Abstract

Background and Objectives

   DNA variations in the NF-kappa-B essential modulator (NEMO) gene are
   linked to incontinentia pigmenti (IP) and also immunodeficiency and
   autoinflammatory conditions. Some patients with IP present with
   neonatal vasculitis-like brain changes, although pathogenesis is
   unclear. We investigated cell-specific gene expression in a neonate
   with IP, who had encephalopathy, seizures, and vasculitis-like brain
   changes, and responded to steroid treatment.

Methods

   Single-cell RNA (ribonucleic acid) sequencing (scRNAseq), using the
   HIVE single-cell system, was performed on a neonate with IP, before and
   after steroid treatment, compared with a sex-matched healthy control
   toddler.

Results

   A total of 20,411 cells were sequenced and clustered into 10 cell
   types. In IP compared with control, upregulated significant gene set
   enrichment analysis gene ontology pathways (FDR <0.05) included defense
   response, complement activation, humoral immune response, and
   phagocytosis across all cell types. After steroid treatment, these
   pathways were predominantly downregulated in monocytes and neutrophils.
   The upregulated genes in IP that became downregulated after steroid
   treatment were interferon-related genes, oligoadenylate synthases, and
   immunoglobulin genes.

Discussion

   IP-associated loss of NEMO function is associated with a
   proinflammatory phenotype, that is moderated by steroids. scRNAseq
   provides a rationale for immune modulation in an n = 1 setting and
   valuable insights into the pathogenesis and therapeutics of this rare
   disease.

Introduction

   Incontinentia pigmenti (IP) is a rare X-linked dominant neurocutaneous
   syndrome that primarily affects female patients with amorphic variants
   in the NF-kappa-B essential modulator (NEMO) gene (also known as the
   IKBKG gene), encoding the NF-κB essential modulator (NEMO/IKKγ)
   protein. Hypomorphic variants in this gene result in milder IP in
   female patients, and ectodermal dysplasia with immune deficiency in
   male patients. The NEMO/IKKγ protein is crucial for NF-κB signaling,
   which governs immune responses, cell survival, and differentiation.
   Phenotypic variability in female patients is due to lyonization and
   skewed X-chromosome inactivation.^[40]1

   IP is marked by inflammatory changes in ectodermal tissues, likely
   triggered by oxidative stress during the transition from fetal to
   postnatal life.^[41]2 The characteristic vesicular skin rash is
   believed to stem from apoptosis of NEMO-deficient keratinocytes,
   triggering NF-κB activation in normal cells.^[42]2 Although skin
   inflammation is often confined to infancy, it can recur, persist, or
   begin later in life, raising questions about whether the inflammation
   is monophasic or chronic.^[43]3[44]–[45]7 Neurologic changes, with a
   greater effect on quality of life, are believed to result from a
   vasculitis-like response in the brain due to increased blood-brain
   barrier permeability, rather than neuronal apoptosis.^[46]8,[47]9
   Neurologic manifestations, including acute seizures or stroke-like
   episodes, typically emerge in early infancy after skin lesions. The
   occurrence of other neurologic changes in IP, including prenatal
   ischemic changes, cortical malformations, or brain atrophy, prompts
   further questions about disease progression.^[48]10

   We report a neonate presenting with encephalopathy and vasculitis-like
   brain changes, who responded to steroid treatment. Single-cell RNA
   sequencing (scRNAseq) was performed on peripheral blood to explore
   therapeutic mechanisms.

Methods

Single-Cell Blood RNA Sequencing—Bioinformatic and Enrichment Analysis

   scRNAseq used the HIVE single-cell platform, as described in eMethods
   on the patient before starting treatment, and 12 days after steroid
   treatment. A 2-year-old healthy girl was recruited as the healthy
   control for scRNAseq.

   scRNAseq data were analyzed in the R statistical environment^[49]11
   with tidyverse,^[50]12 described in eMethods. For HIVE scRNAseq, the
   Seurat package was used for analysis.^[51]13 Pathway enrichment
   analysis was performed through gene set enrichment analysis (GSEA) to
   obtain enriched gene ontology (GO) pathways (false discovery rate [FDR]
   <0.05) using the clusterProfiler^[52]14 package (eMethods).

Ethics Approval

   Ethical approval was granted by the Sydney Children's Hospitals Network
   Human Research Ethics Committee (HREC/18/SCHN/227, 2021/[53]ETH00356).

Data Availability

   Anonymized data not published within this article will be made
   available by request from any qualified investigator.

Results

Clinical Presentation

   The female patient was clinically diagnosed with IP due to vesicular
   eruptions on her arms and abdomen noted from day 3 of life. Skin biopsy
   showed dyskeratotic keratinocytes, with lymphocytic and eosinophilic
   infiltration in the dermis, typical of stage 1 lesions in IP. On day 14
   of life, she presented with generalized clonic seizures, and EEG
   confirmed frequent bifrontal onset seizures, successfully treated with
   levetiracetam. MRI showed multifocal areas of diffusion restriction
   ([54]Figure 1, additional figures in eFigure 1). Lumbar puncture was
   unsuccessful. Blood tests showed eosinophilia (7.0 × 10^9/L) with
   otherwise normal full blood counts, normal creatine kinase levels of
   150 U/L, normal CRP (C-Reactive protein) of 7.9 mg/L (0–10) and
   procalcitonin 0.2 μg/L (0–0.5), and deranged liver function (ALT 238
   U/L, AST 792 U/L, GGT 326 U/L). Liver ultrasound and doppler were
   normal. She received 2 mg/kg/day of oral prednisolone (weaned weekly
   over the ensuing 4 weeks) given the involvement of the NEMO gene in
   NF-κB function ([55]Figure 2A), and previous reports of steroid use for
   neurologic presentations.^[56]15,[57]16 Retinal exudates consistent
   with IP were treated with laser photocoagulation and intravitreal
   ranibizumab on day 18 of life. There were no seizure recurrence and
   normal sequential EEG recordings. Liver function tests normalized by 4
   weeks of age and were believed to likely represent idiopathic neonatal
   hepatitis in the absence of any other abnormal results. Genetic testing
   revealed a de novo recurrent exon 4–10 deletion in the NEMO gene. At
   last follow-up at 8 months of age, the patient had normal liver
   function and blood counts and normal neurologic and developmental
   function.

Figure 1. MRI Brain Changes in NEMO-Associated Incontinentia Pigmenti With
Neonatal Encephalopathy.

   [58]Figure 1
   [59]Open in a new tab

   MRI brain—hyperintense foci on diffusion imaging in (A) bilateral
   centrum semiovale and deep white matter and (B) internal capsules, genu
   and splenium of the corpus callosum, and bilateral frontal and
   occipital cortical gray matter. Areas corresponding to the hyperintense
   foci on diffusion show hypointensity on apparent diffusion coefficient
   (ADC) imaging suggesting cytotoxic edema in the respective areas (C)
   and (D). NEMO = NF-kappa-B essential modulator.

Figure 2. Blood Single-Cell RNA Sequencing in NEMO-Associated Incontinentia
Pigmenti With Neonatal Encephalopathy.

   [60]Figure 2
   [61]Open in a new tab

   (A) On ligand binding, various receptors (toll-like receptor [TLR],
   T-cell receptor [TCR], interleukin-1 receptor [IL1R], tumor necrosis
   factor receptor [TNFR]), transforming growth factor beta receptor
   (TGF-βR) trigger signalling pathways and activate transcription factors
   such as NF-kB. The activity of NF-kB is regulated by the IκB kinase
   (IKK) complex. The inhibitor of the IκB kinase complex includes the
   NF-kappa-B essential modulator (NEMO), also known as inhibitor of
   nuclear factor kappa-B kinase subunit gamma (IKK-γ). The loss of
   function of NEMO results in dysregulation of NF-κB signalling, leading
   to the altered expression of genes involved in inflammation, immunity,
   apoptosis, and other pathways (figure was created by biorender.com).
   (B) Uniform manifold approximation and projection (UMAP) of 3 samples
   (control, patient-presteroid, patient-poststeroid) in single-cell RNA
   sequencing identified 9 unique clusters: neutrophils, classical
   monocytes, nonclassical monocytes, natural killer (NK) cells,
   eosinophils, CD8T cells, CD4T cells, B cells, plasma B cells, and
   HSC-MPP (hematopoietic stem cell and multipotent progenitor) cells. (C)
   Bar chart of top 10 upregulated and downregulated gene set enrichment
   analysis (GSEA) gene ontology (GO) pathways in classical monocytes. In
   the presteroid vs control comparison, upregulated pathways (in red)
   include immune pathways such as defense response to other organism. In
   the poststeroid vs presteroid comparison, downregulated pathways (in
   blue) include immune pathways including defense response to other
   organism. (D) Dot plot of top 5 upregulated and downregulated GSEA GO
   pathways across cell types in the presteroid vs control (on the left)
   and poststeroid vs presteroid (on the right) comparison. Significant
   pathways (FDR <0.05) were simplified,^[62]14 and only those that were
   present in >2 cell types were plotted. Upregulated pathways (red),
   clustered based on similarity scores in GO terms,^[63]22 in “defense
   response”, “complement activation”, “humoral immune response,” and
   “phagocytosis” functions in presteroids vs control showed corresponding
   downregulation (blue) in poststeroid vs presteroids. The cell types
   with the most significant changes across treatment included classical
   monocytes, nonclassical monocytes, and neutrophils.

Single-Cell RNA Sequencing (scRNAseq)

   A total of 20,441 cells were sequenced across 3 samples (patient
   presteroid and poststeroid treatment, and healthy control) (eFigure 2).
   Uniform manifold approximation and projection analysis of biological
   samples revealed 10 distinct cell clusters ([64]Figure 2B).
   Differentially expressed genes (DEGs) with FDR <0.05 ranged from 56 to
   463 DEGs per cell type in presteroid vs control comparison, and 31 to
   1,046 DEGs per cell type in poststeroid vs presteroid comparison. GSEA
   analysis using a ranked gene list derived enriched GO pathways (FDR
   <0.05) for individual cell types.

Top 10 Upregulated and Downregulated Pathways in Classical Monocytes

   In the patient presteroids vs control comparison, top 10 upregulated GO
   pathways (in red) in classical monocytes ([65]Figure 2C) included
   immune pathways such as defense response to other organism. Some of
   these immune pathways were downregulated (in blue) in poststeroid vs
   presteroid comparison (other cell types are presented in eFigure 3, A
   and B).

Enriched Pathways Across Cell Types

   The top 5 upregulated and downregulated GSEA GO pathways of individual
   cell type in presteroids vs control (left) and poststeroid vs
   presteroids (right) were plotted ([66]Figure 2D), showing similarities
   and differences in pathways across cell types. In the presteroid vs
   control comparison, there were upregulated immune pathways in defense
   response, complement activation, humoral immune response, and
   phagocytosis across all cell types, but predominantly in monocytes,
   natural killer cells, and CD4/CD8T cells. The immune pathways that were
   upregulated in presteroid vs control were generally downregulated in
   poststeroid vs presteroid comparison, most evident in the
   classical/nonclassical monocytes.

   A connectivity network (CNET) plot of the most significantly altered
   pathway (defense response to other organism) in the classical monocytes
   was plotted for the presteroids vs control (left) and post vs pre
   comparison (right) ([67]Figure 3A). Genes that were upregulated in
   presteroids vs control (in red) included genes encoding toll-like
   receptors (TLR4, TLR8), complement receptors (CR1, C5AR1), cytokine
   receptors (IL6R, IL7R), interferon receptors (IFNGR1, IFNGR2), and
   tumor necrosis factor receptors (TNFRSF1A). Genes that were upregulated
   in presteroids vs controls and also downregulated in poststeroid vs
   presteroids included interferon genes (ISG15, IFIH1, IFI6, IFI27,
   IFI44L, IFIT1, IFIT2, IFIT3, IFITM1), oligoadenylate synthase (OAS)
   genes (OAS1, OAS2, OAS3), immunoglobulin genes (IGLC2, IGKC, IGHM), and
   S100 genes (S100A8, S100A9) ([68]Figure 3B, full list in eFigure 4,
   additional neutrophils CNET in eFigure 5 reflected similar gene
   changes).

Figure 3. Genes in the “Defense Response to Other Organism” Gene Ontology
Pathway in Monocytes.

   [69]Figure 3
   [70]Open in a new tab

   (A) Connectivity network enrichment plot of the “defense response to
   other organism” GO pathway was upregulated in presteroid vs control (in
   red) and downregulated in poststeroid vs presteroid comparison (in
   blue). Genes that were upregulated in presteroids vs controls and
   downregulated in poststeroid vs presteroids included interferon genes:
   ISG15, IFIH1, IFI6, IFI27, IFI44L, IFIT1, IFIT2, IFIT3, IFITM1; OAS
   genes: OAS1, OAS2, OAS3; immunoglobulin genes: IGLC2, IGKC, IGHM; and
   S100 genes: S100A8, S100A9. (B) Heatmap of genes in the “defense
   response to other organism” GO pathway. Only the top 30 genes that
   showed the most significant differences across cell types, identified
   by clustering, are plotted here. The average fold change of each gene
   across individual cell types was plotted, with red indicating
   upregulation and blue indicating downregulation. The top 30 genes that
   were most different across comparisons included interferon genes:
   IFI27, IFI4LL, IFITM1, IFI5, IFIT1, IFIT3, ISG15; OAS genes: OAS1,
   OAS2, OAS3; and immunoglobulin genes: IGLC2, IGHM, IGKC. GO = gene
   ontology; OAS = oligoadenylate synthases.

   Targeted examination of genes in the NEMO pathway showed general
   downregulation in patient vs control and upregulation (NF-kB, IRAK,
   IL6R, CXCR) after treatment with steroids (eFigure 6).

Discussion

   CNS disorders affect up to a third of individuals with IP,^[71]10 with
   acute neonatal encephalopathy being most common. Vascular abnormalities
   and inflammatory processes likely contribute to microvascular
   occlusions.^[72]9 Animal models show that disruption of TGFβ-TAK1-NEMO
   signalling causes endothelial cell death, brain microvessel occlusion,
   blood-brain barrier disruption, and seizures.^[73]8 Perivascular and
   intravascular eosinophilic infiltration in the CNS, also seen in skin
   and retinal lesions, may play a role, although it is unclear if the
   brain pathology is acute and self-resolving, or chronic and
   progressive.

   NEMO gene variants cause broad immune disorders from immune
   deficiency^[74]1,[75]17 to autoinflammation.^[76]18,[77]19 scRNAseq can
   reveal insights into these processes before symptoms emerge, because
   routine tests, like full blood counts, often appear normal. The HIVE
   system can include granulocytes (neutrophils), typically absent in
   PBMC-based single-cell RNA studies. This technology analyses thousands
   of cells from a single patient and control, providing greater
   statistical power and cell-type comparisons than bulk RNA sequencing.

   In this neonate, scRNAseq showed some downregulation of genes in the
   NEMO pathway at baseline, largely supportive of a loss of NEMO function
   and disrupted NF-kB signalling. Conversely, broad immune activation,
   most notably in interferon, OAS, and immunoglobulin-related genes, was
   identified in both innate and adaptive immune cells. These upregulated
   pathways at baseline were downregulated after steroids, predominantly
   in innate immune cells such as neutrophils and monocytes. Based on
   these data, it is possible the loss of function in NEMO gene results in
   loss of normal proinflammatory cascade, and instead, we see
   compensatory inflammatory response in other immune pathways.

   There is no established treatment for CNS disorders in IP, and our data
   provide a rationale for immune modulation. Steroids have been used for
   acute brain lesions in IP but limited to case
   reports.^[78]15,[79]16,[80]20 Reports of recurrent neurologic disease
   highlight the need for markers, such as a targeted inflammatory marker
   panel, to identify at-risk individuals and explore the potential of
   preventive therapies. Further understanding of CNS disease progression
   in IP could guide the use of treatments that may target an overactive
   innate immune system.

   Limitations of this report include the examination of a single case,
   control was a toddler rather than a neonate, and scRNAseq analysis of
   blood rather than cerebral blood vessels. Comparing scRNA-seq data with
   a child experiencing seizures would help determine whether the gene
   expression changes are specific to the NEMO gene deletion or related to
   seizures. Future studies should include serum and CSF examination,
   omics analysis, and functional immunology assessment in larger patient
   cohorts.^[81]21

   Single-cell RNA sequencing revealed broad immune activation, reversed
   after steroid treatment, in this girl with IP and acute neonatal
   encephalopathy. This case highlights the potential of scRNA sequencing
   in understanding disease mechanisms and therapeutic responses in rare
   disease n = 1 studies.

Acknowledgment