Abstract Objectives Pediatric acute-onset neuropsychiatric syndrome (PANS) is characterized by infection-provoked abrupt-onset obsessive compulsive disorder (OCD) and neurodevelopmental regression. Owing to the neuroimmune hypothesis, we investigated the effects of IV immunoglobulin (IVIg) on cell-specific gene expression. Methods Single-cell RNA sequencing of peripheral immune cells was performed in 5 children with PANS (median age 8 (5.5–16) years), before and after administering open-label IVIg, compared with 4 controls (median age 13.5 [IQR 12–15] years). Results The index PANS event (age 1.8–13 years) involved abrupt eating restriction (n = 5), developmental regression (n = 4), and OCD (n = 3). A total of 144,470 cells were sequenced and clustered into 11 cell types. Children with PANS before IVIg compared with controls showed downregulated immune pathways (defense response, innate immunity, secretory granules) in most cell types, with natural killer (NK) cells showing upregulated immune pathways (response to corticosteroid), supporting baseline “immune dysregulation.” Ribosomal pathways were upregulated in neutrophils and CD8 T cells but downregulated in NK cells. In children with PANS after IVIg, the baseline immune and ribosomal pathway abnormalities were reversed and histone modification pathways (histone methyltransferase, chromatin) were downregulated in neutrophils and NK cells. Discussion We propose that PANS is an epigenetic immune brain disorder with cellular epigenetic, ribosomal, and immune dysregulation. Epigenetic and immune-modulating therapies, such as IVIg, may have disease-modifying effects. Introduction Pediatric acute-onset neuropsychiatric syndrome (PANS) is characterized by acute-onset obsessive compulsive disorder (OCD) or eating restriction and developmental regression.^[42]1 The immune mechanisms driving PANS remain poorly understood.^[43]1 We hypothesize that PANS represents a clinical phenotype where environmental and genetic factors interact, with infection or stress triggers disrupting peripheral-central immune function with secondary inflammation.^[44]2-4 Owing to the hypothesis that PANS is driven by abnormal immune function, immunotherapies such as steroids and IV immunoglobulin (IVIg) have been used with varying degrees of success.^[45]5-7 We explore the effects of IVIg on peripheral immune cells in PANS through single-cell RNA sequencing (scRNA-Seq). Methods Participant Selection Children (<18 years of age) met criteria for PANS with abrupt, dramatic onset (within 48 hours) of OCD or severe food restriction.^[46]1 Patients had ongoing debilitating symptoms and were starting or continuing open-label IVIg. Control Selection Controls (n = 4) included children without neurodevelopmental or neuropsychiatric disorders, autoimmune diseases, or severe allergic conditions (median age 13.5 (IQR 12–15) years, 50% male). Sample Collection In children with PANS, blood samples were collected during cannulation and at the end of IVIg infusion (8–24-hour time interval). Patients and controls had no infections a month before blood collection. HIVE scRNA-Seq scRNA-seq was performed, using the HIVE platform, described in eMethods, and data were analyzed in the R environment with tidyverse and Seurat packages. Pathway enrichment analysis was performed through Gene Set Enrichment Analysis (GSEA) to obtain Gene Ontology (GO) pathways (false discovery rate [FDR] <0.05) using the clusterProfiler package. Ethics Approval Ethical approval granted by the Sydney Children's Hospitals Network (SCHN) Human Research Ethics Committee (HREC/18/SCHN/227, 2021/[47]ETH00356). Data Availability Anonymized data and R code are available on request. scRNA-seq data are accessible through Gene Expression Omnibus (GSE293230). Results Clinical Data Five children (median age 8 (5.5–16) years, 60% male) diagnosed with PANS, who were commencing (2 g/kg) (Patients 1 and 3) or on regular four-weekly IVIg treatment (1.5 g/kg) (Patients 2, 4, and 5), were recruited. The age at the index PANS event was 1.8–13 years, and the mean time since PANS onset to first IVIg treatment was 2 (range 0.5–4) years. Family history and details of PANS events are given in the [48]Table (eAppendix 1). A convincing clinical benefit of IVIg was reported in 4, based on clinician assessment ([49]Table), but headache was a common side effect. All 5 patients had ongoing mood disorders and OCD diagnoses, and the 3 boys had autistic spectrum disorder (ASD) diagnoses (mean follow-up since PANS onset 4.1 [2.8–5.5] years). The 3 boys with ASD had negative trio whole-exome sequencing. Table. Clinical Phenotypes of the 5 Children With PANS, Including Family and Pregnancy History, PANS Phenotype, Treatments, and Treatment Effects Patient number Age at onset/sex Family history Pregnancy Premorbid NDDs Trigger PANS event (order of dominance) Recurrent infection-provoked events Conventional treatment Time from PANS onset to initiation od IVIG (duration of IVIG) Therapeutic benefits of IVIG Current status (follow-up since PANS onset) 1 1.8 y, M Celiac disease, anxiety, postnatal depression (mat) Severe hyperemesis, vanishing twin syndrome — Infections and vaccine Eating restriction, autistic regression, loss of language, incontinence, sleep, emotional control Yes (recurrent) Fluoxetine, psychology 4 y (0.75 y^[50]a); IVIg naïve at time of study Eating, sleep, toileting, OCD ASD, ADHD, OCD, emotional dysregulation, sleep, incontinence (5 y) 2 3 y, M Anxiety, depression (mat), IgA deficiency (mat GM), PANS (sib) Gestational diabetes, cholecystitis, preeclampsia, UTI, and chest infection — Infection OCD, eating restriction, incontinence, regression, self-injury Yes (multiple) Psychology 1.3 y (1.5 y^[51]a) OCD (CYBOCS 32/40 to 9/40), continence, sleep OCD, suspected ASD (2.8 y) 3 5 y, M ADHD, anxiety (mat); ASD, ADHD (sib) — ADHD Infection Eating restriction, autistic regression, incontinence, sleep, emotional control Yes (recurrent) Aripiprazole, psychology 3 y (0.1 y); IVIg naïve at time of study (4.5 y) Eating restriction, emotional control ASD, ADHD, OCD, tics, emotional dysregulation (4.5 y) 4 12 y, F Hypothyroid (mat, mat GM) unexpected death (pat) — — — Separation anxiety, eating restriction, OCD, regression Yes (1) Fluvoxamine, pregabalin, psychology 0.5 y (2.5y) OCD (CYBOCS 34/30 to 10/40), separation anxiety, repetitive motor behavior Ongoing anxiety (5.5 y) 5 13 y, F Ulcerative colitis (mat), ASD (sib) Ulcerative colitis — Infection OCD, eating restriction, incontinence, emotional control Yes (recurrent, strep) SSRI (multiple), ERP/CBT 1.25 y (0.5 y) OCD (CYBOCS 36/40 to 20/40) OCD, anxiety (3 y) [52]Open in a new tab Abbreviations: ADHD = attention deficit hyperactivity disorder; ASD = autistic spectrum disorder; CBT = cognitive behavioral therapy; ERP = exposure response prevention; F = female; FH = family history; GM = grandmother; IVIG = IV immunoglobulin; mat = maternal; M = male; NDDs = neurodevelopmental disorders; OCD = obsessive compulsive disorder; PANS = pediatric acute neuropsychiatric syndrome; pat = paternal; sib = sibling; SSRI = selective serotonin reuptake inhibitor; UTI = urine tract infection. ^a Ongoing. scRNA-Seq A total of 144,470 cells were sequenced across 14 samples from 5 patients before and after IVIg treatment and 4 controls. There were no significant differences in number of cells, number of reads, and percentage of mitochondrial transcripts across samples (eFigure 1). Uniform manifold approximation and projection analysis of samples revealed 11 distinct cell clusters ([53]Figure 1A). Bar charts of differentially expressed genes per comparison are shown in [54]Figure 1B. Figure 1. Single-Cell RNA Sequencing in Children With PANS Before and After IVIg: UMAP, DEGs, Neutrophil Pathways. [55]Figure 1 [56]Open in a new tab (A) Uniform manifold approximation and projection (UMAP) analysis identified 11 distinct cell clusters. (B) Bar chart of differentially expressed genes (DEGs) across 7 cell types in children with pediatric acute neuropsychiatric syndrome (PANS) showing the number of upregulated (red) or downregulated (blue) DEGs. In children with PANS pre-IVIg vs controls (left), differentially expressed genes (DEGs) with FDR <0.05 of the 7 cell types ranged from 49 to 947 genes per cell type. In children with PANS post-IVIg vs pre-IVIg (right), there were 109–2,347 DEGs per cell type. (C) Bar plot of the top 5 upregulated and downregulated GSEA GO pathways in neutrophils from children with pre-IVIg PANS vs controls (left) and children with PANS post-IVIg vs pre-IVIg (right). (D) Connectivity network enrichment plot (CNET) of the upregulated “response to bacterium” GO pathway (left) and downregulated “histone modification” GO pathway (right) in neutrophils from children with PANS post-IVIg vs pre-IVIg. Dot size corresponds to the number of genes enriching that pathway. GSEA used a ranked gene list to identify enriched GO pathways (FDR <0.05) for individual cell types. We focused on results in neutrophils and natural killer (NK) cells (findings for other cell types in eFigure 2). Neutrophils PANS Pre-IVIg vs Control Comparison In children with pre-IVIg PANS vs controls, top 5 upregulated pathways in neutrophils included ribosomal/translation pathways ([57]Figure 1C, left in red) and downregulated pathways included immune pathways such as defense response to symbiont ([58]Figure 1C, left in blue). PANS Post-IVIg vs Pre-IVIg Comparison In children with post-IVIg vs pre-IVIg PANS, top 5 upregulated pathways in neutrophils included immune pathways such as response to bacterium ([59]Figure 1C, right in red). The CNET plot of “response to bacterium” pathway ([60]Figure 1D, left) revealed GO biological process subclusters including response to lipopolysaccharides (TLR4, IRAK3), regulation of cytokine production (CD14, JAK2, FCGR1A), and neutrophil chemotaxis (S100A genes). Top 5 downregulated pathways included ribosomal/translation pathways, RNA processing, and histone modification ([61]Figure 1C, right in blue). The connectivity network enrichment plot (CNET) of “histone modification” pathway ([62]Figure 1D, right) revealed GO cellular component subclusters including histone methyltransferase (KMT genes), histone deacetylase complex (HDAC genes), and chromatin (KDM genes, CHD3, EP300, KAT6A). Neutrophil pathways in the 2 IVIg-naïve patients with PANS (Patients 1 and 3) showed similar baseline downregulation of immune pathways, with post-IVIg changes characterized by upregulated immune and downregulated ribosomal pathways (eFigure 3). However, minor differences were observed at baseline: Patient 1 showed upregulated ribosomal pathways while Patient 3 showed upregulation of mitochondrial pathways, reflecting interpatient heterogeneity. NK Cells PANS Pre-IVIg vs Control Comparison In children with PANS pre-IVIg vs controls, top 5 upregulated pathways in NK cells included immune pathways such as response to corticosteroid and myeloid cell activation ([63]Figure 2A, left in red). Top 5 downregulated pathways included ribosomal/translation pathways ([64]Figure 2A, left in blue). Figure 2. Single-Cell RNA Sequencing in Children With PANS Before and After IVIg: NK Cell Pathways, Dotplot Across Cell Types. [65]Figure 2 [66]Open in a new tab (A) Bar plot of top 5 upregulated and downregulated GSEA GO pathways in NK cells from children with pre-IVIg pediatric acute neuropsychiatric syndrome (PANS) vs controls (left) and children with PANS post-IVIg vs pre-IVIg (right). (B) Connectivity network enrichment plot of the upregulated “cytosolic ribosome” GO pathway (left) and downregulated “histone methyltransferase” GO pathway (right) in NK cells from children with PANS post-IVIg vs pre-IVIg. Dot size corresponds to the number of genes enriching that pathway. (C) Heatmap of genes in “histone methyltransferase” GO pathway by average log2 fold change, with pre-IVIg vs control (left) and post-IVIg vs pre-IVIg comparison (right). (D) Dot plot visualizing the top 5 upregulated and downregulated GSEA GO pathways across cell types in children with PANS pre-IVIg vs controls (left) and children with PANS post-IVIg vs pre-IVIg (right). Significant pathways (FDR <0.05) were simplified, and only those that were present in >3 cell types in PANS pre-IVIg vs control comparison were plotted. Dot size represents the -log10(padj) of the pathway while the color intensity represents the normalized enrichment score (NES). PANS Post-IVIg vs Pre-IVIg Comparison In children with PANS post-IVIg vs pre-IVIg, top 5 upregulated pathways in NK cells included ribosomal/translation pathway, as well as immune pathways such as neutrophil chemotaxis and inflammatory response ([67]Figure 2A, right in red). The CNET plot of “cytosolic ribosome” pathway ([68]Figure 2B, left) revealed GO cellular component subclusters including large ribosomal subunit (RPL genes) and small ribosomal subunit (RPS genes). Top 5 downregulated pathways included cytoskeletal and histone methyltransferase pathways ([69]Figure 2A, right in blue). The CNET plot of “histone methyltransferase” pathway ([70]Figure 2B, right) revealed enrichment of genes including KMT (lysine methyltransferase) and SETD (SET domain), and a heatmap of the expression of these genes showed overall upregulation in children with pre-IVIg PANS vs controls ([71]Figure 2C, in red), followed by downregulation in children with post-IVIg vs pre-IVIg PANS ([72]Figure 2C, in blue). Enriched Pathways Across All Cell Types In children with pre-IVIg PANS vs controls ([73]Figure 2D, left), there were predominantly downregulated immune pathways that were generally upregulated after IVIg ([74]Figure 2D, right) across most cell types. In children with pre-IVIg PANS vs controls, ribosomal pathways demonstrated diverse expression patterns across cell types, with upregulation in neutrophils and CD8 T cells but downregulation in NK cells. These ribosomal pathways showed reversal in direction after IVIg. Discussion In this scRNA-seq study of children with PANS, we identified key baseline alterations in epigenetic, ribosomal, and immune pathways, which were reversed after IVIg. This study was not a clinical trial, and patients received open-label IVIg. Previous IVIg trials in PANS yielded mixed results, with 1 trial failing to show superiority of IVIg over placebo after 6 weeks.^[75]6,7 In our experience, IVIg benefits typically only last 2–3 weeks, with patients requiring monthly treatments because of recurrent infection-provoked events. This highlights the need for further research to understand PANS and identify adjunctive therapies. First, we found that children with PANS exhibited baseline immune dysregulation compared with controls, with predominantly downregulated immune pathways. However, NK cells exhibited upregulated immune pathways at baseline, suggesting that “immune dysregulation” is a better description than “immune deficiency” alone. We hypothesize that an impaired innate immune response, both peripherally and in the CNS, increases infection susceptibility, with recurrent infections and chronic inflammation triggering and sustaining symptoms in PANS. After IVIg, immune pathways that were downregulated at baseline were significantly upregulated, showing broad immune-modulating effects across all peripheral blood cell types. Second, we identified ribosomal dysregulation at baseline in children with PANS compared with controls, which was reversed after IVIg. Epigenetic machinery, including chromatin remodeling, histone modifications, and stress-responsive transcription factors, play a crucial role in regulating ribosome function during stress by modulating gene expression and translation.^[76]8 There is increasing support for the role of epigenetics in the pathogenesis of neurodevelopmental disorders.^[77]9,10 Monogenic epigenetic disorders involving DNA variants in chromatin-related genes have been described in children with PANS-like presentations.^[78]11 Although children with PANS in this study did not have pathogenic DNA variations in these genes, it is plausible that epigenetic mechanisms play a role in PANS, influenced by common gene variants plus environmental factors.^[79]10 In addition, we identified epigenetic effects, mainly histone and chromatin modifications, of IVIg on peripheral immune cells in PANS. This study shows the strength of scRNA-seq technology as a tool to explore therapeutic mechanisms of action in small cohorts.^[80]12 Baseline RNA signatures in IVIg-naïve patients were similar to those in the broader PANS cohort, indicating shared transcriptomic profile independent of IVIg exposure, despite interpatient variability. Limitations of this study include scRNA-seq on peripheral blood immune cells rather than brain tissue. Some children were on psychiatric medications with potential immunomodulatory or epigenetic effects; future studies should account for these confounders. Larger blinded PANS cohorts are needed to identify gene signatures predictive of IVIg response. Targeted epigenetic analyses are needed to explore epigenetic roles in PANS and IVIg response. Conclusion In children with severe PANS who responded to IVIg, we identified baseline abnormalities in epigenetic, ribosomal, and immune regulation. We demonstrate that IVIg exerts both epigenetic and immune effects. Acknowledgment