Immune response profiling identifies autoantibodies specific to Moyamoya patients
© Sigdel et al.; licensee BioMed Central Ltd. 2013
Received: 9 November 2012
Accepted: 14 March 2013
Published: 21 March 2013
Skip to main content
© Sigdel et al.; licensee BioMed Central Ltd. 2013
Received: 9 November 2012
Accepted: 14 March 2013
Published: 21 March 2013
Moyamoya Disease is a rare, devastating cerebrovascular disorder characterized by stenosis/occlusion of supraclinoid internal carotid arteries and development of fragile collateral vessels. Moyamoya Disease is typically diagnosed by angiography after clinical presentation of cerebral hemorrhage or ischemia. Despite unclear etiology, previous reports suggest there may be an immunological component.
To explore the role of autoimmunity in moyamoya disease, we used high-density protein arrays to profile IgG autoantibodies from the sera of angiographically-diagnosed Moyamoya Disease patients and compared these to healthy controls. Protein array data analysis followed by bioinformatics analysis yielded a number of auto-antibodies which were further validated by ELISA for an independent group of MMD patients (n = 59) and control patients with other cerebrovascular diseases including carotid occlusion, carotid stenosis and arteriovenous malformation.
We identified 165 significantly (p < 0.05) elevated autoantibodies in Moyamoya Disease, including those against CAMK2A, CD79A and EFNA3. Pathway analysis associated these autoantibodies with post-translational modification, neurological disease, inflammatory response, and DNA damage repair and maintenance. Using the novel functional interpolating single-nucleotide polymorphisms bioinformatics approach, we identified 6 Moyamoya Disease-associated autoantibodies against APP, GPS1, STRA13, CTNNB1, ROR1 and EDIL3. The expression of these 6 autoantibodies was validated by custom-designed reverse ELISAs for an independent group of Moyamoya Disease patients compared to patients with other cerebrovascular diseases.
We report the first high-throughput analysis of autoantibodies in Moyamoya Disease, the results of which may provide valuable insight into the immune-related pathology of Moyamoya Disease and may potentially advance diagnostic clinical tools.
Moyamoya Disease (MMD) is a rare cerebrovascular disorder that involves progressive bilateral stenosis/occlusion of the supraclinoid internal carotid arteries (ICAs) as well as the proximal anterior and middle cerebral arteries (ACAs and MCAs) [1, 2]. MMD is also characterized by the development of abnormal, fragile collateral blood vessels, usually from enlarged and proliferated lenticulostriate arteries . Early diagnosis is difficult and is usually made angiographically after clinical presentation of cerebral ischemia (stroke or TIA) or hemorrhage, often well after damage to the brain has occurred. The etiology of MMD is currently unknown. Approximately 10–12% of MMD cases are clearly familial with an autosomal dominant inheritance pattern, and several MMD-associated loci have been identified, including 3p24.2-p26, 6q25, 8q23, 12p12, and 17q25.3 [1, 4]. Mutations in individual genes have also been described, including TGFβ1 and PDGFRB , BRCC3 , and RNF213 [1, 4, 7]. There is also some evidence for active vascular remodeling in MMD collaterals, including proliferation of the smooth muscle cell layer in the intima of MMD vessels [1, 4]. The expression levels of several proteins with known roles in angiogenesis have been described, including vascular endothelial growth factor (VEGF), basic fibroblast growth factor (bFGF) and transforming growth factor beta (TGF-β) [1, 4]. There are also reports of increased intercellular adhesion molecule-1 (ICAM-1) and vascular cell adhesion molecule-1 (VCAM-1) in the cerebral spinal fluid (CSF) of MMD patients .
MMD is often described as Moyamoya Syndrome (MMS) when it is associated with other disorders, such as Majewski Osteodysplastic Primordial Dwarfism Type II (MOPD II), Down syndrome, Seckel syndrome, Neurofibromatosis and Graves’s disease, a known autoimmune disorder [9–11]. Although MMD is not traditionally considered an autoimmune disease, several groups have reported elevated levels of circulating autoantibodies (autoAbs), including anti-cardiolipin , anti-thyroid [13, 14] and anti-α-fodrin . In addition, a SEREX profile of CSF obtained from MMD patients identified 5 proteins against which autoAbs were expressed .
To comprehensively characterize the serum autoAb repertoire in MMD patients, our unbiased approach used a commercially-available high-density protein microarray platform to identify circulating serum autoAbs. These arrays have been useful for profiling disease-relevant autoAbs and have led to the identification of novel autoAbs in chronic renal , rheumatologic and autoimmune diseases . Our study identified a number of MMD-associated autoAbs and the biological pathways and networks associated with them.
Demographic data for control and MMD patients
Discovery set (20 sera)
Validation set (68 sera)
Protein arrays analysis
Cross sectional ELISA
n = 10
n = 10
n = 22
n = 46
Gender: males (%)
29.6 ± 12.2
27.8 ± 10.7
49.1 ± 21.6
40.7 ± 13.6
Ethnicity% (1/2/3/4) †
Using standard methods for sera collection , 5.0 mL of blood from each patient was collected, and the serum was separated by centrifugation, aliquoted and immediately stored at −80°C until use. The samples were collected under consent at the Stanford Hospital and Clinics from 2006–2008. The study was approved by the Institutional Review Board of Stanford University.
The ProtoArray® Human Protein Microarray v4.0 (Invitrogen, Carlsbad, CA) was used to characterize IgG reactivity in MMD. V4.0 arrays contain 8268 recombinant human proteins expressed as N-terminal GST fusion proteins. We used our standard protocols as previously described for serum sample preparation, blocking, probing, drying, scanning, and data acquisition . Briefly, ProtoArrays were blocked with 5.0 mL blocking buffer (100 mM sodium phosphate pH 7.4, 200 mM NaCl, 0.08% Triton X-100, 25% glycerol, 20 mM glutathione, 1.0 mM DTT, 1.0% casein) for 1 hour at 4°C. After the blocking step, 5.0 mL of diluted serum sample (1:150 in PBST buffer consisting of PBS, 1% casein and 0.1% Tween 20) was applied and incubated for 90 minutes at 4°C. The slides were washed 5 × 5 minutes with PBST buffer prior to application of secondary antibody (anti-human antibody-Alexa Fluor® 647 conjugate). The slides were incubated for 90 minutes at 4°C, washed for 5 × 5 minutes with PBST, dried and then scanned using the Axon GenePix 4000B Scanner (Molecular Devices, Sunnyvale, CA).
Signal intensities per spot for each individual array were obtained using GenePix pro 6.0. We used Prospector Analyzer® 5.2 (Invitrogen Inc., Carlsbad, CA) to subtract the background, and to normalize and analyze GenePix Results (GPR) files for each array.
We used a standard cut-off Z-Score of 3.0. Individual antigen reactivity was ranked based on Z-score above or below the mean signal for each array. Arrays from patients of a distinct clinical phenotype were analyzed as a group. Group analyses were made by comparing 2 sets of individual antibody levels for every antigen present on the array using M-statistics of the Prospector Analyzer® with the Robust-Linear-Normalization (RLM) method . Differences in significance were displayed as ANOVA and Chebyshev’s Inequality p-Value (≤ 0.05 considered significant). For significant autoAbs in MMD, p-value was calculated with unpaired t-test.
To analyze MMD-associated pathways and major functions we used Ingenuity Pathway Analysis® (http://www.ingenuity.com) for Global Functional Analysis (GFA) and Global Canonical Pathways (GCP) Analysis, which produced a list of major pathways and functions with p-values.
We used a novel data analysis method to identify MMD-specific autoAbs by speculating that those proteins would have a higher chance of being associated with proteins with a higher probability of having functionally-relevant single-nucleotide polymorphisms (SNPs). For this purpose we used a method that employs an integration of genetic and genomic data called functional interpolating SNPs (fitSNPs) . First, we identified three loci (3p26-p24.2, 8q23, and 17q25) previously associated genetically with MMD from Online Mendelian Inheritance in Man (OMIM)  and selected antibodies against proteins that were significantly differentially expressed in the serum samples of MMD patients compared to healthy controls at the autoantibody level. Second, we sub-selected for antibodies against proteins that in general were highly differentially expressed across thousands of publicly-available gene expression mRNA microarray experiments, with a fitSNP differential expression ratio (DER) > 0.45. Third, we identified additional proteins that: (a) contained genetic variants associated with any neurological diseases; (b) were significantly differentially expressed in MMD at the autoantibody level; and (c) were highly differentially expressed at the mRNA level with a fitSNP DER > 0.6.
Independent sera from a different cohort of MMD patients were used for validation of the autoAbs identified in the ProtoArray discovery phase. The control group of patients in this validation phase consisted of age, gender and demographically-matched patients with other causes of cerebrovascular disease, allowing an assessment of the importance and specificity of the autoAbs in MMD. To validate the significant autoAbs identified by ANOVA and FitSNP analysis, we further selected autoAbs that were known to interact with antigens with biological relevance in vascular and autoimmune disease, where the autoAbs could be detected most easily by ELISA, and where ProtoArray measurements indicated a fold change of > 2 in MMD patients. Customized enzyme-linked immunosorbent assays (ELISAs) were developed as previously described  to detect serum immunoglobulin binding to the following array of antigens: amyloid beta A4 protein (APP), catenin, beta 1 (CTNNB1), G protein pathway suppressor 1 (GPS1), stimulated by retinoic acid 13 (STRA13), receptor tyrosine kinase-like orphan receptor 1 (ROR1), and EGF-like repeats and discoidin I-like domains 3 (EDIL3). Briefly, GST-tagged purified proteins (Invitrogen, Carlsbad, CA) were applied to a 96-well immunosorbent plate (NUNC, Rochester, NY) and incubated overnight at 4°C. Standard curves for each protein were prepared using an IgG mouse monoclonal anti-GST tag (Millipore, Temecula, CA) and AP-conjugated AffiniPure goat anti-mouse IgG (Jackson ImmunoResearch, West Grove, PA). The plates were washed with TBST buffer, blocked for 1 hour at room temperature (RT) with 2% dry milk in TBST, and incubated with 50 μL of diluted serum (40-fold with 2% dry milk in TBST buffer) for 1 hour at RT. The plates were then washed with TBST and incubated in AP-conjugated AffiniPure mouse anti-human IgG (Jackson ImmunoResearch, West Grove, PA) and the reactivity was visualized with AP-pNPP Liquid Substrate System for ELISA (Sigma-Aldrich, St. Louis, MO). The signal was measured at 405 nm with a SPECTRAMax 190 microplate reader (Molecular Devices, Sunnyvale, CA). An unpaired t-test p value < 0.05 was considered significant.
Identification of reactive antigens in MMD
CI* P- value
Fold increase in MMD
Calcium/calmodulin-dependent protein kinase (CaM kinase) II alpha (CAMK2A)
B-cell antigen receptor complex-associated protein alpha-chain
EGF-like repeats and discoidin I-like domains 3 (EDIL3)
Ectonucleoside triphosphate diphosphohydrolase 1
Calcium/calmodulin-dependent protein kinase ID (CAMK1D), transcript variant 1
CD7 molecule (CD7)
Na+/K + transporting ATPase interacting 4 (NKAIN4)
CDC-like kinase 4 (CLK4)
Interferon, alpha-inducible protein 6 (IFI6)
Casein kinase 1, delta (CSNK1D), transcript variant 1
Deiodinase, iodothyronine, type III (DIO3)
Fc fragment of IgG, low affinity IIIa, receptor (CD16a) (FCGR3A)
Uncharacterized protein C20orf96
cDNA clone MGC:31944 IMAGE:4878869, complete cds
CD3d molecule, delta (CD3-TCR complex) (CD3D), transcript variant 1
Endoglin (Osler-Rendu-Weber syndrome 1) (ENG)
Solute carrier family 2 (facilitated glucose transporter), member 2 (SLC2A2)
Serine/threonine kinase 25 (STE20 homolog, yeast) (STK25)
Transmembrane protein 32 (TMEM32)
Chromosome 19 open reading frame 39 (C19orf39)
Calcium/calmodulin-dependent protein kinase (CaM kinase) II delta (CAMK2D), transcript variant 3
G protein pathway suppressor 1
LPS-responsive vesicle trafficking, beach and anchor containing (LRBA)
Melanocortin 2 receptor accessory protein (MRAP), transcript variant 1
NIMA (never in mitosis gene a)-related kinase 3 (NEK3), transcript variant 1
Receptor (chemosensory) transporter protein 2 (RTP2)
Ingenuity Pathway Analysis (IPA) software was applied to all of the reactive antigens and revealed over-represented pathways involved in Neurological Disease, Tissue Development, Cellular Growth and Proliferation, and Muscular System Development and Function. The reactive antigen proteins associated with MMD in the aforementioned pathways were: v-abl Abelson murine leukemia viral oncogene homolog 2 (ABL2), activin A receptor, type IIA (ACVR2A), activin A receptor, type IIA (ACVR2B), CAMK1D, calcium/calmodulin-dependent protein kinase II alpha (CAMK2A), CAMK2D, major histocompatibility complex, class II invariant chain (CD74), CD247, CD3d molecule, delta (CD3D), CSNK1D, EDIL3, E74-like factor 2 (ELF2), endoglin (ENG), ectonucleoside triphosphate diphosphohydrolase 1 (ENTPD1), IL2-inducible T-cell kinase (ITK), kaptin (KPTN), LIM domain kinase 1 (LIMK1), mitogen-activated protein kinase kinase kinase 9 (MAP3K9), myosin light chain kinase 2 (MYLK2), p21 protein (Cdc42/Rac)-activated kinase 4 (PAK4), aspartoacylase (ASPA), beta-site APP-cleaving enzyme 1 (BACE1), DEAD (Asp-Glu-Ala-Asp) box helicase 17 (DDX17), disks large-associated protein DLG7 (DLG7), histone deacetylase 1 (HDAC1), homer homolog 2 (Drosophila) (HOMER2), immunoglobulin (CD79A) binding protein 1 (IGBP1), lipoprotein lipase (LPL), microtubule-associated protein 4 (MAP4), and phospholipid scramblase 1 (PLSCR1). The predominant IPA functions of these reactive antigens were Protein Phosphorylation (p value =1.54E-07) and Protein Modification (p value =1.84E-05), with many of the proteins also involved in pathways deployed in Cell Death (p value =3.74E-05). A panel of autoAbs in MMD was also found to have reactive antigens involved in Cardiac Disease, which is a frequent accompaniment of MMD. The increased autoAb expression against CAMK2D is intriguing, as this protein is implicated in cardiac hypertrophy and heart failure, conditions which have also been reported to coexist with MMD . PRKG1, also antigenic in MMD, has been associated with enlargement of cardiomyocytes and smooth muscle cell proliferation [32, 33]. In addition, both LIMK1 and TNFα are associated with cardiac stenosis, while TNFα is specifically associated with cardiac stress response, inflammation, and fibrosis [34–36]. MMD patients exhibited an enrichment of antigenic responses against proteins involved in Immunological Disease and Inflammatory Responses (p value < 2.05E-02), such as AXL receptor tyrosine kinase (AXL), copper metabolism (Murr1) domain containing 1 (COMMD1), DnaJ (Hsp40) homolog, subfamily C, member 12 (DNAJC12) and member 5 beta (DNAJC5B), Fc fragment of IgG, high affinity Ia, receptor (CD64) (FCGR1A) and low affinity IIIa, receptor (CD16a) (FCGR3A), heat shock transcription factor 1 (HSF1), interferon, alpha-inducible protein 27-like 2 (IFI27L2), interferon-induced protein 44-like (IFI44L), interferon, alpha 1/ interferon, alpha 13 (IFNA1/IFNA13), immunoglobulin heavy constant mu (IGHM), immunoglobulin kappa locus (IGK@), interleukin 20 receptor beta (IL20RB), myelin oligodendrocyte glycoprotein (MOG), nuclear factor of kappa light polypeptide gene enhancer in B-cells inhibitor, epsilon (NFKBIE), NFKB repressing factor (NKRF), NLR family, pyrin domain containing 3 (NLRP3), tumor necrosis factor (ligand) superfamily, member 9 (TNFSF9), and ubiquitin-like modifier activating enzyme 7 (UBA7). AutoAbs were also found to be highly enriched against vascular antigens, such as angiotensinogen (AGT), MOG, NLRP3, palmitoyl-protein thioesterase 2 (PPT2), protein kinase C, delta (PRKCD), and SMAD family member 3 (SMAD3).
To determine if there was an acute-MMD autoAb profile, we examined the sera of 5 MMD patients taken prior to surgery and 6 months post-surgery. No significant differences in the majority of MMD-related autoAbs were detected except for 4 autoAbs that were more highly expressed in the sera during active MMD compared to levels 6 months post-surgery. These were Abs against ankyrin repeat and SOCS box-containing 8 (ASB8, p = 0.02), which is known to play a role in intracellular signal transduction ; a zinc transporter solute carrier family 39 (zinc transporter) member 9 (SLC39A9, p = 0.02) ; and secernin 3 (SCRN3, p = 0.02) and cytochrome P450 family 26 subfamily A polypeptide 1 (CYP26A1, p = 0.02), both of which are expressed in the cerebral cortex . It is unclear why the immune reactivity of this panel of antigens diminishes after surgical correction of the vascular stenosis. There was also a trend, although not significant, of an increase in the postoperative levels of the immunoglobulins IGK@, IGHM, IGHV4-31, and IGKV1-5 (p = 0.08) compared to the active disease state pre-surgery.
Identification of key MMD-associated target genes
Enriched p value
ProtoArray p value
Possible role in cell migration and remodeling of cytoskeleton 
Component of the COP9 signalosome, DDR 
Of particular interest are the roles of cell cycle, DNA maintenance and DNA damage response/repair (DDR) pathways. Two highly expressed MMD-associated autoAbs identified in this array with fitSNP and validated by ELISA include those generated against COP9 signalosome complex subunit 1 (GPS1 or CSN1) and stimulated by retinoic acid 13 (STRA13), both of which have well-described functions in DDR. GPS1 is a component of the COP9 signalosome (CSN) complex whose function is to modulate cullin neddylation, which in turn regulates cullin-RING ubiquitin ligase (CRL) activity and the ubiquitination of proteins involved in DDR . The CSN also forms complexes with CRLs containing cullin4 in order to function in DDR, cell cycle regulation and chromatin remodeling . STRA13 (also known as CENP-X or MHF2) is the DNA-binding component of the Fanconi Anemia (FA) core protein complex, through its association with MHF1 and FANCM proteins [50, 51]. The FA complex plays a critical role in DDR, and is specifically recruited to stalled DNA replication forks occurring as a result of blockage or damage . Two other validated MMD-specific proteins, Amyloid beta A4 protein (APP) and β-catenin, also have roles in DDR and DNA maintenance. APP, a protein most widely studied for its role in Alzheimer disease, has recently been identified as a component of the TIP60 DNA repair complex, a complex that modifies histone acetylation at DNA double strand breaks [40, 41]. β-catenin has a key role in the activation of canonical Wnt signaling and has also recently been implicated in chromosomal instability; specifically, Aoki et al. describe defects in β-catenin leading to activation of Wnt signaling and dysregulation of the cell cycle and chromosomal instability . That DDR and DNA maintenance may have a central role is also supported by the identification in the discovery phase of approximately 20 autoAbs, the antigens against which have known roles in DDR and DNA maintenance, including SWSAP1, CSNK1D, MCM10, NUDT1 and UBE2W.
Further evidence that both MMD/MMS may involve the critical DDR comes from the recent discovery of a novel Xq28 deletion in 3 unrelated hereditary MMS families . This deletion results in the loss of BRCC3, an E3 ubiquitin ligase that is a subunit of the BRCA1-BRCA2-containing complex (BRCC), and deletion of this protein in zebrafish results in defects in angiogenesis . The BRCC complex functions in DDR by stabilizing the BRCA1 protein at DNA break points . Intriguingly, both BRCA1 and BRCA2 are also a part of the FA complex of DDR proteins as described above [53, 55]. In addition, MOPD II and Seckel syndrome, diseases strongly associated with moyamoya angiopathy, are sometimes caused by mutations in the pericentrin gene [9, 10, 56]. Pericentrin is an essential component of the centrosome which mediates, among others, cell cycle progression, spindle organization and orientation, cilia formation and DDR .
Given that radiation exposure to the brain is one of the leading environmental causes of moyamoya angiopathy , DDR and DNA maintenance machinery may well be compromised in MMD patients. While further research is required to determine if this is the case, precisely how this leads to a complex disease and syndrome remains purely speculative. It is important to note that while defects in DDR, chromosomal instability and cell cycle disregulation often lead to cancer, there is no evidence of increased cancer rates in MMD patients, aside from an association with NF1 . DNA replication and cellular response to DNA damage involve a vast complex of proteins and protein–protein interactions. Given the clinical heterogeneity of MMD and MMS, it would not be unfounded to expect the disease to involve a spectrum of one to several mutations in multiple genes, as is the case with FA patients .
The over-representation of antigens pertaining to antigen presentation, immune cell trafficking and inflammatory response may also suggest a significant autoimmune component for MMD. STRA13, one of the validated MMD-specific antigens, also plays an essential role in the stabilization of the outer kinetocore, a protein complex that is essential for correct attachment, segregation and movement of chromosomes during cell division . Reports of the expression of autoAbs to other centromeric proteins of the kinetocore are also associated with a number of autoimmune diseases in humans including rheumatoid arthritis and limited cutaneous sclerosis or CREST syndrome . Therefore, MMD may be related in part to an autoimmune condition.
While the presence of autoAbs may not yet be of diagnostic value in MMD, this prospect may be advanced by validating additional autoAbs and using a larger cohort of patients. Our study demonstrates a novel role for the autoimmune response in MMD and may provide insight into the mechanisms of the disease, including such key cellular processes as cell cycle and DNA repair and maintenance. Future studies are required to further elucidate the clinical relevance of these autoAbs in MMD. There are currently no in vitro models of MMD, but recent advances in disease-specific induced pluripotent stem cells (iPSCs) may show some potential as an in vitro model of this complex and rare disease.
Minnie M Sarwal and Gary K Steinberg are Joint Senior Authors.
Anterior cerebral artery
Differential expression ratio
Induced pluripotent stem cells
Internal carotid artery
Middle cerebral artery
Transient ischemic attack.
We thank Cindy H. Samos, Minh-Thien Vu, and Van Dinh for their support during manuscript preparation and members of the Sarwal laboratory for their assistance. We also thank the patients and their families who participated in this study, the Stanford Department of Neurosurgery clinical research team that assisted in obtaining patient consent and samples, and members of the Steinberg laboratory. This research was supported in part by funding from Josef Huber Family Moyamoya Fund, Stanley and Alexis Shin, Reddy Lee Moyamoya Fund, and Child Health Research Fund at Stanford School of Medicine.
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.