An antigen-specific cell therapy for autoimmune disease avoids compromising immunity as a whole.
The field of oncology is abuzz over recent clinical trial results showing that patient T cells engineered in the laboratory are remarkably effective against a few intractable cancers. Chimeric antigen receptor (CAR) T cells express a recombinant receptor that binds a specific tumor antigen, inducing the cell to kill target tumor cells. Writing in Science, Ellebrecht et al.1 have now adapted the approach to autoimmune disease. They developed chimeric autoantibody receptor (CAAR) T cells as an antigen-specific therapy for the autoimmune disease pemphigus vulgaris and showed, both in vitro and in mice, the capacity of the cells to selectively eliminate B lymphocytes that produce autoantibodies to desmoglein (Dsg) 3, the pathogenic mediators of the disease. If these results can be translated to human autoimmune disease, this would represent a major leap toward achieving durable remissions or cures for certain severe and disabling conditions.
Autoimmune diseases can affect almost any organ, and stem from a breakdown in immune tolerance to host-derived or 'self' antigens. Their frequency has risen steadily over the past four decades in industrialized countries2, and together they represent the third leading cause of morbidity and mortality after cardiovascular disease and cancer. Although both autoreactive T and B lymphocytes contribute to some degree to the development of all autoimmune diseases, injury of a given tissue usually results from the predominant action of either one cell type or the other.
Current treatments for autoimmune diseases are based on anti-inflammatory and immunosuppressive agents—including engineered biologics; human or humanized monoclonal antibodies; and fusion proteins selective for certain immune cell subsets or signaling pathways3—but their effect is transient and not antigen-specific. Chronic administration of these agents leads to the common side effects of general immunosuppression, such as an increased incidence of infections.
Major efforts have been devoted to selectively targeting the autoantigen-specific response in various autoimmune diseases (multiple sclerosis, type 1 diabetes, and uveitis) by administering the autoantigen(s) through different routes (subcutaneous, oral, and parenteral). However, the efficacy seen in induced or spontaneous mouse models of autoimmunity has never been successfully translated to the clinic. Among biological agents, CD3 antibodies have shown particular promise for reversing established autoimmunity and durably restoring self-tolerance in animal models. Clinical development of this approach is still in progress, but present data suggest that combination therapies may be needed to achieve a sustained effect in humans3.
Ellebrecht et al.1 have approached the challenge of antigen-specific therapy by designing a modified form of a CAR that leads T cells to kill autoreactive B cells. A conventional CAR fusion protein consists of an extracellular antibody moiety specific to an antigen of interest linked by a transmembrane domain to an intracellular signaling domain that activates the T cell upon antigen binding. The CAAR design is different in that the extracellular domain includes the autoantigen itself, and the T cell is activated when the autoantigen binds cognate autoantibodies on the surface of pathogenic B cells. Ellebrecht et al.1 designed a CAAR in which Dsg3, a key autoantigen in pemphigus vulgaris, is fused to the signaling domains of the T-cell-activating molecules CD3ζ and CD137 (ref. 1) (Fig. 1).
Pemphigus vulgaris is a blistering dermatological disease that is mediated by autoantibodies against the keratinocyte adhesion protein Dsg3. Ellebrecht et al.1 provide robust evidence to support an in vivo effect in two well-designed experimental models. The first model used is the NSG (NOD-scid-gamma) immunodeficient mouse reconstituted with a mixture of bioluminescent human B-cell hybridomas producing anti-Dsg3 antibodies. Injection in these mice of Dsg3 CAAR T cells convincingly showed destruction of the pathogenic B cells, thus avoiding anti-Dsg3 autoantibody production and blistering of the oral mucosa1. In the second model, Dsg3 CAAR T cells were injected into mice bearing human skin xenografts, and no direct toxicity to keratinocytes was observed, supporting the safety of the approach.
In a careful fashion, the authors also demonstrated that anti-Dsg3 antibodies present in the patients' sera do not neutralize the effect of CAAR T cells, an intrinsic problem of targeting antibody-mediated autoimmunity. Finally, they showed that their approach kills B cells bearing antibodies that recognize different domains of Dsg3. Given that anti-Dsg3 autoantibody-producing cells in pemphigus patients are oligoclonal4, this provides further evidence that the approach could be a major step toward restoring self-tolerance in these patients.
The choice of pemphigus vulgaris for an initial application of CAAR T cells was a strategic one, as much is known about the molecular structure of the autoantigen and the antibody repertoire of pathogenic B lymphocytes4. Could CAAR T cells be beneficial in other autoantibody-mediated autoimmune diseases, as the authors suggest? The answer is a tentative yes, provided that some essential prerequisites are met.
First, the pathogenicity of autoantibodies in the disease should be clearly established, as is the case in pemphigus vulgaris. A simple and unambiguous way to approach this question is to consider only autoimmune diseases for which it is proven that autoantibodies can transfer the disease, whether in newborns upon placental transfer during pregnancy or in animal models. Second, the sequence and molecular structure of the autoantigen(s) must be clearly identified to allow engineering of a CAAR expressing the key epitopes recognized by patients' autoantibodies. Currently, only two important autoimmune diseases that fulfill these conditions come to mind.
Myasthenia gravis is a chronic disease that affects the neuromuscular junction. Pathogenic autoantibodies to epitopes of the major immunogenic region of the acetylcholine receptor have been described in 85% of myasthenia gravis patients with generalized muscle weakness and in 50% of those with only ocular involvement5,6. In a minority of patients, autoantibodies instead bind to muscle-specific kinase5. Plasmapheresis, a part of the treatment course in severe cases of myasthenia gravis, has provided clear evidence of the pathogenicity of autoantibodies in this disease.
The other important autoimmune disease that could greatly benefit from the CAAR technology is autoantibody-associated neonatal lupus syndrome (also termed congenital heart block), a passively acquired autoimmune condition. The disease affects the fetus in pregnant women with various autoimmune diseases, such as systemic lupus erythematosus and Sjögren syndrome. Maternal autoantibodies to the Ro/SSA autoantigen, including the unrelated Ro52 and Ro60 proteins7,8,9, cross the placenta and cause disease. Interestingly, in many cases the mother is unaffected by the antibodies as only the fetal heart expresses the target autoantigens7,8,9.
For the many other autoimmune diseases in which B cells contribute substantially, such as systemic lupus erythematosus, one may hope that significant progress in both defining autoantibody pathogenicity and characterizing the autoantigens will permit the development of new therapies based on CAARs.
But why not consider extending the strategy to T-cell-mediated autoimmune diseases? Let's consider a prototypic example, type 1 diabetes, which results from selective destruction of insulin-secreting beta-cells in the pancreas by autoreactive CD4+ and CD8+ T lymphocytes. The autoantigens recognized by the autoreactive T cells are well characterized (including insulin/proinsulin and glutamic acid decarboxylase). In addition, autoantibodies against these antigens are present, from early stages of the disease, and they can be used to identify patients with the beginnings of beta-cell destruction. Although the autoantibodies are not pathogenic, there is a wealth of data from experimental models showing that B cells have a key role in the disease as antigen-presenting cells10. Furthermore, anti-CD20 antibodies, which eliminate B cells, have a therapeutic effect in patients with recent-onset type 1 diabetes11.
The CAAR strategy could therefore be effective at early stages of the disease, when only a few anti-beta-cell autoantibody specificities are present and before full-blown antigen spreading occurs, which provokes massive beta-cell destruction. There is a chance that targeting naive B cells before they become mature B cells capable of potently presenting autoantigens to T cells could do the trick. Establishing peripheral tolerance with CAAR T cells while the autoreactive B cell response is in an early phase could provide durable prevention of B-cell autoantigen presentation. Thus, CAAR T cells that selectively target pathogenic autoantigen-presenting B cells could be advantageous over CD20-specific antibodies that deplete all B lymphocytes independently of their fine specificity.
We are at an important crossroads where advances in biotech give us a glimpse of the possibility of providing a real cure for autoimmune diseases in a not-too-distant future.
Ellebrecht, C.T. et al. Science 353, 179–184 (2016).
Bach, J.-F. N. Engl. J. Med. 347, 911–920 (2002).
Chatenoud, L. Curr. Opin. Pharmacol. 23, 92–97 (2015).
Hammers, C.M. et al. J. Invest. Dermatol. 135, 742–749 (2015).
Phillips, W.D. & Vincent, A. F1000 Res. 5, 1513 (2016).
Luo, J. et al. J. Neurosci. 29, 13898–13908 (2009).
Ambrosi, A., Sonesson, S.E. & Wahren-Herlenius, M. Exp. Cell Res. 325, 2–9 (2014).
Gleicher, N. & Elkayam, U. Autoimmun. Rev. 12, 1039–1045 (2013).
Skog, A., Lagnefeldt, L., Conner, P., Wahren-Herlenius, M. & Sonesson, S.E. Acta Obstet. Gynecol. Scand. 95, 98–105 (2016).
Serreze, D.V. et al. J. Exp. Med. 184, 2049–2053 (1996).
Pescovitz, M.D. et al. N. Engl. J. Med. 361, 2143–2152 (2009).
The author declares no competing financial interests.
About this article
Cite this article
Chatenoud, L. Precision medicine for autoimmune disease. Nat Biotechnol 34, 930–932 (2016). https://doi.org/10.1038/nbt.3670
Cellular & Molecular Immunology (2020)
Comparing convolutional neural networks and preprocessing techniques for HEp-2 cell classification in immunofluorescence images
Computers in Biology and Medicine (2020)
Chemical Reviews (2019)
Frontiers in Immunology (2019)