Quantifying chronic inflammatory burden from transcriptomes in viral and immune-mediated pathologies

Human T-Cell Leukemia Virus Type 1, HTLV-1, is a pathogenic retrovirus infecting approximately 10 million individuals worldwide. The virus causes two distinct pathologies: Adult T-cell leukemia/lymphoma (ATL) and HTLV-1 Associated Myelopathy / Tropical Spastic Paraparesis (HAM/TSP). The common treatment of ATL currently consists of combination therapy with interferon (IFN) α and zidovudine. However, early reports showed IFN-β was also an effective treatment strategy, though IFN-α treatment became the standard based on empirical results. To explore the potential viability of IFN-β treatment in ATL, we tested the differential effects of IFN-α and -β on short term PBMC cultures of ATL patients and concluded that IFN‑β has superior anti‑proliferative and pro‑apoptotic effects. Additional meta‑analysis in four ATL gene expression datasets revealed a consistent decrease in RORC transcript abundance. In addition, a robust negative correlation exists between IL17C gene expression and proliferative gene expression in ATL and in other lymphoid leukemias. The transcriptomic experiments used in these studies also showed that inflammation could serve a protective role in ATL. As HTLV-1's other major pathology, HAM/TSP, is a neuroinflammatory disorder, we aimed to find a robust way of quantifying the inflammatory burden in transcriptomic experiments. Glycoprotein Acetylation (GlycA) is a novel biomarker for inflammation quantified in blood serum or plasma using Nuclear Magnetic Resonance (NMR) spectroscopy. This marker is a summary measure associated with a broad range of inflammatory processes and can be interpreted as a patient's chronic inflammatory burden. Using various machine learning algorithms on a large collection of paired NMR measurements and blood gene expression profiles, we constructed a predictive model which quantifies relative GlycA concentration from the gene transcript abundance in a patient's blood. This predictive model was first shown to replicate published GlycA associations. Then, novel predictions were made using publicly available third‑party data, which were tested, and confirmed to be accurate, using new NMR experiments. The GlycA measurements in Inflammatory Bowel Disease (IBD) and Systemic Lupus Erythematosus (SLE) were studied in greater detail. In IBD, GlycA concentration in patient serum samples was found to be higher than what was measured in healthy controls. In patients that responded to treatment and achieved mucosal healing, GlycA fell back down to the levels observed in healthy controls. Patients that showed an endoscopic treatment response but did not achieve full mucosal healing showed a GlycA decrease but fell short of returning to the healthy control GlycA levels. Considering our data shows that GlycA tracks disease activity even in patients without elevated C-reactive protein, our results demonstrate that GlycA holds great promise as a serological biomarker for disease activity in IBD. In SLE, our results show that GlycA levels are higher in SLE patients than those observed in healthy controls and even in nephritic controls without lupus, despite the altered renal function of the latter. We find that GlycA is associated to the SLE disease activity index and that proliferative lupus nephritis patients have higher GlycA concentrations than non‑proliferative patients at time of renal biopsy. When comparing performance of GlycA to traditional biomarkers, we show that GlycA is the more informative biomarker. (IWT project 141614: A transcriptomic approach to determine immunomodulatory therapeutic strategies in current and novel viral epidemics)

Jaar van publicatie:2019

Toegankelijkheid:Open