Cholesterol is an apparently indispensable lipid for numerous processes required for cell proliferation. Levels of this molecule are primarily regulated at the transcriptional level by the SREBPs (sterol-regulatory-element-binding proteins) and LXR (liver X receptor). In this issue of the Biochemical Journal, Rodríguez-Acebes et al. show that a cholesterol precursor, desmosterol, can support cell proliferation in the absence of cholesterol in a murine macrophage-like model (J774-D cells). These cells are defective in DHCR24 (sterol-Δ24-reductase, or 3β-hydroxysterol Δ24-reductase), leading to desmosterol accumulation, and yet sterol homoeostasis appears to be normal with respect to SREBP processing and LXR activation. Other potentially cholesterol-dependent processes which were not the focus of this study are briefly discussed, such as lipid-raft-dependent cell signalling.
Cholesterol may receive a bad press from the mass media and in food advertising, but it is a critical lipid for mammalian life. It maintains membrane fluidity and integrity, allowing lipid bilayers to form a permeability barrier with mechanical strength, while also enabling lateral diffusion of proteins and curvature of the membrane, required for processes such as endocytosis. Cholesterol also stabilizes the structure of lipid rafts, cholesterol and sphingolipid-enriched microdomains that can act as platforms for the association of proteins and lipids to regulate signalling and trafficking events. Cholesterol also plays a part in signalling due to its covalent attachment to the protein Hedgehog, required for embryonic development and cellular differentiation, and is a precursor for bile acids, oxysterols and steroid hormones. But exactly how essential is cholesterol? Can ‘any old sterol’ do the same job?
In this issue of the Biochemical Journal, Rodríguez-Acebes et al.  show that, in the absence of cholesterol, desmosterol is sufficient for the support of cell proliferation and the regulation of cholesterol homoeostasis. Desmosterol is the immediate precursor of cholesterol and differs structurally from cholesterol by only an additional double bond on the side chain (Figure 1). The conversion of desmosterol into cholesterol is catalysed by DHCR24 (sterol-Δ24-reductase, or 3β-hydroxysterol Δ24-reductase), an FAD-dependent oxidoreductase which has been generating quite a ‘buzz’ of late. Initially identified as being critical for plant growth (the plant gene was dubbed DIMINUTO or DWARF1), the human orthologue was originally described to be down-regulated in affected neurons in Alzheimer's disease, and thus named Seladin-1 (the selective Alzheimer's disease indicator-1). The fact that it also catalyses the penultimate step in cholesterol synthesis came as somewhat of a surprise, but in the context of Alzheimer's disease probably relates to the importance of the desmosterol–cholesterol balance in the brain. Significantly, high desmosterol levels are found in astrocytes and in the developing brain. The enzyme was also the target of MER-29 (Triparanol), the first cholesterol biosynthesis inhibitor used clinically to treat hypercholesterolaemia. This drug became mired in scandal in the 1960s after causing severe side effects including hair loss and cataracts, which had already been observed in animal studies, but which the manufacturer Richardson-Merrell failed to mention to the Food and Drug Administration, eventually resulting in massive payouts and discouraging further drug development in this area until the advent of the statins .
Conversion of desmosterol into cholesterol is blocked in J774-D cells
Desmosterol has previously been shown to substitute for cholesterol in mutant mouse L-cell fibroblasts as early as 1970 using growth in delipidated serum , although this culturing condition was not confirmed to be cholesterol-free. More recently, studies using DHCR24 knockout mice , or cells derived from them [5,6], indicated that desmosterol will suffice as a cholesterol replacement. These mice were viable, albeit small and infertile. Maternal cholesterol can be obtained during development in mice, so these animals are not cholesterol-free, at least for the first several months of life .
Rodríguez-Acebes et al.  used J774 cells, a murine macrophage-like cell line which lack a full-length transcript for DHCR24, and consequently accumulate desmosterol. Cholesterol made up approx. 20% of total sterols after 10 days of growth in media containing LPDS (lipoprotein-deficient serum), which in any case contains a low concentration of cholesterol, so the maintenance media was changed to one that was cholesterol-free. Even when the cells, designated J774-D cells, were passaged for more than a year prior to analysis, cellular cholesterol and normal DHCR24 transcript levels remained undetectable. This observation seems to contrast with the contention that small amounts of cholesterol are essential for roles that support life, besides fulfilling a bulk membrane requirement .
Treatment with a statin to inhibit sterol synthesis halted proliferation in J774-D cells, which could be averted through supplementation with either cholesterol or desmosterol. In contrast, proliferation was only partially rescued by incubation with cholest-5,22-trans-dien-3β-ol, where the double bond is closer to the steroid nucleus, at C22 rather than C24. This is somewhat surprising: although a wide range of sterols and cholesterol analogues such as β-sitosterol, ergosterol and 5β-cholestan-3β-ol are unable to support cell proliferation, 5α-cholestan-3β-ol (dihydrocholesterol) and desmosterol can [1,9]. This suggests that even slight alterations to the side chain can have a deleterious effect on cell processes, and highlights the extremely fine functional discrimination between similar sterols in mammalian systems.
This cholesterol-free system is an excellent model for investigating the requirement for cholesterol in specific systems. One set of processes include those involved in maintaining lipid homoeostasis, including the regulatory mechanisms which control cellular levels of cholesterol itself. Regulation is primarily achieved by two transcription factors: SREBPs (sterol-regulatory-element-binding proteins) and LXR (the liver X receptor). SREBPs are master regulators of lipid homoeostasis, controlling the expression of a whole suite of genes required for the synthesis or uptake of lipids including cholesterol, fatty acids, triacylglycerol and phospholipids, such as HMG-CoA (3-hydroxy-3-methylglutaryl-CoA) reductase and the low-density lipoprotein receptor. In order to maintain cellular sterol levels within a narrow range, proteolytic processing required for SREBP activation is prevented by binding of cholesterol to Scap (SREBP cleavage-activating protein) in the endoplasmic reticulum, or alternatively through binding of oxygenated forms of cholesterol, so-called oxysterols, to Insig (insulin-induced gene product) [10,11]. There are multiple isoforms of SREBP, with SREBP-2 primarily regulating cholesterol and SREBP-1a and (particularly) SREBP-1c regulating fatty acid metabolism. SREBPs can sometimes activate their own expression, as well as that of Insig-1, but SREBP-1c is additionally a target of LXR. Specific oxysterols appear to be the major natural ligands for this nuclear receptor, whose targets include genes involved in cholesterol efflux from the cell, such as ABCA1 (ATP-binding cassette, subfamily A, member 1).
So what happens to lipid homoeostasis in the absence of cholesterol? Rodríguez-Acebes et al.  found that the J774-D cells demonstrated similar decreases in SREBP processing and SREBP target gene expression when either cholesterol or desmosterol were added. Desmosterol is known to be registered by the sterol-sensing protein Scap  and to inhibit SREBP-2 processing in Chinese hamster ovary cells [11,12], but the study by Rodríguez-Acebes et al. took special care to ensure that the J774-D cells were cholesterol-free (although the authors cautiously use the term ‘virtually free’). Furthermore, desmosterol treatment yielded a marked increase in Srebf1c and Abca1 mRNA, supporting the contention that this sterol can serve as an LXR ligand [5,12].
For a more physiologically relevant system, SREBP and LXR target–gene expression was studied in response to endogenously synthesized desmosterol in J774-D cells. To observe the restoration of homoeostasis, the cells were pre-treated with statin to up-regulate SREBP target genes, and the drug removed to allow synthesis of sterols which would then interact with the regulatory machinery. mRNA levels of SREBP target genes returned to control levels hours after drug removal. In contrast, expression of the LXR target gene, Abca1, followed the opposite trend, falling with statin treatment before rising after the removal of the statin, in line with the restored synthesis of LXR sterol ligand(s) . This key experiment suggests that endogenously produced desmosterol can regulate SREBPs and LXR in the absence of cholesterol.
However, it is important to note, particularly for this last experiment, that 24(S),25-epoxycholesterol will most certainly also be produced in this model, since this oxysterol is not derived from cholesterol. Instead, 24(S),25-epoxycholesterol is synthesized in parallel by a shunt in the same pathway, independently of DHCR24, with oxidation at the position of the double bond found in desmosterol (Figure 1). This oxysterol is approx. five-fold more potent as an LXR ligand than desmosterol . Research from our laboratory has shown that 24(S),25-epoxycholesterol, at levels encountered endogenously and under similar experimental conditions, is a potent activator of LXR and a strong suppressor of SREBP processing . The relative contributions of endogenous 24(S),25-epoxycholesterol compared with desmosterol to sterol homoeostasis in J774-D cells could be addressed by measuring their respective levels at multiple time points. Additionally and ideally, endogenous production of 24(S),25-epoxycholesterol could be excluded as a potential confounder by employing a molecular approach shown to selectively inhibit synthesis of this oxysterol .
The apparently normal sterol homoeostasis found in J774-D cells does not appear to translate in vivo, with perturbed sterol homoeostasis observed in DHCR24 knockout mice [5,6]. This was attributed to the accumulated desmosterol strongly activating LXR target genes in the livers of these animals. The liver is an important determinant of whole-body cholesterol homoeostasis, and makes bile acids from cholesterol. Other organs may also have specific requirements for cholesterol (for example, in producing steroid hormones). Two reported cases of DHCR24 deficiency in humans (desmosterolosis) resulted in catastrophic developmental defects, as occurs in other lesions in the distal cholesterol-biosynthetic pathway . These defects are caused by impaired Hedgehog signalling (requiring cholesterol for optimal functioning), providing a clear demonstration of the essentiality of cholesterol during mammalian development. Therefore it is likely that the ability of a cell or organism to tolerate the substitution of cholesterol with desmosterol will be cell-type-specific, and certainly different in vivo compared with cultured cells.
J774 cells are used extensively for studying cholesterol metabolism, particularly reverse cholesterol transport, which may seem curious since they do not synthesize the sterol. In contrast with the current model, these systems are generally far from being cholesterol-free, with an ample exogenous supply. Other characteristics of the cell line are probably of greater importance than increased desmosterol levels, such as a lack of caveolin-1 and apolipoprotein E , which can prove extremely useful or confounding, depending on the precise research question.
A cholesterol-free system is a good model for studying the importance of cholesterol in signalling, and possibly for determining which processes have a strict requirement for lipid rafts, although this line of investigation was beyond the scope of the study by Rodríguez-Acebes et al. Cyclodextrins such as methyl-β-cyclodextrin have been commonly utilized to disrupt rafts through cholesterol depletion, but they exert additional unwanted effects, such as general toxicity and on the lateral diffusion of proteins. Consequently, cells which already lack cholesterol may provide a valuable alternative. Desmosterol has been shown to be poor at forming and maintaining lipid rafts compared with cholesterol, and as a result impairs insulin signalling [6,15]. Other critical raft-mediated processes may be abnormal in J774-D cells, but sufficiently functional to sustain indefinite proliferation in the absence of cholesterol. Along similar lines, raft-dependent β-secretase activity and amyloid-β generation were impaired in the brains of DHCR24 knockout mice when desmosterol levels were low, whereas function was restored when desmosterol levels matched those of cholesterol in a normal mouse, since it allowed formation of necessary detergent-resistant membrane domains, identified as lipid rafts .
It was Konrad Bloch who first proposed that the order of sterols in the cholesterol-biosynthetic pathway reflects their distance in evolutionary time, with evolution perfecting the molecule for optimal membrane function. Even without the benefit of a time machine, researchers can use this concept of a ‘molecular fossil’ to investigate the evolutionary fitness of cholesterol precursors in processes that were likely to have been vital in the development of multicellular organisms. For example, is desmosterol sufficient for endocytosis? In conclusion, with the benefit of hindsight, it may not be too surprising that desmosterol could replace cholesterol in supporting proliferation and sterol homoeostasis in the study by Rodríguez-Acebes et al. . Still, while cholesterol itself may not be unique in its ability to sustain mammalian life, sterols very much like cholesterol, such as desmosterol, appear to be absolutely required.
ATP-binding cassette, subfamily A, member 1
sterol Δ24-reductase or 3β-hydroxysterol Δ24-reductase
liver X receptor
SREBP cleavage-activating protein
selective Alzheimer's disease indicator-1