A key connection between gut microbiome and experimentally-induced diabetes
by Adam Burrack, PhD
Readers of our series will be familiar with the concept of the gut microbiome. In short, this term refers to the observation that the number of bacteria in our intestines outnumbers the human cells of our bodies by a factor of 10-to-1. In our “science of diabetes” blog series we have previously described how the gut microbiota may play a role in development of type 1 diabetes in our post about Mark Atkinson’s research program at the University of Florida. As much as 70% of the cells of our immune system’s reside in the gut and gut-associated tissues (additional locations include the spleen, bone marrow, thymus, and blood), so responses to either pathogens or commensals in the gut play a big role in shaping the overall immune repertoire. In general the gut is a very tolerance-promoting environment, because the vast majority of bacteria in our gut are commensals – they aid in digestion. In our post about zonulin, we described how this molecule may play a key role in maintaining tight junctions between epithelial cells in the intestine, keeping gut-resident bacteria in the gut. It’s important to note that some of the bacteria that are commensals in the gut would be pathogens in the blood or other peripheral tissues. Maintaining the boundary between outside (the gut tube) and inside (the peritoneum) is the key role of the tight junctions between epithelial cells of the intestine.
To summarize, (1) most of the cells of our immune system reside in our gut, and (2) people who go on to develop T1D may not maintain the boundary between gut (outside the body) and inside the body as well as they should. As described previously, Zonulin – the physical barrier of connections between epithelial cells – may be part of the answer to this puzzle. Another part of the answer may be the degree of activation of signaling molecules of the innate immune system in the gut. It’s possible that the innate arm of the immune is over-active – in the absence of infection – in people who go on to develop autoimmunity. Alternatively, the innate immune response may respond overly strongly to infection in people prone to autoimmunity, leading to collateral damage to normal cells, which could seed immune responses against those damaged normal cells (since those proteins would be associated with inflammation and proteins derived from infections). To explore this concept, today I will describe a paper investigating diabetes onset in mouse models, in which the role of one of the key signaling molecules was thoroughly investigated.
In a recent paper in the Journal of Experimental Medicine, researchers led by a group at Yale’s immunology department studied chemically-induced diabetes in mice and the role the signaling molecule NOD2 plays in beta cell death. NOD2 (nucleotide-binding oligomerization domain-containing 2) is a protein within cells whose role is to recognize a specific component of bacterial cell walls (MDP, or muramyl dipeptide). The function of NOD2 is to alert cells that they are infected. When NOD2 detects MDP, signaling events are triggered which result in the production of type 1 interferon (IFN), the classical signal of infection. This signal promotes a multi-faceted immune response against bacterial infection. In other words, there is probably some interesting biology happening at the interface of the gut – including the pancreas – and the innate immune system.
Traditionally, immunology researchers have focused on the behavior of T cells since they are easily measureable – in type 1 diabetes T cells kill beta cells. However, T cell behavior is the result of priming by antigen-presenting cells and the innate response occurring during that priming. Not all NOD mice develop diabetes, and not every person with the HLA DR4/DQ8 haplotypes develops type 1 diabetes. I wonder about the (probably subtle) differences in T cell priming in the pancreatic lymph nodes that lead to these large differences in beta cell death – and diabetes – outcomes.
Our article-of-interest today demonstrates that gut microbiota move from the intestine lumen to the pancreatic lymph nodes following treatment of mice with the beta cell toxin streptozotocin. Here are their basic results. First, expression of NOD2 in the pancreatic lymph nodes of diabetes-prone mice was increased at 10-12 weeks of age, which is the time at which key antigen-presentation events to T cells occur. Second, NOD2-deficient mice treated with the beta cell toxin streptozotocin (STZ, a free radical generator), did not develop diabetes. The researchers then checked whether NOD2 played a key role in beta cell death mediated by STZ in mice expressing this protein. A higher proportion of NOD2-sufficient CD4 T cells produced pro-inflammatory molecules than did NOD2-deficient T cells. This establishes that NOD2 likely does have a role promoting strong responses within T cells themselves. The next question was how does the absence of NOD2signaling change the behavior of macrophages, which serve a key “clean-up” function and play a role in ramping up T cell responses against damaged beta cells? Intriguingly, following STZ treatment, macrophages from NOD2-deficent mice increased expression of proteins associated with wound healing, rather than with pro-inflammatory responses. This suggests that NOD2 has a key role in “skewing” macrophage behavior toward a viral clearance phenotype and away from a “clean-up” or wound healing phenotype. The experiment that made the title of the paper came next: following STZ treatment, researchers found bacterial DNA normally found only in the lumen of the intestine within the pancreatic lymph nodes. This suggests that intestinal permeability was increased by STZ treatment, allowing bacteria to migrate across the epithelial barrier of the intestine. In that scenario, you absolutely want the immune system to respond to new bacteria – just not also to “self” beta cells in the pancreas. Next, if researchers treated the mice with broad-spectrum antibiotics prior to STZ, bacterial translocation to the pancreatic lymph nodes was prevented. Finally, NOD2-deficient mice (not treated with antibiotics) still had bacteria in their pancreatic lymph nodes following STZ treatment. Therefore, STZ treatment appears to increase intestinal permeability, whereas NOD2 signaling appears to promote macrophage activation leading to subsequent T cell activation. In this experimental model, beta cell destruction appears to be a consequence of “guilt by association” with the new bacteria in the lymph node following STZ treatment.
This paper represents an important step forward in our understanding of the interactions between normal gut bacteria and the immune system. STZ treatments in this study served the role of changes in zonulin expression and activity levels in earlier reports. When autoimmunity develops, it appears that the immune system might be responding to something to which it should respond (bacteria in this case). Unfortunately, for both genetic reasons and – it appears – environmental reasons (NOD2 signaling in the pancreas lymph node would promote enhanced responses to any protein shown to T cells in that lymph node at that time), beta cells are also targeted. We applaud this body of work. This study represents one more piece in the puzzle of determining how beta cells are targeted and destroyed. Only through a comprehensive understanding of these processes can rational preventative or curative therapies be developed.