LAUSR

Dear editor, In an interesting letter published recently in this journal, Drs Israël and Schwartz hypothesize that the breakdown of ketone bodies is a tumor cell's “unique source of mitochondrial acetyl CoA”. They propose to inhibit the enzyme succinyl-CoA:3-oxoacid-CoA transferase (SCOT, also known as 3-oxoacid CoA-transferase 1 or OXCT1) which is specific for the ketolytic pathway. Therefore, its inhibition would “deprive tumor cell mitochondria of acetyl CoA” which could be a tumor cell's most vulnerable metabolic pathway. As a corollary of their hypothesis, a ketogenic diet (KD) would be inappropriate to treat cancer patients as it increases ketone body supply to tumor cells and hence supports acetyl CoA production and tumor growth. While we think their main hypothesis that SCOT inhibition could target tumor cells is worth further studying, we argue that the presented preclinical and clinical data provide at best weak evidence for it and in particular do not support the notion that KDs would be contraindicated in cancer treatment. As the only clinical evidence in their case against the KD, the authors refer to a review by Klement about the seminal studies of Brünings during World War II in which cancer patients were treated with high doses of insulin and a very low carbohydrate diet. Brünings initially observed tumor shrinkage, but in most cases, there was a rebound effect after 2-3 months. Israël and Schwartz suggest that after 2 months, a “ketogenic adaption” would have lead to production of ketone bodies which in turn would have supported tumor regrowth. This is highly speculative since it is unclear why a rise in ketone bodies should occur after exactly this time interval despite continuing high insulin injections. As Israël and Schwartz stated, insulin would inhibit ketogenesis (although no conclusions can be drawn as Brünings did not measure ketone bodies). Further, Israël and Schwartz cite one preclinical study to support their hypothesis that ketone bodies drive tumor growth. However, this study needs to be interpreted with caution, as the results stem from a genetically modified tumor model with OXCT1 overexpression and the transferability to naturally occurring tumors is uncertain. Next, the authors mention Carney triad cancers with putatively high SCOT activity due to high succinyl CoA production. This does nothing to support the notion that ketone bodies fuel tumor growth. First, there is no reference showing that Carney triad cancers are indeed driven by SCOT activity. Second, to the best of our knowledge there exists no study showing that a KD accelerates growth or development of Carney triad cancers. The findings that KDs have antitumor effects in preclinical models of brain tumors present an anomaly for the author's hypothesis which they explain by proposing that SCOT is downregulated in these tumors through a “defense mechanism” which slows ketolysis. They do not explain why this should only apply to brain tumors, nor do they mention that antitumor effects have also been found in many extracranial tumor models. Indeed, tumor cells with intrinsically low SCOT expression have been described, and not only in brain tumor cells as Israël and Schwartz acknowledge. The work of Zhang et al contains many examples of tumor cells with low OXCT1 expression, for example, PANC-1 pancreatic cancer or ARO thyroid cancer cell lines. To the best of our knowledge, current evidence does not support the notion that ketone bodies drive tumor growth. If ketone bodies would support tumor growth, we would also expect calorie restriction and fasting to accelerate tumor growth, but in general, the opposite effect is observed. At the same time, some clinical studies even indicate beneficial effects of KDs on clinical outcomes in cancer patients. Further, it is not likely that ketone bodies are the tumor cell's unique source of acetyl CoA, since several alternative pathways are known to supply the tumor cell with acetyl CoA. If ketone bodies indeed constitute a vital supply of tumor cells’ acetyl CoA via SCOT, we would expect positive correlations between SCOT expression, ketone body uptake and cell proliferation. This has not been observed. Bartmann et al, studying seven human breast cancer cell lines, found no correlation between SCOT mRNA expression, β-hydroxybutyrate or acetoacetate uptake and proliferation. Maurer et al found that despite SCOT expression in five glioma cell lines, β-hydroxybutyrate was not able to rescue these cells from cell death under low glucose conditions; it was not utilized as a substantial substrate nor did it influence growth, proliferation, motility or invasiveness. Further, SCOT was downregulated under hypoxic conditions that would frequently be found in all solid tumors of a certain size; this strongly questions the notion of SCOT as an ubiquitous “Achilles heel” of cancer cells. Zhang et al observed that under low glucose conditions β-hydroxybutyrate did not show an effect on proliferation in tumor cells with low expression of OXCT1 and BDH1 (PANC-1), while it increased proliferation in cells with high expression of OXCT1 and BDH1 (HeLA). Remarkably, under high glucose conditions, β-hydroxybutyrate did not show an effect on proliferation in HeLa cells, which points towards glucose as a more important substrate for proliferation. in vivo experiments showed that a KD in mice with Received: 13 April 2020 Accepted: 15 April 2020