Background: Immune reconstitution inflammatory syndrome (IRIS) occurs with potent treatment of HIV in the setting of co-infections such as M. tuberculosis (Mtb). IRIS is associated with substantial morbidity and mortality. We sought to better understand this syndrome and to develop a metabolic biosignature in plasma from subjects that developed IRIS. Samples were obtained from subjects enrolled in the immediate treatment arm of ACTG study A5221.
Methods: Subjects from the IRIS and Non-IRIS were matched for age, gender, and CD4 cells by the ACTG Data Center. Plasma samples from as many as 3 time points per subject (entry, IRIS window and follow-up) were analyzed. Plasma was extracted in triplicate with methanol and analyzed by liquid chromatography-mass spectrometry (LC-MS). LC-MS data were analyzed with Mass Profiler Professional (MPP) software (Agilent, Inc.) and XCMS online (Scripps Institute) to select molecular features (MFs) that distinguished IRIS and non-IRIS samples based on statistical significance.
Results: Each triplicate revealed 2,000-5,000 MFs and these were down selected based on a FC of 1.5 between IRIS and non-IRIS at each time point with a significance of p<0.05. The greatest difference between the IRIS and non-IRIS samples was observed with the IRIS window time point and included 59 MFs for all replicates analyzed. Four metabolomics databases (METLIN, HMDB, MTB LIPID MAPS and LIPID MAPS) were searched to provide a putative identification of the MFs. MF presumed to be small peptides and lipids (triglycerides, glycosphingolipids, vitamin D metabolites) dominated the signature. Some of the putative lipids are in well-known inflammatory pathways. Each significant MF was manually crossed checked against our in-laboratory Mtb lipid database and no matches were found, leading us to believe that the lipid molecules identified are of host origin. Targeted LC-MS analyses are currently underway to confirm identities and further assess changes in lipid metabolism of IRIS subjects.
Conclusions: These studies provide evidence that a biosignature based on metabolites can separate IRIS from non-IRIS early in the course of disease and may be able to predict those at risk for IRIS. Lipid species appear in our discriminatory signature. Efforts are ongoing to formally identify each of the significant MF we discovered. Ultimately we aim to confirm and validate our findings using samples collected in prospective trials or sites studying HIV/Mtb co-infected individuals who develop IRIS.