Spatial Network Mapping of Pulmonary Multidrug-Resistant Tuberculosis Cavities Using RNA Sequencing

Abstract

<b>Rationale:</b> There is poor understanding about protective immunity and the pathogenesis of cavitation in patients with tuberculosis.<b>Objectives:</b> To map pathophysiological pathways at anatomically distinct positions within the human tuberculosis cavity.<b>Methods:</b> Biopsies were obtained from eight predetermined locations within lung cavities of patients with multidrug-resistant tuberculosis undergoing therapeutic surgical resection (<i>n</i> = 14) and healthy lung tissue from control subjects without tuberculosis (<i>n</i> = 10). RNA sequencing, immunohistochemistry, and bacterial load determination were performed at each cavity position. Differentially expressed genes were normalized to control subjects without tuberculosis, and ontologically mapped to identify a spatially compartmentalized pathophysiological map of the cavity. <i>In silico</i> perturbation using a novel distance-dependent dynamical sink model was used to investigate interactions between immune networks and bacterial burden, and to integrate these identified pathways.<b>Measurements and Main Results:</b> The median (range) lung cavity volume on positron emission tomography/computed tomography scans was 50 cm³ (15-389 cm³). RNA sequence reads (31% splice variants) mapped to 19,049 annotated human genes. Multiple proinflammatory pathways were upregulated in the cavity wall, whereas a downregulation "sink" in the central caseum-fluid interface characterized 53% of pathways including neuroendocrine signaling, calcium signaling, triggering receptor expressed on myeloid cells-1, reactive oxygen and nitrogen species production, retinoic acid-mediated apoptosis, and RIG-I-like receptor signaling. The mathematical model demonstrated that neuroendocrine, protein kinase C-θ, and triggering receptor expressed on myeloid cells-1 pathways, and macrophage and neutrophil numbers, had the highest correlation with bacterial burden (<i>r</i> > 0.6), whereas T-helper effector systems did not.<b>Conclusions:</b> These data provide novel insights into host immunity to <i>Mycobacterium tuberculosis</i>-related cavitation. The pathways defined may serve as useful targets for the design of host-directed therapies, and transmission prevention interventions.