Introduction

Restless leg syndrome (RLS) is a neurological condition characterized by an uncontrollable urge to move the legs, typically accompanied by discomfort and unpleasant sensations in the lower extremities.¹ The severity of symptoms typically follows a circadian pattern, worsening in the evening or at night and improving with movement of the lower limbs.^1–4 Although peripheral iron metabolism disorders are common, only 10–20% of RLS patients present with decreased serum ferritin levels.⁵ Therapeutic options for RLS include iron supplements, dopamine agonists, calcium channel ligands and opioids.^6,7 Notably, dopamine agonists can worsen the condition over time, a phenomenon known as augmentation, characterized by a gradual exacerbation of symptoms beyond pretreatment levels and the loss of the circadian pattern typically observed in RLS.^8–10 Dopaminergic dysfunction and brain iron deficiency (BID) are widely believed to play key roles in the pathophysiology of RLS,¹¹ along with the involvement of adenosinergic, GABAergic, opioid and glutamatergic neurotransmission systems.¹² Additionally, inflammatory mechanisms have been implicated in RLS,¹³ as evidenced by elevated levels of inflammatory cytokines¹⁴ and changes in oxidative stress.¹⁵ However, the precise mechanisms remain to be further validated. Elucidating these pathways remains crucial for developing more effective therapies, particularly given RLS’s significant impact on sleep quality and quality of life.

Alterations in dopaminergic signaling are considered one of the primary pathological mechanisms of RLS, evidenced by symptom improvement with dopamine D2/3 receptor (D2/3R) agonists,¹⁶ and worsening with dopaminergic antagonists.^17,18 However, the simplistic view of reduced DAergic function in RLS does not fully explain the circadian rhythm of RLS.¹⁹ Genetic studies have revealed significant associations between RLS and cell adhesion molecules (CAMs). Linkage studies in RLS families have identified at least eight susceptibility loci,²⁰ and genome-wide association studies (GWASs) have so far identified 21 susceptibility genes.^20–22 Among these genes, the myeloid ecotropic viral integration site one homolog (MEIS1), BTB domain containing 9 (BTBD9), and protein tyrosine phosphatase receptor type delta (PTPRD) genes are linked to dopaminergic transmission and iron metabolism.²³ Notably, PTPRD²⁴ may regulate synaptic states by mediating microcellular connectivity in dopaminergic neurons. Thus, revealing the circuits and synaptic connections involving CAMs is crucial. Over 60% of RLS patients report a positive family history, underscoring the significant role that genetic factors play in the etiology of this condition.²⁵ Having a first-degree relative increases the risk of developing RLS by 6–7 times.²⁵ Additionally, pathway analysis from a recent GWAS meta-analysis suggested a potentially pathological role for neurodevelopmental processes,²² which are essential for properly forming and maintaining functional neural circuits in the nervous system. Neurophysiological studies implicate that the dysfunction of complex cortico-subcortical networks may underlie RLS, highlighting the presence of hyperactive excitatory mechanisms and impaired inhibitory control in RLS.^26,27 Such excitability alterations have been associated with altered sensory processing in RLS. Elena et al²⁸ reported increased excitability in the primary somatosensory cortex (S1) and impaired somatosensory gating, which play a critical role in aberrant sensory information processing. Supporting this, deficits in cortical plasticity and sensorimotor integration have also been observed in RLS patients.²⁹

Chronic inflammatory states and oxidative stress may contribute to the pathophysiology of RLS.³⁰ Some studies have evaluated peripheral inflammatory markers in RLS patients and healthy controls^31–34 and identified inflammatory cytokines and immune markers as potential biomarkers.¹⁵ Although proteomic studies have reported upregulation or downregulation of proteins involved in inflammatory processes, these findings should be considered preliminary.¹⁵ The potential role of inflammatory factors in RLS pathogenesis or their value as biomarkers remains to be determined.

Next-generation sequencing offers a powerful high-throughput approach to investigate transcriptomic alterations and associated signaling pathways in RLS patients. Several previous transcriptomic studies have provided valuable insights into RLS genetics and pathophysiology.^33,35 Therefore, we hypothesize that differential mRNA expression patterns and associated signaling pathways in RLS patients, independent of familial genetic aggregation, may play a role in the pathogenesis of this disease. This could provide clues for new diagnostic and therapeutic approaches.

Materials and Methods

Subjects

From September 2022 to September 2023, we consecutively enrolled a total of 95 participants, including 37 RLS patients, 39 unrelated healthy controls (HCs) and 19 healthy relatives of RLS patients with a positive family history (familial healthy controls, fHC) from Shanghai General Hospital affiliated to Shanghai Jiao Tong University School of Medicine.

37 RLS patients were diagnosed by movement disorder specialists using International RLS Study Group (IRLSSG) criteria,³⁶ exhibiting characteristic symptoms (urge to move/worsening at rest with lower-limb sensory disturbances relieved by movement), with an average disease duration of 13.0 years. As in our previous study,⁴ we conducted face‒to‒face interviews to collect the demographic information, disease duration, symptoms, medications, and findings from general neurological and medical examinations. Exclusion criteria included: (1) sleep disorders other than RLS; (2) psychiatric disorders; (3) neurological complications; or (4) use of any medications that could alter the circadian rhythm within three weeks before the study. All patients had no RLS comorbidities and were not taking dopaminergic or α2δ-ligand drugs.

Healthy controls were age-matched, medication-free, and had no history of sleep, physical, neurological, or psychiatric disorders, and primarily were the spouses of patients. 19 fHCs were used as a separate comparison group to explore potential familial or genetic contributions and excluded from the differential expression analysis between RLS patients and HCs. Detailed information on these RLS families is shown in Figure S1.

Peripheral iron status was evaluated by obtaining clinical laboratory data on TfR, ferritin, transferrin saturation (TSAT), and serum iron levels. Written informed consent was obtained from all participants, including those in the pedigree tree. This study was conducted under the principles of the Declaration of Helsinki and was approved by the Ethics Committee of Shanghai General Hospital affiliated to Shanghai Jiao Tong University School of Medicine (Ethics number: 2017–176).

Isolation of Cells, Total RNA Extraction and Quality Control

Total RNA extraction process included blood sample collection after assessments, PBMC isolation within 30 minutes, total RNA extraction and RNA storage. Venous blood was collected in the morning after all clinical assessments after a 12-hour fasting period. Blood samples were drawn into EDTA anticoagulant tubes and promptly transported on ice to the laboratory within 30 minutes to preserve cell viability. PBMCs were isolated via density gradient centrifugation via Ficoll‒Paque (GE Healthcare, USA), and the intermediate layer was extracted. Red blood cell lysis (Sigma‒Aldrich, USA) buffer was used at room temperature for five minutes, and the mixture was centrifuged at 250 ×g and 4°C for 5 min to remove any residual erythrocytes in the cell pellet. TRIzol reagent and DNase (Cat R401-01, Vazyme, Nanjing, China) were used to extract total RNA and ensure purity and accurate quantification according to the manufacturer’s protocol. The same procedures were followed for sample collection, storage, extraction and purification. RNA extraction for PBMC samples in the same group was performed simultaneously and completed within 7 days to ensure meaningful transcriptomic changes. The RNA was stored at −80°C, and its quality was evaluated via a NanoDrop™ 2000 spectrophotometer (Thermo Fisher Scientific, USA). Only samples with a 260/280 absorbance ratio between 1.8 and 2 were considered acceptable for further analysis.

Sequencing Library Construction

MRNA was extracted from total RNA for each group within 7 days using VAHTS^® mRNA Capture Beads (N401-02, Human, Vazyme). Double-stranded cDNA was synthesized from purified polyadenylated mRNA templates using the PrimeScript™ Double Strand cDNA Synthesis Kit (TaKaRa), following the manufacturer’s protocol. cDNA libraries for Illumina sequencing were prepared using the VAHTS Universal V6 RNA-seq Library Prep Kit for Illumina^® (NR604-02, Vazyme). The library concentration was quantified using a Qubit 2.0 Fluorometer (Thermo Fisher Scientific, USA), and library fragment distribution was assessed with an Agilent 4200 TapeStation (Agilent Technologies, China).

Transcriptomic and Functional Enrichment Analysis

Before sequencing, the concentration and size distribution of the cDNA library were assessed using a Qubit 2.0 fluorometer (Thermo Fisher Scientific) and an Agilent 4200 TapeStation (Agilent Technologies) to ensure RNA integrity (RIN ≥ 9). Sequencing was then performed on an Illumina NovaSeq 6000 platform, generating raw image files that were converted into reads. To improve data quality, low-quality reads were filtered out using Seqtk (v1.0) based on the following criteria: (1) removal of adapter sequences, (2) trimming of bases with a quality score below 20 at the 3′ end, (3) exclusion of reads shorter than 25 nucleotides, and (4) removal of ribosomal RNA reads, resulting in a set of clean reads. The spliced mapping algorithm of Hisat2 (v2.0.4) was applied to perform Genome mapping on the preprocessed reads. To ensure comparability of gene expression levels across different genes and samples, read counts were normalized to fragments per kilobase of transcript per million mapped reads (FPKM) values, following the method described in Mapping and quantifying mammalian transcriptomes by RNA-Seq. Gene fragments were counted using StringTie (v1.3.0) after alignment with HISAT2, and normalization was performed using the trimmed mean of M values (TMM) method, as described in A scaling normalization method for differential expression analysis of RNA-seq data. Sample relationships were assessed through correlation analysis and principal component analysis (PCA), with batch effects corrected accordingly. Differential gene expression analysis was conducted using edgeR (v3.2.0), with age and sex included as covariates. Q value was corrected after false discovery rate (FDR) correction applied to control the P-value threshold, and fold-change values were calculated to identify differentially expressed genes. Differentially expressed mRNAs were defined as those identified with a q value ≤ 0.05 and a fold change (FC) ≥ 1.5. The average level of mRNAs was calculated via the average expression function. The selected differentially expressed genes were mapped to GO and KEGG terms. Terms with a q-value ≤ 0.05 after multiple hypothesis correction (FDR correction) were considered significantly enriched. GO enrichment analysis, KEGG enrichment analysis, and visualization were performed via the ClusterProfiler (v4.0.0) R package.

Preparation of cDNA and Real-Time Quantitative PCR Analysis

cDNA synthesis was conducted with 200 ng of total RNA via a TB Green Premix Ex Taq II (Tli RNaseH Plus) Kit (Cat RR820B, TaKaRa, Dalian, China) following the manufacturer’s protocol. Real-time quantitative PCR (RT-qPCR) was conducted via the LightCycler^® 480 System (Roche, Basel, Switzerland) to assess the mRNA expression levels. Three technical and biological replicates were performed to verify PCR efficiency (90–110%), and melting curve analysis was checked for non-specific amplification. The primer sequences were as follows: for SPARC-like protein 1 (SPARCL1), forward: 5′- ACGGTAGCACCTGACAACAC-3′ and reverse: 5′- ATGGTGGGAATCGTCTTCTGT-3; for proinflammatory C–C motif chemokine ligand 2 (CCL2), forward: 5′- CAGCCAGATGCAATCAATGCC-3′ and reverse: 5′- TGGAATCCTGAACCCACTTCT-3′; for proinflammatory C–C motif chemokine ligand 8 (CCL8), forward: 5′- TGGAGAGCTACACAAGAATCACC-3′ and reverse: 5′- TGGTCCAGATGCTTCATGGAA-3′; for endothelial adhesion molecule 1 (SELE), forward: 5′- CAGCAAAGGTACACACACCTG-3′ and reverse: 5′- CAGACCCACACATTGTTGACTT-3′; for Homo sapiens apolipoprotein A-II (APOA2), forward: 5′- CTGTGCTACTCCTCACCATCT-3′ and reverse: 5′- CTCTCCACACATGGCTCCTTT-3′. Target gene expression levels were calculated compared to housekeeping gene 18S rRNA (forward: 5′-GCAATTATTCCCCATGAACG-3′ and reverse: 5′-GGGACTTAATCAACGCAAGC-3′). Relative mRNA expression levels were described using the 2 − ΔΔCt method (ΔCt = Cttarget − Ctreference, − ΔΔCt = sample ΔCT − 18S ΔCT).

Statistical Analysis

Quantitative variables are described as either the median with interquartile range (25th–75th percentiles) or the mean ± standard deviation (SD). Group comparisons were performed using Welch’s t-test or Mann–Whitney U-test. For RNA sequencing data, differential gene expression analysis was performed using edgeR, and p-values were adjusted by FDR correction (q-values). Genes with q < 0.05 and fold-change (FPKM-based) were considered significant. Spearman correlation corrected for age was used to test the association between levels of mRNA levels and clinical index. Statistical significance was defined as *P < 0.05; **P< 0.01; ***P< 0.001; ****P < 0.001; ns, not significant. SPSS version 20.0 (IBM Corporation, USA) was utilized to conduct statistical analyses, and Prism software (version 8.0; GraphPad Software, USA) was used for data visualization.

Result

Clinical Characteristics of the Participants

During the screening phase, we recruited 37 RLS patients, 39 matched HCs, and 19 familial healthy controls (fHCs) to minimize the impact of familial genetic susceptibility (Figure S1). In the validation phase, a subset of 10 RLS patients and 19 HCs were included and matched by sex and age. The demographic and disease characteristics of all participants are summarized in Table 1. Notably, serum iron levels were significantly lower in RLS patients than in HCs (P = 0.0198).

Table 1 Demographic and Clinical Characteristics of Participants

mRNA Expression Profiles of RLS Patients and Healthy Controls

The research design is shown in Figure 1A. Hierarchical cluster analysis performed using the PBMC mRNA results revealed significant differences in mRNA expression between RLS patients and HCs (Figure 1B). In total, 71 differentially expressed mRNAs were identified, with 18 significantly elevated and 53 significantly reduced in RLS patients (q value < 0.05) (Figure 1C; the detailed information is listed in Table S1). Among them, MTND1P23 exhibited the highest upregulation with a fold change of 5.6, while genes such as APOA2 and SPARCL1 showed the most significant downregulation (each with a fold change below −5) (Table S1).

Figure 1 Transcriptomic analysis of RLS patients and healthy controls. (A) Study design and timeline. (B) Heatmap illustrating the differential expression levels of mRNAs identified in restless legs syndrome (RLS) patients and healthy controls (HCs). The X-axis represents the patient and control groups from the cluster analysis, whereas the Y-axis lists the 71 differentially expressed mRNAs. Red and green indicate upregulation and downregulation, respectively. (C) Volcano plot depicting the differences in mRNA expression between patients and HCs. The X-axis represents the log2 (fold change) values, whereas the Y-axis represents the -log10 (q value) values. The gray dots indicate mRNAs without significant expression differences, the red dots represent mRNAs with upregulated expression, and the blue dots indicate mRNAs with downregulated expression. Differentially expressed genes were identified using edgeR, with a false discovery rate (FDR)-adjusted q value ≤ 0.05 and an absolute log2 fold change (|log2FC|) ≥ 1.5. Abbreviations: RLS, restless legs syndrome; HC, healthy control; FDR, false discovery rate.

Functional and Pathway Enrichment Analysis of Differentially Expressed mRNAs

To elucidate the biological roles of the differentially expressed mRNAs, GO and KEGG analyses were performed on 71 genes (q value ≤ 0.05 and |FC| ≥ 1.25) (Figure 2A and B). GO enrichment analysis indicated that these mRNAs were associated with various molecular functions, biological processes, and cellular components (Tables S2–S4). Notably, upregulated genes such as MTCO1P40, carbonic anhydrase 1 (CA1), CBY2, and ubiquitin-associated protein 2 (UBBP2) were found to be associated with the following molecular functions: binding, protein binding and catalytic activity. Among the downregulated genes, APOA2 and SPARCL1 were involved in binding, protein binding, and ion binding. KEGG analysis identified 22 signaling pathways, with the q value threshold set at 0.05, among which the interleukin 17 (IL-17) signaling pathway, tumor necrosis factor (TNF) signaling pathway and CAMs were the most prominent. The differentially expressed mRNAs involved in each pathway, along with the distribution of these pathways across various processes, functions, or organs, are presented in Table S5. Additionally, functional enrichment analysis was performed separately for upregulated and downregulated genes. GO analysis revealed that upregulated genes were primarily enriched in immune response pathways (Figure 2C), while downregulated genes were mainly associated with the regulation of leukocyte migration and humoral immune response (Figure 2D). KEGG analysis further showed that downregulated genes were enriched in cytoskeleton-related pathways in muscle cells (Figure 2E and F).

Figure 2 GO and KEGG analysis of differentially expressed mRNAs in RLS patients. (A) Bubble plot illustrating selected enriched GO terms. The color gradient represents q values, with red indicating smaller q values. Larger bubbles correspond to a greater number of genes involved in each pathway. Different shapes represent molecular functions (MF), biological processes (BP), and cellular components (CC). (B) Bubble plot displaying selected enriched KEGG pathways. The color gradient reflects the q values, with red signifying smaller q values. Larger bubbles indicate that more genes are involved in each pathway. (C) GO analyses of upregulated genes. (D) GO analyses of downregulated genes. (E)KEGG analyses of upregulated genes. (F) KEGG analyses of downregulated genes. For (C–F), the bubble size and color gradient follow the same conventions as in (A and B). GO and KEGG pathway enrichment analyses were performed based on differentially expressed genes (DEGs). P-values were corrected for multiple hypothesis testing using the false discovery rate (FDR) method, with a significance threshold set at q-value ≤ 0.05. GO terms and KEGG pathways meeting this criterion were considered significantly enriched. Abbreviations: RLS, restless legs syndrome; GO, Gene Ontology; KEGG, Kyoto Encyclopedia of Genes and Genomes; MF, molecular function; BP, biological process; CC, cellular component; DEG, differentially expressed gene; FDR, false discovery rate.

mRNA Expression Profiles of Related Familial Healthy Controls

To further investigate the distinct differences in mRNA expression in RLS patients, we compared the mRNA expression profiles of 19 fHC and 39 HCs. In the fHC group, we identified 247 mRNAs with upregulated expression (q < 0.05) and 894 with downregulated expression (q < 0.05) (additional data are listed in Table S6). To minimize the influence of familial genetic aggregation, we filtered out those mRNAs with similar up- or downregulation trends in both the RLS patient and fHC groups. Among them, nine mRNAs were consistently upregulated, and 25 were downregulated in both RLS patients and fHCs (Table S7). Notably, PTN exhibited an increasing trend across the RLS, fHCs, and HCs groups in the sequencing data (Figure S2). The specific and shared DEGs were demonstrated in Figure 3A and B. This process allowed us to focus on the mRNAs demonstrating uniquely upregulated (9 mRNAs) or downregulated (28 mRNAs) expression in RLS patients (q < 0.05, Table 2).

Table 2 37 mRNAs Specifically Differentially Expressed in RLS Patients Compared to Both fHCs and HCs

Figure 3 Venn diagram of specific and shared DEGs between RLS patient and fHC. (A) Upregulated DEGs in RLS patients and fHCs. The numbers indicate unique and shared DEGs between the two groups. (B) Downregulated DEGs in RLS patients and fHCs. The numbers in each section represent the DEGs that are specific to RLS patients, specific to fHCs, or shared between the two groups. Statistical significance determined by an adjusted q-value ≤ 0.05 and |log2FC| ≥ 1. Abbreviations: DEGs, differentially expressed genes; RLS, restless legs syndrome; fHC, familial healthy controls.

Validation of Candidate mRNAs Screened by RT-qPCR

We selected SPARCL1, CCL2, CCL8, SELE and APOA2 from the 37 significantly downregulated mRNAs in RLS patients. The expression levels of these genes were validated in a subset of 10 RLS patients and 19 HCs via RT-qPCR (Table 3). The results confirmed significantly lower levels of SPARCL1, CCL8 and SELE in RLS patients compared to HCs (Figure 4A, C and D), while CCL2 and APOA2 showed no significant differences (Figure 4B and E).

Table 3 Differences in the Expression of the 4 Candidate mRNAs in the Validation Cohort (N = 29)

Figure 4 Validation of differential expression of five candidate mRNAs in an independent cohort. (A)Relative expression levels of SPARCL1 were compared between RLS patients and HCs. (B) Relative expression levels of CCL2 were compared between RLS patients and HCs. (C) Relative expression levels of CCL8 were compared between RLS patients and HCs. (D) Relative expression levels of SELE were compared between RLS patients and HCs. (E) Relative expression levels of APOA2 were compared between RLS patients and HCs. Statistical significance was determined using an unpaired two-tailed t-test. Data are presented as mean ± standard deviation. **P < 0.01; ****P < 0.0001; ns, not significant. Abbreviations: RLS, restless legs syndrome; HC, healthy controls; SPARCL1, SPARC-like protein 1; CCL2, proinflammatory C–C motif chemokine ligand 2; CCL8, proinflammatory C–C motif chemokine ligand 8; SELE, endothelial adhesion molecule 1; APOA2, Homo sapiens apolipoprotein A-II.

Relationships between Five Selected mRNAs and Clinical Indices

Correlation analysis of five selected mRNAs, including SPARCL1, CCL2, CCL8, APOA2 and SELE in patients, as presented in Table 4, revealed no significant associations between their downregulated expression and RLS severity, as assessed by IRLSS or peripheral iron deficiency-related laboratory parameters (TfR, ferritin, TSAT, and iron).

Table 4 Correlation Analysis of Four Candidate mRNAs with Clinical Indicators

Discussion

In this study, we investigated the transcriptomic characteristics of the mRNA expression profiles in RLS patients’ PBMCs. A total of 71 candidate mRNAs were identified by comparing RLS patients and HCs. After excluding the influence of familial aggregation, we identified 9 upregulated and 28 downregulated mRNAs specific to RLS patients. RT-qPCR confirmed significant downregulation of SPARCL1, CCL8, and SELE in RLS PBMCs. GO and KEGG analyses indicated potential involvement of the IL-17 and TNF signaling pathways in the pathogenesis of RLS. While these findings suggest that inflammation and immunity may be important contributing factors to the pathophysiology of RLS, further research is needed to confirm these associations and elucidate the underlying mechanisms.

Among genes with a consistent upregulation trend between RLS patients and fHCs, PTN showed an increasing trend across RLS, fHCs, and HCs, after excluding pseudogenes and novel protein-encoding genes. PTN is known for its role in promoting neurite outgrowth.³⁷ Although this trend was not successfully validated, its dysregulation may still represent a potential genetic risk factor for RLS progression.

Inflammation and Immunity

Our research findings suggest that inflammatory mechanisms are involved in RLS, as indicated by the involvement of the TNF and IL-17 signaling pathways. This aligns with a previous transcriptomic study supporting the involvement of inflammatory network mechanisms and immune processes in RLS. That study identified IL-17, chemokines, and inflammatory mediators of TRP channels, further suggesting that these mediators may be associated with sensory symptoms in RLS patients.³³ Proteomic studies in RLS patients have revealed elevated levels of reactive oxygen species, activation of complement components,¹⁵ and increased levels of inflammatory proteins,¹⁴ further highlighting the role of inflammatory mechanisms. Previous studies have reported significantly higher levels of serum or plasma tumor necrosis factor α (TNF-α), interleukin-1β (IL-1β) and interleukin-6 (IL-6) in RLS patients than in controls, indicating systemic inflammatory changes.³⁸ Our findings reinforce the involvement of the TNF signaling pathway in the pathogenesis of RLS,^31,38,39 suggesting a potential possible link between RLS symptoms and comorbid psychosomatic disorders.³⁹ Interestingly, Trotti et al⁴⁰ reported that among 137 untreated RLS patients, those with frequent periodic limb movements in sleep (PLMS) were more likely to have elevated levels of hypersensitive C-reactive protein (hs-CRP), although the IL-6 and TNF-α levels were not significantly increased.

Increasing evidence indicates that an abnormal immune regulatory response contributes to RLS, as shown by its higher prevalence in multiple sclerosis (MS)⁴¹ and rheumatoid arthritis (RA).⁴² Additionally, an RNA transcriptomic study has implicated the role of inflammatory cytokines in Parkinson’s disease (PD) pathogenesis, further supporting a potential immune-RLS link, given RLS is a common comorbidity in patients with PD.⁴³ We observed an enrichment of differentially expressed mRNAs in the IL-17 pathway in RLS patients, further emphasizing the contribution of immune mechanisms to the pathogenesis of RLS. Autopsy studies on RLS patients showed microvascular T-cell infiltration,^44,45 which may mediate the inflammatory infiltration of microvessels through mechanisms involving leukocyte transendothelial migration. Additionally, alterations in IL-17 may also be associated with RLS by affecting impaired dopamine synthesis, leading to altered dopaminergic signaling.⁴⁶ We also detected the downregulation of CCL2 and CCL8 expression in RLS, which may play a role in the immunomodulatory function of sensory neurons. The anti-inflammatory molecule interleukin 10 (IL-10) has been reported to bind to receptors expressed on somatosensory neurons on the dorsal root ganglion (DRG)⁴⁷ and trigeminal ganglion (TG),⁴⁸ thereby downregulating CCL2.⁴⁹ Given that DRG neuron-derived CCL2 is involved in immune cell infiltration and activation,⁵⁰ the restriction of the ability of somatosensory neurons to respond to these alterations through immune modulation is reasonable. Our findings increase the understanding of neuroimmunity in regulating important sensory symptoms in RLS. In conclusion, our findings strengthen the evidence linking inflammation and immunity to RLS. However, the specific functions of inflammatory and immune factors in the pathogenesis and their potential as biomarkers require further investigation in future studies.

Neurotransmission and Synapse

Dopamine dysfunction in RLS is characterized by dysregulation, including increased presynaptic DA synthesis and release, decreased DA uptake, and reduced postsynaptic receptor density.¹⁹ This alteration in the dynamic relationship between presynaptic and postsynaptic functions involves complex changes in multiple elements. One perspective is that differences in microscopic intercellular connectivity among dopaminergic neurons may explain these changes.⁵¹ Pathway analysis of the differentially expressed mRNAs suggests that CAMs play a role. CAMs are described as potential candidates for playing a central role in determining synaptic connectivity and neural development.⁵² PTPRD, a CAM, has been identified repeatedly associated with RLS in several GWAS.^53–55 Among these studies, single-nucleotide polymorphisms (SNPs) in PTPRD are associated with reduced mRNA expression.^56,57 Axonal PTPRD is involved in motor neuron axon termination and synapse formation during embryonic development⁵⁸ via postsynaptic ligand interactions.^59–65 PTPRD knockout mice exhibit impaired neuronal differentiation and cortical disruption, underscoring its role in neural connectivity.⁶⁶ Notably, PTPRD is highly expressed in dopaminergic neurons, suggesting its involvement in dopaminergic neural circuit remodeling. Although whether PTPRD mutations directly impair dopaminergic neurotransmission remains unclear, these findings support its critical role in the synaptic connectivity, potentially leading to altered susceptibility to RLS.^24,67 Dopaminergic network abnormalities in RLS patients and worsening symptoms with prolonged dopamine agonist use may be linked to CAM-mediated circuit dysfunction.⁶⁸ A detailed elaboration of CAM-regulated synaptic connections could clarify treatment-related complications and identify new therapeutic targets.

A novel finding in our study was the downregulation of SPARCL1 in RLS patients. SPARC, a matricellular protein (MCP), modulates tissue development and remodeling modulator by influencing key cellular processes such as cell adhesion and proliferation.⁶⁹ SPARCL1, expressed primarily in the central nervous system, is reported to be associated with the formation and maturation of excitatory synapses.^70,71 Specifically, SPARCL1 facilitates the development of synaptic N-methyl-d-aspartate (NMDA) receptors, which can be essential for synaptic connectivity.⁷² Clinical evidence suggests the involvement of NMDA receptor-mediated alterations in glutamatergic neurotransmission in RLS pathogenesis. Noncompetitive NMDA receptor inhibitors, such as ketamine, have shown efficacy in alleviating RLS symptoms,⁷³ and another NMDA antagonist, antamethadone, has also proven effective.⁷⁴ Although a direct link between SPARCL1 and RLS has not yet been established, our results may suggest the potential role of SPARCL1 in NMDA receptor-mediated glutamatergic dysregulation in RLS. A recent optogenetic study in rats with BID demonstrated hypersensitivity in corticostriatal glutamatergic terminals.⁷⁵ This hyperglutamatergic state may be related to the hyperarousal often observed in RLS patients.⁷⁶ Given that BID is prevalent in the majority of RLS patients, our results could provide valuable insights into SPARCL1’s role in synaptic adaptation and its potential contribution to the hyperarousal state in RLS.

Additionally, NMDA receptor-mediated glutamatergic neurotransmission may also be associated with maladaptive synaptic plasticity in the brain,^77,78 which could explain RLS-related sensory symptoms such as chronic pain. A study of neuropathic pain in SPARCL1 knockout mice suggested that SPARCL1-mediated enhancement of NMDA currents plays a key role in chronic pain maintenance.⁷⁵ In particular, sensory neurons expressing SPARCL1 have been identified in human sensory ganglia, indicating the participation of SPARCL1 in the transmission of nociceptive signals.⁷⁹ Therefore, one hypothesis is that SPARCL1 may modulate synaptic modifications to regulate the transsynaptic transmission of nociceptive signals, potentially alleviating the discomfort associated with sensory symptoms in RLS. The possible role of SPARCL1 in sensory adaptations is a perspective that has not been previously addressed. However, given the scarcity of relevant studies, the role of SPARCL1 in RLS needs to be further investigated. These observations are clinically significant, as inhibitors targeting inflammatory pathways and synaptic regulation may be promising therapeutic options for RLS.

Although we implemented strict inclusion and exclusion criteria, certain conditions have not been extensively explored concerning RLS, such as nonanemic iron deficiency⁸⁰ and undiagnosed cardiovascular abnormalities.⁸¹ Residual effects of prior medications (eg, immunosuppressants/antidepressants) may have influenced the observed gene expression variability, although none of the participants were actively taking dopaminergic or α2δ-ligand medications. While our findings provide transcriptional insights, limitations include the lack of protein-level validation, a relatively small sample size, and potential discrepancies between peripheral blood transcriptomics and CNS pathology. Expanding the cohort size and incorporating more comprehensive clinical assessments, such as sleep parameters, disease severity, and peripheral iron status, are needed to clarify potential correlations and enhance the observations of our findings.

Conclusion

To summarize, the pathophysiological mechanisms of RLS are intricate and multifaceted. Our bioinformatics analysis identified 37 differentially expressed mRNAs (9 upregulated, 28 downregulated) in RLS patients after accounting for familial genetic aggregation. Notably, we validated significant downregulation of SPARCL1, CCL8 and SELE, whereas no significant difference was observed in CCL2 and APOA2 levels. Correlation analysis did not reveal significant associations between the expression of these five mRNAs and RLS severity or peripheral iron status. Functional analyses implicated the IL-17 and TNF signaling pathways as well as CAMs in RLS pathogenesis, suggesting two potential therapeutic avenues: anti-inflammatory therapies and synaptic intervention targeting connectivity in RLS management. These findings reinforce the growing evidence for inflammatory and synaptic adhesion mechanisms in RLS while highlighting the need for functional validation of identified pathways, RLS subtyping based on transcriptomic profiles and development of precision medicine strategies.