HIF is a heterodimeric complex consisting of an oxygen-sensitive α subunit and an oxygen-insensitive β subunit. There exist three distinct α and two β subunits. In the presence of adequate oxygen concentration, several amino acids of the α subunit undergo post-translational hydroxylation, which drives HIFα to ubiquitinylation and proteasomal destruction (Annex C). A decrease in the availability of oxygen induces a decrease of this hydroxylation and a relocalisation of the α subunit to the nucleus; in the nucleus, the α subunit can be associated to the β subunit and the complex can stimulate the transcription of target genes (Fig. 16.1). There are many signalling pathways which activate HIF transcription, especially the proliferation pathways opened by MAP kinases and PI3 kinase (Chaps.​ 2 and 3). Several transcription factors, such as MYC or NFκB, thus activate the synthesis of HIF mRNAs.

The α subunits contain a domain called ODDD (oxygen–dependent degradation domain), with two proline residues that can be hydroxylated by prolyl 4-hydroxylases (PHD, genes EGLN) and a C-terminal domain containing an asparagine residue that can be hydroxylated in the nucleus by asparagine hydroxylase, an enzyme also known as FIH (factor inhibiting HIF, gene HIF1AN). These enzymes are non-haem dioxygenases, which use molecular oxygen, α-ketoglutarate and Fe²⁺ as a cofactor. Their activity is decreased when their substrate O₂ is not abundant enough (hypoxia). The hydroxylated proline residues are recognised by the VHL (von Hippel–Lindau) protein, which belongs to an E3 ubiquitin ligase complex called VBC (VHL–elongin B–elongin C), which enables HIF1α to be directed to the proteasome.

In the nucleus, asparagine hydroxylation prevents HIFα binding to a coactivator of transcription, CBP (CREB–binding protein), which acts as a histone acetyltransferase for chromatin relaxation (Annex B). In the nucleus, the HIFα and HIFβ subunits are combined through their N-terminal domains and bind DNA, via a HLH (helix–loop–helix, Annex B) motif, on a hypoxia-responsive element (HRE) belonging to the promoter of the target genes. In the C-terminal domain lays the transactivation domain, TAD, divided, for HIF1α and HIF2α, into a hydroxylation-independent domain (N-TAD) and a domain inhibited by the FIH-mediated hydroxylation of an asparagine residue, C-TAD. The HIF3α subunit could rather play a negative dominant role on the transcription induced by the other HIFα subunits. The main target genes of HIF proteins are those encoding the VEGFs (vascular endothelial growth factors), especially VEGFA (Chap.​ 1). HIFs appear, therefore, as fundamental regulators of angiogenesis. Several dozens of other target genes are under the dependence of HIFs, involved especially in energy metabolism, erythropoiesis, vasodilatation and autophagy: for all these processes, oxygen availability is a crucial limiting factor.

Hypoxia constitutes the more important angiogenesis-promoting signal. Angiogenesis is indispensable to tumour growth beyond some millimetres in diameter, because only vessels can bring to the tumour the required nutriments and oxygen. In the absence of vascularisation, the deep tumour regions do not benefit of this supply and become rapidly necrotic, which limits tumour growth. Neo-angiogenesis is one of the major oncogenic processes and its targeting, which had been suggested 40 years ago as susceptible to bring original anticancer weapons, has eventually allowed the development of therapeutic approaches.

HIF overexpression is commonly observed in human cancers, sometimes in association with poor prognosis. Whereas there are no known activating mutations of the genes encoding the various HIF subunits, a loss-of-function mutation frequently occurs in the VHL gene, especially in renal cancers in which it appears recurrent. In the absence of this E3 ubiquitin ligase activity, the stability of the HIFα subunits is increased and their effects on tumour growth are increased. Antiangiogenic agents such as bevacizumab, an anti-VEGFA monoclonal antibody (Chap.​ 1), have brought proof of activity in this type of cancer as in several others.

Among the target genes of HIFs, several genes are involved in glucose transport and anaerobic metabolism. In order to provide ATP, glucose can be metabolised according either to the mitochondrial pathway of oxidative phosphorylation, which requires oxygen, or to the cytosolic anaerobic glycolysis pathway. This pathway is less energetically profitable (two molecules of ATP are formed per molecule of glucose instead of 38), but it can present an utmost interest for hypoxic cells. By favouring glucose uptake and commitment to glycolysis, HIFs enable hypoxic cell survival. It has been known for a long time that tumour cells have a high rate of anaerobic glycolysis, even in normoxic conditions (this is the Warburg effect); this may represent a preadaptation to the risk of hypoxia, which may have been genetically selected during repetitive hypoxia events followed by re-oxygenation.

Whereas there are presently no ways to target tumour cells by agents that would selectively kill the cells using anaerobic glycolysis rather than oxidative phosphorylation, this peculiarity of tumour cells is utilised by an imaging technique, positron-emission tomography: 2-deoxy-2-¹⁸ F-glucose, a non-metabolised glucose analogue, serves as a glucose substitute to detect tumour cells.

Among the various hypoxia-regulated genes is DDIT4 (DNA damage–inducible transcript 4), which encodes REDD1/RTP801 (regulated in development and DNA damage response). This protein activates the complex TSC1–TSC2, an inhibitor of mTOR activation by the small G-protein called RHEB (Chap.​ 3); mTOR is involved in protein synthesis, lipogenesis, cell growth and autophagy. Hypoxia appears therefore as a way to reduce protein synthesis and to protect cell survival by autophagy.

Other potential targets of HIFs are the genes involved in apoptosis triggering (Chap.​ 18), via the opening of the transition pore that enables the efflux of cytochrome c from the mitochondria: proteins of the BNIP (BCL2–interacting protein) family are induced by HIFs. It seems that apoptosis induction can be operated when the hypoxic cell can no longer envisage survival.

Finally, many proteins playing a fundamental role in cell adhesion and migration are also transcriptionally induced by HIFs: vimentin (VIM), fibronectin (FN), keratins (KRT), matrix metalloproteinases (MMPs), cathepsins (CTSs), etc. Indirectly, E-cadherin (CDH1, Chaps.​ 7 and 11), which plays a major role in inhibiting epithelial-to-mesenchymal transition, is repressed during HIF activation. Hypoxia is, therefore, heavily implicated in metastatic processes.

Since tumour cells generally live and grow in a hypoxic environment, it would be interesting to develop drugs which would selectively target hypoxia. An old method for targeting hypoxic cells is the use of bioreductive drugs, which are activated by hypoxia by monoelectronic reduction catalysed by DT-diaphorase (NQO1) or cytochrome P450 reductase (POR). Mitomycin C is a good example of such drugs. Another approach is the use of nitroimidazoles, which have a radiosensibilising effect at the level of hypoxic regions of tumours; after reduction, these compounds stabilise DNA lesions. Misonidazole is used in positron emission tomography to visualise hypoxic tumour regions.

Upstream HIF1α, hypoxic cells targeting may consist in mTOR inhibition, which would induce a decrease in HIF1α expression (Chap.​ 3). Rapalogs and mTOR serine/threonine kinase inhibitors can be considered as antiangiogenic treatments and are especially active in renal cancers for which VHL mutations is an oncogenic driver. Targeting HIF1α itself is a difficult challenge. Indirect approaches with heat shock proteins (HSPs) inhibitors, such as geldanamycin, are considered. Direct approaches are represented by antisense oligodeoxynucleotides or by small molecules or monoclonal antibodies targeting HIF1α-induced proteins, such as carbonic anhydrase IX (CA9) or the glucose transporter GLUT1.

ROS are endogenously produced from molecular oxygen in several organelles, especially mitochondria, which is the main place where oxygen is utilised. ROS is the generic name for three chemical entities: the superoxide radical O₂ ^– ●, the hydroxyl radical OH^● and hydrogen peroxide H₂O₂. Superoxide ions are formed by reduction of molecular oxygen during a reaction catalysed by several enzymes, especially the oxidoreductases of the mitochondrial electron transport chain and membrane NADPH oxidases (NOX):
Superoxide ions are detoxified by superoxide dismutases (SOD), Cu²⁺-Zn²⁺−SOD in the intermembrane mitochondrial fluid and Mn²⁺−SOD in the mitochondrial matrix. The dismutation reaction generates hydrogen peroxide as follows: