Using intravital 2-photon imaging methods developed in the LBS over the past several years, we are now able to routinely image peripheral organs and tissues such as the liver, kidney, bone marrow, gut, and skin so that immune effector cell behavior and to some extent effector functions during infectious processes can be observed. Using these new methods, we previously described lymphoid and myeloid cell dynamics in BCG-induced granulomas in the liver and showed that only a small fraction of the antigen-specific T cells within a granuloma undergo migration arrest at any time. This small fraction of such arrested cells correlates quantitatively with the fraction of specific cells making the key effector cytokine IFNgamma, as assessed by isolation and intracellular staining of the cells from the infected liver. Confocal imaging data showed that these T cells polarize their secretion of the effector cytokine towards sites of antigen load (bacteria). The limited migration arrest of antigen-specific T cells and the low number of cytokine secreting cells making just detectable levels of IFNgamma is not an artifact of the analytic method, because injection of high levels of the cognate antigenic peptide into the infected animals results in both migration arrest of nearly all the T cells and production of 1-2 logs more IFNgamma per cell by 80-90% of the T cells under these conditions. These data suggested that during normal immune responses to mycobacteria in liver granulomas, there is very limited antigen presentation just sufficient at any moment to activate a small fraction of all available effector cells into a cytokine-secretory state, and to do so just at the margin of quantitative response potential. These findings provide entirely new insights into the way in which effector T cells operate in the natural in vivo setting and point to the large differences between in vitro evoked responses and the actual behavior of effector cells at sites of infection. Many of these observations have been repeated in M tuberculosis-infected animals, arguing that it is not the low pathogenicity of the BCG that leads to such limited effector responses, or the differential dynamic behavior of macrophages or T cells within granulomas. The finding that only a small fraction of the cells being imaged in a tissue are actually engaged in effector function at any time and that these have a dynamic behavior distinct from the bulk of the imaged cells raises critical questions about many existing and ongoing studies in other laboratories using intravital 2-photon imaging, in which the bulk or average behavior of clonally-related cells is taken as representative of the functional population;our data suggest this is a very dangerous way to interpret these images and that in many cases, investigators may be missing the distinct behavior of a small percent of the cells that contribute all or most of the relevant biological activity in the system under study. These findings also point to the possibility that using therapeutic vaccination in Mtb-infected individuals to generate more effector T cells may not be as effective as desired unless ways to enhance the display of antigen needed to trigger these cells in the infected tissue sites (granulomas) can be devised to accompany the vaccination response. In addition to the liver granuloma model, during the past year we have completed studies of T-cell motility in infected lungs. Recent progress in lung explant imaging has permitted us to compare the response of CD4+ T cells to influenza vs. BCG infection in the lungs and address the question of whether in this site, the same limited activation of effector function is seen as in the liver and whether this varies with the pathogen. We suspected that it would, as different organisms have different mechanisms for evading the immune system;mycobacteria try to diminish antigen presentation, among other immune manipulations, whereas influenza mainly seeks to circumvent the innate immune response and does not seem to target antigen presentation. These studies confirm the conclusions of our liver granuloma model, in that only a very small fraction of the mycobacteria-specific T cells arrest movement and produce cytokine (IFN-gamma) in this infected tissue setting. In striking contrast, the influenza-expressed antigen specific T cells showing much greater fractional migration arrest and a correspondingly greater fraction of cytokine producing T-cells. These latter results both confirm the relatively close relationship we previously reported between antigen-induced stopping and effector activity of T-cells in tissues, and also emphasize that the extent of the effector response varies greatly denuding on the microbe, in apparent concordance with the density of presented antigen (influenza >>mycobacteria). In other studies, we are examining how the structure of lymph nodes is organized to limit access of invading organisms to the blood stream. Because viruses and bacteria can access lymph nodes via the lymph and from there, enter the blood stream, there are cellular and structural features of lymph nodes that capture and eliminate these organisms as they enter in the afferent lymph. Our recent data have identified both myeloid cells in the subcapsular sinus that play an important filtering role, as well as innate and adaptive lymphoid cells that take up residence near the sinus to mediate effective and rapid responses to organisms reaching this site. Loss of these barriers results in substantial systemic dissemination of pathogens such as P aeruginosa. In another model, we are using uropathogenic E coli in a model of bladder infection in which we are assessing the location and function of various myeloid cell populations, including macrophages, dendritic cells, and neutrophils, in an effector to better understand how these organisms are sensed and handled by the innate immune system. In our Systems Biology program, we also analyzed the interaction of mice with lethal vs. non-lethal influenza virus and used a top-down system approach in combination with automated imaging and flow cytometry to derive evidence for a fatal feedfoward loop of innate inflammation involving neutrophils. The predictions of the model was that attenuation of this feedback loop would prevent infection-associated lethality and this proved correct.
Showing the most recent 10 out of 13 publications
