Abstract Background The study of genetic predisposition to pediatric systemic lupus erythematosus (pSLE) has brought new insights into the pathophysiology of SLE, as it is hypothesized that genetic predisposition is greater in children. Furthermore, identifying genetic variants and linking disrupted genes to abnormal immune pathways and clinical manifestations can be beneficial for both diagnosis and treatment. Here, we identified genetic alterations in a patient with childhood-onset SLE and analyzed the immunological mechanisms behind them to support future diagnosis, prognosis, and treatment. Methods Whole exome sequencing (WES) was adopted for genetic analysis of a patient with childhood-onset SLE. Gene mutations were confirmed by Sanger sequencing. Clinical data of this patient were collected and summarized. Ingenuity Pathway Analysis was used to provide interacting genes of the perturbed genes. Online Enrichr tool and Cytoscape software were used to analysis the related pathways of these genes. Results We present a case of a 2-year-old girl who was diagnosed with idiopathic thrombocytopenic purpura (ITP) and SLE. The patient was characterized by cutaneous bleeding spots on both lower extremities, thrombocytopenia, decreased serum complements levels, increased urinary red blood cells, and positive ANA and dsDNA. The patient was treated with methylprednisolone and mycophenolate, but clinical remission could not be achieved. The genomic analysis identified three novel mutations in this pSLE patient, a double-stranded missense mutation in ACP5 (c.1152G>T and c.420G>A) and a single-stranded mutation in SAMHD1 (c.1423G>A). Bioinformatic analysis showed that these two genes and their interacting genes are enriched in the regulation of multiple immune pathways associated with SLE, including cytokine signaling and immune cell activation or function. Analysis of the synergistic regulation of these two genes suggests that abnormalities in the type I interferon pathway caused by genetic variants may contribute to the pathogenesis of SLE. Conclusion The combined complexity of polymorphisms in the coding regions of ACP5 and SAMHD1 influences the susceptibility to SLE. Alterations in these genes may lead to abnormalities in the type I interferon pathway. Our study extends the spectrum of mutations in the ACP5 and SAMHD1 genes. The identification of these mutations could aid in the diagnosis of SLE with genetic counseling and suggest potential precise treatments for specific pathways. Keywords: novel mutation, ACP5/TRAP, SAMHD1, systemic lupus erythematosus (SLE), whole-exome sequencing (WES), bioinformatics Introduction Systemic lupus erythematosus (SLE) is a chronic multisystem autoimmune disease characterized by the production of autoantibodies and overactivation of the type I interferon (IFN) pathway, with alternating exacerbations and remissions ([31]1). SLE is a heterogeneous disease and the diagnosis of SLE requires the presence of 4 of the 11 defining criteria. Clinical manifestations of SLE patients include arthritis, rash, serositis, various cytopenias, renal disease, psychiatric, neurological, and other manifestations ([32]2, [33]3). Despite the wide variability, approximately 50% of SLE patients have persistently elevated blood levels of type I IFN over time. In addition, 60–80% of SLE patients have increased expression of interferon-stimulated genes (ISG) in peripheral blood, which is referred to as an IFN signature ([34]1, [35]4). As with other heterogeneous diseases, the pathogenesis of SLE remains unclear. However, heritability has long been recognized as an important causative factor, affecting approximately 66% in SLE ([36]5). To date, genome-wide association studies (GWAS) have identified more than 80 loci closely associated with lupus, and single-gene susceptibility to SLE has been suggested to be caused by single-nucleotide mutations in the coding regions of nearly 30 genes ([37]6–[38]8). The current view linking many susceptibility genes to key immune pathways is consistent with previous experimental studies including the role of immune complexes, host immune signaling, and interferon pathways in the pathogenesis of SLE ([39]8, [40]9). Monogenic type I interferon diseases such as spinal chondrodystrophy (SPENCD) and Aicardi-Goutières syndrome (AGS) clinically overlap with SLE, both of which are associated with spontaneous type I IFN responses with no detectable exogenous viral infection ([41]10). SPENCD is driven by an autosomal recessive mutation in ACP5. AGS can have different patterns of inheritance, mostly caused by single mutations in the ADAR, TREX1, RNASEH2A, RNASEH2B, RNASEH2C or SAMHD1 genes ([42]11). These studies, together with laboratory observations, suggest that IFN plays an important role in the pathogenesis of SLE. Given the complexity and heterogeneity of SLE, identifying genetic alterations and linking them to underlying immunologic pathogenesis could improve individualized diagnostic and therapeutic approaches to SLE. Here, we identified a pediatric-onset SLE patient with germline mutations in both ACP5 and SAMHD1.This patient carried three novel mutations from both father and mother, resulting in a double missense mutation in ACP5 (c.1152G>T in the mother and c.420G>A in the father) and a single missense mutation in SAMHD1 (c.1423G>A in the mother). We found that these two genes and their regulatory networks were enriched in multiple immune pathways associated with SLE, including cytokine signaling and immune cell activation or function. We also analyzed the synergistic regulation of these two genes and demonstrated their combined contribution to type I IFN, which suggests a pathogenic role of this genetic variant in this pSLE patient. Materials and Methods Human Patients We collected and summarized the clinical data of a patient who was diagnosed with lupus and carried specific gene mutations. Laboratory results before treatment were recorded for analysis. Written informed consent to participatein this study was provided by the participants' legal guardian/next of kin. Protocol of this study was approved by the Renji Hospital Biobank is funded by the National Human Genetic Resources Sharing Service Platform (2005DKA21300). DNA Sequencing DNA from probands and their family members were isolated and purified from blood and prepared for whole-exome sequencing (WES). DNA samples were enriched with Human SureSelect XT2 All Exon V4 Kit and sequenced by Illumina HiSeq 2000 (Illumina, Inc.). WES had 21% low or uncovered exon bases. Bioinformatic analysis was performed at JCSMR, ANU. Raw sequence reads were aligned to the reference genome (Hg19) and single-nucleotide variants and small insertions and deletions called using GATK. All SNVs of interest in ACP5 and SAMHD1 were confirmed by Sanger sequencing. Amplifluor to detect ACP5G290V, ACP5R46Q, and SAMHD1R408H in the APOSLE cohort was performed using the CHEMICON Amplifluor SNPs HT Genotyping System Fam-Joe kit S7909 (Merck-Millipore). The 45 and Up ([43]12, [44]13) and ASPREE ([45]14) datasets were used as reference healthy controls, accessed through the MGRB Collaborative ([46]http://sgc.garvan.org.au/mgrb/initiatives). WES Data Processing and Batch Correction Probes were filtered out if the detection P value was greater than 0.01 for at least 100% of the samples. All data values <10 was set to 10 and then the data were log2 transformed. An additional filter selecting the 75% most variable transcripts was performed, leaving a total 18,004 probes for analysis. Principal variance component analysis (PVCA) was conducted to identify undesirable sources of technical variability within the data and batch correction was applied to correct for this technical variation. Both PVCA and batch correction were conducted using JMP Genomics 7.0 (SAS Institute) analysis software. Protein Preparation and Mutations Interactions Analysis The X-ray resolved crystal structures of wild type ACP5 (PDB: 2BQ8) and SAMHD1 (PDB: 7A5Y) were download from RCSB PDB database, the mutations' structures were generated by pymol2.2. Water molecules, ions, heteroatoms, and all ligands were eliminated. Finally, the protonation states of the wild and mutated structures were then deliberated using a H++ server. Additionally, all missing hydrogen atoms were inserted by pymol2.2 The interactions analysis was used pymol2.2. Statistical and Bioinformatics Analysis With IPA, Cytoscape We used BioProfiler searches the Ingenuity Pathway Analysis (IPA, Qiagen) database to explore associations gene and phenotypes. IPA was used to construct pathways and networks ([47]15, [48]16), to quickly identify biological relationships, mechanisms, pathways, functions and diseases most relevant to experimental datasets. Then we obtain two comprehensive lists of proteins implicated in either the ACP5 or SAMHD1 by IPA. Detailed information of the expressed proteins and endogenous biochemical compounds that have been associated with genes by using the “Annotation” module in IPA. The protein-protein interaction (PPI) network based on ACP5 and SAMHD1 was obtained via String (V.11.5) to predicting the protein interactions and obtaining protein connection scores, and the interaction scores beyond 0.400 were uploaded to Cytoscape for visualization and analysis of biological networks ([49]17, [50]18). Protein overlapping network identification by using CentiScaPe to investigate node centrality in both networks, and important module in the protein network related to the hub genes were abstracted by using Molecular Complex Detection (MCODE) ([51]19). The Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis from ClueGO were used to perform biological pathway enrichment analysis for the core cluster identified by MCODE ([52]20). Results Case Presentation The patient was a 2-year-old girl who had recurrent upper respiratory tract infections and oral thrush since birth. She initially presented with cutaneous bleeding spots on both lower extremities, mainly on the calves and dorsum of the feet ([53]Figure 1A). There was no fever, cough, sputum, joint pain, or other signs of discomfort. She was seen at a local pediatric outpatient clinic, which showed thrombocytopenia (4 × 10^9/L) and decreased hemoglobin (98 g/L). The renal function report showed decreased blood creatinine (Bcr) level (19.3 μmol/L), urine red blood cells (373.4/ul) and urine protein (2+). Serum complement levels were decreased (C3 387 mg/L, C4 44 mg/L). No signs of hematological malignancy were found in bone marrow smear and bone marrow biopsy sections. Systemic lupus erythematosus and idiopathic thrombocytopenic purpura (ITP) were diagnosed based on skin symptoms, thrombocytopenia and decreased serum complement levels. She was then started on methylprednisolone (MP). However, the abnormalities in urine red blood cells and serum complement levels persisted. Figure 1. [54]Figure 1 [55]Open in a new tab Identification of gene mutations in patient with pSLE. (A)The patient initially presented skin bleeding spots in both lower limbs. (B) Pedigrees showing affected probands and parents. Sanger sequencing confirmation of the c.1152G>T (p.Gly290Val), c.420G>A (p. Arg46Gln) at ACP5 (C) and c.1423G>A (p.Arg408His) at SAMHD1 (D) variant identified in our patient using exome sequencing. Chromatograms represent the reference (top) and the mutant sequence. The variant is indicated by the red arrow. (E) Top: Structure modeling of wild type and p.Gly290Val mutation of ACP5. Top left: Gly290 form H-bond with Ala309; Asn50 form H-bonds with Arg46 (Distance: 2.1Å) and Ala 260 (Distance: 2.8 Å). Top right: The mutant model shows the change of H-bonds, Asn50 form H-bonds with Gly261 (Distance: 2.1 Å) and Ala 260 (Distance: 2.4 Å). Bottom: Wild type and p.Arg46Gln mutation of ACP5. Bottom left: Arg46 form H-bonds with Ser288 (Distance: 1.8 Å and 2.3 Å). Bottom right: The mutant shows the loss of two H-bonds, this might lose the ability to form any polar interaction with other residues. (F) Wild type and p.Arg408His mutation of SAMHD1. Left: Arg408 form salt bridge with Asp394. Right: The mutant shows the loss of salt bridge with Asp394, and form H-bond with Ile397 (Distance: 2.3 Å). P, patient; P-M, mother; P-F, father. Three months later, she presented to our pediatric clinic. Laboratory tests showed elevated white blood cell count (14.63 × 10^9/L), lymphocyte count (6.45 × 10^9/L), monocyte count (1.6 × 10^9/L), calcitonin (0.28 ng/ml), and decreased Bcr (16 μmol/L). Anti-nuclear antibody (ANA, >1/80), anti-ribosomal antibody (ANuA, >1) and anti-dsDNA (31.79 IU/ml) were positive. Serum complement (C3 0.66 g/L, C4 0.06 g/L), immunoglobulin (Ig) A (0.31 g/L) was decreased and IgE (308.47 IU/ml) was increased. The percentage of T lymphocytes (CD3^+ 35.61%, CD4^+ 19.78%, CD8^+ 14.73%) in PBMC was decreased and B lymphocytes were increased (CD19^+% 57.46%). Mycoplasma pneumoniae IgM antibodies were positive (1:80). Erythrocyte sedimentation rate (ESR), C-reactive protein (CRP) and renal function were normal. Cytomegalovirus (CMV) and Epstein-Barr virus (EBV) antibodies were negative. Rheumatoid factor, IgG, IgM and urine protein were all negative. Antibodies to chlamydial pneumonia, tuberculosis, TORCH, and anti-streptozotocin O were negative. Her abdominal ultrasound scan, skeletal radiography, EEG and brain MRI were normal. Immune markers such as serum immunoglobulin, complement and lymphocyte subpopulation values are listed in [56]Table 1. Table 1. Immune indices of patient. Items Patient Reference WBC (× 10^9/L) 14.63↑ 3.69–9.16 Hb (g/L) 137 113–151 PLT (× 10^9/L) 122 101–320 CRP (mg/L) 3.48 0–8 ESR (mm/h) 5 0–20 PCT(ng/mL) 0.280↑ <0.05 Anti-dsDNA(IU/mL) 31.79↑ 0–7.0 ANA 1:640 (+) 0 ANuA 4.702 (+) <1 C3 (g/L) 0.66↓ 0.90–1.80 C4 (g/L) 0.06↓ 0.10–0.40 IgG (g/L) 11.5 7.00–16.00 IgA (g/L) 0.31↓ 0.70–4.00 IgM (g/L) 1.3 0.40–2.30 IgE (IU/mL) 308.47↑ 1.31–165 CD3^+ (%) 35.61↓ 61.1–77 CD4^+ (%) 19.78↓ 25.8–41.6 CD8^+ (%) 14.73↓ 18.1–29.6 CD16^+ CD56^+ (%) 8.14↓ 8.7–38.3 CD19^+ (%) 57.46↑ 4.7–19.3 NK cytotoxic index (%) 8.5↓ 15–25 AST(U/L) 42↑ 10–40 ALT(U/L) 22 0–75 Urinary protein (mg/24h) 45.1 31–120 EBV-IgG (INDEX) 6.07↑ <1 EBV-IgM (INDEX) 0.08 <1 CMV-IgG (AU/mL) 79 (+) 0–6 CMV-IgM (INDEX) 0.32 0–1 [57]Open in a new tab ↑, Increase; ↓, decrease; (+) positive. Therefore, she met the diagnostic criteria for SLE and growth retardation (weight: 10.5 kg, −1.20 standard deviation (SD); height: 80 cm, −2.20 SD). She was then started on MP and mycophenolate mofetil (MMF) for 1 year and measures associated with SLE activity remained abnormal (increased dsDNA, positive ANA). At the same time, we found that the patient's IFN-α (31.56 pg/ml), IFN-γ (10.92 pg/ml), TNF-α (8.93 pg/ml), and IL-17A (23.78 pg/ml) increased dramatically during the active phase. Therefore, methotrexate (MTX) was given, but she developed leukopenia (2.76 × 10^9/L). Considering the inhibitory effect of MTX on bone marrow hematopoiesis, MTX was discontinued at that time and the regimen of MP, HCQ and MMF was changed. However, the patient's indicators such as dsDNA, serum complement and Ig remained unstable. Given the early onset, multi-organ involvement and the refractory nature of alternating episodes and remissions, we should further explore the etiology to improve the clinical diagnosis and treatment of this patient. The Germline Rare Mutations of ACP5 and SAMHD1 Identified in This Patient Next-generation sequencing brings great convenience for accurate diagnosis. High-throughput sequencing can detect genetic alterations in an entire genome, an exome, or a group of genes. Patients with early-onset lupus often carry a high frequency of pathogenic variants, highlighting the importance of genetic testing for pSLE ([58]21, [59]22). We then applied WES and identified three new mutations in this patient, all inherited from her father and mother. This patient carried a bilateral missense mutation in the ACP5 gene ([60]NM_001111034.2) at 2 different loci (c.1152G>T [p. Gly290Val, G290V]; c.420G>A [p. Arg46Gln, R46Q]) and a unilateral missense mutation in the SAMDH1 gene (NM_15474.3) (c.1423G>A [p.ARG408HIS, R408H]). The combined annotation-dependent depletion (CADD) scores representing the deleteriousness of these three mutations were 21 (ACP5, c.1152G>T), 22 (ACP5, c.420G>A) and 6.649 (SAMDH1, c.1423G>A) ([61]Figures 1B–D). The sites of these mutations have not been reported previously. Protein structure predictions showed that the G290V mutation resulted in weaker hydrogen bond (H bond) strength, the R46Q mutation in ACP5 resulted in the loss of two H bonds ([62]Figure 1E), and the R408H mutation in SAMHD1 resulted in the loss of the salt bridge and the formation of an H bond ([63]Figure 1F). We know that protein structure is closely related to function, and changes in protein structure caused by mutations may lead to disruption of biological function. Therefore, in combination with the high CADD score and altered protein structure, mutations in ACP5 may have greater pathogenicity. Mutations in ACP5 and SAMHD1 are known to drive the pathogenesis of SPENCD and AGS, both of which are monogenic interferon diseases characterized by increased type I IFN signaling leading to vasculopathy, autoinflammation, and SLE-like disease ([64]8, [65]23, [66]24). To date, 30 cases of SPENCD with SLE have been reported in the literature ([67]23, [68]25–[69]35) ([70]Figure 2A; [71]Table 2), and 9 cases of SLE with SAMHD1 mutations have been reported in the literature ([72]21, [73]36–[74]41) ([75]Figure 2B; [76]Table 3), including the patients in this article. Patients with SPENCD or AGS with SLE have a homogeneous clinical presentation, including autoantibody positivity, hematologic involvement, and neurologic symptomatic manifestations. Interestingly, this patient did not show clinical manifestations of epiphyseal lesions of SPENCD and neurological dysfunction of AGS in addition to autoimmune disease, implying that complex protein functional alterations due to polymorphisms in the ACP5 and SAMHD1 genes lead to different clinical symptoms. Here, we used high-throughput sequencing to identify three novel mutations in the coding regions of ACP5 and SAMHD1 in this patient that have not been previously reported to be associated with SLE. Figure 2. [77]Figure 2 [78]Open in a new tab ACP5 mutations cause SPENCD and SAMHD1 mutations cause AGS both diseases clinically overlap with SLE, and both are associated with a spontaneous type I IFN response. (A) Summary of ACP5 mutation associated with SLE data. (B) Summary of SAMHD1 mutation associated with SLE data. Table 2. Cases of SPENCD associated with SLE. Case Gender Mutation Amino-acid alteration ANA Anti-dsDNA Thrombocytopenia Proteinuria Hematuria Autoimmunehaemolytic anemia LN Renal involvement Raynaud's /vasculitis Hypocomplementemia (C3 and C4) Drug References