1 Therapeutic Oligonucleotides — The Road Not Taken

In this issue of Clinical Cancer Research, Hong, et al (1) report on the safety and tolerability of LY2275796, a second generation antisense oligonucleotide (ASO) targeted to the eukaryotic translational initiation factor eIF-4E mRNA, a gene that has been reported to play a role in tumor initiation. In the study, ASO MTD and biologic effective dose was set at 1000 mg, due to a DLT of grade 3 fatigue at the 1200 mg dose, and target inhibition was assessed by serial tumor biopsies. At first look, the ASO appeared to achieve target silencing, but closer inspection reveals that the housekeeping genes (e.g., β-actin) were also downregulated (by 64%, compared to 80% for the target mRNA), despite in vitro results suggesting excellent target specificity (2). But even so, why (since the target was in fact silenced) didn’t “target inhibition” achieve “tumor inhibition” in this phase I trial? The problem, in part, is that human cancer is a multitude of processes, pathways, and “cross talk” with profound redundancy; cells can often “bypass” the inhibited pathway with minimal detriment to their malignant potential (3). LY2275796 therefore fails in part because the preclinical models in which it was evaluated do not accurately recapitulate human clinical cancer. Mice and their xenografts, it has oft been suggested, are not men. Nevertheless, as the authors suggest, combining gene-targeted with cytotoxic therapy is clearly the way forward. We strongly agree, but though this point has been understood for decades, we still do not have a single FDA-approved approved ASO therapeutic agent in cancer. What has happened? The answer is that each clinical trial in cancer of an ASO, in addition to its own unique problems, also suffers from the problems shared by all ASO therapeutics, including RNAi. Unfortunately, despite the passage of near 25 years for ASOs, our level of understanding about fundamental processes that govern in vivo efficacy of therapeutic ASOs, particularly in cancer, is almost nil. DNA is negatively charged, and in an 18mer antisense molecule such as, for example, the anti-Bcl-2 ASO Oblimersen (4), there are 17 negative charges. Substitution of a sulfur for an oxygen atom at each phosphorus, forming a phosphorothioate ASO, the type of ASO employed in virtually all clinical cancer trials, maintains the negative charge. However, the melting temperature (Tm) of the duplex formed between the target mRNA and a phosphorothioate antisense ASO will almost always be significantly depressed (5). In addition, while first generation