Abstract:
Tumour hypoxia results in aggressive tumours with increased metastatic potential and resistance to therapy. Information is limited on cellular responses to hypoxia in the tumour microenvironment primarily due to lack of suitable research tools. After hypoxia-dependent metabolism and entrapment of adducts in hypoxic cells, the ‘clickable’ 2- nitromidazole hypoxia marker (SN33267) bearing a terminal alkyne group can be derivatised via copper(I)-catalysed alkyne-azide cycloaddition (CuAAC) with azides. Labelling of hypoxic cells with fluorophore azides, after exposure to SN33267, was optimised in vitro by simultaneously improving hypoxia-selectivity and preserving the integrity of cellular RNA. With the optimised protocol, the CuAAC-based method was able to separate pre-mixed aerobic (20% O2) and hypoxic (0% O2) cells (2-h exposure with 100 μM SN33267) by fluorescenceactivated cell sorting (FACS) according to the cellular fluorescence intensity in every human cancer line tested (N=4) without prior cell fixation and permeabilisation. The RNA extracted from FACS-sorted fractions was of sufficient quality (RIN>6) for measurement of gene expression. The method was independently validated by evaluating transcript abundance (qPCR) of 15 clinicallyvalidated hypoxia marker genes. Mice bearing subcutaneous human tumour xenografts were administered SN33267 (60 mg/kg; intraperitoneal) and excised 2 h later. Dissociated tumour cells were fluorescently labelled via the CuAAC-based approach and sorted by FACS into four fractions. The transcript expression of 15 hypoxia marker genes was sequentially increased with increasing fluorescence intensity of the four FACS-sorted cell populations. This was mirrored by protein expression of one of the genes (BNIP3). The expression of 10 oxidoreductases involved in the activation of hypoxia-activated anticancer prodrugs was assessed in the same fractions. In FaDu xenografts, TXNRD1 was the only oxidoreductase gene with upregulated transcription (P<0.05) in response to elevated hypoxic stress. In SiHa xenografts, the reductase genes were only moderately upregulated (up to 2-fold) in the fractions of intermediate hypoxia and the most hypoxic fraction showed no change in gene expression relative to the least hypoxic fraction, as did protein expression of the key reductase gene POR. This novel CuAAC-based method is much faster and more versatile than the established antibodybased hypoxic cell detection techniques and provides at least the same sensitivity. The new method was validated independently using a hypoxia gene signature. It provides a powerful and user-friendly tool for pre-clinical study of the hypoxia-mediated effects in the tumour microenvironment by gene and/or corresponding protein expression in tumours in situ.