Clinical strains of mycobacterium tuberculosis induce strain-specific patterns of cytokine production, gene expression and pathway changes in pulmonary alveolar epithelial cells.

Abstract

The epidemiological success of M. tuberculosis strains, dominant in different geographic regions globally, may be ascribed to a subversion of the host‟s protective immune response. The increasing prevalence of F15/LAM4/KZN, Beijing, F11 and F28 Mycobacterium tuberculosis strain families, coupled with rapidly evolving drug resistance within the KwaZulu-Natal province of South Africa population has resulted in a need to characterize host response associated with infection by these strains. Therefore, in this study, cytokine/chemokine production and host transcriptomics were investigated in A549 pulmonary epithelial cells infected with the F15/LAM4/KZN, Beijing, F28, F11, Unique and H37Rv strains. Cytokines/chemokines were quantified using the Bio-Plex Pro Human Cytokine 27-Plex assay at 0, 24, 48 and 72 hr post-infection. Changes in host gene expression were determined by whole genome RNA Sequencing (RNA-Seq) using the Illumina HiSeq 2000 platform. The 50 bp reads were mapped to the human genome (hg19) using Tophat (2.0.10). Differential expression was quantified using Cufflinks (2.1.0) with false discovery rate (FDR) of 0.05 and a log fold change cutoff of ≥2. R commands (Bioconductor), MeV and Ingenuity Pathway Analysis (IPA) were used to generate heat maps, network and pathways analysis. Twenty-three out of 27 analytes were detected. All strains, except the F28 strain induced an increased production of 18, and a decrease in 5 cytokines/chemokines at 24, 48 and 72 hr post-infection, compared to the uninfected control. Increased production of all 23 analytes by the F28 strain occurred at 48 and 72 hr. Among the 23 cytokines/chemokines that were detected, anti-inflammatory and pro-inflammatory cytokines, as well as chemokines were produced at the different time intervals. Compared to the other strains, high cytokine levels were induced by the F28 strain at 48 hr and F15/LAM4/KZN strain at 72 hr for most analytes. A lower cytokine production was induced by the Beijing and Unique strains at all time intervals. In the case of the laboratory strain H37Rv either a higher or lower cytokine/chemokine production was observed, compared to the clinical strains. RNA-Seq revealed differential gene expression that varied among the strains with respect to both up- and down-regulated genes: F15/LAM4/KZN (1187), Beijing (1252), F11 (1639), F28 (870), Unique (886) and H37Rv (1179). A total of 292 genes were commonly induced by all strains, of which 52 were down-regulated and 240 were up-regulated. Different strain combinations induced different genes that were involved in a variety of pathways, including immune response and apoptosis pathways. Furthermore, strain specific genes were activated by each strain as follows: F15/LAM4/KZN (138), Beijing (52), F11 (255), F28 (55), Unique (185) and H37Rv (125). The F15/LAM4/KZN, Unique and H37Rv were the only strains that had molecular signatures with overlapping functional Kegg and Reactome pathways for their specific genes. IPA analysis revealed canonical pathways that differed among the strains, with the interferon signalling and hepatic fibrosis/hepatic stellate cell activation pathways being among the top 5 pathways in all the strains. Cholesterol biosynthesis and immune related pathway enrichment was similar in the Beijing and Unique strains whilst the F15/LAM4/KZN strain showed closer relatedness to the F11 strain, and the F28 strain closely clustered to the H37Rv strain. The Beijing and Unique strains highly enriched cholesterol biosynthesis pathways compared to other clinical and laboratory H37Rv strain. The top scoring networks induced by these clinical strains varied among the strains with the associated functions. These gene networks were involved in antimicrobial response, developmental disorder, organismal injury, infectious disease and cellular development. Among the transcriptional factors, only EHL, IRF7, PML, STAT1, STAT2 and VDR were induced by all clinical strains, while other factors were strain specific. In conclusion, low cytokine/chemokine production and activation of immune associated pathways by the Beijing and Unique strains suggest a higher virulence for these strains compared to the F15/LAM4/KZN, F11 and F28 strains. These characteristics may explain the high transmissibility and prevalence of the Beijing strains. A similar pattern exhibited by the less prevalent, non-clustering Unique strain, may suggest some virulence attributes in common with the Beijing strain. Findings in this study have the potential to reveal useful biomarkers that can be used as targets for alternative TB therapeutics including immunomodulators that take into consideration network regulations and strain-specific pathways and molecular signatures.