A Comprehensive Review on the Role of Genetic Factors in Neuromyelitis Optica Spectrum Disorder

Neuromyelitis optica spectrum disorders (NMOSD) comprise a variety of disorders being described by optic neuritis and myelitis. This disorder is mostly observed in sporadic form, yet 3% of cases are familial NMO. Different series of familial NMO cases have been reported up to now, with some of them being associated with certain HLA haplotypes. Assessment of HLA allele and haplotypes has also revealed association between some alleles within HLA-DRB1 or other loci and sporadic NMO. More recently, genome-wide SNP arrays have shown some susceptibility loci for NMO. In the current manuscript, we review available information about the role of genetic factors in NMO.


INTRODUCTION
Neuromyelitis optica spectrum disorders (NMOSD) comprise a variety of disorders being described by acute inflammatory responses in the optic nerve and spinal cord, i.e., optic neuritis and myelitis, respectively (1). NMO is mostly triggered by IgG autoantibodies against aquaporin 4 (AQP4) (2). AQP4 monomers comprise six transmembrane helical domains and two small helical parts around a thin aqueous pore (3). These monomers lump together to make corresponding tetramers with the ability of being aggregated in cell plasma membranes. The constructed supramolecular collections are named as orthogonal arrays of particles (OAPs) (3). AQP4 is the supreme ample water-channel protein in the central nervous system (CNS) (1). A number of NMO patients do not have AQP4-IgG, yet they have IgG antibodies against myelin oligodendrocyte glycoprotein, a glycoprotein in the outer myelin sheath of CNS neurons (4).
Following the discovery of AQP4-specific proliferative T cells in NMO patients, it has been recognized that AQP4-specific T cells exhibit Th17 features and display molecular mimicry with a peptide sequence encoded by the commensal bacterium Clostridium perfringens. Further studies have revealed distinct features of gut microbiota in NMO cases versus both multiple sclerosis (MS) cases and healthy subjects (5).
Although this disorder has some similarities with MS, it is important to distinguish between these two conditions, particularly at early stages of the disorder, since therapeutic modalities for these disorders are different (6). Most importantly, a number of prescribed agents for MS might be harmful for patients with NMO (7,8). NMO and MS can be differentiated through assessment of NMO antibody. Although the existence of cerebral lesions has been formerly regarded as a criterion for differentiation between these two conditions, it is currently acknowledged that these lesions do not exclude NMO. In fact, with the advent of NMO antibody assessment techniques, some cases diagnosed as MS for a long time have been found to have NMO (9).
Typically, NMO manifests around the ages of 35 to 45 years, yet less than 20% of cases occur in children, and elderlies account for 18% of cases. NMO is recognized as a condition with female predominance. Although 70% to 90% of total NMO patients are female, such sex bias is not seen in children (6,10). In NMO-AQP4 cases, gender influences both age at disease onset and site of attack (11).
NMO is most probably a complex multifactorial disorder. Most cases of this disorder are sporadic, yet 3% of cases are familial (12). A previous meta-analysis of whole-genome association studies in NMO has shown association of AQP4-IgG positive NMO with two independent signals in the MHC region. Notably, one of these signals has been suggested to be related with structural variations in the complement component 4 region. Moreover, a significant causal effect has been found between AQP4-IgG positive NMO and recognized risk variant for systemic lupus erythematosus (SLE). Most notably, such causal link has not been observed with MS risk variants (13). A number of other studies have reported an association between genetic variants and gene expressions alterations and NMO. In the current manuscript, we review available information about the role of genetic factors in NMO.

FAMILY STUDIES
Familial and sporadic NMO are similar in terms of clinical manifestations, age onset of disease, gender-based effects, and proportion of AQP4-IgG positive cases (12). A pioneer study in this field has reported occurrence of NMO in identical twin sisters at the ages of 24 and 26, respectively (14). A subsequent study reported NMO manifestations such as sudden loss of vision and transverse myelopathy in two sisters at the age of 3. Notably, HLA haplotyping revealed a shared haplotype between these two sisters, yet an unaffected sib also had this haplotype (15). More recently, a group of researchers described a series of familial NMO cases including siblings, parent-child, and auntniece pairs, more than 80% of them being female. A number of reported cases had either maternal or paternal transmission. More than 75% of cases had AQP4-IgG. About half of cases had clinical manifestations or serologic markers of another immunerelated condition. The observed familial transmission of NMO suggested a complex genetic etiology for this disorder (12).
A number of other studies also reported familial clustering of NMO cases, with some of them reported the presence of a shared haplotype among affected cases. Table 1 summarizes the results of family studies in NMO.

HLA STUDIES
An HLA genotyping study in seropositive Brazilian NMO patients has revealed some susceptibility loci for NMO, most importantly HLA-DRB1*04:05 and *16:02. A number of alleles within HLA class I showed association with NMO, yet this association did not remain significant after corrections for multiple comparisons (22). Another study in Afro-Caribbean NMO cases has shown higher frequency of HLA-DRB1*03 in NMO patients. On the other hand, HLA-DRB1*15, but not DRB1*03 allele has been recognized as a susceptibility locus for MS. In brief, distribution of HLA-DRB1 and DQB1 has been different among NMO and MS cases in this population (23). Another study in seropositive Brazilian NMO patients has shown overrepresentation of the HLA-DRB1*03 allele group in NMO cases compared with unaffected individuals. On the other hand, MS patients have shown higher frequency of the HLA-DRB1*15 allele group. DRB3 and DRB5 have had higher frequencies in NMO and MS cases, respectively (24). Another study has confirmed overrepresentation of HLA-DRB1*03 and HLA-DRB1*10 alleles in another group of Brazilian NMO patients compared with controls, in spite of no significant overrepresentation of MS-associated alleles (25). In addition, the DR3 and DR15 haplotypes have been found to be more common in NMO and MS, respectively. The association between HLA-DRB1*03:01 allele and NMO has not been dependent on seropositivity (26). In a study in Japanese patients, HLA-DRB1*08:02 and HLA-DRB1*16:02 have been found as risk loci, while HLA-DRB1*09:01 has been a protective allele (27). Table 2 shows the results of HLA studies in NMO cases in different populations.

GENOMIC STUDIES
Whole-exome sequencing (WES) has facilitated identification of risk loci for NMO. Application of this method in addition to HLA sequencing in seropositive NMO cases of Chinese origin has shown significant association between HLA-DQB1*05:02 and NMO. Additionally, the frequency of "HLA-DQB1*05:02-DRB1*15:01" haplotype has been higher in the NMO group compared with controls. Besides, this study has shown higher frequency of loss-of-function mutations in NOP16 in these patients compared with healthy subjects. The G390R of IgG1, which decreases the threshold for BCR activation, has been another NMO-associated variant. Notably, most of the NMOassociated genetic factors have been enriched pathways related with nervous system and immune responses (43).
Another genome-wide study using an SNP array has identified the rs1964995 in the MHC region as a risk locus for NMO. Notably, three MS-associated variants have also been found to be associated with NMO. A variant within KCNMA1 gene has been associated with disability score as well as presence of transverse myelitis (27).
The importance of copy number variations (CNVs) in conferring risk of NMO has been previously assessed using a genome-wide method. The majority of identified CNVs have been located at TCRg and TCRa regions. These CNVs have been mostly deletions with sizes of 5 to 50 kb. Since they have been only in the peripheral blood T cells, it has been deduced that they are most probably somatically acquired CNVs. Moreover, it has been an association between the presence of CNVs in NMO cases and seronegativity for AQP4-IgG or low antibody titer (44).
Several SNPs within AQP4 gene have been genotyped in NMO cases to find possible risk loci for this condition in different ethnic groups. For instance, Matiello et al. have compared genotype frequencies of 8 SNPs within AQP4 gene in sporadic and familial NMO cases as well as healthy controls. One of these SNPs has been found to be associated with risk of NMO. Moreover, two missense mutations at Arg19 have been found in three NMO patients. The authors have reported that apart from one infrequent SNP, no other examined SNP or   (46). Table 3 shows the results of genomic studies in NMO cases.

EXPRESSION STUDIES
Expressions of several immune-related genes have been assessed in NMO cases at transcript or protein levels. Moreover, a number of high-throughput sequencing strategies have been employed to assess expression of different subtypes of transcripts. For instance, lncRNA and mRNA profile has been assessed in these patients using microarray technique. Such type of analysis has led to the identification of more than 1,300 lncRNAs with differential expression between NMO cases and normal controls. Moreover, more than 700 mRNAs have been found to be differentially expressed between NMO cases and normal subjects. These genes have been functionally correlated with IL-23-related cascades, IFN-g signaling, natural killer-kB pathway, and a number of other immune-related mechanisms (74). Another RNA expression profiling experiment has shown possible contribution of T-cell-related genes and the TNF/NF-kB cascade in the pathogenesis of NMO. Notably, IL7Ra (CD127) has been found to be downregulated in the circulation of NMO patients compared with control subjects. Moreover, transcription factors located in the upstream of CD127 and survival pathways in its downstream have been considerably downregulated. These expression changes have been accompanied by decrease in the quantities of naïve T cells, reduction of BID-mediated Tcell survival signaling and activation of cell apoptosis. Taken together, these observations indicate the importance of IL7Ra signaling in the pathoetiology of NMO (75). A highthroughput expression profiling in brain tissue samples obtained from an NMO patient as well as patients with Parkinson's disease and amyotrophic lateral sclerosis has shown upregulation of more than 200 genes in brain lesions of NMO patients with the mostly upregulated ones being associated with immune response. Upregulation of IFI30, CD163, and SPP1 has also been confirmed by further RNA and protein-based techniques. Genes with high expression in NMO brain lesions has been functionally related with NF-kB and Blimp-1, indicating the importance macrophage-mediated inflammatory responses in the pathoetiology of NMO brain lesions (76).
With the aim of finding effective markers for the assessment of response of NMO patients to therapeutic options, Vaknin-Dembinsky et al. have assessed miRNAs profile in the blood of NMO patients before and following treatment with rituximab. They have reported upregulation of 14 miRNAs and downregulation of 32 miRNAs in NMO patients after treatment with rituximab. Moreover, they have shown higher levels of 17 miRNAs and lower levels of 25 miRNAs in untreated cases compared with healthy controls. Notably, rituximab could normalize expression of a number of these miRNAs, among them have been brain-specific or brain-enriched miRNAs. Cumulatively, circulatory miRNA profile can be used as a biomarker for therapeutic response (77).
The pleiotropic cytokine IL-6 is also implicated in the pathogenesis of NMO through enhancement of survival of plasmablasts, induction of release of antibodies against AQP4, disruption of integrity of blood-brain barrier and its functionality, as well as increasing differentiation and activity of proinflammatory T cells (78). Expression of this cytokine has been reported to be elevated in CSF and blood samples of NMO patients (79). Table 4 shows the results of expression studies in NMO.

IN VITRO STUDIES
A number of in vitro studies have appraised the functional mechanisms of development of NMO. In an effort to find the impact humoral factors on astrocyte injury in NMO, Haruki et al. have conducted a series of experiments on immortalized human primary astrocytes. Moreover, they assessed the effect of TY09 human brain microvascular endothelial on the quantity and localization of AQP4 protein in astrocytes. Serum samples of NMO patients have been shown to induce cytotoxic effects on AQP4-expressing astrocytes. Moreover, these serum samples could decrease AQP4 expression at both mRNA and protein levels, while increasing release of TNF-a and IL-6 from astrocytes. Experiments in an in vitro BBB model has shown  In SLC28A3 gene, rs10868138 and rs12378361 were correlated with higher and lower erythrocyte concentration of 6-TGNs, respectively. rs507964 in SLC29A1 was associated with lower erythrocyte concentration of 6-MMPNs and 6-MMPNs:6-TGNs ratio.
(Continued)  localization of AQP4 protein at the astrocytic membrane following co-culture with TY09, in contact with these cells (132). Sera samples of these patients or even NMO-IgG have also been shown to rapidly downregulate AQP4 levels on the surface of astrocytes. Astrocytes treated with NMO-IgG, IL-6/R, and NMO-IgG + IL-6/R have shown over-production of IL-6 transcripts. Moreover, NMO-IgG could elicit alterations in gene transcription via the JAK/STAT3 pathway. Cumulatively, NMO-IgG has been reported to induce the JAK1/2/STAT3 pathway in astrocytes, representing a crucial event in the pathoetiology of NMO. Besides, suppression of JAK1/2 signaling might be a therapeutic modality for NMOSD (133).
Another in vitro study has shown similar magnitude of lymphoproliferation and cytokine profiles in peripheral blood mononuclear cells of NMO cases and healthy controls in reponse to Staphylococcus aureus and Candida albicans. However, NMOoriginated Escherichia coli-induced cell cultures have exhibited higher proliferation of CD4+ T cells in association with higher production of IL-1b, IL-6, and IL-17. IL-10 release has been lower in NMO-derived cells compared with controls. Notably, the in vitro E. coli-stimulated expressions of IL-6 and IL-17 have been correlated with neurological debilities. Overproduction of Th17-associated cytokines has been associated with the production of IL-23 and IL-6 by LPS-stimulated monocytes. Consistently, LPS levels have been higher in the plasma samples of NMO cases. Therefore, increase in Th17 type response to E. coli might contribute in the pathogenesis of NMO (134). Table 5 shows the results of in vitro mechanistical studies in NMO.

DISCUSSION
NMO comprises a group of immune-meditaed conditions with complex etiology. While family studies have shown clustering of NMO cases in some familites, the exact genetic background of this disorder has not been clarified yet. Since the first report of familial NMO cases in 1936 (14), several studies have attempted to find susceptibility loci for NMO. The first attempts have been focused on the HLA region, based on the importance of this region in the regulation of immune responses and their association with MS, a disorder that clinically resembles NMO. However, various studies have shown that HLA-related susceptibility loci for NMO is distinct from MS. The HLA-DRB1*03 allele has been the mostly appreciated risk locus for NMO. Several other HLA-DRB1, DQB1, and DPB1 alleles have been found to be associated with NMO. Yet, the results of these studies have not been validated in independent cohorts from different ethnic backgrounds.
Exome sequencing and genome-wide SNP arrays have also validated the significance of the HLA region in conferring risk of NMO. In addition, they have shown other risk loci within AQP4, CYP27B1, CYP7A1, CD226, CD58, CD6, FCRL3, GPC5, MIF, ATG5, PD-1.3, IL2RA, IL7RA, and IL17A. With the exception of AQP4 and CD58, almost other genes have been assessed in single studies, needing confirmation in independent cohorts. Moreover, a number of variants, particularly within SLC28A3 and SLC29A1, have been associated with clinical course or some immune markers in patients with NMO.  Japanese CSF and serum/ The CSF levels of IL-1 receptor antagonist, IL-6, IL-8, IL-13, IL-10, g-csf, and IP-10 were significantly higher in NMO, while only IL-6 level in serum has upregulation. CSF IL-6 level correlated with CSF cells and glial fibrillary acidic protein.
(Continued)    Higher level of IL-6 was identified in sera and SCF samples of patients, particularly in seropositive AQP4-ab than negative type. CSF IL-6 level also correlated with disease severity and AQP4-ab levels.  Deletion-type CNVs can also been regarded as predisposing factors for NMO. Notably, these CNVs have been found to occur as somatic changes.
In addition to several cytokines that are altered in the course of NMO development, expressions of numerous mRNAs, lncRNAs, and miRNAs have been found to be deregulated in the peripheral blood or brain lesions of NMO patients. Not surprisingly, these genes are mostly enriched in pathways related to functions of the immune system.
Finally, in vitro studies have shown the effects of NMO sera on deregulation of function of astrocytes, suggesting the impact of humoral responses on pathoetiology of this condition. Moreover, these circulatory markers could negatively affect permeability of the blood-brain barrier.
Taken together, NMO has a complex genetic background with prominent roles of immune-related genes, particularly cytokine coding genes and those coding cytokine receptors. Future genome-wide studies in NMO patients from different ethnic background would facilitate identification of risk loci for this condition. Finally, systematic review and meta-analysis studies are recommended to produce quantitative results without any bias along with an overview of genetic aspects of Patients with poor recovery had higher AQP4-Ab serum level. Furthermore, AQP4-Ab in good recovery patients was even lower than poor group after treatment. CXCL12 level was significantly lower in poor recovery group and negatively correlated with AQP4-Ab level. It was also related to TNFa and GFAP CSF levels.

NMOSD and 8 healthy controls
Spanish Serum/ Immunofluorescence Assay and ELISA According to the results, only anti-AQP4 antibodies could act as a biomarker in NMOSD diagnosis, and its level was not correlated with disease progression. disease. Also, further studies should assess treatment responses in association with distinct genetic backgrounds. Finally, a limitation of studies conducted in this filed is that the expression profiles of genes and cytokines have not been assessed in association with different treatment options.

AUTHOR CONTRIBUTIONS
MT and SG-F wrote the draft and revised it. TA collected the tables and data. All authors contributed to the article and approved the submitted version.
Eosinophil NMO patients __ Eosinophils cultured from mouse bone marrow exposed to NMO sera Eosinophils induced antibody-dependent cell-mediated cytotoxicity in AQP4-expressed cells and through complementdependent cell-mediated cytotoxicity led to killing cells.
27 cytokines/chemokines 20 NMOSD patients and 10 healthy controls Japanese BMECs treated with human sera/multiplexed fluorescent bead-based immunoassay system and ELISA IL-6, MCP-1, and IP-10 were significantly upregulated in BMECs treated with NMOSD acute phase sera. IP-10 levels were correlated with CSF/serum albumin ratio.
T-cell functions 20 NMO patients and 20 healthy controls Brazilians PBMC, CD4-free PBMC, and purified CD4+ T cells cultured and exposed to glucocorticoid inhibitor/flow cytometry and ELISA T-cell proliferation and Th1 cytokine production were significantly lower in NMO cell cultured, while Th17-like phenotype, IL-6, and IL-23 production were increased. IL-6, IL-21, and IL-23 secretion were less sensitive to glucocorticoid inhibitor. (138)