Autoantibodies against cytokines: back to human genetics

Anne Puel, Jean-Laurent Casanova

In this issue of Blood, Chi et al have discovered that the occurrence of autoantibodies against IFN-γ, which trigger mycobacterial diseases and mimic inborn errors of IFN-γ immunity, is genetically determined.1

It has long been known that infections can trigger autoimmunity. Paradoxically, autoimmunity has also recently been shown to precipitate infectious diseases. Indeed, autoantibodies against certain cytokines have been shown to underlie the pathogenesis of specific infectious diseases. This is best illustrated by autoantibodies against IFN-γ, which were first described in 2004-2005 in 5 adults with no significant medical history who developed disseminated disease caused by environmental and tuberculous mycobacteria.24 A 25-year-old woman from Thailand2 and a 47-year-old man from the Philippines3 died of mycobacterial disease. The other 3 patients, 2 British women aged 46 and 59 years, respectively, and a 32-year-old South African man, slowly improved after treatment with antibiotics and, in 1 case, recombinant IFN-γ.4 Neutralizing anti–IFN-γ autoantibodies have been reported now in approximately 130 individuals. Most patients initially presented as adults with disseminated environmental mycobacterial infections.25 Strikingly, these patients are late-onset clinical phenocopies of patients with Mendelian susceptibility to mycobacterial disease (MSMD) carrying inborn errors of IFN-γ immunity.6 This observation alone strongly suggested that autoantibodies against IFN-γ were the cause rather than a consequence of mycobacterial disease. Intriguingly, some patients with autoantibodies against IFN-γ also display other opportunistic infectious diseases, not all of which are seen in MSMD patients,5 suggesting that these anti–IFN-γ autoantibodies have a broader impact or that other pathogenic mechanisms, including the production of other autoantibodies, for example, are at work.

Three other types of pathogenic autoantibodies against cytokines described before autoantibodies against IFN-γ have not been found to underlie infectious diseases. Some patients with pure red-cell aplasia display autoantibodies against erythropoietin. Autoantibodies against granulocyte colony-stimulating factor (G-CSF) have been found in neutropenic individuals with Felty syndrome or systemic lupus erythematosus. Autoantibodies against granulocyte macrophage colony-stimulating factor (GM-CSF) were found in patients with pulmonary alveolar proteinosis,7 which may also be caused by mutations in the gene encoding the GM-CSF receptor. It was not until 2008 that autoantibodies against IL-6 were found in a single patient with severe staphylococcal skin infection and low serum C-reactive protein concentration.8 However, in the absence of known inborn errors of IL-6 or its receptor, it is difficult to determine whether these autoantibodies caused or resulted from staphylococcal disease. Finally, another type of anti-cytokine autoantibody was unambiguously shown to underlie infectious diseases in a study carried out in 2010, on patients with autosomal recessive autoimmune polyendocrinopathy syndrome type 1 (APS-1). These patients had previously been shown to display autoantibodies against IFN-α and IFN-ω, which were clinically silent, probably due to the existence of multiple antiviral IFNs. These patients were shown to display autoantibodies against IL-17A, IL-17F, and IL-22.9,10 The autoantibodies against IL-17A and IL-17F are probably responsible for chronic mucocutaneous candidiasis (CMC), the only infection seen in these patients, because other patients with inborn errors of IL-17 immunity display a pure phenotype of CMC.11 Overall, the occurrence of autoantibodies against 2 key T-cell cytokines, IFN-γ and IL-17, has been shown to underlie clinical phenocopies of the corresponding inborn errors, MSMD and CMC, respectively.

In APS-1 patients, the early production of autoantibodies against IL-17 and many other autoantigens reflects the disruption of tolerance due to biallelic mutations of AIRE. But what provokes the production of autoantibodies against IFN-γ in previously healthy adults in the first place? A genetic origin was first suggested by the observation that autoantibodies against IFN-γ were found mostly in individuals of East Asian descent.2,3,5 In the accompanying ground-breaking report, Chi and colleagues describe their study of 17 Chinese adults with environmental mycobacterial disease due to autoantibodies against IFN-γ.1 They elegantly provide the first molecular evidence for genetic control of anti–IFN-γ autoantibody roduction in humans. They found that 2 linked HLA-II alleles defining a haplotype—DRB1*16:02 and DQB1*05:02—were associated with disease, with an odds ratio of approximately 8.1 In other words, the relative risk of developing this illness was almost 10 times higher in individuals carrying at least 1 such haplotype. HLA genes are among the strongest known genetic factors underlying autoimmunity. In some populations, they can increase the relative risk from 10 to more than 150 times (eg, the HLAB27 alleles for ankylosing spondylitis). In this superb study, Chi et al bring the autoimmune phenocopy of inborn errors of IFN-γ back into the domain of human genetics. The penetrance of anti–IFN-γ autoantibody production in patients bearing this HLA haplotype is undoubtedly lower than that of anti–IL-17 antibody production in APS-1 patients, and the autoantibodies occur much later in life, but the authors nonetheless provide clear evidence of genetic determinism. Further investigations are now required to determine the mechanism by which either or both HLA class II alleles contribute to the production of autoantibodies against IFN-γ.


  • Conflict-of-interest disclosure: The authors declare no competing financial interests. ■