Potent Antiproliferative Activity of Steroidal Natural Products is Mediated by Oxysterol-Binding Proteins

          The bis-steroidal pyrazine cephalostatin 1 (structure shown below) is a potent antiproliferative natural product. The average GI₅₀ of cephalostatin 1 against a collection of sixty human cancer cell lines (often referred to as ‘the NCI-60’) is 1.8 nM. Two structurally related molecules, ritterazine B and OSW-1, exhibit a similar cytotoxicity pattern against the NCI-60, suggesting that all three molecules share a biological efficacy target or modulate the same cellular pathway. Of course, an important property of a potential anticancer therapeutic agent is selective toxicity to cancer cells compared to nonmalignant ones.  In 2005, Zhou and co-workers published a study that examined the selectivity of the potent anticancer activity of OSW-1. The authors also initiated preliminarily investigations into the mode of action of this highly cytotoxic cholestane glycoside natural product.

          The 2005 study employed the MTT reductase assay to determine the cytotoxic potency of OSW-1 against various (cancerous and normal) cultured cell lines. In brief, this is a colorimetric assay to assess cell viability by quantifying mitochondrial enzymatic reductase activity which often (but not always) correlates with proliferation. The yellow tetrazolium salt (structure below) is reduced in living cells to a purple formazan dye that absorbs light between 500-600 nm. Interestingly, OSW-1 was found to display a degree of selectivity with regard to antiproliferative activity against cancer cells compared to nonmalignant cells. For example, the IC₅₀ of OSW-1 against malignant brain tumor cells is 0.05 nM versus 7.13 nM for normal astrocytes. While OSW-1 is still very toxic to normal cells, this amounts to a selectivity index of ~150. Transmission electron microscopy was then used to examine mitochondrial morphology changes associated with the observed cytoxicity. Evidence of damage to the mitochondrial membrane and cristae suggested that OSW-1 affects mitochondrial respiration and/or metabolic function. Structural and functional damage to the mitochondria was shown to trigger activation of a calcium-dependent apoptosis pathway. Notably, the cellular target of OSW-1 and the dimeric steroids shown above was recently identified by Shair and co-workers as a cytoplasmic mammalian receptor known as oxysterol-binding protein (OSBP).

          The authors used affinity purification to identify cellular proteins that bind to analogues of OSW-1, modified on the eastern disaccharide moiety. The affinity reagent shown below is covalently linked to a Sepharose polymer resin. Affinity chromatography of HeLa-S3 (for the story of HeLa cells, see this) cell lysates with this solid-supported steroid derivative followed by protein purification by gel electrophoresis generated an enriched protein band that was identified as OSBP by mass spectrometry techniques. Pretreatment of the lysate with soluble OSW-1 competed away the band, providing evidence of a specific ligand/receptor binding event. A related paralog of OSBP (OSBP-related protein 4L or ORP4L) was also ‘pulled down’ by the affinity reagent. Competition binding assays with tritiated 25-hydroxycholesterol (structure shown above), a high-affinity ligand of OSBP, indicated that cephalostatin 1, OSW-1 and ritterazine B all directly bind to OSBP with double-digit nanomolar K_i values.

          Small hairpin RNA (shRNA) interference technology can be utilized in a cellular context to silence gene expression. In this type of experiment, shRNA is cleaved by cellular machinery into small interfering RNA (siRNA), which then binds to the RNA-induced silencing complex (RISC). The activated complex can now bind to and cleave mRNAs that match the siRNA already bound to the RISC. Interestingly, when reduction of OSBP expression is effected by shRNA knockdown, inhibition of cancer cell (HCT-116) growth is not observed. However, diminished OSBP expression did sensitize cells from two cell lines to sterols (cephalostatin 1 and OSW-1) by four to nine-fold. The authors account for this apparent discrepancy by hypothesizing that a complete inhibition of OSBP cellular function is required for cell growth inhibition, while shRNA knockdown alone reduces OSBP levels by only about 85%. Given that OSBP has already been implicated in a variety of lipid transport and lipid metabolism processes (listed in the graphic below, left side), the Harvard study has uncovered yet another interesting link between lipid processing and cancer cell proliferation. The research raises questions about the existence of an endogenous cytotoxic OSBP ligand that might be implicated in tumor growth and/or metastasis. It will also be interesting to characterize and compare the specific ligand/receptor binding site interactions (binding modes) for monomeric steroids such as OSW-1 as compared with those of the bis-steroidal dimers.