T3 Integrated transcriptomic analysis of human tuberculosis granulomas and a biomimetic model identifies sphingosine kinase 1 as a potential therapeutic target

The global burden of tuberculosis (TB) continues at pandemic proportions, currently with a quarter of the world’s population infected with Mycobacterium tuberculosis (Mtb), and 1.4 million people dying from TB in 2019 (WHO, 2020). Mtb has undergone prolonged co-evolution with humans, with the balance between protective and pathological host responses likely to play a key role in determining clinical severity. Sarcoidosis is another granulomatous condition primarily affecting the lung and lymph nodes. TB and sarcoidosis share histological and clinical features, which can be indistinguishable, including immune-related phenomena which suggest shared immunological processes. Despite scientific advances, the immunopathology of TB and sarcoidosis remain incompletely understood. Here, we show the combination of unbiased analysis of patient samples and a biomimetic model has established a translational pipeline to identify new therapeutic approaches.We hypothesised that whole transcriptome analysis of human TB granulomas isolated by laser capture microdissection could identify therapeutic targets, and that comparison with sarcoidosis would identify disease-specific mechanisms. Treatment-naïve biopsies were analysed from seven TB patients, ten sarcoidosis patients and seven control samples. Bioinformatic analysis of RNAseq data identified one third of differentially expressed genes were communal to TB and sarcoidosis relative to control samples (absolute log2 fold change ≥ 1.5, adjusted P value < 0.05), with overlap of numerous pathways, including the extracellular matrix and cytokine signalling. Importantly, a TB unique cluster demonstrates TB results from a dysregulated inflammatory immune response, whereas a sarcoidosis predominant cluster relates to elevated lysosomal activity.To translate these insights, we compared three primary human cell culture models: classical 2D; 3D alginate; and 3D collagen model. We demonstrated the Mtb-infected 3D collagen model most closely reflected clinical TB biopsies. We investigated signalling pathways shared between human disease and the 3D model, and used a systematic selection process to identify twelve intracellular enzymes as potential therapeutic targets. Sphingosine kinase 1 (SphK1) inhibition controlled Mtb growth in a dose-dependent manner, concurrently lowering intracellular pH in infected monocytes and suppressing inflammatory mediator secretion. Immunohistochemical staining confirmed that SphK1 is expressed in human