Journal of Molecular Medicine

, Volume 87, Issue 3, pp 229–234

Association of the co-stimulator OX40L with systemic lupus erythematosus


  • Harinder Manku
    • Molecular Genetics and RheumatologyImperial College London
  • Deborah S. Cunninghame Graham
    • Molecular Genetics and RheumatologyImperial College London
    • Molecular Genetics and RheumatologyImperial College London
    • Rheumatology Section, Division of Medicine, Hammersmith CampusImperial College

DOI: 10.1007/s00109-008-0431-2

Cite this article as:
Manku, H., Graham, D.S.C. & Vyse, T.J. J Mol Med (2009) 87: 229. doi:10.1007/s00109-008-0431-2


The archetypal systemic autoimmune disease systemic lupus erythematosus (SLE) has incompletely understood pathogenesis, although evidence suggests a strong genetic component. Unlike organ-specific autoimmune diseases such as type 1 diabetes, the genetics of lupus are not as dominated by the effect of a single locus. Undoubtedly, the major histocompatibility complex is the greatest and most consistent genetic risk factor in SLE susceptibility; however, recent candidate gene and whole genome association (WGA) studies have identified several other genes that are likely to advance our understanding of this complex disease. One of these, the TNF superfamily member OX40L, interacts with its unique receptor OX40, to maintain T cell memory by providing a late-stage co-stimulatory signal to sustain the survival of activated T cells. The precise immunological consequences are yet to be determined; however, signalling through OX40-OX40L is bidirectional and the reverse signalling pathway via OX40L may quantitatively enhance B cell proliferation to augment the B cell hyperactivity found in SLE. Like OX40L, several genes recently identified in WGA studies are components of B cell pathways. Collectively, these genes will help us to unravel the mechanisms by which aberrant B cell signalling results in lupus pathogenesis.


Lupus erythematosus SLEGeneticsAutoimmune diseaseMHC whole genome association WGA

Genetics of OX40L

OX40L (chromosome 1q25) is located within an interval showing genetic linkage to systemic lupus erythematosus (SLE). The genetic contribution of this gene to SLE was determined by genotyping a dense map of single nucleotide polymorphism (SNPs) across the entire gene and its flanking sequence, stretching 90 kb upstream of the transcription start site to a point 24 kb downstream of the poly A tail [1]. This enabled definition of the 5′ and 3′ limits of haplotype blocks from 80 kb upstream to 2 kb downstream of the gene. We found evidence that SNPs in the 5′ region of OX40L were associated with SLE in two family-based studies (Pp = <1 × 10−5, OR = 1.61), and these findings were then further replicated in two case–control studies (Pu = 7 × 10−3, combined). None of the variants analysed in OX40 reached significance. The associated OX40L SNPs form a single GCTAATCATTTGA haplotype that is overtransmitted in lupus families, and both populations also contain an undertransmitted ACAGTCAAGCCC haplotype [1]. Figure 1 shows risk and protective OX40L haplotypes superimposed over the LD structure of the genomic region around the gene, as constructed from CEU data exported from the HAPMAP database.
Fig. 1

a Pattern of LD in CEU HAPMAP samples. b Human OX40L (accession number AL022310.1) consisting of three exons. The translated exons are illustrated as grey boxes and the 5′ and 3′ UTRs as black boxes. The four potential poly A signals in the 3′ UTR are denoted by asterisks. In the diagram, the SNPs used in the analysis have been re-coded as numbers from 1 to 36: SNP 1 (rs10798267), SNP 2 (rs12118748), SNP 3 rs4916318, SNP 4 (rs10912580), SNP 5 (rs844665), SNP 6 (rs844654), SNP 7 (rs844648), SNP 8 (rs2795288), SNP 9 (rs12039904), SNP 10 (rs844644), SNP 11 (rs844643), SNP 12 (rs2205960), SNP 13 (rs1234317), SNP 14 (rs3861953), SNP 15 (rs7535152), SNP 16 (rs1234315), SNP 17 (rs1234314), SNP 18 (rs3850641), SNP 25 (rs7518045), SNP 26 (rs6661173), SNP 27 (rs4113832), SNP 28 (rs7513384), SNP 29 (rs3861950), SNP 30 (171420977C<A), SNP 31 (rs7514229) and SNP 33 (rs3900307). The markers which failed quality control are place-marked in the sequence by grey lines. SNPs 4 (rs10912580), 9 (rs12039904), 12 (rs2205960) and 13 (rs1234317) are shown in red because they have overtransmitted minor alleles which tag overtransmitted haplotypes. c The haplotype block structure across OX40L constructed from 413 European Caucasian parent-proband trios in the UK study cohort and in 262 US Minnesota parent-proband trios. There are three haplotype blocks across the gene. The haplotypes are numbered on the left of each haplotype in brackets from 1 to 13, with the haplotype frequencies shown to the right of each haplotype. Only haplotypes with a frequency of greater than 2.5% are shown. The SNP numbers across the top of the haplotypes correspond to those in the gene diagram. The red-boxed haplotypes are those showing overtransmission in both populations

Genes closest to OX40L in the human genome are Pyridoxine 6 (PRDX6), which is 270 kb upstream and GITRL (TNFSF18), a 9.6-kb TNF superfamily member which is 132.8 kb downstream. Since OX40L, PRDX6 and GITRL are in separate LD blocks with no extended long-range LD between them, we are convinced that any observed associations within the gene are discrete effects limited to OX40L. GITRL is also an activation-inducible TNF superfamily member involved in activated T cell interactions; we precluded the possibility of long-range LD by association analysis. Several SNPs were typed across GITRL in our UK lupus cohort and none of these reached significance. In an earlier study relating OX40L genotypic variation to SLE, Tokunaga and colleagues found no association between three newly defined OX40L SNPs and SLE [2]. These non-coding SNPs (in the promoter region, 5′ untranslated region and intron 2) were not identified after sequencing our screening cohort. Within the last 12 months, an independent study in a collection of 251 childhood-onset SLE families reported association for SNP rs1234314 (P = 1.14 × 10−4). The G allele of this variant was carried on the risk haplotype from our UK SLE study [3].

The overtransmitted and undertransmitted haplotypes are both tagged by unique SNPs, allowing us to pursue the genetic association at a functional level. We examined the gene product on activated Epstein–Barr virus (EBV) lymphoblastoid cell lines (LCLs) and on activated peripheral blood lymphocytes (PBLs). There was increased expression of cell surface OX40L in individuals homozygous for the risk haplotype compared to homozygotes for what appeared to be a protective haplotype in both LCLs and PBLs, with significantly more OX40L expression in patient PBLs [1]. Quantification of OX40L mRNA levels by quantitative real-time polymerase chain reaction in LCLs showed that cells homozygous for the risk haplotype had a 6.7-fold increase in the level of OX40L transcript compared to LCL homozygous for the ‘protective’ haplotype [1]. Signalling by OX40 through OX40L is likely to be enhanced in homozygotes for the risk haplotype. Increased levels of cell surface OX40L may augment the B cell differentiation and proliferation to cause disease pathology in SLE [4]. Alternatively, homozygotes for the risk haplotype may quantitatively augment APC–T cell interactions to enhance cellular hyperactivity in this disease.

OX40/OX40L as SLE biomarkers

Several TNF superfamily co-stimulatory markers have been measured as biomarkers in SLE, and the percentage of CD4+OX40L+ T cells observed in SLE patients is significantly higher compared to healthy controls. Specifically, an increased number of these cells are found in the peripheral blood of SLE patients with lupus nephritis, although disease activity in all SLE patients correlates significantly with OX40 expression [5]. An interesting observation made by Aten et al. [6] is that OX40L appears to be abundantly present in the glomerular capillary wall in biopsy specimens of SLE nephritis.

OX40L protein structure and expression

OX40L exists as an atypical tumour necrosis factor superfamily (TNFSF) homotrimer. Each compact OX40L protomer has no discernable proteolytic site and is therefore predominantly found in a membrane-bound form [7]. In contrast to the distinctive structure of OX40L, OX40 is a relatively conventional multidomain TNFSF [7]. The ligand is expressed on the surface of professional antigen-presenting cells including activated B cells [8] and dendritic cells [9] (DCs) 1–3 days after initial antigen encounter. OX40L is also expressed on memory B cells [10], vascular endothelial cells [11], activated CD8+ intraepithelial T cells [12] and DC-like accessory cells which are unable to process antigen but which are CD3−CD4+CD30L+OX40L+ [13, 14]. Its cognate receptor, OX40 (TNFRSF4) is primarily expressed on activated CD4+ T cells [15]. SLE is characterised by inappropriate activation of peripheral memory B lymphocytes; therefore, aberrant OX40L expression on these cells is likely to be important in disease pathogenesis. A feature of human and murine OX40L is that they are less similar to each other than many other TNFSF orthologs, with only ∼40% sequence homology, yet they share the same distinctive features [7]. Unusually, mOX40L will bind hOX40. This makes the mouse a useful model with the aim of further dissecting the role of OX40L in human SLE.

OX40L and immune-mediated disease

A large body of evidence from animal models of disease suggest that OX40L plays a role in T-cell-specific autoimmunity: Expression of this ligand is associated with paralytic episodes of EAE [16] and selective depletion of myelin-reactive T cells with anti-OX40 mAb ameliorates this disease [17]. The administration of an anti-OX40L mAb into DBA/1 mice dramatically improves collagen-induced arthritis [18], and a SCID model of Th1-mediated colitis is characterised by an overwhelming increase in OX40L+DCs compared to normal Balb/c mice [19]. In diabetes, crossing an OX40L KO mutation onto the NOD or B6-H2g7 background results in complete protection in NOD mice and a strong reduction in autoreactive T cells in BDC2.5/B6g7 mice [20]. OX40L transgenic mice on a C57BL/6 background spontaneously develop inflammatory bowel disease, and this is accompanied with a significant production of anti-DNA antibodies in the sera [21].

Human OX40L appears to have multiple immune system effects: Seshasayee et al. [22] have provided in vivo evidence that OX40L causes CD4+ T cells to deviate towards a Th2 phenotype. By using a fully human monoclonal antibody against hOX40L on TSLP-activated DCs, in a nonhuman primate model of atopic inflammation, they found the blocking antibody inhibited antigen-driven Th2 inflammation. OX40L inhibits the generation and suppressive effects of type 1 regulatory (Tr1) cells. The generation of Tr1 cells, induced by dexamethasone and vitamin D3 from naïve and memory CD4+ T cells, is completely inhibited by the ligand [23]. Tr1 cells inhibit the development of autoimmunity; however, few studies have directly investigated the role of OX40L in human autoimmune disease.

OX40L and other complex genetic diseases

OX40L influences atherosclerosis susceptibility and is expressed in all major cell types in atherosclerotic lesions, suggesting that it has a pivotal role in the inflammatory response in this disease [24]. The minor G allele of the upstream SNP rs3850641 is overtransmitted in two independent cohorts with myocardial infarction (MI). Individuals carrying rs3850641-G have an increased risk of MI compared to individuals carrying rs3850641-A (P = 0.05, OR = 1.4) [24]. It is likely that the risk haplotype for MI and in the UK SLE cohorts are similar because the G allele of rs3850461 is carried on the risk haplotype in both our Minnesota and UK SLE populations and this SNP is in perfect linkage with other risk variants [1]. The increased incidence of cardiovascular disease in individuals affected by lupus is likely to be due to the presence of the OX40L risk haplotype and the resultant increased expression of OX40L.

OX40-OX40L signalling

In the early stages of T cell stimulation, the co-stimulatory molecule CD28 is required for T cell proliferation and survival through recognition of CD80/CD86 molecules expressed by antigen-presenting cells. TNF receptor family members such as 4-IBB, CD27, CD30 and OX40 represent co-stimulatory molecules required by T cells at later stages of activation to complete their differentiation and exert more specific immune functions [25]. A strong co-stimulatory signal through OX40 at the same time as TCR engagement allows T cell activation and survival at effector and memory stages of the response. This is followed by recruitment of TNF receptor-associated factor 2 (TRAF2) and, via protein kinase B activation, sustained Survivin expression and then T cell division and clonal expansion [26].

OX40-OX40L is bidirectional, and reverse signalling via OX40L enhances B cell proliferation and differentiation [25]. OX40L hyperexpression could therefore augment the B cell hyperactivity found in SLE (Fig. 2). It is clear that OX40-OX40L co-stimulation is essential for high levels of nuclear factor of activated T cells c1 to accumulate in T cell nuclei [4]; however, the precise sequence of signalling events has yet to be determined. The immunological consequences of disruption of signalling through OX40-OX40L may preclude SLE pathogenesis in either direction, as increased numbers of both CD4+OX40L+ T cells and CD86+OX40L+ B cells are found in lupus patients (unpublished data).
Fig. 2

The complex sequence of events that lead to development of SLE are emphasised by genes identified in recent candidate gene and whole genome association scans. Many of these SLE susceptibility genes, including OX40L (TNFSF4) are involved in B cell and dendritic cell activation, and the complexity of the diagram reflects the multitude of immunological abnormalities found in lupus. Prolonged OX40-OX40L interaction can result in amplification of activated autoreactive B and/or T lymphocytes, as signalling is bidirectional. The resulting autoantibodies and immune complexes can mediate pathology in multiple systems in lupus individuals. An additional potential mechanism linking OX40L to SLE pathogenesis is via its effect on T regulatory cells. OX40L provides a potent signal to negatively regulate the generation and function of IL-10-producing T-regulatory (Tr1) cells which play a critical role in maintaining peripheral tolerance. Another recently identified SLE susceptibility gene, TNFAIP3, functions downstream of TNFR, including OX40, as a negative regulator of NF-κB signalling. Although OX40L and TNFAIP3 fall within the same pathway, it is apparent that other signalling pathways are also affected in lupus: Ligation of FCyRIIb inhibits B cell activation to upregulate apoptosis; SLE can occur when B cells lack this mechanism. Other B cell signalling events may also be disrupted to contribute to lupus pathogenesis, as expected of this heterogenous disease, and functional variants in both B lymphoid tyrosine kinase (BLK) and the tyrosine kinase LYN, which transduce signals downstream, can predispose individuals to lupus

Novel targets in B cell signalling from candidate gene and GWA studies

Within the last year, genome-wide association (GWA) studies and candidate-gene-based association studies have identified a series of novel non-major histocompatibility complex SLE susceptibility genes, including OX40L. Each gene produces only a moderate increased risk of disease (odds ratio < 1.8), so it is likely that interactions between multiple associated genes result in lupus susceptibility. OX40L is expressed on the surface of B lymphocytes; multiple abnormalities in B cells are thought to initiate SLE pathogenesis. Recently published GWA studies identify ways in which disruption of B cell signalling cascades preclude the development of SLE: Several new targets identified are involved in these signalling pathways. The tyrosine kinase LYN and the B lymphoid tyrosine kinase (BLK) and are both components of B cell signaling, and the B cell scaffold protein with ankyrin repeats (BANK1) promotes LYN-mediated tyrosine phosphorylation [27]. The low affinity IgG receptor FCyRIIb (CD32) inhibits B cell signalling, and functional variation in this gene has also been described to contribute to the risk of SLE [28].

LYN is a member of the src family of tyrosine kinases, and the precise role of this gene in SLE remains uncertain, although it may have both stimulatory and inhibitory functions in B cell signalling. The stimulatory role is dependent on interaction of the SH2 domain with ITAM molecules including BCR-associated Igα/β, whereas the inhibitory role of LYN depends on the phosphorylation of ITIM-containing receptors such as CD22 and FCγRIIb [29]. A functional genetic variant in FCGR2b is also a susceptibility factor in SLE: The FCGR2B-T232 allele is a non-synonymous variant in exon 6 which promotes a decrease in calcium dephosphorylation to reduce anti-Fas antibody-induced apoptosis; this is associated with SLE susceptibility in Asians [30]. Co-ligation of FCyRIIb and the B cell receptor inhibits B cell activation via signalling events to upregulate anti-Fas antibody-induced B cell apoptosis. This cell death control mechanism is important in maintaining peripheral tolerance, and it has been proposed that SLE can occur when B cells lack this mechanism [31, 32].

BLK is a less well-defined member of the src family of tyrosine kinases. Engagement of the preBCR/BCR complex results in rapid activation of BLK to transduce the signal downstream [33]. GWA and case–control replication in Swedish subjects found a BLK-containing locus which contributes to the risk of SLE [34]. Enrichment of the A allele of rs13277113 in both analyses was found in lupus individuals, and this allele was associated with lower levels of BLK mRNA from EBV-transformed cell lines of healthy HAPMAP subjects. The BLK protein may influence tolerance levels in B cells; low levels may predispose individuals to SLE [34].

BANK1 was identified as a lupus susceptibility gene by GWA in a Scandinavian cohort, followed by replication in four independent case–control sets [27]. The gene promotes association with Lyn and IP3R, causing the release of calcium ions from the endoplasmic reticulum on BCR ligation. The variants in BANK1 associated with SLE could contribute to prolonged BCR signalling and the increased B cell activity characteristic of this disease [27]. A recent whole genome association (WGA) scan has identified variants in or near theTNFAIP3 gene (also known as A20) on chromosome 6q23 to be associated with SLE [35]. TNFAIP3 functions as a negative regulator of NF-κB signalling through ubiquitin modification of adaptor proteins downstream of TNFR, including OX40. The gene plays a role in B cell activation, signaling and apoptosis and is required to prevent inflammation. Reduced TNFAIP3 activity may result in diminished anti-inflammatory events resulting in the excessive cellular inflammation characteristic of several autoimmune diseases, including SLE. Recently, Graham et al. have found association of the rs5029939 variant in TNFAIP3 with SLE in the 431 SLE cases and 2155 controls which comprise the GWAS dataset. An additional SNP, rs6930330, associated with rheumatoid arthritis (RA) and located 185 Kb upstream of TNFAIP3, showed modest association in this dataset (P=0.01), although the association was strengthened on replication in a family-based study (P=8.92 x 10−5) [35]. In an independent analysis of this gene, Musone et al. found association of a perfect substitute for another RA-associated SNP, rs13192841, with SLE, amongst individuals of European ancestry [37]; evidence from both studies points to TNFAIP3 as a common susceptibility gene in systemic autoimmunity. In addition to the rs13192841 association, Musone et al. found independent association of a coding variant, rs2230926, which results in a non-synonymous amino acid substitution, reducing the anti-inflammatory activity of TNFAIP3 and permitting an excessive response to TNF.

Over the past year, there has been a massive increase in the identification of new SLE susceptibility genes. Large-scale candidate gene association and WGA studies have been a powerful approach for detecting genetic effects underlying susceptibility to lupus, as the effects are moderate relative to those found in organ-specific autoimmune diseases such as type 1 diabetes. Many of the newly identified SLE genes, including OX40L, are expressed on B cells, and it is increasingly apparent that multiple genetic abnormalities in these cells, especially in memory B cells, are implicated in SLE pathogenesis. Although some lupus susceptibility genes fall into clusters within the same signalling pathway, others affect very different aspects of B cell function, and this is in keeping with the heterogeneous nature of this complex disease. We have progressed our understanding of SLE pathogenesis measurably in the recent past; however, it is still relatively incomplete. It is anticipated with the advent of the one million SNP array that we will fine map SLE susceptibility loci, including the OX40L promoter region, allowing better definition of causal genetic variants in lupus.

Copyright information

© Springer-Verlag 2008