The proteomic profile was obtained by liquid chromatography mass spectrometry (LC-MS/MS). We identified several significantly differentiated proteins between HC and the PM groups at Visit 1 (V1), Visit 2 (V2), and between the visits. We further reported the dysregulated protein pathways, including sphingolipid and amino acid metabolism. Our findings are in agreement with previous studies showing that pathways involved in mitochondrial bioenergetics, as observed in other neurodegenerative disorders, are significantly altered and appear to contribute to the development of FXTAS. Lastly, we compared the blood proteome of the PM who developed FXTAS over time with the CSF proteome of the FXTAS patients recently reported and found eight significantly differentially expressed proteins in common. To our knowledge, this is the first report of longitudinal proteomic profiling and the identification of unique biomarkers and dysregulated protein pathways in FXTAS.


Introduction
The prevalence of various neurodegenerative diseases, such as Alzheimer's dementia and Parkinson's disease, has risen in recent years among many populations due to the increase in the aging population.Developing effective treatments for these complex disorders is challenging due to the complex underlying molecular mechanisms involved, the lack of biomarkers for early diagnosis, the broad spectrum of symptoms, limited natural history, data, and the difficulty in conducting clinical trials with small patient populations.Identifying biomarkers and changes in the associated pathways, particularly in assays, that can quickly and objectively indicate changes in disease pathology is crucial for improving patient outcomes.
Fragile X-associated Tremor/Ataxia Syndrome (FXTAS) is a late-onset neurodegenerative disorder with an average age of onset of 62 that affects carriers of a premutation (PM) allele (55-200 CGG repeats) in the fragile X messenger ribonucleoprotein 1 (FMR1) gene, usually presenting with a more severe clinical phenotype in males, likely due to the presence of a second X chromosome in females [1,2].The high prevalence of the premutation allele among the general population (1:430 males and 1:110-200 females) leads to an estimate of approximately 1.5 million individuals in the general US population who are at risk for FMR1 premutation associated disorders over their life spans.In addition, among the PM population, an estimated 8-16% of females and 40-60% of males are at risk of developing FXTAS [2,3].
Currently, there is no effective specific treatment for FXTAS, and the motor/cognitive symptoms progressively worsen over time, causing reduced quality of life, increased medical expenses, and eventually premature death.FXTAS is clinically distinguished by the presence of intention tremor, cerebellar ataxia, global brain atrophy and white matter disease, autonomic dysfunction, progressive Parkinsonism, and ubiquitin-positive intranuclear inclusions in brain astrocytes, neurons, and Purkinje cells [4].It is caused by the expanded CGG repeats (55-200 CGG) in the 5 UTR of the FMR1 gene.In those with the normal FMR1 gene, the number of CGG repeats lies between 5 and 44, while individuals carrying alleles with a repeat expansion greater than 200 develop fragile X syndrome (FXS), the most common form of intellectual disability and monogenic cause of autism spectrum disorder (ASD) [5].At the molecular level, the eight-to tenfold increase in the level of FMR1 mRNA in a PM containing the expanded CGG repeats [6] leads to RNA toxicity and ultimately to neurodegeneration.Three main mechanisms have been proposed to explain the pathogenesis of FXTAS, including the sequestration of CGG-binding proteins amplified by the elevated levels of FMR1 mRNA, the production of toxic FMRPolyG proteins due to RAN translation, and the chronic activation of the DNA damage response [7,8].
Mass spectrometry (MS)-based proteomics, which involves the advance of data mining and bioinformatic analysis to examine protein structure and function, can be used as an effective technology to quickly analyze large amounts of clinical and biological information within a given sample [9].Recent advances in proteomic profiling technology and processing have also made it possible to efficiently analyze hundreds of proteins, precisely obtain a snapshot of the altered pathways in an organism and identify biomarkers for disease development and progression [10].Although these MS-based proteomic workflows for biomarker discovery and profiling are well established, studies focused on blood proteome profiling and, importantly, on samples collected at different time points have not been carried out in PM at risk of FXTAS.
Recently, Ma and colleagues (2019) performed LC-MS/MS-based proteomics of the intranuclear inclusion isolated from postmortem FXTAS brain tissue.Their work highlighted the presence of more than 200 proteins within the inclusions, including a high abundance of SUMO2 and p62/sequestosome-1 (p62/SQSTM1), supporting a model where the inclusion formation results from increased protein loads and elevated oxidative stress [11].Later, based on these observations, a proteomic profile was characterized in the FXTAS cortex as compared to those obtained from healthy controls (HC) [12].Specifically, a significant decrease in the abundance of proteins including tenascin-C (TNC), cluster of differentiation 38 (CD38), and phosphoserine aminotransferase 1 (PSAT1) was observed in these samples.In addition, the authors confirmed the significantly high abundance of novel neurodegeneration-related proteins and of the small ubiquitin-like modifier 1/2 (SUMO1/2) in the FXTAS cortex as compared to HC [12].Finally, a recent study reported changes in the levels of many proteins, including amyloid-like protein 2, contactin-1, afamin, cell adhe-sion molecule 4, NPC intracellular cholesterol transporter 2, and cathepsin, by comparing the cerebrospinal fluid (CSF) proteome of FXTAS patients with the CSF of HC patients.Changes in acute phase response signaling, liver X receptor/retinoid X receptor (LXR/RXR) activation, and farnesoid X receptor (FXR)/RXR activation pathways were observed [13].
Importantly, no study evaluating predictive biomarkers by blood proteomic alterations in PM, who developed symptoms of FXTAS over time has been reported to date.Here, we present our findings on global profiling derived from male participants enrolled in an ongoing longitudinal study carried out at the UC Davis MIND Institute.The participants have been followed for at least two longitudinal time points: Visit 1 (V1) and Visit 2 (V2).At each time point, neuroimaging, neuropsychological, and molecular measurements, as well as medical and neurological examinations, were collected.A fraction of the premutation participants, all symptom-free or not meeting criteria for FXTAS diagnosis at the time of enrollment (V1), developed symptoms later on (V2) that warranted a diagnosis of FXTAS during the study and were defined as converters (CON).The remaining premutation participants who did not develop symptoms that warranted a diagnosis of FXTAS by the time of the follow-up visit at (V2) are here defined as non-converters (NCON).In the current work, we performed the blood proteome profiling of PM, including CON and NCON, at both V1 and V2 and compared it to HC.We identified a number of potential predictive proteomic biomarkers for early diagnosis, as they showed significant changes in expression levels over time only in the converter group, and we also reported the altered protein pathways among the groups, suggesting their role in the pathogenies of the disorder.

Demographics
DNA testing confirmed the presence of a premutation allele in the PM group, with the participants who converted at V2 (CON; n = 17) and PM who did not convert at V2 (NCON; n = 19), and the absence of a premutation allele in the healthy control (HC; n = 12) group.Participant ethnicity did not differ significantly between the three groups.CGG repeat numbers were significantly lower in healthy controls compared to the other two groups (p < 0.001 in both comparisons) but not significantly different between CON and NCON.Healthy controls were significantly younger than non-converters (p = 0.0319), as reported in Table 1.

Differential Protein Expression between Healthy Control and Premutation Groups
To identify biomarkers potentially associated with the development and progression of FXTAS, we compared the blood proteomic profile of HC to the PM, including CON and NCON.The groups display a separation trend, as shown in Figure 1.A sparse partial least squares discriminant analysis (s-PLSDA) was performed, which showed that all samples from each group aggregated, and the separation between groups indicated differences in the proteomic characteristics between PM and HC and between CON and NCON.A total of 79 proteins were identified by s-PLSDA analysis to be features that separate the groups.Out of these, 78 were among the list of significantly differentially expressed proteins in differential expression analysis using limma.Their expression profile is summarized in Table 2 and Figure 2.
from each group aggregated, and the separation between groups the proteomic characteristics between PM and HC and between C of 79 proteins were identified by s-PLSDA analysis to be features t Out of these, 78 were among the list of significantly differentiall differential expression analysis using limma.Their expression p Table 2 and Figure 2.   from each group aggregated, and the separation between groups indicat the proteomic characteristics between PM and HC and between CON and of 79 proteins were identified by s-PLSDA analysis to be features that sepa Out of these, 78 were among the list of significantly differentially expre differential expression analysis using limma.Their expression profile is Table 2 and Figure 2.

Identification of Proteomic Biomarkers of FXTAS
From this untargeted proteomic profiling, we identified 227 proteins that showed significant changes in expression (adjusted p < 0.05) in pairwise comparisons of the CON as compared to the NCON at V1 (Table 3) and 196 proteins at V2 (Table 4).Between the CON and NCON, we observed 67 proteins that were consistently differentially expressed (adjusted p < 0.05) at V1 and kept changing at V2 (Table 5).While comparing the visits, we identified 170 differentially expressed (adjusted p < 0.05) proteins between V1 and V2 in the converter group, suggesting their role as biomarkers for the progression of FXTAS (Table 6).We further identified the pathways that are altered from V1 to V2 in CON and NCON, including protein lipids and amino acids.Upon examination of protein pathways that were altered between visits in NCON and CON (Figure 3), we found that pathways associated with cell signaling, immune function, cellular organization growth and proliferation, and inflammatory response were those that were more significantly altered from V1 to V2 in the CON group.Similarly, when investigating the protein pathways altered between NCON and CON at V1 or V2, we found that pathways related to synapse signaling (retrograde endocannabinoid signaling pathway) and lipid metabolism were more significantly altered between NCON and CON at V2 (Figure 4).Interestingly, when investigating the list of consistently differentially expressed proteins between CON and NCON groups at V1 and V2, we observed that the pathways related to neurodegeneration are ranked among the top enriched pathways, including the pathways of neurodegeneration, Huntington's disease, and Alzheimer's disease (Figure 5), which provides confidence that the potentially relevant biomarkers may be among these proteins.From the gene ontology point of view, the proteins that are consistently differentially expressed between CON and NCON at both visits are enriched in mitochondrial functions, protein synthesis machinery, and transport, as well as positive regulation of the BMP signaling pathway (Figure 6).These suggest the association of this list of proteins with FXTAS development, similar to other neurodegenerative disorders.Further, upon development of FXTAS at V2 (Figure 7), we observed a high level of dysregulation in retrograde endocannabinoid signaling pathways, mRNA surveillance pathways, cancer, cGMP-PKG signaling, calcium, sphingolipid, and lipid pathways, as observed in other neurodegenerative disorders such as Alzheimer's disease, dementia, and Parkinsonism [14].Further investigating the lipid and amino-acid metabolism, we detected various associated proteins that were differentially expressed in CON as compared to NCON at V2, suggesting their role in the progression of FXTAS (Figure 8).

Differentially Expressed Common Proteins Identified from CSF and Blood Proteomic Profiling
We compared the blood proteome profile of CON at V2 with the recently reported cerebrospinal fluid (CSF) proteome of FXTAS patients.The CSF proteome identified 414 proteins, out of which 46 were identified to be significantly altered between FXTAS patients and controls [13].In the present study of the blood proteome, we identified a total of 2166 proteins, of which 97 were found to be common with the CSF proteome, and eight proteins were significantly altered in both studies, including Complement C3, Alpha-2-HS-glycoprotein, Pigment epithelium-derived factor, Inter-alpha-trypsin inhibitor heavy chain H2, Retinol-binding protein 4, Alpha-2-macroglobulin, Prothrombin, and Lumican (Figure 9).We compared the blood proteome profile of CON at V2 with the recently reported cerebrospinal fluid (CSF) proteome of FXTAS patients.The CSF proteome identified 414 proteins, out of which 46 were identified to be significantly altered between FXTAS patients and controls [13].In the present study of the blood proteome, we identified a total of 2166 proteins, of which 97 were found to be common with the CSF proteome, and eight proteins were significantly altered in both studies, including Complement C3, Alpha-2-HS-glycoprotein, Pigment epithelium-derived factor, Inter-alpha-trypsin inhibitor heavy chain H2, Retinol-binding protein 4, Alpha-2-macroglobulin, Prothrombin, and Lumican (Figure 9).[13].Blood proteome profile of CON at V2 (in pink) identified a total of 2166 proteins of which 97 were found in common.By looking at significantly altered proteins from blood proteomic profile (n = 110) and CSF proteomic profile (n = 46), 8 proteins were found to be in common (indicated in red ink), including Complement C3, Alpha-2-HS-glycoprotein, Pigment epithelium-derived factor, Inter-alpha-trypsin inhibitor heavy chain H2, Retinol-binding protein 4, Alpha-2-macroglobulin, Prothrombin, and Lumican.

Discussion
The identification of protein biomarkers and altered molecular pathways in FXTAS is a crucial requirement for both the research and clinical communities as it improves our  [13].Blood proteome profile of CON at V2 (in pink) identified a total of 2166 proteins of which 97 were found in common.By looking at significantly altered proteins from blood proteomic profile (n = 110) and CSF proteomic profile (n = 46), 8 proteins were found to be in common (indicated in red ink), including Complement C3, Alpha-2-HS-glycoprotein, Pigment epithelium-derived factor, Inter-alpha-trypsin inhibitor heavy chain H2, Retinol-binding protein 4, Alpha-2-macroglobulin, Prothrombin, and Lumican.

Discussion
The identification of protein biomarkers and altered molecular pathways in FXTAS is a crucial requirement for both the research and clinical communities as it improves our ability to identify individuals most at risk for the disease as well as to create novel targeted therapies.There are multiple proteins and pathways that were found to be highly implicated in FXTAS.SUMO2 and p62/sequestosome-1 (p62/SQSTM1) proteins have been observed to accumulate in intranuclear inclusions isolated from postmortem FXTAS brain tissue [11], while tenascin-C (TNC), cluster of differentiation 38 (CD38), and phosphoserine aminotransferase 1 (PSAT1) have been observed in FXTAS cortex [12].Furthermore, it is worth noting that previous studies have examined the proteomic profile of cerebrospinal fluid (CSF) in individuals with Fragile X-associated Tremor/Ataxia Syndrome (FXTAS), highlighting alterations in proteins and pathways when compared to healthy controls [13].However, to the best of our knowledge, our study represents the first longitudinal investigation of blood proteomic changes specifically in PM, some of whom exhibit progressive symptoms of FXTAS over time.These findings provide valuable insights into the potential role of these proteomic alterations as biomarkers for early diagnosis, disease progression, and the overall development of FXTAS.
We observed a number of important proteins altered between HC and PM, including both CON and NCON (Table 2).Further, we found that a number of those proteins associated with various important pathways are dysregulated between CON and NCON at V1 (Table 3), V2 (Table 4), and even between visits (Tables 5 and 6).Interestingly, most of these significantly dysregulated proteins are linked to essential pathways and reported to be involved in the development of other age-related neurodegenerative disorders like Alzheimer's disease, dementia, and Parkinsonism.
In our previous study, we reported lipid and amino acid metabolism dysregulation along with mitochondrial dysfunction in individuals developing FXTAS over time.Specifically, we reported on the clear involvement of different types of lipids in FXTAS and provided evidence of the role that their dysregulation plays in the development and progression of FXTAS [15,16].Specifically, we have identified altered sphingolipid metabolic pathways, including increased levels of sphingosine, sphinganine, and ceramides, in PM who developed FXTAS over time.Further, we reported on decreased levels of the hexosylceramides and lactosylceramides (LCER), both implicated in neuroinflammatory diseases and mitochondrial dysfunction [17,18], common features observed in FXTAS.In this study, we confirmed and validated the previous finding as we observed abrupted sphingolipids and amino acid metabolism (Figure 8) along with mitochondrial dysfunction in PM, including both CON and NCON at the protein level (Figure 6).Indeed, proteomic profiles clearly show a different protein signature among the groups (CON vs. NCON at both V1 and V2), and enrichment pathway analysis demonstrates the involvement of key pathways, including lipids, mitochondria, neurodegeneration, and others, as illustrated in Figures 5-8.Among these proteins, the cytochrome c oxidase subunit Va (COX5A) and the mitochondrial electron transport chain associated protein MT-CO2 were differentially expressed in the CON group (Table 6).As COX5A is involved in maintaining normal mitochondrial function and plays a vital role in aging-related cognitive deterioration via BDNF/ERK1/2 regulation [19], it could represent a potential target for anti-senescence drugs.The mitochondrially encoded cytochrome C oxidase II (MT-CO2) is located in the mitochondrial inner membrane is part of the respiratory chain complex IV, which is defective in individuals with FXTAS.It is a biomarker of Huntington's disease [20] and associated with cerebellar ataxia and neuropathy [21], both clinical features observed in FXTAS.Further, a recent metabolomic study of patients with mitochondrial disease demonstrated elevated acylcarnitine levels, suggesting that an altered fatty acid oxidation pathway may represent a downstream mitochondrial respiratory chain dysfunction [22].Interestingly, we reported high levels of plasma acylcarnitines in the CON group but not in the NCON group [15].
Neural degeneration is a key contributor to the development of neurodegenerative disorders, and we observed a differential expression of the VASP protein in CON.Downregulation of VASP leads to neuronal cell death through an apoptotic pathway and is implicated in the establishment and maintenance of the axonal structure; changes in the expression level can trigger neuronal degeneration [23].In addition, we identified the RNA and mRNA protein pathway dysregulation in CON at V2 (Figure 7), including snRNA-associated Sm-like protein (LSm3), a critical activating factor for mRNA removal in eukaryotic cells participating in RNA metabolism, silencing, and degradation.Abnormal expression of LSM3 has been found to be associated with mild cognitive impairment (MCI) and Alzheimer's disease (AD) [24].
In one of the recent studies, Abbasi and colleagues characterized the cerebrospinal fluid (CSF) proteome of FXTAS patients and reported 317 proteins, among which the expression levels of 38 were significantly altered between FXTAS patients and controls [13].We looked at the overlap with our dataset of 2069 identified proteins and found 97 proteins in common, along with eight significantly altered proteins in both studies (Figure 9).Retinol-binding protein 4 (RBP4) is one of the 8 proteins altered in the CSF as well as in the blood of individuals who developed FXTAS over time.RBP4 is the sole specific transport protein for vitamin A (retinol), and it has been reported that RBP4 can directly induce retinal neurodegeneration in mice through microglia [25].In the CON group, we also observed increased levels of the C3 protein, a key component of the complement cascade signaling pathway and of the immune system that plays a crucial role in inflammation and host defense [26].Overactivation of C3 has also been reported in AD, leading to neuronal damage [27], suggesting its contribution to neurodegeneration in various neurological diseases, including AD and FXTAS.Our findings demonstrate that overactivation of C3 could be contributing to neurodegeneration and that perhaps blocking C3 function could be protective and might lead to the development of strategies for future target treatments.
Among the other proteins, which were commonly differentially expressed in this study using blood from FXTAS (CON V2) and in the study using CSF, were the pigment epithelium-derived factor (PEDF), a unique neurotrophic protein that decreases with aging, the acute-phase protein alpha-2 macroglobulin (A2M), which is a significant component of the innate immune system; the serine protease inhibitor (SERPIN), associated with diverse thrombosis disorders, the interalpha-trypsin inhibitor heavy chain 2 (ITIH2), the small leucine-rich proteoglycan (LUM), a member of the small leucine-rich proteoglycan family playing a role in cancer, adhesion, and migration [28][29][30].These neurodegeneration-associated proteins have also been linked to inflammatory processes [28,31], which are observed in FXTAS pathogenesis and may be promising target pathways for pharmacology.
Finally, blood-brain barrier (BBB) abnormalities have been reported across multiple neurodegenerative disorders such as vascular dementia, MS, Lewy body disease, and spinal muscular atrophy and may contribute to the neurological pathology that often enhances neurodegenerative disorders [32,33].The CSF/serum quotient of albumin (QAlb) is an indirect measurement of the permeability of the BBB, and [13] highlighted the significant correlations between patients' QAlb and their respected CGG repeat length and FXTAS rating scale score.They suggested that the observed higher QAlb levels in their study and also in the CON group from our study presented here were associated with a more severe clinical phenotype and proposed dysregulations in BBB permeability as a clinical prognostic measure for disease severity for patients diagnosed with FXTAS.Of relevance, our findings that disruption in protein levels and associated pathways, detected in both blood and CSF, argue in favor of the use of a less invasive tissue, blood, to be utilized to identify molecular biomarkers, predictors of disease development, severity, and progression.
One of the limitations of this study is the small sample sizes; however, it is important to acknowledge that FXTAS is a disease that has been understudied and is not common, making it challenging to obtain a larger sample pool.Despite these obstacles, longitudinal and additional studies with a larger sample size should be conducted to confirm our findings, identify the most robust and predictive biomarkers, and gain further insight into the disease pathogenesis.Despite these limitations, our study offers valuable information on the proteomic differences between PM, who developed the disorder over time and controls, which can lead to more comprehensive research into the disease's underlying mechanisms and potential therapeutic interventions.

Study Participants
As part of a continuing longitudinal study, male participants PM, over the age of 45 years and male participants with non-carrier age-matched healthy controls (HC) were recruited as detailed in [34].All participants were white in ethnicity, with the exception of three Hispanic participants in the HC group, one in the CON group, and none in the NCON group.The studies and all protocols were carried out in accordance with the Institutional Review Board at the University of California, Davis.All participants gave written informed consent before participating in the study, in line with the Declaration of Helsinki.FXTAS stage scoring was based on the clinical descriptions as previously described [35].Three categories were used in the diagnosis of FXTAS as explained in Zafarullah and Tassone [36] and termed "definite", "probable" and "possible FXTAS".Three age-matched groups were included in this study: CON, NCON, and HC.Using the data from two brain scans, neurological assessment, FXTAS stage, and CGG repeat length, 17 participants were classified as "CON" as they developed clear FXTAS symptomology and thus met criteria of diagnosis between visits (FXTAS stage score was 0-1 at V1 and ≥2 at V2); 19 were defined as "NCON" because they continued to show no signs of FXTAS at V2 (FXTAS stage score was 0-1 at both V1 and V2); and 12 as HC (normal FMR1 alleles/non-PM).The stages of FXTAS range from no tremor at stage 0 to significant tremor that interferes with activities of daily living and intermittent falls at stage 3 [35].

CGG Repeat Length
Genomic DNA (gDNA) was isolated from 5 mL of peripheral blood leukocytes using the Gentra Puregene Blood Kit (Qiagen, Hilden, Germany).CGG repeat allele size and methylation status were assessed using a combination of Southern blot and PCR analysis.Details of the protocols are as previously reported [37,38].

Sample Handling and Preparation
Peripheral blood was collected in cell preparation tube (CPT) vacutainers with sodium citrate (Becton Dickinson, Singapore) and centrifuged according to the manufacturer's recommendations for separating mononuclear cells from whole blood.PBMCs were washed with Dulbecco's phosphate-buffered saline (PBS) and frozen in RPMI 1640 media with 10% fetal bovine serum and 10% dimethyl sulfoxide.Frozen, isolated PBMCs were quickly thawed in a 37 • C water bath, transferred to a 1.5 mL tube, and spun for 20 min to pellet the cells.The freezing medium was removed, and proteins were extracted in 5% SDS in 50 mM triethyl ammonium bicarbonate (TEAB).Protein concentration was determined by BCA assay (Pierce, Appleton, WI, USA), and 150 ug of proteins was digested on an S-Trap™ (ProtiFi, New York, NY, USA) Digestion column plate.Initially, 10 mM dithiothreitol (DTT) was added, incubated at 50 • C for 10 min, and rested at room temperature for 10 min.Next, 5 mM iodoacetamide (IAA) was added and incubated at room temperature for 30 min in the dark.The samples were acidified with 12% phosphoric acid, followed by the addition of freshly made S-trap buffer (90% methanol, 100 mM TEAB, pH 7.1), and mixed immediately by inversion.
The entire acidified lysate buffer mix was transferred to the S-trap plate and pushed through with a Tecan Resolvex A200 (Tecan, Männedorf, Switzerland) until all the solution passed through.Columns were washed with 400 µL of S-trap buffer.Trypsin enzyme digest buffer was carefully added (1:25 enzyme: total protein in 120 µL of 50 mM TEAB, pH 8.0) to the column.After two hours of incubation at 37 • C, the same amount of trypsin and TEAB was added to the S-trap as a boost step, and the reaction continued overnight at 37 • C. The following day, peptides were eluted from the S-trap.Peptide elution steps included 80 µL of 50 mM TEAB (pH 8.0) and 80 µL of 0.5% formic acid 80 µL of the solution containing 50% acetonitrile and 0.5% formic acid.The final pooled elution was dried down in a speed vacuum.Peptides were resuspended in 0.1% TFA and 2% ACN and quantified using the Pierce™ Quantitative Fluorometric Peptide Assay (Thermo Fisher Scientific, Waltham, MA, USA).

Liquid Chromatography Mass Spectrometry (LC-MS/MS)
LC separation was carried out on a Dionex Nano Ultimate 3000 (Thermo Scientific) with a Thermo Easy-Spray source fitted with a PepSep emitter.The digested peptides were reconstituted in 2% acetonitrile/0.1% trifluoroacetic acid, and 5 µL of each sample was loaded onto a Thermo Scientific PepMap 100 C18 5 µm 0.3 mm × 5 mm reverse phase trap, where they were desalted online before being separated on a PepSep 8 cm ID 150 1.5 µm reverse phase column.Peptides were eluted using a 90 min gradient with a flow rate of 0.500 µL/min.The samples were run on an Orbitrap Exploris 480 (Thermo Scientific) in data-independent acquisition (DIA) mode; mass spectra were acquired using a collision energy of 30, resolution of 30 K, maximum inject time mode on auto, and an AGC target of 1000%, using an isolation window of 45.7 Da in the m/z range 350-1200 m/z.Raw spectrometry data and analysis are available from the Massive and Proteome Exchange repositories using the respective ID numbers (MSV000092680, PXD044608).

Data Analysis
DIA data were analyzed using Spectronaut 15 (Biognosys Schlieren, Schlieren, Switzerland), using the direct DIA workflow with the default settings.Briefly, trypsin/P-Specific was set for the enzyme, allowing two missed cleavages.Fixed modifications were set for carbamidomethyl, and variable modifications were set to acetyl (protein N-term) and oxidation.For DIA search identification, PSM and Protein Group FDR were set at 0.01%.A minimum of 2 peptides per protein group were required for quantification.A report was exported from Spectronaut using the reporting feature and imported into SimpliFi (https://simplifi.protifi.com/)for QC and statistical analysis (Protifi, Farmingdale, NY, USA).
For the age and CGG repeats, the p-values are from an ANOVA F-test followed by Tukey HSD pairwise comparisons.Differential expression analyses were conducted using limma-voom.For comparisons between PM and HC participants at baseline, the model used in limma included PM/HC as the only factor.For analyses of PM among participants at V1 and V2, the model used in limma included factors for conversion status, time, and the interaction between conversion status and time, and estimates and standard errors of log fold changes were adjusted for within-participant correlations.Multiple testing corrections were carried out using the Benjamini-Hochberg (BH) approach.Pathway enrichment analysis was carried out using the Wilcoxon rank-sum test on the raw p-values from the differential expression analysis on individual comparisons.Pathway enrichment analysis was carried out using Fisher's exact test on the overlapping list of proteins that are significantly different between CON and NCON at V1 and at V2 using the BH adjusted p-value cutoff of 0.05.Pathway enrichment visualization uses the continuous raw p-values from the enrichment analysis.sPLS-DA analysis was carried out using the R package mixOmics version 3.17 [39].

Conclusions
Currently, there is no effective treatment for FXTAS, and the only options available focus on managing the symptoms.So, a deep understanding of the FXTAS pathogenesis requires the identification of proteins that can be used to understand the altered pathways, serve as biomarkers for early identification of the most at-risk carriers to develop the syndrome, and lead towards the development of targeted therapeutics.However, the investigation of neurodegenerative disorders, including FXTAS, is limited by the availability

Figure 1 .
Figure 1.Blood proteome analysis of the present study.The sparse partial analysis (sparse PLSDA) score plot based on the data of the blood prot non-converters (V1 and V2) and healthy controls.

Figure 2 .
Figure 2. Differential protein expression levels among the HC and the P 79 most significantly altered proteins (adjusted p < 0.05) in the PM group V1 and V2.Heatmap was generated using R code; red indicates high an expression.

Figure 1 .
Figure 1.Blood proteome analysis of the present study.The sparse partial least squares discriminant analysis (sparse PLSDA) score plot based on the data of the blood proteome from converters and non-converters (V1 and V2) and healthy controls.

Figure 1 .
Figure 1.Blood proteome analysis of the present study.The sparse partial least squ analysis (sparse PLSDA) score plot based on the data of the blood proteome fro non-converters (V1 and V2) and healthy controls.

Figure 2 .
Figure 2. Differential protein expression levels among the HC and the PM group 79 most significantly altered proteins (adjusted p < 0.05) in the PM group as comp V1 and V2.Heatmap was generated using R code; red indicates high and blue i expression.

Figure 2 .
Figure 2. Differential protein expression levels among the HC and the PM groups.Heatmap of the 79 most significantly altered proteins (adjusted p < 0.05) in the PM group as compared to HC at both V1 and V2.Heatmap was generated using R code; red indicates high and blue indicates low gene expression.

Figure 3 .
Figure 3. Protein pathways altered from V1 to V2 in CON and NCON groups.Heatma tein pathways that are altered between Visit 2 and Visit 1 in NCON and CON groups.H generated using R code; the color from blue to red indicates the increase in statistical s

Figure 3 .
Figure 3. Protein pathways altered from V1 to V2 in CON and NCON groups.Heatmap of the protein pathways that are altered between Visit 2 and Visit 1 in NCON and CON groups.Heatmap was generated using R code; the color from blue to red indicates the increase in statistical significance.

Figure 3 .
Figure 3. Protein pathways altered from V1 to V2 in CON and NCON groups.Heatmap of tein pathways that are altered between Visit 2 and Visit 1 in NCON and CON groups.Heatm generated using R code; the color from blue to red indicates the increase in statistical signif

Figure 4 .
Figure 4. Protein pathways altered between CON and NCON groups.Heatmap of the prote ways that are altered (p < 0.05) between CON and NCON at V1 and V2.Heatmap was ge using R code; the color from blue to red indicates the increase in statistical significance.

Figure 4 . 31 Figure 5 .
Figure 4. Protein pathways altered between CON and NCON groups.Heatmap of the protein pathways that are altered (p < 0.05) between CON and NCON at V1 and V2.Heatmap was generated using R code; the color from blue to red indicates the increase in statistical significance.Int.J. Mol.Sci.2023, 24, x FOR PEER REVIEW 23 of 31

Figure 5 .
Figure 5. Enriched pathways for the proteins that are consistently differentially expressed between CON and NCON from V1 to V2. Protein-protein interactions from STRING database are represented as edges between proteins.

Figure 5 .
Figure 5. Enriched pathways for the proteins that are consistently differentially expressed between CON and NCON from V1 to V2. Protein-protein interactions from STRING database are represented as edges between proteins.

Figure 6 .
Figure 6.Molecular functions altered between CON and NCON at V1 and V2.Gene ontology molecular functions, including the mitochondrial, protein synthesis machinery and transport, and positive regulation of BMP signaling pathways enriched in the proteins (enclosed in orange circle) that are consistently differentially expressed between CON and NCON at V1 and V2.The blue color represents the proteins that are up-regulated in CON.Yellow represents the down-regulated ones.While Grey is representing the non-differential proteins and Red is FMR1.

Figure 6 . 31 Figure 7 .
Figure 6.Molecular functions altered between CON and NCON at V1 and V2.Gene ontology molecular functions, including the mitochondrial, protein synthesis machinery and transport, and positive regulation of BMP signaling pathways enriched in the proteins (enclosed in orange circle) that are consistently differentially expressed between CON and NCON at V1 and V2.The blue color represents the proteins that are up-regulated in CON.Yellow represents the down-regulated ones.While Grey is representing the non-differential proteins and Red is FMR1.Int.J. Mol.Sci.2023, 24, x FOR PEER REVIEW 24 of 31

Figure 7 .
Figure 7. Significantly altered pathways comparing CON to NCON at V2. Protein-protein interactions from STRING database are represented as edges between proteins.

Figure 7 .
Figure 7. Significantly altered pathways comparing CON to NCON at V2. Protein-protein interactions from STRING database are represented as edges between proteins.

Figure 8 .
Figure 8. Sphingolipid and amino acid metabolism altered in CON.Proteins associated with the sphingolipid and amino-acid pathways, including glycine, serine, and threonine metabolism (enclosed in orange circle), are found to be enriched in the comparison between CON and NCON at V2.The blue color represents the proteins that are up-regulated in CON.Yellow represents the down-regulated ones.The level of the significance is indicated with the intensity of the color.

Figure 8 .
Figure 8. Sphingolipid and amino acid metabolism altered in CON.Proteins associated with the sphingolipid and amino-acid pathways, including glycine, serine, and threonine metabolism (enclosed in orange circle), are found to be enriched in the comparison between CON and NCON at V2.The blue color represents the proteins that are up-regulated in CON.Yellow represents the down-regulated ones.The level of the significance is indicated with the intensity of the color.

Figure 9 .
Figure 9.Comparison of CSF and blood proteomic profile.Cerebrospinal fluid (CSF) proteome of FXTAS patients identified 414[13].Blood proteome profile of CON at V2 (in pink) identified a total of 2166 proteins of which 97 were found in common.By looking at significantly altered proteins from blood proteomic profile (n = 110) and CSF proteomic profile (n = 46), 8 proteins were found to be in common (indicated in red ink), including Complement C3, Alpha-2-HS-glycoprotein, Pigment epithelium-derived factor, Inter-alpha-trypsin inhibitor heavy chain H2, Retinol-binding protein 4, Alpha-2-macroglobulin, Prothrombin, and Lumican.

Figure 9 .
Figure 9.Comparison of CSF and blood proteomic profile.Cerebrospinal fluid (CSF) proteome of FXTAS patients identified 414[13].Blood proteome profile of CON at V2 (in pink) identified a total of 2166 proteins of which 97 were found in common.By looking at significantly altered proteins from blood proteomic profile (n = 110) and CSF proteomic profile (n = 46), 8 proteins were found to be in common (indicated in red ink), including Complement C3, Alpha-2-HS-glycoprotein, Pigment epithelium-derived factor, Inter-alpha-trypsin inhibitor heavy chain H2, Retinol-binding protein 4, Alpha-2-macroglobulin, Prothrombin, and Lumican.

Table 2 .
Differential expression statistics (BH adjusted p-values) among converters, non-converters, and healthy controls.

Table 3 .
Differentially expressed proteins between converter and non-converter at Visit 1.

Table 4 .
Differentially expressed proteins between converter and non-converter at Visit 2.

Table 5 .
Differentially expressed proteins between converter and non-converter at Visit 1 and Visit 2.

Table 6 .
Differentially expressed proteins between Visit 1 and Visit 2 in converter and non-converter.