ABSTRACT
Hypoxia influences nearly all steps in the metastatic cascade, and is an independent adverse indicator for cancer
prognosis. Cellular response to hypoxia is tightly regulated, and is mediated by hypoxia-inducible factors
(HIFs), with HIF-1 being the ubiquitously expressed homologue. Canonical response to hypoxia involves
stabilization of HIF-1α, which acts as a key transcriptional factor, regulating the expression of more than 1000
gene products indirectly, influencing key steps in cancer progression. However, we discovered that the
population response to hypoxia is more complex than the canonically understood response, with a small
subpopulation displaying oscillations in HIF-1α stabilization and transcriptional activity in a lactate dependent
manner. Lactate is a byproduct of glycolysis, which is itself increased due to HIF-1α activity, and can cause
degradation of HIF-1α by chaperone mediated autophagy, driving oscillations. Owing to the centrality of HIF-
1α in transcriptional regulation in hypoxic tumors, oscillations in HIF-1α in a subset of cells could have profound
consequences in gene expression, and cancer progression. Our preliminary data show that oscillatory hypoxic
input can drive large scale transcriptomic changes, resulting in increased metabolic activity, cell proliferation,
as well as altered regulation of pathways related to circadian rhythms, and invasion. These data suggest that
possibly emergent oscillations in HIF-1α activity may provide a selective advantage to these cells to escape
hypoxia induced stress response.
Our preliminary data present a strong rationale to investigate this emergent phenotype in cancer populations,
the mechanisms driving these oscillations and the phenotypic consequence of these oscillations. Furthermore,
many of the genes responded to oscillating hypoxia as a qualitatively different signal, suggesting presence of
regulatory motifs, incoherent feedforward loops (IFFLs) which could distinguish between oscillatory and
sustained HIF-1α signal. Using an integrated approach involving computational modeling, bioinformatics, and
experimentation, we will systematically identify and validate these IFFLs, as well as the co-factors necessary to
form these IFFLs along with HIF-1α. Our aim will not only shed light on the fundamental regulatory
mechanisms of decoding of oscillatory signaling in cancer, but also provide a targeting strategy to contain the
phenotypic consequences of emergent HIF-1α oscillations. Finally, we will test the consequence of emergent
HIF-1α oscillations in vivo in a mouse model of breast cancer tumorigenesis, and test if oscillating HIF-1α confers
increased tumorigenicity, proliferation, and survival. Our proposed method will facilitate mechanistically
understanding the genesis of oscillations, generation of phenotypic heterogeneity in cancer, as well as
understand and target the consequence of this emergent subpopulation in influencing cancer progression.
