Bifurcation analysis of a tuberculosis progression model for drug target identification.

Tuberculosis (TB) is a major cause of morbidity and mortality worldwide. The emergence and rapid spread of drug-resistant M. tuberculosis strains urge us to develop novel treatments. Experimental trials are constrained by laboratory capacity, insufficient funds, low number of laboratory animals and obsolete technology. Systems-level approaches to quantitatively study TB can overcome these limitations. Previously, we proposed a mathematical model describing the key regulatory mechanisms underlying the pathological progression of TB. Here, we systematically explore the effect of parameter variations on disease outcome. We find five bifurcation parameters that steer the clinical outcome of TB: number of bacteria phagocytosed per macrophage, macrophages death, macrophage killing by bacteria, macrophage recruitment, and phagocytosis of bacteria. The corresponding bifurcation diagrams show all-or-nothing dose-response curves with parameter regions mapping onto bacterial clearance, persistent infection, or history-dependent clearance or infection. Importantly, the pathogenic stage strongly affects the sensitivity of the host to these parameter variations. We identify parameter values corresponding to a latent-infection model of TB, where disease progression occurs significantly slower than in progressive TB. Two-dimensional bifurcation analyses uncovered synergistic parameter pairs that could act as efficient compound therapeutic approaches. Through bifurcation analysis, we reveal how modulation of specific regulatory mechanisms could steer the clinical outcome of TB.

The human amniotic epithelium confers a bias to differentiate toward the neuroectoderm lineage in human embryonic stem cells.

Human embryonic stem cells (hESCs) derive from the epiblast and have pluripotent potential. To maintain the conventional conditions of the pluripotent potential in an undifferentiated state, inactivated mouse embryonic fibroblast (iMEF) is used as a feeder layer. However, it has been suggested that hESC under this conventional condition (hESC-iMEF) is an artifact that does not correspond to the in vitro counterpart of the human epiblast. Our previous studies demonstrated the use of an alternative feeder layer of human amniotic epithelial cells (hAECs) to derive and maintain hESC. We wondered if the hESC-hAEC culture could represent a different pluripotent stage than that of naïve or primed conventional conditions, simulating the stage in which the amniotic epithelium derives from the epiblast during peri-implantation. Like the conventional primed hESC-iMEF, hESC-hAEC has the same levels of expression as the 'pluripotency core' and does not express markers of naïve pluripotency. However, it presents a downregulation of HOX genes and genes associated with the endoderm and mesoderm, and it exhibits an increase in the expression of ectoderm lineage genes, specifically in the anterior neuroectoderm. Transcriptome analysis showed in hESC-hAEC an upregulated signature of genes coding for transcription factors involved in neural induction and forebrain development, and the ability to differentiate into a neural lineage was superior in comparison with conventional hESC-iMEF. We propose that the interaction of hESC with hAEC confers hESC a biased potential that resembles the anteriorized epiblast, which is predisposed to form the neural ectoderm.