An Update Evolving View of Copy Number Variations in Autoimmune Diseases

Autoimmune diseases (AIDs) usually share possible common mechanisms, i.e., a defect in the immune tolerance exists due to diverse causes from central and peripheral tolerance mechanisms. Some genetic variations including copy number variations (CNVs) are known to link to several AIDs and are of importance in the susceptibility to AIDs and the potential therapeutic responses to medicines. As an important source of genetic variants, DNA CNVs have been shown to be very common in AIDs, implying these AIDs may possess possible common mechanisms. In addition, some CNVs are differently distributed in various diseases in different ethnic populations, suggesting that AIDs may have their own different phenotypes and different genetic and/or environmental backgrounds among diverse populations. Due to the continuous advancement in genotyping technology, such as high-throughput whole-genome sequencing method, more susceptible variants have been found. Moreover, further replication studies should be conducted to confirm the results of studies with different ethnic cohorts and independent populations. In this review, we aim to summarize the most relevant data that emerged in the past few decades on the relationship of CNVs and AIDs and gain some new insights into the issue.


INTRODUCTION
Copy number variations (CNVs), as a main type of structure variation (SV) caused by genomic rearrangement, mainly include deletion and duplication of sub-microscopic but large genomic segments ranging from 1 kb to 3 Mb (Redon et al., 2006). Single-nucleotide polymorphisms (SNPs) have been recognized to be involved in many autoimmune diseases (AIDs) (Song et al., 2021a;Song et al., 2021b;Jiang et al., 2021); however, CNVs containing more nucleotide content per genome than SNPs are responsible for a large proportion of human genetic variation and show an importance in genetic diversity and evolution (Redon et al., 2006;Cleynen et al., 2016). Thus, more attention has been paid to the research of CNVs in diseases. Nowadays, the genome-wide assays for CNV study include array-based comparative genomic hybridization (aCGH), SNP genotyping microarrays, next-generation sequencing, and long-read sequencing techniques (Hehir-Kwa et al., 2018). There are several main categories in the molecular mechanisms during the process of CNV formation, DNA recombination, rearrangement, and error replication. Besides, CNVs also have several types, such as insertions, deletions, inversions, and translocations Human Genome Structural Variation Working Group et al., 2007). Numerous reports imply that CNV is one of the main genetic factors underlying human diseases, including AIDs (Hauptmann et al., 1974;Tomer  TaqI southern blots (Yang et al. 2007); real-time PCR (Pereira et al. 2019) Four copies of C4 genes (majority) Lower copy number of C4 and C4A (p 0.00002; Yang et al. 2007); low total copy number of C4 (p < 0.001), C4A (p < 0.001), and C4B (p 0.03) Consistent Hauptmann et al. (1974), Yang et al. (2007), and Pereira et al. (2019) Iafrate et al., 2004;Mack et al., 2004;Redon et al., 2006;Yang et al., 2007;Yim et al., 2010). The mechanisms underlying the involvement of CNVs in clinical phenotypes are mainly gene disruption and rearrangement (McCarroll and Altshuler, 2007;Yim et al., 2010). Further deep studies on CNVs have shed new light on human genome structure, genetic variations between individuals, and genetic pathogenic factors of human AIDs. AIDs usually share possible common mechanisms, i.e., a defect in the immune tolerance exists due to diverse causes from central and peripheral tolerance mechanisms. Although majority of human genetic variations do not contribute to overt diseases (Tomer and Davies, 2003), some genetic variations including SNPs, nucleotide insertions/deletions, structural variations, and CNVs are known to link to several AIDs during the past few decades and are of importance in the susceptibility to AIDs and the potential therapeutic responses to medicines (Iafrate et al., 2004;Redon et al., 2006). Studies have revealed that some SNPs are related to AIDs and could be genetic mechanisms underlying the development of AIDs (Song et al., 2021a;Song et al., 2021b;Jiang et al., 2021). Structural variations including complex rearrangement of segments with sizes of thousands to millions of base pairs have been recognized as a rich source of genetic diversity. CNVs as a crucial source of genomic diversity caused by the rearrangement of genome are ubiquitously presented in human genome and may affect the susceptibility to many diseases (Iafrate et al., 2004). During the past years, thousands of gene CNVs have been reported (Redon et al., 2006). However, most studies focus on their relationship with tumors and chronic diseases (Jia et al., 2011;Saadati et al., 2016;Voll et al., 2017). Moreover, CNVs are very common in genomic regions encoding immune-related genes, which are closely related to the etiology of AIDs. Thus, they potentially impact polygenic autoimmunity and may lead to the imbalance of the autoimmune system and the development of some AIDs. Additionally, some common CNVs have also been reported to be correlated to several specific AIDs (Mamtani et al., 2008;Mamtani et al., 2010;Liu et al., 2011). Yim et al. (2015) reviewed studies on the clinical implications of copy number variations in autoimmune disorders in detail. However, the relationship between CNVs and the pathogenesis of AIDs has not been fully revealed. Furthermore, in recent years, as new

SYSTEMIC LUPUS ERYTHEMATOSUS
Systemic lupus erythematosus (SLE), with a prevalence rate of eight-nine times higher in females than in males during childbearing age, is a typical systemic autoimmune inflammatory disease with a strong genetic susceptibility and is characterized by the production of autoantibodies and the existence of chronic inflammation. There is a wide range of clinical features in SLE patients, such as discoid lesions, nephritis, arthritis, and malar rash (Barbosa et al., 2018). So far, the exact genetic physiology of SLE remains an open question. Complement is well known to be involved in many immunemediated diseases. Among them, complement component C4 is a pivotal effector of the immune system. There are two common isoforms of C4: C4A and C4B, and 95% of C4A and 54% of C4B contain an endogenous retroviral sequence in their ninth intron, HERV-K (C4) (Wouters et al., 2009). HERV-K (C4) can cause the antisense transcription of C4 as it is oriented opposite to C4 (Mack et al., 2004). According to the presence or absence of HERV-K (C4), there are two different size varieties of isotypes for both C4A and C4B: C4L (long) and C4S (short) (Mack et al., 2004;Wu et al., 2008). Hauptmann et al. (1974) first found the deficiency of C4 copy number in SLE patients. Later, Yang et al. (2007) uncovered that SLE patients had lower copy number of total C4 and C4A genes than healthy volunteers. Pereira et al. (2019) observed that the risk of developing SLE was 2.62 times higher in subjects with low total C4 copy number and 3.59 times higher in subjects with low C4A copy number. These consistent results imply that deletion or deficiency of C4 or C4 isotypes will increase the risk for SLE. The potential mechanism is like this, the decrease of C4 CNVs will cause C4 deficiency, bring the impairment of autoantigene clearance and the negative selection of auto-reactive B cells, and then favor the onset of SLE (Pereira et al., 2019). In addition, there was another study aiming at the relationship between C4 CNVs and the drug response to treatment of SLE. The study of Mulvihill et al. (2019) found that higher copy number of complement C4B and elevated serum complement levels were associated with hypertension and effective response to statin therapy in childhood-onset SLE patients. C-C chemokine ligand 3 like-1 (CCL3L1) is a potent ligand for the HIV coreceptor, and C-C chemokine receptor 5 (CCR5) is an important factor in immune response (Gonzalez et al., 2005). Mamtani et al. (2008) found that the CNVs of CCL3L1-CCR5 were strong predictors for the overall risk of SLE and high autoantibody titers and lupus nephritis, and subjects with lupus nephritis differentially recruit leukocytes.
Receptors for the Fc portion of IgG are involved in the handling and clearance of immune complex and in the regulation of B cell activation during SLE development (Niederer et al., 2010). Willcocks et al. (2008) have found that low copy number of Fc gamma receptor 3B (FCGR3B), which is correlated with protein expression and immune complex uptake, was associated with SLE, implying that the association of this gene CNVs with SLE may influence protein expression and function and further confer risk for the predisposition of AIDs. Molokhia et al. (2011) identified that low copy number of FCGR3B was associated with the risk of SLE in the Afro-Caribbean population, but not in the African ancestry population. Chen et al. (2014) further showed that low copy numbers (less than two copies) of FCGR3A or FCGR3B were significantly associated with SLE, and high copy numbers (more than two copies) of FCGR3A were also related to SLE onset in the Taiwanese population. Another team, Barbosa et al. (2018), detected CNV at whole-genome level using a case-control design and showed that increased FCGR3B/ ADAM3A copy number was a protective factor against SLE development. In addition, they, for the first time, uncovered heterozygous deletions overlapping the CFHR4, CFHR5, and HLA-DPB2 genes in SLE patients. Notably, different genetic manifestations attributing to different backgrounds may present different trends of gene CNV association with diseases.
Heat shock proteins 90 (HSP90) is a pivotal modulator of multiple innate and adaptive inflammatory processes (Tamura et al., 2012). HSP90 has two major cytosolic isoforms, HSP90AA1 and HSP90AB1. Zhang et al. (2019) showed that HSP90AB1 CNV was correlated with SLE in the Chinese Han population, especially in females, implying that HSP90AB1 CNV is involved in the pathomechanism and development of SLE.

RHEUMATOID ARTHRITIS
Rheumatoid arthritis (RA) is characterized by a massive tissue infiltration of inflammatory cells and affects approximately 1% of the adult population worldwide. Clinically, it mainly causes the chronic inflammation of synovial joints, which will result in the progressive destruction of the cartilage and bone (Achour et al., 2018). Many lines of data have implied the association of RA with genetic variations including SNPs and CNVs of several immunerelated genes (McKinney et al., 2008;Graf et al., 2012;Olsson et al., 2012).
It is intriguing whether FCGR3B CNV is involved in RA. Graf et al. (2012) found a significant association between low FCGR3B copy number and RA (Rahbari et al., 2017). Chen et al. (2014) revealed a significant association of RA with low copy number of FCGR3A, but not FCGR3B in a Chinese cohort. Then, Rahbari et al. (2017) verified that RA patients from the UK had decreased copy number of FCGR3B. More recently, Ben Kilani et al. (2019) reported that genotypes without null allele of FCGR3B gene (copy numbers range from 1 to 3) were significantly associated with RA. In addition, increased FCGR3B copy number was only found in RA, and deletion of FCGR3B may have a protective effect on RA. The discrepancies on the correlation of RA with FCGR3B CNVs may partly be due to different genetic backgrounds, and this requires further investigation in different ethnic populations.
CNVs of CCL3L1 are potentially associated with RA. McKinney et al. (2008) uncovered that a higher than two copy number of CCL3L1 gene was a risk factor for RA in the New Zealand population (Wouters et al., 2009). Mohamed et al. showed that CCL3L1 copy number varies from 0 to 4 in the French population and from 0 to 7 in the Tunisian population. In addition, five copies of CCL3L1 gene protect people from RA development in the Tunisian population (Ben Kilani et al., 2016). The results showing ethnic heterogeneity of CCL3L1 among different populations suggest that CCL3L1 is a susceptible factor for RA, and these differences of results may be due to genetic background and/or environmental differences existing among various populations and the different technical detection methods for genotyping multi-allelic CNVs. In the future, more reliable technologies are needed to explore the relationship between CCL3L1 CNVs and RA in more populations to identify the results.
Inflammation is the central trait of RA, accompanied with the production of reactive oxygen species (ROS), which induces oxidative damages to cellular molecules including DNA and lipids, causing diverse cytotoxic products (Fahmi et al., 2002). The neutrophil cytosolic factor 1 (NCF1) gene encodes one of five sub-units of the NADPH oxidase (NOX2) complex, which could produce ROS in the immune cells, including antigen presenting cells (APCs), phagocytes, etc. Olsson et al. (2012) found that RA patients are less likely to have an elevated copy number of NCF1 compared to controls, which may suggest that a higher copy number of NCF1 could be a protective factor for RA. Glutathione S-transferase (GST), a multifunctional enzyme that is required for the detoxification mechanism against ROS products, exerts a vital protective antioxidant function in cells against ROS aggression (Blackburn et al., 2006). In 2011, a Swedish study found that more than one copy of glutathione S-transferase M1 (GSTM1), a member of the GST family, seems to be a risk factor for autoantibody-positive RA in non-smoking females of age older than 60 years, and GSTM1 acts as a protective factor in ACPAnegative smoking men, suggesting that the copy number of the GSTM1 gene is correlated with the development and severity of RA (Lundström et al., 2011). In 2016, it was also found that GSTM1 is deleted in Tunisian anti-cyclic citrullinated peptide (anti-CCP)-positive RA patients, although a genetic association of GSTM1 CNV with predisposition to RA was not detected (Achour et al., 2018). The above data indicate that GSTM1 CNVs do not influence the susceptibility to RA, but may have an effect on its severity because deletion of GSTM1 could increase the risk of anti-CCP-positive RA. These interesting results showed that the CNVs of inflammatory-related genes may be involved in the development of RA through influencing the production of ROS and oxidative damages to cellular molecules.

INFLAMMATORY BOWEL DISEASE
Inflammatory bowel disease (IBD) is a kind of organ-specific inflammatory disease and has two main clinical subtypes, ulcerative colitis (UC) and Crohn's disease (CD). The occurrence of CD is 300 per 100,000 people in the population with European ancestry and increasing in other ethnic population. CD affects the gastrointestinal tract and has such symptoms, like diarrhea, abdominal pain, and aberrant weight loss (Cleynen et al., 2016). Studies have demonstrated the associations between C4 gene and several AIDs (Hauptmann et al., 1974;Hou et al., 2013), and the team led by Cleynen et al. (2016) showed that CD cases tended to have lower C4L and higher C4S copies. They also found that serum C4 protein level was not significantly different between CD patients and controls, but CD patients with higher C4 copy number may have higher serum C4 concentration (Cleynen et al., 2016). These results suggest that more C4 copy number may lead to higher C4 expression and there may be a dose-efficiency correlation between C4 copy number and protein expression in CD patients. Asano et al. (2013) studied polymorphisms of FCGR genes in the Japanese population and found that FCGR3B copy number was related to susceptibility of UC.
Defensins are endogenous antimicrobial peptides to protect the intestinal mucosa against bacterial invasion. Fellermann et al. (2006) showed that healthy volunteers as well as UC patients have 2 to 10 copies of the human beta-defensin 2 (HBD-2) gene with the median of 4 copies. However, patients with colonic CD have lower HBD-2 copy than the controls. In addition, they also found that less than four copies of HBD-2 gene were correlated with diminished mucosal HBD-2 mRNA expression (Fellermann et al., 2006). The DEFA1A3 gene encodes alpha-defensins 1-3. Jespersgaard et al. (2011) found that a higher DEFA1A3 copy number was related to CD, especially to colonic CD.

PSORIASIS
Psoriasis is a serious inflammatory disease of the skin, scalp, nails, and joints and has a prevalence of about 2% in the populations of developed countries (McKinney et al., 2008). Multiple studies have identified a strong genetic component in the development of psoriasis and demonstrated the relationship between CNV of some genes and psoriasis. Wu et al. (2014) found that Chinese patients with psoriasis vulgaris had a higher copy number (more than two copies) of FCGR3B compared to controls through a case-control study. Carpenter et al. (2011) found no association between CCL3L1 copy number and psoriasis. Hollox et al. studied psoriasis in Dutch and German populations and found significant associations between higher genomic copy number for betadefensin genes (DEFB4, SPAG11, DEFB103, DEFB104, DEFB105, DEFB106, and DEFB107) and the risk of psoriasis in both of these cohorts.
IL-22, which belongs to IL-10 cytokine family, has a significant proliferative effect on different cell lines and a role of immune regulation. Prans et al. (2013) showed that the copy number variation in exon 1 but not exon 5 of IL-22 gene was significantly correlated with the severity of psoriasis.
With the emergence of new technologies, more loci with CNVs have been identified to be associated with psoriasis. Bergboer et al. (2012) utilized a pooling approach, genome-wide CNV analysis, and array comparative genomic hybridization to detect CNV variability in psoriasis and found that the absence of the late cornified envelope (LCE) gene cluster members LCE3B and LCE3C (LCE3C-LCE3B-del) was significantly associated with the predisposition of psoriasis in populations from Netherlands.

BEHCET'S DISEASE
Behcet's disease (BD) is an immune-mediated systemic inflammatory disorder involving non-granulomatous uveitis, recurrent oral and genital ulcers, as well as skin lesions (Yang et al., 2008a). It has been verified to be associated with the HLA-B51 gene (Xu et al., 2015). Further studies have provided new insight into the pathogenesis of BD and attracted more attentions on whether complement is involved in BD. Hou et al. (2013) investigated the relationship of CNV of C4A and C4B with BD and found that the frequency of more than two copies of C4A was significantly increased in BD patients, and C4A CNV was an independent risk factor for BD. Moreover, C4A expression was significantly increased in BD patients with high C4A copy number than with low C4A copy number (Hou et al., 2013). Xu et al. (2015) found that BD patients have increased frequencies of more than two C3 copies, and C5 CNV was associated with BD. Furthermore, interleukin-17 (IL-17) and IFN-gamma expressions were upregulated in BD patients with high C3 copy number, but not in BD patients with high C5 copy number (Xu et al., 2015). The results imply that there are indeed CNVs in complement-related genes, and the CNVs in these genes may be involved in the development of BD.
There were also some studies targeting the relationship between FCGR3B CNVs and BD. Black et al. (2012) found that CNV of the FCGR3B gene was associated with the risk of BD in the Iranian population. The risk of BD was decreased by 40% in people with less than two copies of FCGR3B and by 25% in people with more than two copies of FCGR3B, although these tendencies were not statistically significant. They concluded that no association exists between high or low copy number of FCGR3B and BD or its clinical features (Black et al., 2012). Of course, further studies are need to identify this result in other populations.
Toll-like receptors (TLRs), along with RIG-I-like receptor and NOD-like receptors, belong to the family of pattern recognition receptors and contain 10 functional members in human beings, namely, TLR1-10 (Medzhitov et al., 1997). TLRs have been known to play important functional roles in several AIDs and inflammatory disorders (Chang et al., 2004;Chang et al., 2006). Fang et al. (2015) uncovered that more than 98% of people tested have two copies of all TLRs except for TLR7. In addition, they found that compared to healthy controls, male BD patients had an increased frequency of more than one copy of TLR7, and female BD patients had an increased frequency of more than two copies of this gene. This research suggested that a high TLR7 copy number may contribute to the pathogenesis of BD . Therefore, from this research, we included that TLRs also play a potential role in BD, and the mechanism underlying this still needs to be clarified.
Many interleukins play vital roles in BD development. Hou et al. (2015) showed that BD patients have increased frequencies of more than two copies of IL-17F and IL-23A, and after stratified by sex, the association just exists in male BD patients. IL-17F protein expression is positively correlated with its gene copy number, and higher IL-17F copies are associated with enhanced proliferation of peripheral blood mononuclear cells (PBMC) . This suggests that not only SNPs but also CNVs of interleukins are involved in the pathogenesis of BD. Liao et al. (2015) investigated the association of some transcript factors with BD and showed that high CNV of related orphan receptor (RORC) was associated with BD susceptibility, and low Foxp3 CNV was correlated with female BD. In addition, individuals with high RORC copy number seem to have relatively high mRNA levels of RORC, IL-1β, and IL-6, but not Foxp3 .
The disturbed apoptosis has been reported to be involved in BD development. Yu et al. (2015)investigated whether CNVs of apoptosis-related genes, including FAS, caspase8, caspase3, and BCL2, were associated with BD in the Chinese population and showed that BD patients had an increased frequency of high FAS copy number, and BD patients with more than two copies of FAS had an increased mRNA expression of FAS in anti-CD3/ CD28 antibody-stimulated CD4 + T cells. Their results provided important evidence that high FAS copy number is involved in the pathogenesis of BD (Yu et al., 2015).
It is well-known that BD may be triggered by infectious agents in some genetically susceptible people. DEFB4 CNV can affect the level of human beta-defensin 2, which is an inducible antimicrobial peptide. Park et al. (2011) found that DEFB4 copy number was lower in BD samples than in controls without statistical significance, and DEFB4 copy number was not associated with the clinical characteristics of BD. This suggests that DEFB4 CNVs confer no risk for the susceptibility of BD (Park et al., 2011). In 2012, Ahn et al. (2012) found that 31.1% of samples had five copies of DEFA1 with a mean of 5.4 ± 0.2. Although the distribution of DEFA1 copy number is not different between BD patients and the controls, high DEFA1 copy number is related to the intestinal involvement in BD, suggesting that a high DEFA1 copy number may be associated with the development of intestinal involvement in BD (Ahn et al., 2012). Hitherto, the genes related to infectious agents may be involved in BD through changing the variation of copy number, and further researches are needed to be carried out to identify this.
There was another study that investigated the relationship between microRNA CNVs and BD, like miR-23a, miR-146a, and miR-301a. As a result, no association of CNVs of the abovementioned miRNAs was observed in BD patients . However, whether other microRNA CNV is associated with BD still needs further studies to explore and clarify.

ANKYLOSING SPONDYLITIS
Ankylosing spondylitis (AS) is an inflammatory AID causing spondyloarthritis of the spine and sacroiliac joints and prevalent Frontiers in Genetics | www.frontiersin.org January 2022 | Volume 12 | Article 794348 mainly in men with a ratio of 10:1 at the age of 20-30 years old in respect to women (Chimenti et al., 2021). Because the exact pathology of AS is still unclear, there are incoming data about the relationship between CNVs of these genes and AS, including CCL3L1, FCGR3A, FCGR3B, TLR7, UGT2B17, BMP8A, and so on. Wang et al. (2016) found that AS patients had low copies (&3) of FCGR3A and FCGR3B in the Chinese population, implying that a lower copy number of these two genes confers risk for the susceptibility of AS. The study of Dahmani et al. (2019) found that the proportion of AS patients with less than two copies of FCGR3A was higher in the Algerian population and that less than two copies of FCGR3A was only associated with HLA-B27-negative AS patients, suggesting that FCGR3A deletion has an independent effect on AS regarding HLA-B27 status. Their results also showed that CCL3L1 and FCGR3B CNVs may not be involved in the predisposition of AS in the Algerian population (Dahmani et al., 2019). Wang et al. (2018) found that one copy of TLR7 was related to AS in the Chinese population after Bonferroni correction and adjustment of age and sex, and less than one copy of TLR7 confers risk for AS susceptibility in male patients, but is a protective factor in female AS patients. Uddin et al. (2013) conducted a genome-wide CNV analysis and found that UGT2B17 copy number was increased in a large AS multiplex family. The UGT2B17 gene encodes an enzyme that metabolizes steroid hormones such as testosterone and selected xenobiotics (Xue et al., 2008). UGT2B17 copy number has been shown to be related to bone mineral density and involved in the pathogenesis of osteoporosis (Yang et al., 2008b). It is known that AS in patients is often accompanied with osteoporosis (Vosse et al., 2009). This may, to some extent, explain the underlying mechanism that gain in UGT2B17 copy number could increase the risk for AS. Bone morphogenic protein 8A (BMP8A) plays multiple functions in the formation of a bone. Shahba et al. (2018) reported that the expression of BMP8A in PBMCs is decreased in AS patients, and BMP8A CNVs do not influence its transcription in PBMCs and are not associated with AS susceptibility in the Iranian population. Cai et al. (2015) found that the copy number of defensing-related gene DEFB103 was in the range of two to six in both AS patients and controls, and it was not associated with AS. More studies in different populations are needed to further identify the relationship between these gene CNVs and BD. CD4 + T cells play pivotal roles in many AIDs, but whether CNVs of transcription factor genes in CD4 + T cells are involved in AS remains poorly defined. Bai et al. (2016) investigated whether CNV of transcription factor genes in CD4 + T cells including T-bet, GATA binding protein 3 (GATA-3), RORC, and fork-head box protein 3 (FOXP3) are associated with acute anterior uveitis (AAU) in the presence or absence of AS. They found that a higher T-bet copy number is more common in AAU+, AS+ and AAU+, AS− cases compared with healthy controls. Additionally, the frequency of AAU+, AS+ patients with high GATA-3 copy is higher, and the proportion of female AAU+, AS+ patients with high FOXP3 copy number is also higher than that of other populations, but the copy number of RORC is not correlated with AAU+, AS+ or AAU+, AS− patients .
People are also interested in the relationship between CNVs of various microRNAs (miRNAs). Yang et al. (2017) studied the association between CNVs of miRNAs and AS and found that the frequencies of AAU+, AS+ patients with low copy numbers of miR-143, miR-146a, miR-9-3, and miR-205 as well as high copy numbers of miR-301a and miR-23a all increased, and the frequencies of patients with AAU+, AS− with low copy number of miR-146a and high copy numbers of miR-23a and miR-205 are significantly different. In addition, they found that miR-9-3 mRNA expression is significantly decreased in AAU+, AS+ patients and positively correlated with its copy number (Yang et al., 2017).

PRIMARY SJOGREN'S SYNDROME
Primary Sjogren's syndrome (pSS) is characterized by the presence of circulating autoantibody (anti-Ro/SSA and anti-La/ SSB), as well as the involvement of the exocrine glands (salivary and lacrimal gland), joint, and muscle Haldorsen et al., 2013). There were few studies on the issue of FCGR3B CNVs and pSS. Mamtani et al. (2010) showed that the median FCGR3B gene copy is two in the cohort of Spanish ancestry. The risk of pSS would increase if people carry less or more than two copies of FCGR3B (Mamtani et al., 2010). Nossent et al. (2012) conducted a case-control study and found that less than two copies of FCGR3B can confer risk for pSS in the Australian population, and low FCGR3B copy number is associated with the levels of rheumatoid factor (RF) titers and serum IgG, but not with anti-Ro ± La autoantibodies. They further identified that FCGR3R CNV is a genetic susceptibility factor for pSS . However, Haldorsen et al. (2013) showed no association of FCGR3B CNV with pSS in the Norway and Switzerland populations. To clarify, these controversial results need more studies in more populations with different ethnicities and regions.

TYPE 1 DIABETES MELLITUS
Type 1 diabetes mellitus (T1DM) is characterized by β cell destruction in the pancreas and the production of antibodies against β cells, with a high prevalence of 1 in 350 teenagers in the UK (Bluestone et al., 2010). There is a series of symptoms in T1DM, such as polydipsia, polyuria, polyphagia, and weight loss. C4 is a gene of the highly variable complement pathway situated ∼500 kb from DRB1 and DQB1 and strongly associated with diverse AIDs (Hauptmann et al., 1974;Hou et al., 2013;Cleynen et al., 2016), so some scientists also carried out numerous researches to uncover whether C4 is associated with T1DM development. Kingery et al. (2012) found that higher C4A copy number tends to be correlated with the protection of residual β-cell function in new-onset T1DM patients, while lower C4B copy number is related to the end of disease remission at 9 months post diagnosis. Mason et al. (2014) explored the relationship between C4 CNV and T1DM and found that individuals with T1DM have significantly fewer copies of HERV-K (C4), one notable component of C4. About the relationship between CCL3L1, people also did a lot of work. McKinney et al. (2008) revealed an association between CCL3L1 CNVs and T1DM in the Caucasian population from New Zealand, but the association was not statistically significant with P of 0.064. Then, the Wellcome Trust Case Control Consortium found no association between CCL3L1 CNV and T1DM (Wellcome Trust Case Control Consortium et al., 2010). These work are consistent and implies that CCL3L1 CNVs may be not associated with T1DM. FCGR3B CNVs are involved in the pathogenesis of several AIDs, such as SLE and RA, because its role in the clearance of immune complexes is impaired in these disease settings (Willcocks et al., 2008;Graf et al., 2012). Almal and Padh (2015) found that FCGR3B copy number in the Indian population varies significantly from zero to two per diploid genome in other populations, which helps us to understand the potential role of FCGR3B CNV and its association with AIDs in the Indian population.
Killer immunoglobulin-like receptors (KIRs) reside on the surface of natural killer cells to bind to their corresponding human leukocyte antigen (HLA) class I ligands. It is noted that KIRs are vital candidates for HLA-associated AIDs, including T1DM. Although Pontikos et al. (2014) did not find a relation of KIR3DL1/3DS1 copy number to T1DM in the white European population, Grayson et al. (2010) utilized a more powerful genome-wide CNV analysis and found 39 CNVs either enriched or depleted in T1DM patients, including a deletion on chromosome 6p21, near an HLA-DQ allele. Their results indicated that both enrichment and depletion of these genes are high risk factors for developing T1DM, and genetic variants such as CNVs may contribute to the development of islet autoimmunity in T1DM (Grayson et al., 2010).

AUTOIMMUNE THYROID DISEASE
Autoimmune thyroid disease (AITD), which mainly includes Graves' disease (GD), Hashimoto's thyroiditis (HT), and Graves' ophthalmopathy (GO), is a kind of organ-specific AID with a prevalence of 5% of the overall population (Jin et al., 2018). GD is characterized by hyperthyroidism caused by positive autoantibodies against thyroid-stimulating hormone receptors (TSHR), and HT is often characterized by positive anti-thyroid peroxidase antibody (TPOAb), anti-thyroglobulin antibody (TgAb), and hypothyroidism. There have been many studies focusing upon the relationship between the CNVs of immunerelated genes and the development of AITD, and lots of interesting results have been reported. The two earliest studies on CNVs and AITD were both published in 2011. One study aimed to explore the association between CNVs of immune regulatory genes and AITD and found no CNVs of CD40 and CTLA4 genes in GD cases and few PTPN22 CNVs in two GD individuals (Huber et al., 2011). Liu et al. (2011) found that CNVs of C4, C4A, and C4B contribute to the predisposition of GD, but not GO. Liao et al. (2014) revealed that TSHR CNVs harbor the etiology of GD, but not GO. Besides, they also observed that only female GD patients have fewer TLR7 copies, and there is no significant association between sex and TLR7 CNVs (Liao et al., 2014). GD is a disease predominantly in females (Manji et al., 2006); therefore, they believed that TLR7 CNVs affect the pathogenesis of female GD due to the gender-dependent immune response (Liao et al., 2014). In 2017, we conducted microarray to explore the profile of genes with CNVs in GD and found some genes with copy number gain. In addition, seven of these genes including CFH, CFHR1, KIAA0125, UGT2B15, UGT2B17, TRY6, and CCL3L1 were chosen to further validate these findings in an expanded cohort. The results showed no correlation between CNVs of these genes and GD (Song et al., 2017). Jin et al. (2018) assessed CNVs of two immune-related genes SIRPB1 and TMEM91 in AITD and found that the distributions of SIRPB1 copy number were different between AITD patients and the controls, implying SIRPB1 is a risk factor for AITD. Guan et al. (2020) showed that the distribution of copy numbers of cell growth-related genes glypican-5 (GPC5), B9 domain-containing 2 (B9D2), as well as ankyrin repeat and suppressor of cytokine signaling (SOCS) box-containing protein 11 (ASB11) is different in AITD patients and the controls, and GPC5 CNVs are risk factors for AITD. However, they did not find any association between their CNVs and the occurrence of AITD (Guan et al., 2020). The different relationship between CNVs of these genes and different sub-types of AITD implies the diverse genetic mechanisms underlying AITD. Overall, researchers have found that CNVs of several thyroid-susceptible genes are correlated with the development of AITD. Studies on the correlation between CNVs and AITD susceptibility inevitably deepen our understanding of the pathogenetic mechanism of AITD and further promote molecular diagnosis and therapies of AITD. These studies were mainly done in Chinese population; in the future, we will need further studies in more populations to identify these results.

FUTURE OUTLOOK
To further gain new and deep insights on the genetic mechanism of AIDs, we reviewed the association between different AIDs and CNVs of some genes with potential pivotal roles in the development of AIDs. This review provides an update evolving the view of copy number variations in AIDs. For the first, several CNVs are very common in diverse AIDs, implying these AIDs may potentially possess a similar genetic pathomechanism. Therefore, more association studies should be done on some other diseases when a certain link is identified between some CNVs and a specific AID. For the second, some CNVs are differently distributed in various diseases in different ethnic populations, suggesting that AIDs have their own different phenotypes and different genetic and/or environmental backgrounds among diverse populations. Herein, more researches aiming to uncover the relationship between environmental factors and diseases and the influences of environmental elements on immunity should be encouraged. For the third, with the continuous advancement in genotyping technology such as the high-throughput whole-genome sequencing method, more susceptible variants will be found.
Thus, further replication studies should be conducted to confirm the results of studies with different ethnic cohorts and independent populations.
For that, CNVs may be an important pathogenesis of AIDs; CNVs will also become an effective way to study the molecular mechanism of AIDs; and we can find molecular markers for genetic diagnosis or judgment of prognosis of this kind of disease. At the same time, it opens a new page for the research of AIDs and is becoming a new research hotspot. We believe that genetic diagnosis or judgment of prognosis based on CNVs will cover more AID spectrum and benefit a wider population. It is believed that with the high-throughput genome-wide CNV scanning platform and the new development of statistical calculation method, GWAS based on CNVs, a new genetic susceptibility marker, will become a powerful tool to study the genetic susceptibility of AIDs, just like the traditional GWAS based on SNPs and their haplotypes. These two complementary genetic markers will help us to understand the molecular mechanism, identify susceptible genes, and understand the relationship between genetic variations and disease phenotype of complex diseases, such as AIDs, which is of great significance. Simultaneously, to our hope, more functional experiments and more replication studies should be done and collected, and an entire autoimmune CNVs database should be set up, which can be searched easily and help us to understand the pathogenesis of AIDs much better. GWAS has already pointed toward genetic susceptibility loci and potential mechanisms of pathogenicity. Chromosomal microarrays have added little further information. It therefore seems unlikely that whole-genome sequencing alone will answer the necessary questions. Rather, genomic DNA approaches will likely need to be combined with other measures, such as RNA sequencing and/or proteomics to solve many of the remaining questions at hand.
Although CNVs describe the pathogenesis of AIDs from a new perspective, it is still early to explain the occurrence and development of complex diseases only by CNVs at the genomic level because the molecular mechanism of complex diseases at the chromosomal level is not completely clear. Due to the existence of multiple mechanisms of CNVs and the different effects of CNVs on molecular phenotype and gene expression, therefore, the interpretation of clinical significance and genetic mode of CNVs must be done more carefully, and it should be based on the comprehensive assessment of genomic variation.

AUTHOR CONTRIBUTIONS
R-HS designed the study. C-QG and JZ extracted the data. R-HS analyzed the data and wrote the first draft of the report. J-AZ revised the manuscript. All authors contributed to the article and approved the submitted version.