Spatially mapping and modeling tuberculosis granulomas to define the immunological determinants of infection outcome

National Institute Of Allergy And Infectious Diseases

Investigators

Linked publications & trials

Abstract

Mtb has an acrobatic ability to evade host immunity. Instead of clearing infection, early immune responses to Mtb induce spatially organized hubs of immune cells called granulomas. While granulomas contribute to pathogen containment, they can also permit sustained bacterial persistence and immune evasion. Identifying the features of the host immune response that promote bacterial clearance versus those that enable persistence will be critical to developing effective host-directed therapies and vaccines. Yet, several challenges have impeded these efforts: 1) The vast majority of research on human immunity to Mtb comes from analysis of peripheral blood. However, as the sites of infection, granulomas hold the data needed to identify immune pathways that subvert or promote protective immunity during TB. 2) Unlike in humans or non- human primates (NHP), mice do not generate granulomas in response to Mtb, yet much of the established dogma in TB immunology comes from murine models. Like HIV and COVID-19, TB is a uniquely human infectious disease that necessitates human-like experimental systems. 3) The limitations of historically utilized imaging techniques prevent complete mapping the cellular interactions and functional microenvironments in TB granulomas and 4) Prior work has not deeply explored the relationship between immune responses in the granuloma and those in lymph nodes (LNs), where adaptive immunity is primed to either tolerate or resist infection. Our work aims to overcomes these challenges by 1) Conducting deep tissue mapping of the spatial proteome and transcriptome of TB granulomas in the NHP model of human TB 2) Defining the drivers of protective versus tolerogenic immune responses to Mtb in a human TB granuloma organoid and 3) Providing the first human in vitro system to simultaneously model human TB granuloma formation and LN responses. 1) Tissue mapping of the spatial proteome and transcriptome of TB granulomas in human and NHP TB: In this reporting period, we pursued tissue mapping studies of TB granulomas from humans and non-human primates (in collaboration with Daniel Barber of NIAID and JoAnne Flynn of University of Pittsburgh). For our human imaging work, we continued with analyses of a previously collected dataset of human TB and sarcoidosis granulomas from the Auria Tissue Biobank. These specimens were imaged via Multiplexed Ion Beam Imaging by time-of-flight for 64 protein targets to uncover information on the organization and state of immune cells in TB granulomas from across a clinical spectrum of human TB disease and pulmonary sarcoidosis. For our non-human primate work we have three studies ongoing. The first is in collaboration with Daniel Barber of the Laboratory of Parasitic Diseases. With his group and Laura Via of the TB Imaging Program we used a protocol previously developed at NIAID to label hypoxic granuloma environments in situ by infusing a small molecule called pimonidazole into rhesus macaques experimentally infected with Mtb. To define define microenvironments based on metabolic pathways, hypoxia, and Mtb presence in these granulomas we have developed an NHP optimized antibody panel for “Iterative bleaching extends multiplexity” (IBEX), a high dimensional imaging platform. This is in collaboration with Ron Germain. Furthermore, we are analyzing these specimens with a single-cell spatial transcriptomics approach called Xenium in collaboration with NIAID’s Research & Technologies Branch (RTB). Additionally, we have generated a paired spatial transcriptomic and glycomic dataset of NHP TB granulomas with the NIAID RTB and our extramural collaborator, Michael Angelo at Stanford University. Using Visium and MALDI-MS, we have produced spatial transcriptomic and N-glycan maps of 12 TB granulomas from the cynomolgus macaque NHP model of human TB in collaboration with JoAnne Flynn. We have also optimized a pipeline for analysis of this data. This cohort represents early TB 4-6 weeks post-infection (n = 6) and established TB 9-12 weeks post-infection (n = 6) allowing us to evaluate how granuloma organization and pathways related to immunometabolism, cytokine signaling, inflammation, and tolerance evolve over time with infection. This study is currently being prepared for publication. Lastly, in collaboration with Clifton Barry (NIAID, LCIM), Laura Via (NIAID, TBIP), and Daniel Barber (NIAID, LPD) we are pursuing analyses of immune responses to BCG vaccination and Mtb challenege in the common marmoset NHP model of TB. To achieve this, we are preparing analyses of clinically relevant variables (CFU, PET/CT imaging) with flow cytometry, spatial transcriptomics, and immunohistochemistry. 2) Defining the drivers of protective versus tolerogenic immune responses to Mtb in a human TB granuloma organoid: The remainder of work we completed during the reporting period was with our in vitro granuloma organoid system. Studying human TB has been challenged by the relevance and tractability of model systems, including mice, which do not generate granulomas in response to Mtb. Analyzing the mechanistic contributions of metabolism to granuloma activity in TB requires a system that accurately recapitulates human pathogenesis. Using the comprehensive single cell spatial mapping of human TB granulomas I generated during my Ph.D., I evaluated all published in vitro granuloma models. I found just one that met the criteria of being both technically feasible and immunologically accurate. In this model—developed by Damien Portevin’s lab, where I trained during my postdoctoral period—human peripheral blood mononuclear cells (PBMCs) are cultured in a 3D collagen matrix. When Mtb is introduced to the culture, granulomas form autonomously and recapitulate many aspects of human TB, including variable granuloma outcomes between donors and macrophages with an immunometabolic profile akin to tumor macrophages. Importantly, this model recapitulates granuloma hypoxia. This system offers an unprecedented opportunity to link observations from high-dimensional imaging data to deep mechanistic investigation of host and microbial factors that drive TB granuloma activity. We now have the BSL2 and BSL3 version of this model established in the lab. We are acquiring input cells from healthy human donor peripheral blood mononuclear cells from NIH’s blood bank and are preparing to conduct a large single cell RNA sequencing study of the determinants of bacterial control in this assay. We have also been collaborating with Clif Barry’s group (NIAID, LCIM) to develop two different reporter strains of Mtb to allow us to evaluate the viability of bacteria with both live-imaging and flow cytometric assays.