T follicular helper cells, T cell activation and function, Translational Research
by Adam Burrack, PhD
T cells of the immune system are a complicated beast. To study the events leading to the destruction of insulin-producing beta cells and the development of type 1 diabetes (T1D) immunologists often study differences in proteins expressed on the surface of autoreactive T cells in mouse models of the disease compared to cells that are not autoreactive or cells that are not activated. These surface molecules tell us about the activation status of the cells (activated or recently activated) and the specific peptides (like insulin or other beta cell-derived proteins) that the autoreactive T cells specifically target. Since T cells in mammals operate in broadly similar ways, we hope that these models faithfully recapitulate the details of T cell activation and T cell function in the human disease scenario. To know with certainty we must study human T cells from human patients with T1D.
The study I highlight below takes the important proof-of-principle step of comparing gene expression and protein expression data between a mouse model of T1D with T cells isolated from clinical samples from dozens of subjects with T1D. This work is by necessity correlational – the researchers are trying to establish correlation between protein expression level and disease status in mice, and between mouse and human T cells. Correlation however, is not causation. Please note that additional caveats I present are by-and-large inherent to either human studies or mouse models of human diseases, or in comparing similar cells between species. The results in this manuscript represent a heroic amount of work and are a major step forward in our understanding of how T1D develops.
A recent article from Lucy Walker’s research group at the Institute of Immunity and Transplantation at University College, London investigated the role of a sub-type of CD4+ T cells called follicular helper T cells (TFH cells) in diabetes onset in both a mouse model of T1D and human subjects. Follicular helper T cells are a sub-type of CD4+ T cells that are thought to play an important role in generating B cell responses within the lymph nodes. In particular, T follicular helper cells are usually found in the germinal center of the lymph node and participate in a process called the germinal center reaction. In this process, TFH cells promote B cell activation, antibody class-switching, along with promoting effector T cell responses. For these reasons, TFH cells are seen as central to developing effective adaptive immune responses.
This paper studied a mouse model of T1D other than the NOD mouse. These researchers used a model in which a protein called ovalbumin which is normally expressed in birds but not in mice is expressed by cells that makes insulin (so, beta cells). This other protein is used because there are T cells which can be used – transferred into the mouse at defined time points – which specifically target this other peptide. In this way researchers model an autoimmune T cell response, but are not studying the actual disease or bona fide autoreactive T cells.
Autoreactive T cells can mediate beta cell death through either cell contact-dependent mechanisms. Contact-dependent beta cell death can be incited by molecules including granzyme B and perforin from CD8 T cells and FasLigand from CD4 T cells which promote signaling pathways within beta cells that lead to beta cell death. T cells can also promote beta cell death through contact-independent mechanisms by secreting beta cell-toxic chemicals called cytokines. Two cytokines in particular IFN-γ and TNF-α have been shown to promote beta cell death. These cytokines are known to be relevant in mouse models of T1D and promote beta cell death in test tube experiments (in vitro). Today’s highlight paper demonstrated that T cells associated with diabetes onset in the mouse model used express both of these beta cell-toxic cytokines, along with IL-21 (which promotes germinal center reactions and is expressed by TFH cells). These cells do not express markers of regulatory T cells (described in our post about Jeff Bluestone’s work) or IL-17 (which is characteristic of T cells that are important in autoimmune arthritis and in clearance of fungal infections). See figure 3 of the paper for the visualization of these differences. Overall, these results suggest that diabetes-causing CD4+ T cells in mouse models may be TFH cells, not Th1 (IFN-γ and TNF-α only) or Th17 (IL-17 and other cytokines). An important follow-up experiment performed was that enriching T cells for cell-surface molecules expressed by TFH cells made transfer of diabetes from diseased to non-diseased mice more efficient (figure 4). These results are important because TFH cells play roles early in the generation of an immune response, polarizing both B cells and T cells to be autoreactive, and therefore represent an early target to therapeutically inhibit or deplete in clinical attempts to stop or reverse T1D. If the germinal center reaction could be prevented, perhaps T1D could be prevented in susceptible individuals.
Importantly, the researchers then extended these experiments to human samples to see if TFH cells play an important role in the human disease. In figure 5, researchers quantified gene expression levels on memory T cells (defined by cell surface markers) – but not specifically on autoreactive T cells – to demonstrate a general enrichment for TFH phenotype among CD4+ memory T cells in the blood of patients with T1D compared to controls. This result is key because it demonstrates a broad phenotype among memory T cells in patients with an autoimmune disease and supports the mouse data (figures 3 and 4) that a tendency toward developing TFH responses – an “inappropriately over-active” T cell response in general – may promote development of autoimmune conditions. In the case of T1D, one might surmise that HLA associations may play a role in determining what peptides those potentially autoreactive T cells are exposed to, and have the potential to react against, during the germinal center reaction.
In light of these results, studies to (1) find, and (2) remove or inhibit autoreactive T cells and B cells as they accumulate would be key to prevent development of clinical disease. A third option would be to inhibit the general polarization toward TFH phenotype among memory T cells. In my opinion, addressing the phenotype – determining what host genetic and environment factors create the tendency toward too aggressive TFH cell development – would be an even more beneficial, but much more difficult, clinical approach. In practical terms, it has not yet been demonstrated that memory CD4+ TFH cells in human patients are the actual disease-associated autoreactive T cells. This must be shown before depletion studies focused on TFH cells can be considered.