An Autoimmune Disease-Associated Risk Variant in the TNFAIP3 Gene Plays a Protective Role in Brucellosis That Is Mediated by the NF-κB Signaling Pathway

ABSTRACT Naturally occurring functional variants (rs148314165 and rs200820567, collectively referred to as TT>A) reduce the expression of the tumor necrosis factor alpha-induced protein 3 (TNFAIP3) gene, a negative regulator of NF-κB signaling, and predispose individuals to autoimmune disease. In this analysis, we conducted a genetic association study of the TT>A variants in 1,209 controls and 150 patients with brucellosis, an infectious disease, and further assessed the role of the variants in brucellosis. Our data demonstrated that the TT>A variants were correlated with cases of brucellosis (P = 0.002; odds ratio [OR] = 0.34) and with individuals who had a positive serum agglutination test (SAT) result (titer of >1/160) (P = 4.2 × 10−6; OR = 0.23). A functional study demonstrated that brucellosis patients carrying the protective allele (A) showed significantly lower expression levels of the TNFAIP3 gene in their peripheral blood mononuclear cells and showed increased NF-κB signaling. Monocytes from individuals carrying the A allele that were stimulated with Brucella abortus had lower mRNA levels of TNFAIP3 and produced more interleukin-10 (IL-10), IL-6, and IL-1β than those from TT allele carriers. These data showed that autoimmune disease-associated risk variants, TT>A, of the TNFAIP3 locus play a protective role in the pathogenesis of brucellosis. Our findings suggest that a disruption of the normal function of the TNFAIP3 gene might serve as a therapeutic target for the treatment of brucellosis.

B rucellosis is a zoonosis caused by Gram-negative bacteria of the genus Brucella that infect many farm animals, including cattle, sheep, goats, and pigs. There are approximately half a million human cases of Brucella infections per year, even though humans are only incidental hosts (http://data.stats.gov.cn/index.htm). There is no approved human vaccine available. In humans, brucellosis typically presents with a high, undulating fever. However, chronic brucellosis may affect multiple host organs, leading to arthritis, orchitis, encephalomyelitis, and endocarditis. Osteoarticular disease represents the most common complication (1). The diverse manifestations of brucellosis make the diagnosis of this disease even more complicated. Brucellosis in humans and livestock is relatively uncommon in industrialized nations. However, brucellosis is endemic in many developing regions, including China. As shown in Fig. 1, the annual incidence rate of brucellosis in China has increased from 2002 to 2015 (http://data.stats .gov.cn/index.htm).
The interaction of Brucella with the human immune system is critical for the development of chronic parasitism and the clearance of infection (2)(3)(4). NF-B signaling plays a vital role in the immune system by regulating innate and adaptive immunity (5). Bacterial and viral infections rapidly induce the activation of NF-B signaling, producing a potent inflammatory response. It is known that immune responses vary among individuals. These differences between individuals can be partly explained by the presence of genetic polymorphisms within immune-response-related genes that regulate the activities of inflammatory signaling pathways.
The tumor necrosis factor alpha-induced protein 3 (TNFAIP3) gene encodes the ubiquitin-editing enzyme A20, a key negative regulator of NF-B activity. Genetic studies have suggested a role for TNFAIP3 in susceptibility to complex genetic autoimmune disorders (6)(7)(8)(9)(10)(11)(12), including systemic lupus erythematosus (SLE) (13)(14)(15)(16)(17)(18)(19)(20)(21)(22)(23)(24). In our previous studies, we described a pair of tandem polymorphic dinucleotides (rs148314165 and rs200820567, collectively referred to as TTϾA) that decrease the gene expression level of TNFAIP3 (25)(26)(27). Further characterization of the molecular mechanisms involved revealed that the TTϾA variants reside in an enhancer element that binds the transcription factors NF-B and SATB1, enabling the interaction of the TTϾA enhancer with the TNFAIP3 promoter through long-range DNA looping. Impairment of NF-B binding to the TTϾA risk alleles inhibits the looping interaction, resulting in reduced A20 expression. Additionally, transcription activation-like effector nuclease (TALEN)-mediated knockout of the TTϾA enhancer in HEK293T cells enhanced the activity of NF-B signaling.
Cell-mediated immunity serves a crucial role in host protection against Brucella. Among the cells responsible for this, macrophages are key elements in the cellular immune response against intracellular Brucella (28,29). Activated macrophages use a broad range of phagocytic and inducible bactericidal functions to kill Brucella. Moreover, infected macrophages produce proinflammatory cytokines and chemokines. A previous study showed that A20 knockdown in macrophages promoted Brucella abortus-induced NF-B activation and macrophage cell death, which suppressed B. abortus intracellular replication (30). Additionally, several regulatory molecules secreted by monocytes may contribute to the intracellular survival of Brucella (4, 31). plication, and the presence of high titers of specific antibodies. According to the ELISA results, the patients were divided into IgM-positive (n ϭ 136) and IgG-positive (n ϭ 120) groups. An additional 1,209 matched population controls were enrolled in the same geographic location. Arthralgia, fever, fatigue, and sweating were common in most of the patients. Other relevant patient clinical characteristics at presentation are listed in Table 1.
The rs7749323 variant (G>A) is a perfect proxy for the TT>A polymorphism downstream of the TNFAIP3 gene in a Chinese population. The TTϾA variants each include a deletion of T followed by a T-to-A transversion at positions 138272732 and 138271733 on chromosome 6, 42 kb downstream from the TNFAIP3 gene. To obtain accurate genotypes of the TTϾA variants (which include both an SNP and a deletion), a perfect proxy SNP (rs7749323) of the TTϾA variants was selected and genotyped in our cohort. The proxy SNP is in complete linkage disequilibrium (LD) with the TTϾA variants (r 2 ϭ 1) in European and Korean individuals (25,27). To assess the frequency of the TTϾA variants in our cohort, we resequenced an 850-bp DNA fragment, centered on the TTϾA variants, in 50 brucellosis cases and 50 healthy controls. As shown in Fig. 2 and Table S1 in the supplemental material, the TTϾA variants are in complete LD with the rs7749323 variant, with an r 2 value of 1. The SLE-associated risk allele (T) plays a protective role in brucellosis. The demographics of 150 brucellosis cases and 1,209 matched population controls enrolled in the study are shown in Table S2 in the supplemental material. There were no significant differences between patients and control subjects in terms of mean age or gender distribution. Single-marker association was performed by using logistic regression. We observed a significant negative association (P ϭ 2.0 ϫ 10 Ϫ3 ; OR ϭ 0.34) between the minor A allele of the rs7749323 variant and brucellosis in our cohort, indicating that this variant plays a protective role (Table 1). Next, we assessed the role of rs7749323 in 221 SAT-positive samples; interestingly, we observed a stronger association of rs7749323 (P ϭ 4.2 ϫ 10 Ϫ6 ; OR ϭ 0.23) with SAT positivity than with the occurrence of brucellosis, suggesting a more important role of the variant in SAT-positive samples. We further assessed the genetic association between variant rs7749323 and IgM-or IgG-positive patients with brucellosis. As shown in Table 1, we observed significant associations of the variant with both subgroups (P ϭ 0.0029 for IgG-positive patients and P ϭ 0.0021 for IgM-positive patients). To further evaluate the role of rs7749323 in brucellosis, we stratified cases based on clinical phenotypes. As shown in Table 1, we found that the rs7749323 polymorphism was associated with multiple clinical phenotypes of brucellosis.
rs7749323 is associated with reduced expression of the TNFAIP3 gene, decreased sensitivity to stimulations with B. abortus, and increased expression of NF-B1. To assess the role of the rs7749323 variant in regulating the expression of the TNFAIP3 gene, we isolated PBMCs from 39 healthy individuals with different genotypes at the rs7749323 variant (13 G/G, 13 G/A, and 13 A/A) and determined the mRNA expression levels of the TNFAIP3 gene using real-time quantitative PCR assays. In line with data from previously reported studies (25,27), individuals carrying the protective A allele of rs7749323 displayed significant reductions in mRNA levels of the TNFAIP3 gene (as shown in Fig. 3A). We further assessed the expression of NF-B1 in monocytes with different genotypes at the rs7749323 variant. Our data showed that individuals carrying the protective A allele demonstrated a significant increase in NF-B1 expression compared to that of carriers of the G allele (Fig. 3D). We further assessed mRNA expression levels of the TNFAIP3 gene in monocytes stimulated with lipopolysaccharide (LPS) and B. abortus. As shown in Fig. 3B, the mRNA levels of TNFAIP3 were significantly increased compared to those under unstimulated conditions. Interestingly, when samples were stratified by their genotypes at rs7749323, we observed that individuals carrying the G/G genotype are more sensitive to such stimulations than are individuals carrying the G/A genotype (Fig. 3C).
The brucellosis-associated protective A allele significantly upregulates the expression of multiple cytokines in monocytes. Our data showed that the protective A allele of the rs7749323 variant reduces the expression of TNFAIP3 and increases the expression of NF-B1. We further assessed expressions of multiple cytokines in monocytes with various genotypes at the rs7749323 variant. As shown in Fig. 4, we detected expressions of IL-1␤, IL-6, and IL-10 in monocytes stimulated with E. coli-derived LPS and B. abortus. When comparing G/G and G/A genotypes, we observed that monocytes with the G/A genotype showed significant increases in the expression levels of IL-1␤, IL-6, and IL-10. These data suggest that the brucellosis-associated protective A allele reduces the gene expression level of TNFAIP3, leading to the increased activity of NF-B signaling and increased expression levels of multiple cytokines in monocytes.

DISCUSSION
Increasing evidence shows that human diseases are influenced by genetic variants, but our understanding of the mechanisms that link a DNA sequence to a disease

Role of TNFAIP3 in Brucellosis
Journal of Clinical Microbiology phenotype is limited. Large numbers of disease-susceptible nucleotide variants fall within the noncoding region of the human genome, making it likely that they act by modifying regulatory sequences for genes. In the present study, we investigated the TTϾA functional variants, 42 kbp downstream of the TNFAIP3 gene, in our cohort of 1,209 healthy controls and 150 brucellosis patients. To our knowledge, this is the first evidence that SLE-associated risk variants play a protective role in an infectious disease, namely, brucellosis. Brucellosis is a well-known condition with importance in animal husbandry and in human disease in many regions of the world. Since Brucella species bacteria cause intracellular infection, the robust immune response against them is mediated by cell-mediated immunity, with a focus on macrophage activation (28). TNFAIP3 encodes A20, a ubiquitin-editing enzyme with a pivotal role in negatively regulating NF-B signaling downstream of multiple cell surface receptors (35). A20-deficient cells excrete high levels of proinflammatory cytokines and immediately activate lymphoid and myeloid cells. Genome-wide association studies (GWASs) of various autoimmune diseases have reported significant associations with sequence variants in the vicinity of the TNFAIP3 gene (36). We previously described a pair of functional variants (TTϾA) located in a distal enhancer of the TNFAIP3 gene that were associated with SLE (25,27). The enhancer element in which the TTϾA variants are found binds to the NF-B transcription factors p50, p65, and cREL and delivers them to the promoter region of the TNFAIP3 gene, activating NF-B-mediated transcription in multiple types of cells. The SLE-associated risk A allele disrupts the binding of NF-B transcription factors and significantly reduces the expression of TNFAIP3. In our previously reported study, we also showed that TALEN-mediated TTϾA enhancer knockout leads to reduced expression levels of TNFAIP3 and significant increases in NF-B signaling activity (26). These data have positioned the TTϾA variants as NF-B signaling modifiers, and they might play a role in infectious diseases. As shown in this study, the SLE-associated risk allele was significantly associated with a decreased risk of brucellosis.
Toll-like receptor (TLR) activation leads to intracellular signaling via MyD88 and IRAK-4, resulting in the activation of NF-B and mitogen-activated protein kinases (MAPKs) and the production of inflammatory cytokines (37). The TTϾA functional variants led to reduced expression levels of TNFAIP3 and enhanced NF-B signaling, which predisposes individuals to autoimmune diseases while protecting them from brucellosis. We therefore assessed the role of TTϾA variants in regulating the gene expressions of multiple cytokines. Our data showed that the increased activity of NF-B in monocytes results in significant increases in the expression levels of IL-1␤, IL-6, and IL-10.
IL-1␤ is produced by monocytes, tissue macrophages, and dendritic cells (38). Increased levels of production of IL-1␤ were detected in the sera of untreated brucellosis patients and in the supernatants of Brucella suis-infected macrophage cell cultures, suggesting an important role of IL-1␤ in the pathogenesis of brucellosis (39,40). After treatment, the level of IL-1␤ in the serum of brucellosis patients was significantly decreased and normalized (40). However, more studies are required to further understand the mechanism by which IL-1␤ influences brucellosis. Although the function of IL-6 in brucellosis is not clear, previous studies showed a significant association between an IL-6 polymorphism and brucellosis patients in Turkey (41,42). Further evaluation of both the role of this IL-6 genetic variant and the function of IL-6 in brucellosis is required. IL-10 is a cytokine synthesis inhibitory factor because of its negative regulatory effect on cytokines, specifically IL-2 and gamma interferon (IFN-␥). In a murine model, IL-10 has been shown to influence the development of brucellosis by downregulating the response of CD4 ϩ T helper cells and the secretion of IFN-␥ (43,44). On the other hand, IL-10 also promotes B cell antibody production. Furthermore, a genetic variant of the IL-10 gene has been reported to influence susceptibility to brucellosis (41,42,45,46).
In addition to the TTϾA polymorphic dinucleotide, the functional TϾG (p.Phe127Cys) variant rs2230926 at exon 3 of TNFAIP3 has been reported in different populations (47).
Follow-up studies showed that the coding variant rs2230926 reduces the activity of A20, resulting in decreased NF-B signaling (47). Further study of the role of this TNFAIP3 gene coding variant in brucellosis is required for a better understanding of how functional genetic variants influence susceptibility to brucellosis.
In summary, we demonstrated for the first time that autoimmune disease-associated risk variants (TTϾA) in the TNFAIP3 gene play a protective role in brucellosis. The TTϾA variants are associated with reduced expression levels of TNFAIP3 and increased NF-B activity. Functional studies revealed that individuals carrying the protective A allele display increased expression levels of multiple inflammatory cytokines.

ACKNOWLEDGMENTS
We thank all individuals, including the brucellosis patients and the controls, who participated in this study. We are grateful to the research assistants, coordinators, and physicians who helped in the recruitment of participants.
L.L., W.B., and X.L. performed the genomic DNA isolation and SNP genotyping; L.L., H.S., and H.L. performed the functional studies; W.B., Y.W., K.Z., and W.G. participated in obtaining the clinical samples and in data collection; S.W. and K.T. jointly directed this project; S.W. wrote the manuscript, and all authors participated in proofreading.
We declare no competing financial interests. This work was supported by the National Natural Science Foundation of China (grants 81373143 and 81571535 to Z.T.). The funders had no role in the study design, data collection and analysis, decision to publish, or preparation of the paper.