Tissue-of-origin for cancers determines HIF-1 activation induced phenotypic heterogeneity.

Hypoxia-inducible factor-1 (HIF-1) is the master regulator of cellular response to hypoxia, and is activated in many cancers contributing to many steps in the metastatic cascade by acting as a key transcription co-regulator for a large number of downstream genes. Presence of hypoxia within a tumor is spatially nonuniform, and can also by dynamic. Further, although HIF-1 is primarily stabilized and activated by lack of molecular O2, its stability is also affected by other factors present in the tumor microenvironment. HIF-1 also crosstalks with other transcription factors in co-regulating gene expression. Consequently, it is nontrivial to predict the gene expression patterns in cells in response to hypoxia, or HIF-1 activation. Additionally, cancers originating from tissue origins with different basal level of partial oxygen tension may activate HIF-1 at different threshold of hypoxia. We analyzed large published single cell RNAseq data for colorectal, lung, and pancreatic cancers to investigate the phenotypic outcome of HIF-1 activation in cancer cells. We found that cancers from tissues with different partial O2 tension levels exhibit HIF-1 activation at different stages of metastasis, and phenotypically respond differently to HIF-1 activation, likely by contextual co-option of different transcription factors. We experimentally confirmed these predictions by using cell lines representative of colorectal, lung, and pancreatic cancers, finding that while hypoxia enhances growth of colorectal cancer, it induces increased invasion of lung, and pancreatic cancers. Our analysis suggest that HIF-1 activation may act as a rheostat regulating downstream gene expression towards phenotypic outcomes differently in various cancers.

Oscillatory Hypoxia Induced Unfolded Protein Folding Response Gene Expression Predicts Low Survival in Human Breast Cancer Patients.

Hypoxia is one of the key factors in the tumor microenvironment regulating nearly all steps in the metastatic cascade in many cancers, including in breast cancer. The hypoxic regions can however be dynamic with the availability of oxygen fluctuating or oscillating. The canonical response to hypoxia is relayed by transcription factor HIF-1, which is stabilized in hypoxia and acts as the master regulator of a large number of downstream genes. However, HIF-1 transcriptional activity can also fluctuate either due to unstable hypoxia, or by lactate mediated non-canonical degradation of HIF-1. Our understanding of how oscillatory hypoxia or HIF-1 activity specifically influence cancer malignancy is very limited. Here, using MDA-MB-231 cells as a model of triple negative breast cancer characterized by severe hypoxia, we measured the gene expression changes induced specifically by oscillatory hypoxia. We found that oscillatory hypoxia can specifically regulate gene expression differently, and at times opposite to stable hypoxia. Using The Cancer Genome Atlas (TCGA) RNAseq data of human cancer samples, we show that the oscillatory specific gene expression signature in MDA-MB-231 is enriched in most human cancers, and prognosticate low survival in breast cancer patients. In particular, we found that oscillatory hypoxia, unlike stable hypoxia, induces unfolded protein folding response (UPR) in cells resulting in gene expression predicting reduced survival.