Editorial: HLA–B27, Cytokines, and Spondyloarthritis: Noncanonical Functions of a Curious Class I Major Histocompatibility Complex Gene

Authors

  • Eric Gracey,

    1. Toronto Western Research Institute, University Health Network, and University of Toronto, Toronto, Ontario, Canada
    Search for more papers by this author
  • Nigil Haroon,

    1. Toronto Western Research Institute, University Health Network, and University of Toronto, Toronto, Ontario, Canada
    Search for more papers by this author
  • Robert D. Inman

    Corresponding author
    1. Toronto Western Research Institute, University Health Network, and University of Toronto, Toronto, Ontario, Canada
    • Department of Medicine, Toronto Western Hospital, 399 Bathurst Street, 1E-423, Toronto, Ontario M5T 2SB, Canada. E-mail: Robert.Inman@uhn.ca

    Search for more papers by this author

Despite the prevalence and health impact of spondyloarthritis (SpA), the fundamental pathogenesis of this disease continues to be an enigma. Insights into the cell biology of SpA have been hampered by the relative inaccessibility of the target sites (the spine and sacroiliac joints). This has led to an increased emphasis on imaging, particularly magnetic resonance imaging, as a surrogate marker of inflammation (both as a diagnostic method in the assessment of back pain and as an outcome measure in clinical trials). Because of reliance on axial radiographic changes, diagnosis of SpA may be delayed until years after the onset of symptoms. It is clearly beneficial to identify early immunologic markers of disease prior to the development of target organ damage.

Studies of inflammatory lesions in the spine have proved to be challenging, since biopsy of lesions in the spine is technically difficult and rarely clinically indicated. Patients undergoing spinal osteotomies, usually for postural correction, can serve as an informative source of tissue. These studies have demonstrated that local inflammation is characterized by persistence of CD4+ T cells, monocytes, and B cells, with local production of tumor necrosis factor α (TNFα) and interferon-γ (IFNγ) ([1]). Recent studies have implicated proinflammatory CD4+ Th17 cells in SpA, since those cells appear elevated in patient blood; however, the exact role of CD4+ Th17 cells remains unknown because levels of Th17 do not correlate with standard measurements of disease progression. Aside from the importance of TNFα, the cytokine correlates of cellular infiltrates that are evident in sacroiliac joint biopsy samples have been less well characterized and thus the quest for a distinctive cytokine signature has continued. This search has been energized by the elusive nature of the fundamental immune basis of the chronic inflammation. If there is an autoimmune component in the pathogenesis of SpA, it has been difficult to demonstrate. Classically, SpA has been characterized as seronegative because of the lack of autoantibodies, in contrast to seropositive diseases such as rheumatoid arthritis. The recent finding of autoantibodies and circulating immune complexes in SpA ([2]) indicates that these nosologic distinctions will have to be revised.

In whatever way the immune response underlies SpA, the class I major histocompatibility complex (MHC) must play a central role. This concept is based on the strength of the association of HLA–B27 with SpA, which remains among the strongest links that have been established between any gene and a human disease. The mechanism by which B27 confers susceptibility to SpA continues to be an area of intense research, even now, on the 40th anniversary of the discovery of the B27–SpA association ([3, 4]). Herein lies a central paradox of SpA: how does a class I MHC lead to an aberrant class II MHC–dependent CD4+ T cell response? Clearly, a paradigm shift is required to demonstrate nonclassic biologic activity of B27. Provocative hypotheses have been proposed to address this paradox, including that an arthritogenic peptide is uniquely presented by B27, that an unfolded protein response is induced by misfolded B27 in the endoplasmic reticulum (ER), and that aberrant class I MHC dimer expression leads to distinctive interaction with natural killer cells. Recent studies have linked genetic variants of additional aminopeptidases to SpA ([5]), thus bringing the peptide processing and presentation pathway into the equation, but the question remains unanswered. In this regard, the development of the B27-transgenic rat has provided a much-needed experimental system to rigorously test these mechanistic theories ([6]).

B27-transgenic rats recapitulate a number of the clinical features of SpA, including inflammation in the joints, gut, skin, and eyes. Importantly, B27 appears to impart a T cell polarization bias on antigen-presenting cells (APCs), with a skew toward the development of Th17 cells. Furthermore, CD8+ T cells do not appear to play a major role in disease progression in B27 rats ([7]). Exactly how B27 affects the phenotype of APCs remains to be elucidated; however, it has been demonstrated that B27-positive APCs produce more interleukin-23 (IL-23) than controls ([8, 9]). These results were replicated in human studies ([10]); however, subsequent work with human APCs has contested this observation ([11]).

In this issue of Arthritis & Rheumatology, Fert et al ([12]) shed light on the intrinsic effects of B27 by demonstrating that this class I MHC allele imparts defects in the IFNγ signaling pathway within rat CD103+CD4+ dendritic cells (DCs). This subset of DCs is reported to produce less IFNα than plasmacytoid DCs, as well as no TNFα, but it can induce Th1 and Th2 polarization ([13]). Fert and colleagues hypothesize that this altered IFNγ signature, which is required for the priming of Th1 cells, may inadvertently result in a Th17 skew. Previous work identified a similar distinct IFNγ signature in human macrophages ([14]), but the current analysis was restricted to DCs.

Some key insights from this study may have arisen unexpectedly. In initial DC transcriptome experiments, the authors used impure DC isolates. This resulted in the B27-related up-regulation of genes typical of non-DC innate immune cells, such as macrophages, neutrophils, and mast cells. Although the authors validated these as genes from contaminating cells using ultrapure DC isolates, this raises the intriguing question of whether the noncanonical effects of B27 could be impacting a wide range of innate immune cells. Indeed, one of these contaminating genes, the macrophage marker CD163, is uniquely up-regulated in clinical SpA synovium and gut tissues, and is more highly expressed in B27-positive patients than B27-negative patients ([15, 16]). Another element not addressed in the study was the lack of a B27-dependent effect on IL-23 expression, which is not surprising since CD103+CD4+ DCs produce very little of the IL-23p19 transcript ([13]). This suggests that this subset of APCs may not be responsible for the Th17 skewing observed in both rat and human SpA.

Despite these unanswered questions, this research has provided some clarity with regard to recent studies of B27 and IFN biology. Previous studies suggested that a negative IFN signature was present in clinical SpA macrophages due to a reduced production of IFNα ([14]). Upon stimulation with exogenous interferon, the gene signature could be reversed. Thus the possibility of an inherent defect in IFN gene regulation in SpA was suggested. The current study demonstrated that B27 itself can mediate this change. Based on what is known about the function of B27 and current proposed mechanisms underlying SpA, there is no clear explanation as to how this might be a direct effect. The unfolded protein response is associated with increased rather than decreased IFN production. Furthermore, in the (21-3 × 283-2)F1 rat model there was minimal unfolded protein response up-regulation and yet the negative interferon signature was still observed. Finally, a clear link between B27 and suppressor of cytokine signaling 3 (SOCS-3) expression cannot be deduced at this time, and SOCS-3 suppresses TNFα and IL-6 in addition to IFNγ. The authors address this conundrum by proposing aberrant reverse signaling by B27, with further experiments proposed to address this hypothesis.

From a broader perspective, this research importantly sets the stage to address some fundamental unanswered questions in our understanding of SpA. Unlike most autoimmune rheumatic diseases, SpA affects males more than females. In the study, Fert et al examined only male SpA-prone rats. Sex is known to influence T cell skewing, with elevation of Th17 cells in males as compared to females ([17]). This raises the question of whether female rats also demonstrate the reverse IFNγ signature, or if they are protected against it. Furthermore, subclinical gut inflammation is believed to commonly occur in SpA, and a germ-free environment protects against disease in B27-transgenic rats ([18]). Since CD103 is a well-defined gut homing marker, it will be important to see if there is any link between aberrant splenic CD103+CD4+ DCs and gut inflammation. Do these DCs originate from the gut as suggested by previous studies ([9]), and can interruption of their trafficking to non-gut tissue prevent disease? Finally, environmental factors such as microbiota composition and obesity are thought to influence SpA in humans and rodents ([19, 20]), and these factors are easily manipulated in animal models of disease. Can the B27-induced aberrant IFNγ signal be modulated with a high-fat diet or with antibiotics?

The availability of B27 and HLA–B7–transgenic rats also presents a unique research opportunity to advance our understanding of how these alleles interact when codominantly expressed. It is now recognized that B7 confers protection against SpA ([21]). How B7 is protective is unknown, but immunodominance, the phenomenon of peptide competition by coexpressed MHCs, may play a critical role. A recent study demonstrated diminished capacity of B27 to present viral peptides when coexpressed with B7 ([22]).

In conclusion, Fert et al further validate the B27-transgenic rat as an informative model of SpA. Continued efforts should be directed at utilizing this model to further investigate immune mechanisms leading to SpA and elucidate environmental factors that exacerbate these mechanisms. With respect to understanding and treating SpA, these studies raise important questions about aberrant cytokine signatures that might pinpoint key events in the pathogenesis of the disease. Such insights always raise the possibility of novel therapeutic interventions, as was the case when an increased TNFα signal was detected in sacroiliac joint biopsy samples. Developing new approaches to correct the aberrant IFNγ signature may be close at hand. It would be ironic if, in the era when Red Sox fans saw a way to reverse the curse at Fenway Park, novel immune interventions might find a way to curse the reverse.

AUTHOR CONTRIBUTIONS

All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published.

Ancillary