Purpose: In the early 1920s, Warburg published experimental data on the enhanced conversion of glucose to pyruvate (followed by lactate formation) even in the presence of abundant oxygen (aerobic glycolysis, Warburg effect). He attributed this metabolic trait to a respiratory injury and considered this a universal metabolic alteration in carcinogenesis. This interpretation of the data was questioned since the early 1950s. Realistic causative mechanisms and consequences of the Warburg effect were described only during the past 15 years and are summarized in this article. Conclusions: There is clear evidence that mitochondria are not defective in most cancers. Aerobic glycolysis, a key metabolic feature of the Warburg phenotype, is caused by active metabolic reprogramming required to support sustained cancer cell proliferation and malignant progression. This metabolic switch is directed by altered growth factor signaling, hypoxic or normoxic activation of HIF-1α- transcription, oncogene activation or loss-of-function of suppressor genes, and is implemented in the hostile tumor microenvironment. The “selfish” reprogramming includes (a) overexpression of glucose transporters and of key glycolytic enzymes, and an accelerated glycolytic flux with subsequent accumulation and diversion of glycolytic intermediates for cancer biomass synthesis, (b) high-speed ATP production that meets the energy demand, and (c) accumulation of lactate which drives tumor progression and largely contributes to tumor acidosis, which in turn synergistically favors tumor progression and resistance to certain antitumor therapies, and compromises anti-tumor immunity. In all, the Warburg effect is the central contributor to the cancer progression machinery.

Contrary to Warburg's original thesis, accelerated aerobic glycolysis is not a primary, permanent and universal consequence of dysfunctional or impaired mitochondria compensating for poor ATP yield per mole of glucose. Instead, in most tumours the Warburg effect is an essential part of a ‘selfish’ metabolic reprogramming, which results from the interplay between (normoxic/hypoxic) hypoxia‐inducible factor‐1 (HIF‐1) overexpression, oncogene activation (cMyc, Ras), loss of function of tumour suppressors (mutant p53, mutant phosphatase and tensin homologue (PTEN), microRNAs and sirtuins with suppressor functions), activated (PI3K–Akt–mTORC1, Ras–Raf–MEK–ERK–cMyc, Jak–Stat3) or deactivated (LKB1–AMPK) signalling pathways, components of the tumour microenvironment, and HIF‐1 cooperation with epigenetic mechanisms. Molecular and functional processes of the Warburg effect include: (a) considerable acceleration of glycolytic fluxes; (b) adequate ATP generation per unit time to maintain energy homeostasis and electrochemical gradients; (c) backup and diversion of glycolytic intermediates facilitating the biosynthesis of nucleotides, non‐essential amino acids, lipids and hexosamines; (d) inhibition of pyruvate entry into mitochondria; (e) excessive formation and accumulation of lactate, which stimulates tumour growth and suppression of anti‐tumour immunity – in addition, lactate can serve as an energy source for normoxic cancer cells and drives malignant progression and resistances to conventional therapies; (f) cytosolic lactate being mainly exported through upregulated lactate–proton symporters (MCT4), working together with other H⁺ transporters, and carbonic anhydrases (CAII, CAIX), which hydrate CO₂ from oxidative metabolism to form H⁺ and bicarbonate; (g) these proton export mechanisms, in concert with poor vascular drainage, being responsible for extracellular acidification, driving malignant progression and resistance to conventional therapies; (h) maintenance of the cellular redox homeostasis and low reactive oxygen species (ROS) formation; and (i) HIF‐1 overexpression, mutant p53 and mutant PTEN, which inhibit mitochondrial biogenesis and functions, negatively impacting cellular respiration rate. The glycolytic switch is an early event in oncogenesis and primarily supports cell survival. All in all, the Warburg effect, i.e. aerobic glycolysis in the presence of oxygen and – in principle – functioning mitochondria, constitutes a major driver of the cancer progression machinery, resistance to conventional therapies, and poor patient outcome. However, as evidenced during the last two decades, in a minority of tumours primary mitochondrial defects can play a key role promoting the Warburg effect and tumour progression due to mutations in some Krebs cycle enzymes and mitochondrial ROS overproduction. image

Data from 125 studies describing the pretreatment oxygenation status as measured in the clinical setting using the computerized Eppendorf pO2 histography system have been compiled in this article. Tumor oxygenation is heterogeneous and severely compromised as compared to normal tissue. Hypoxia results from inadequate perfusion and diffusion within tumors and from a reduced O2 transport capacity in anemic patients. The development of tumor hypoxia is independent of a series of relevant tumor characteristics (e.g., clinical size, stage, histology, and grade) and various patient demographics. Overall median pO2 in cancers of the uterine cervix, head and neck, and breast is 10 mm Hg with the overall hypoxic fraction (pO2