Microarray and whole-exome sequencing analysis of familial Behçet’s disease patients

Behçet’s disease (BD), a chronic systemic inflammatory disorder, is characterized by recurrent oral and genital mucous ulcers, uveitis, and skin lesions. We performed DNA microarray analysis of peripheral blood mononuclear cell (PBMC) mRNA from 41 Japanese BD patients and revealed elevated levels of interleukin (IL) 23 receptor (IL23R) mRNA in many BD patients. DNA sequencing around a SNV (Rs12119179) tightly linked to BD revealed an elevated frequency of the C genotype, consistent with a previous report that IL23R is a susceptibility locus for BD. Notably, four of these BD patients are members of familial BD; a whole-exome sequencing (WES) of these BD patients identified 19 novel single-nucleotide variations (SNVs) specific to these patients. They include heterozygous SNVs in the genes encoding IL-1 receptor-associated kinase 4 (IRAK4), nucleotide-binding oligomerization domain (NOD)-like receptor family pyrin domain-containing 14 (NRP14) and melanoma antigen-encoding gene E2 (MAGEE2); IRAK4 harbors a missense mutation, whereas NRP14 and MAGEE2 harbor nonsense mutations. These SNVs may serve as genetic markers that characterize BD.

Bardet-Biedl syndrome 5 (BBS5) is one of 18 causative genes (BBS1-18) for an autosomal recessive ciliopathy disorder with eye, kidney, and heart defects 31 . The BBSome (a complex of seven BBS proteins) localizes at or near the primary cilium and the centrosome, associates with the centriolar satellite protein CEP131 via BBS4, and plays a role in transporting ciliary membrane proteins 32 . BBS5, which localizes along the axonemes of rods and cones, is phosphorylated by protein kinase C to regulate the translocation of arrestin1 (Arr1) between the inner and outer segments according to the light conditions in photoreceptor cells 33 . Eye site defects observed in BD patients may be a consequence of improper translocation of Arr1 due to overexpression of BBS5.
Bile acid CoA:amino acid N-acyltransferase (BAAT) is involved in bile acid metabolism by catalyzing the hydrolysis of long-and very long-chain saturated acyl-CoAs into the free fatty acid and coenzyme A, and also conjugates glycine to these acyl-CoAs.
Defective mutations in BAAT are involved in familial hypercholanemia.
Cysteinyl leukotrienes (CysLT) mediate bronchoconstriction and play important roles in the development of asthma; CysLT receptor 2 (CysLTR2), a G-protein-coupled seven-transmembrane receptor for CysLTs, is expressed in multiple cell types including macrophages, vascular smooth muscle, brain, and B and T lymphocytes, in which it regulates allergic inflammation 34,35 . Notably, sequence variants of the CYSLTR2 gene are associated with aspirin-intolerant asthma 36 . Given that augmented expression of CysLTR1 in nasal inflammatory leukocytes is also related to aspirin-intolerant asthma 37 , elevated CysLTR2 expression may be involved in inflammation of various tissues in BD patients.
By contrast, no report to date has suggested any clinical relevance of several other up-regulated genes, including those encoding neuromedin U (NMU), guanylate cyclase 2F (GUCY2F), pyrophosphatase 2 (PPA2), and phosphoglucomutase 5 (PGM5). Therefore, it remains elusive why these genes were up-regulated in PBMCs of BD patients.
Down-regulated genes: mRNA levels of interferon (IFN) alpha-inducible protein 27 (IFI27), olfactomedin 4 (OLFM4), phosphatidylinositol N-acetylglucosaminyltransferase subunit C (PIGC), and membrane-spanning 4-domains, subfamily A, member 2 (MS4A2) genes were down-regulated by >˗3.0-fold in both male and female BD patients relative to healthy volunteers ( Fig. 1b, Supplementary Fig. S2b, S3b). IFI27 encodes a mitochondrial protein that sensitizes cells to apoptotic stimuli via mitochondrial membrane destabilization, which may influence the innate immune response to IFNs 38 . Olfactomedin proteins are essential for early development and functional organization of the nervous system and hematopoiesis 39 . PIGC is an endoplasmic reticulum-associated protein involved in the first step of glycosylphosphatidylinositol lipid anchor biosynthesis 40 . MS4A2 is a subunit of the high-affinity IgE receptor (the allergic response involves the binding of allergen to receptor-bound IgE), and SNVs in the MS4A2 gene are potential risk factors for asthma 41 .
However, scatter plot showed that average values of these genes were almost unaltered ( Supplementary Fig. S5). Thus, we conclude that these down-regulated genes are not important.

Pathway analysis
Ficolin-1 (FCN1) functions as a recognition molecule in the complement system. We reported previously that the FCN1 mRNA level is elevated in PBMCs from patients with systemic autoimmune diseases such as Takayasu arteritis (TA), a rare vasculitis affecting the aorta and its associated branches 42 , and microscopic polyangiitis (MPA), a small-vessel vasculitis with fatal symptoms 43 . Hence, we used the Ingenuity Pathway Analysis software to investigate whether FCN1 and/or other members of the complement system are activated in BD patients. We found that the upstream factors of the classical pathway including C1r and C1s, and downstream factors including C5 and C7, are indeed up-regulated in BD patients (red circles and squares in Supplementary Fig. S6a), whereas other complement factors were down-regulated (green circles and diamonds in Supplementary Fig. S6a). In contrast to a previous study of a Chinese BD population 11 , C4 protein level was not elevated in the PBMCs of our BD patients. A heat map revealed that the magnitudes of up-or down-regulation of genes encoding complement factors ( Supplementary Fig. S6b), including FCN1 (Fig. S6c), were smaller than those of the genes listed in Fig.1, suggesting that the complement system plays a less significant role in the pathogenesis of BD.
Analysis of other immune-related pathways revealed no notable genes, except for a modest up-regulation of defensin beta 1 (DEFB1) and mitogen-activated protein (MAP) kinase kinase 6 (MAP2K6) in many BD patients (Supplementary Fig. S7). DEFB1 encodes an antimicrobial peptide that is a basic component of human innate immunity; DEFB1 plays a protective role against allergies and infection, and is up-regulated by inflammatory or microbial stimuli 44 . Given that MAP2K6 plays a role in inflammatory diseases via phosphorylation-mediated activation of p38 MAP kinase in response to inflammatory cytokines, and MAP2K6-deficient mice exhibit reduced severity of arthritis 45 , it is reasonable to speculate that its activation may influence the arthritis-related inflammatory abnormalities in some BD patients.

Human subjects and ethical considerations
Forty-three BD patients were enrolled at Osaka University Hospital (BD1-41 plus the mother and daughter of BD20) between 2001 and 2009; the genders and ages of BD patients are provided in Figure S1. The study was reviewed and approved by the Research Ethics Committee of Osaka University, and written informed consent was obtained from all participants. The diagnosis of BD was established according to standard criteria proposed by the Japan BD Research Committee; actual study methods were carried out in accordance with the approved guidelines. Serum samples were obtained from patients regardless of their symptoms and level of disease activity/inactivity.

Target selection and sequencing
Exome sequencing was conducted on four DNA samples from patients with familial BD: patient BD47 and her aunt (BD20), grandmother (BD26), and mother (BD50). Genomic DNA was extracted from PBMC using the PAXgene Blood DNA Kit (QIAGEN), sheared into 150-200 bp fragments, and used to make a library for multiplexed paired-end sequencing (Illumina). The resultant library was hybridized to biotinylated cRNA oligonucleotide baits from the SureSelect Human All Exon 50Mb kit (Agilent Technologies) for exome capture. Targeted sequences were purified using magnetic beads, amplified, and sequenced on an Illumina HiSeq2000 platform in paired-end 101 bp configuration. The raw sequence data were submitted to the NCBI SRA database under accession No. SRP059981 (NCBI BioProject PRJNA288379).

Mapping and SNV/indel calling
The quality of DNA reads was checked using software developed in-house. Adapter sequences and sequences of inadequate quality were removed. After quality control, reads were mapped to the reference human genome (UCSC Genome Browser, hg19) using BWA

Figure S1. Gender and age distributions of healthy volunteers (a) and BD patients (b).
Females and males are indicated by circles and triangles, respectively. "Unknown" indicates uncharacterized genes. Agilent's sample ID is presented for identification of probes used for the microarray analysis. Mosaic tile representation of each gene is also shown, with intensity gradients indicating the mean value of the expression level (log 2 ratio): blue (down-regulation) and crimson (up-regulation) relative to the average value in healthy volunteers (gray). Names of notable genes are highlighted in larger font with blue arrows; genes also appearing in Figure S2 are shown in pink font. Vertical arrows indicate familial BD patients (see Fig. 2A). (c) Bar represents the standard intensity gradient. List of bottom 50 genes down-regulated in most BD patients is shown in increasing order of fold-change values for female patients. "Unknown" indicates uncharacterized genes. Agilent's sample ID is presented for identification of probes used for the microarray analysis. Mosaic tile representation of each gene is also shown, with intensity gradients indicating the mean value of the expression level (log 2 ratio): blue (downregulation) and crimson (up-regulation) relative to the average value in healthy volunteers (gray). Names of notable genes are highlighted in larger font with blue arrows; genes also appearing in Figure S2 are shown in pink font. Vertical arrows indicate familial BD patients (see Fig. 2a). (c) Bar represents the standard intensity gradient.    Fig. 1, S2 and S3. The y-axis shows the log value of hybridization signal intensity obtained from the microarray data for BD patients (average values). The x-axis shows the log value of signal intensity obtained from samples of healthy volunteers. Dots for IL12RB2 and IL10 are also shown to compare with IL23R dot. Figure S6. Expression profiling of microarray data for complement system genes expressed in PBMCs of 41 BD patients. (a) Ingenuity Pathway Analysis of microarray data for complement system genes identified C5 and C7 as the most conspicuously up-regulated genes in all 41 BD patients. The network shows a graphical representation of the molecular relationships between molecules, which are represented as nodes. Edges (lines) represent the biological relationship between two nodes; all edges are supported by at least one reference from the literature, a textbook, or canonical pathway information stored in the Ingenuity Knowledge Base. Node shapes represent the functional classes of the gene products, and node color intensity indicates the degree of up-(red) or down-(green) regulation. Edges are labeled in various ways that describe the nature of the relationships between nodes. (b) Mosaic tile representation of genes involved in the complement system. Notable genes are highlighted in large font and by red (up-regulated) or green (down-regulated) arrows. (c) Mosaic tile representation of three independent FCN1 probes. Tile colors indicate the mean relative transcript levels in PBMCs from BD patients and normal controls. Blue corresponds to a log 2 ratio of ˗5 (down-regulation), red corresponds to a log 2 ratio of 5 (up-regulation), and intermediate values are represented by shades of red (pink) or blue. Gray indicates no change.