Thomas Tedder, B cell depletion, clean mice and dirty people
by Adam Burrack, PhD
A majority of the blogs in this series have focused on T cell biology and methods to inhibit T cell-mediated destruction of insulin-producing beta cells in the pancreas. Today I will take a slight tangent into B cell biology. B cells are a key component of the adaptive immune system. B cells have a well-known role in the immune response in the production of antibodies against foreign proteins, which is a key component to the adaptive immune response. In fact, most effective childhood vaccines work through promoting an antibody response, not a T cell response. In addition to their role in producing pathogen-specific antibodies, B cells also present (or “show”) pathogen-derived (or self-derived) peptides to T cells, which promote T cell responses (or T cell tolerance, depending on the situation). Through their B cell receptor – which is peptide-specific and generated in the same random genetic process as the T cell receptor – B cells can be much more specific in the peptides they show to T cells than can the two other major types of antigen-presenting cells (macrophages and dendritic cells). Through this specificity of peptide presentation to T cells, B cells can strongly influence T cell responses. This influence can have both beneficial effects (for example in potential treatments for cancer) and deleterious effects, as in the case of T cell-mediated autoimmune diseases like type 1 diabetes (T1D). Because autoantibodies against insulin and other beta cell-derived proteins do not transfer disease (in the mouse model of T1D), and are therefore not considered directly pathogenic, research interest in B cells in T1D has focused instead on their role as antigen-presenting cells leading to disease onset.
Thomas Tedder is a Professor of Immunology at Duke University. Early in his career Dr Tedder was one of the first researchers to characterize the structure and function of the CD20 molecule, a key protein on the surface of B cells that plays an important role in B cell activation and function. CD20 is expressed on the surface of B cells throughout development, but unfortunately is not expressed on plasma cells, which are long-lived memory B cells which produce antibody. Therefore, in order to influence T1D, target therapy toward the CD20 molecule must precede disease development, before memory B cells could develop into plasma cells secreting autoantibodies.
Dr Tedder was among the first to show that B cell depletion in NOD mice, targeting the CD20 molecule, could prevent type 1 diabetes. Another group pursuing this line of inquiry at approximately the same time were Mark Schlomchik and colleagues at Yale. Importantly, mice were treated prior to development of disease, which these authors concluded prevented disease onset and the Schlomchik group concluded could both prevent and reverse new-onset diabetes. In addition, Dr. Tedder’s group demonstrated that removing B cells as an antigen-presenting cell inhibited the activation of autoreactive CD4 T cells in vivo, perturbing another immune system pathway critical to the development of T1D. Taking these observations together, CD20-targeted B cell depletion appeared to be a very promising clinical treatment for T1D.
Therefore, CD20-focused depletion of B cells in at-risk human patients was rapidly applied in the clinic. The New England Journal of Medicine article describing this clinical trial, was, however, underwhelming. Over 12 months of treatment, patients showed higher C-peptide levels, lower IgM (antibody) levels, and improved HbA1c levels, along with lower CD19+ B cell totals out to 9 months after the beginning of treatment. However, B cell depletion did not stop disease in human patients, and B cell numbers recovered. In short, B cell depletion in human subjects did not appear to have as broad or as beneficial effects in the human disease. In addition, people live much longer than mice, and the mice used in these single-center research studies were housed in specific-pathogen-free conditions, while the humans, of course, live in the real world and in the presence of many infections and their own underlying immune memory. Therefore, the long-term consequences of B cell depletion in childhood/adolescence are unclear, especially in terms of development of B cell memory against viral and bacterial challenge, and in response to vaccines against common pathogens. In hind-sight, a clinical protocol depleting/inhibiting B cells broadly may have represented more risk to patients than was appreciated at the outset.
In a follow-up study in NOD mice, a 4-center consortium including diabetes centers at Yale, La Jolla, the University of Florida, and the University of Colorado found no disease prevention benefit of CD20-based B cell depletion in NOD mice. What these researchers did, which was not done prior to the clinical trial, was apply clinical trial methods (same treatment in parallel across several research centers) to the NOD mouse, using CD20 treatments to deplete B cells. So, unfortunately, despite early success in the mouse model and positive early results in the human clinical trial, anti-CD20-based therapy to prevent or stop T1D onset may not be sufficient.
This negative result highlights an important issue in “translating” basic science findings to the clinic. Validating the efficacy and dosing regimen in multiple independent sites using the relevant mouse model before moving into human patients may represent a critical step when “translating” protocols from mouse to man, and is rarely performed. Many interesting research results are published using groups of 5-10 mice treated with a candidate therapy, at one institution. Taking results from this small scale to the large-scale clinical trial level is risky, at best. Performing the intermediate step of a multi-center mouse-based study would likely save time and effort in later clinical trials.
A recent review and opinion article co-authored by Kevin Herold explores reasons the NOD mouse is not a perfect model for human T1D and ways in which the mouse might be modified to better model the human disease. One potentially relevant difference between mouse and human is the gut-derived microbiome, or the genetic diversity of intestinal bacteria and how this diversity may differ between species. As previously described in this series, it appears that bacterial diversity in the intestine influences the immune system’s likelihood to develop autoimmunity along the gut tube. Another relevant, and related, difference between research mice and real-world human subjects is the relative absence of memory T and B cells (besides diabetes-causing autoreactive cells) in NOD mice and the presence of a high proportion of memory cells in humans. The presence of T cell and B cell memory for pathogens, potentially comprising >50% of total T cells in humans, may negatively influence the efficacy of therapies developed in “immune-naive” animals. Long story short, Dr Herold posits that therapies designed to inhibit “naïve” T cells in young mice may not be as effective when applied to “memory” T cells in adult humans. This is a current challenge in biomedical research and will require innovative thinking and new approaches to address.