Seeing the Future of Cancer-Associated Transcription Factor Drug Targets

Abstract

IN THE ERA OF GENOMIC MEDICINE, DISEASES ARE PERceived as manifestations of abnormal gene expression patterns. Cancer is the prime example of this concept. Transcription factors, the terminal effectors of the malignant gene expression patterns, occupy a central role in all 6 classic hallmarks of carcinogenesis: self-sufficiency in growth stimuli, insensitivity to antigrowth signals, endless replication, angiogenesis, tissue invasion, and metastasis. An ever-increasing number of pathognomonic tumor-specific genetic and epigenetic events have been shown to directly inactivate tumor suppressor or activate oncogenic transcription factors (TABLE). Alternatively, modulation of transcription factor activity may be the critical consequence of genetic and epigenetic alterations that affect tumor suppressor genes or oncogenes implicated in upstream signal transduction pathways. For example, all distinct chromosomal translocations found in mucosa-associated lymphoid tissue lymphomas culminate in induction of nuclear factor B (NFB) transcriptional activity without individually involving NFB directly. This plethora of tumor-specific genomic alterations that directly involve transcription factors or indirectly modulate transcription factor activity highlights the potential of transcription factors as anticancer drug targets. However, with the exception of drugs targeting transcription factors of the nuclearhormone-receptor superfamily (eg, tamoxifen targeting estrogen receptors), pharmacological manipulation of transcription factors using Ehrlich’s classic magic-bullet concept remains elusive, and transcription factors are generally considered not good drug targets or inferior drug targets. Even diseases with well-defined genetic alterations, implicating transcription factors are currently managed with chemotherapy. The modular composition of transcription factors renders them ideal candidates for direct and selective pharmacological interference. Although drugs could target transcription factors indirectly by modulating upstream signaling pathways, this approach lacks specificity and has limited potential given the redundancy and extensive cross-talk between upstream signaling cascades. Transcription factors have structurally and functionally distinct domains (ligand-binding, dimerization, transeffector, DNA-binding, nuclear-localization, and regulatory domains), which may be targeted directly by smallmolecule drugs or by targeting the posttranslational modifications (ie, phosphorylation) that modulate their function. The presence of chimeric fusion proteins consisting of functional transcription factor domains in certain malignancies provides an additional therapeutic opportunity because these chimeric proteins are absent in nonmalignant cells, thereby providing specificity to the action of small-molecule drugs. Nevertheless, traditional drug discovery approaches such as high-throughput chemical screening of compound libraries or de novo drug synthesis by organic chemistry have been only modestly successful in the case of transcription factors. These approaches are limited because unlike enzyme active sites, protein-ligand, protein-DNA, or protein-protein interfaces generally lack hot spots or deep binding sites for small molecules, and their surfaces greatly exceed the binding area of small-molecule drugs. Another challenge is the nuclear localization of transcription factors and the associated difficulty to selectively target nuclear biochemical events. These limitationsof traditionaldrugdiscoveryapproacheshave largely contributed to the doctrine of transcription factors being inferior drug targets, whereby only transcription factors of the nuclear-hormone-receptor superfamily, which are surfacecytoplasmic receptors that translocate to the nucleus after binding to their ligands, are considered good drug targets. Novel approaches in drug design and delivery hold promise for overcoming many of the aforementioned challenges. Structure-based design combines techniques such as nuclear magnetic resonance spectroscopy, protein– ligand x-ray crystallography, and molecular modeling with powerful computational algorithms of conformational analysis and ligand-docking for design of novel small-molecule inhibitors. These approaches have been successfully applied to develop inhibitors of the p53–MDM2 protein– protein interaction, which was previously considered undruggable. Examples of such inhibitors are the nutlins, the spirooxindoles (discovered via molecular modeling), and JNJ-26854165, a tryptamine-derived MDM2 inhibitor with activity against glioblastoma and lung cancer models, which is undergoing phase 1 trial evaluation in solid tumors. Innovative chemical genomic approaches have also identifiedcandidatetranscriptionfactor–basedanticancerregimens. Gene expression–based high-throughput screening initially defines gene expression signatures that distinguish between