Abnormal cellular metabolism is a hallmark of cancer cells, from alterations in the pathways that use glucose to aberrant activation of lipid metabolism. Lipids are a highly complex class of molecule with many cellular functions1, one of the most important of which is to provide the building blocks for the synthesis of cellular lipid membranes. Most tissues in the adult body rely on lipids obtained from the diet or those made in the liver, but many cancer cells instead activate lipid-synthesis pathways to support their rapid proliferation2. This difference between normal and cancerous cells suggests a possible tumour-cell vulnerability that might be exploited therapeutically. Indeed, preventing the synthesis or modification of fatty acids (the building blocks for lipids) can reduce tumour growth in several animal models of cancer2, although this approach has not been successful in the clinic yet. Writing in Nature, Vriens et al.3 report results that might indicate a way forward.
One reason that anticancer strategies targeting lipid metabolism have been ineffective in the clinic could be that alternative pathways compensate for the pathway that is blocked by a given drug. Vriens and colleagues have identified one such compensatory pathway in cancer cells that enables the cells to make monounsaturated fatty acids if the pathway that they normally use is blocked. This alternative pathway is known to act in oil-producing sebaceous glands in human hair follicles, and the authors’ discovery has revealed that cancer cells can also harness this pathway to meet their metabolic demands.
The enzyme stearoyl-CoA desaturase (SCD) catalyses the formation of a specific double bond in palmitate, a saturated fatty acid (a fatty acid without a double bond), and this type of desaturation reaction is needed in the pathway that generates the monounsaturated fatty acids palmitoleate and oleate. These fatty acids are key building blocks for the phospholipids that are components of cellular membranes. The authors investigated the effect of an SCD inhibitor on human cancer cells grown in vitro. They found that some of the types of cancer cell tested were highly sensitive to SCD inhibition, and either stopped dividing or died, whereas others were insensitive and continued to divide. This was unexpected, because the predicted outcome of SCD inhibition would be the accumulation of saturated fatty acids that are toxic to cells at high concentrations4.
Vriens et al. found that cancer cells that are insensitive to SCD inhibition contain high levels of sapienate, a type of a monounsaturated fatty acid that is usually produced in the sebaceous gland. Sapienate is produced5 from palmitate by an enzyme called FADS2 (Fig. 1). FADS2 is also required6 in mammalian tissues for the processing of omega-3 and omega-6 essential fatty acids, which are those obtained from the diet.
The authors report that, relative to its expression in normal tissue, FADS2 expression is elevated in samples from human liver and lung tumours. They note that sapienate is detectable in tumours from mouse models of liver cancer, and that, in these tumours, FADS2 expression correlates with resistance of the cancer cells to SCD inhibition. Interestingly, in an analogous manner to how oleate is formed from the elongation of palmitoleate, the monounsaturated fatty acid cis-8-octadecanoate is formed from the elongation of sapienate. The authors found that both sapienate and cis-8-octadecanoate are incorporated into the membrane lipids of cancer cells.
Vriens and colleagues next investigated whether the FADS2-dependent pathway for the synthesis of monounsaturated fatty acids could compensate for the lack of these compounds that usually occurs when SCD is inhibited. They indeed found that either engineering human cancer cells to express FADS2 or adding sapienate to cells enabled the survival of cancer cells grown in vitro that would usually die if SCD was inhibited. However, human cancer cells grown in vitro that were insensitive to SCD inhibition were killed by a combination of SCD inhibition and depletion of FADS2. In a mouse model of liver cancer that the authors tested, inhibition of both SCD and FADS2 caused a moderate reduction in tumour growth compared with tumour growth in animals in which neither enzyme was inhibited.
Experiments using human cells grown in vitro indicated that the activities of SCD and FADS2 are interdependent. The production of sapienate by FADS2 increased if SCD was inhibited. Conversely, when FADS2 activity was blocked, the synthesis of palmitoleate by SCD was enhanced. This flexibility in lipid-production pathways is highly beneficial for rapidly dividing cancer cells that require a constant supply of monounsaturated fatty acids. However, the authors observed that depletion of FADS2 in the absence of SCD inhibition increased the proliferation of cancer cells, indicating that, although FADS2 might offer a way of generating monounsaturated fatty acids, it comes at a cost in terms of the cells’ proliferative ability.
Accumulation of palmitate can shift the activity of FADS2 towards favouring palmitate as its substrate, and can thereby promote sapienate production7. This could therefore provide a fail-safe mechanism for producing monounsaturated fatty acids when SCD is blocked. Indeed, cis-8-octadecanoate was undetectable in samples of phospholipids from cancer cells in the absence of SCD inhibition, suggesting that no more than a low level of sapienate is generated in cells in which SCD is active.
Vriens and colleagues’ work raises a number of questions. For example, which mechanisms control the level of expression of FADS2 in cancer cells? Considering that FADS2-dependent production of sapienate is relevant only in the absence of SCD, it seems unlikely that sapienate production is the reason for high FADS2 expression in human cancer. It is probable instead that the main function of FADS2 in such cells is to perform its usual role in processing omega-3 and omega-6 fatty acids to generate lipid-signalling molecules involved in functions such as immune evasion8. Switching to sapienate production when SCD is inhibited might prevent FADS2 from performing its usual role and block the production of these signalling molecules.
It is not known whether the monounsaturated fatty acids produced by FADS2 functionally replace those produced by SCD. Incorporation of sapienate and cis-8-octadecanoate into membrane lipids could result in differences in membrane fluidity, curvature or the association of membrane proteins, compared with the corresponding characteristics of membrane lipids made with palmitoleate and oleate.
Another question arising from this study is whether the tumour microenvironment influences the dependence of cancer cells on SCD and FADS2. Cells can also obtain monounsaturated fatty acids through the uptake of a type of phospholipid called a lysophospholipid9. Hence, the levels of such molecules in the tumour microenvironment might determine whether inhibiting both SCD and FADS2 would be an effective way of killing cancer cells. Vriens et al. found that human liver cancer cells implanted in the livers of mice treated with an SCD inhibitor take up sapienate from the tumour microenvironment. This suggests that sapienate synthesis by FADS2 in the tumour is insufficient to satisfy its need for monounsaturated fatty acids. Moreover, consistent with this possibility, the inhibition of tumour growth observed after combined depletion of SCD and FADS2 in mice was only moderate. Perhaps inhibiting fatty acid uptake from the tumour microenvironment might help to block tumour growth when SCD and FADS2 are inhibited.
Vriens et al. provide a thought-provoking example of how cancer cells evolve to meet their metabolic needs. Tackling the complexity of the mechanisms involved remains a challenge for effectively targeting lipid metabolism in cancer therapy.
Nature 566, 333-334 (2019)