Drugging the Undruggable: Transcription Factors

As we battle to expand the reach of modern therapeutics to treat diseases with high unmet need, we confront the limitations in our armamentarium. Most often we are limited by biology – we don’t understand the molecular drivers of disease and consequently don’t know which targets to aim at. But the more frustrating situation is to know the biology but to be unable to do anything about it because the target is “undruggable,” simply not amenable to intervention with our current therapeutic approaches. What if we could unlock these undruggable targets and zero in precisely on the molecular drivers of disease? Where would we start? In this and subsequent posts here, we’d like to share our thinking on this question.

One attractive class of biologically potent but largely undruggable targets is transcription factors (TF). TFs are master regulators of biology, capable of profoundly altering cell phenotype and behavior by coordinately regulating the expression of large sets of gene. Consequently, they punch considerably above their weight. Moreover, they act with an exquisite degree of selectivity by converting what is often a generic intracellular signal into a cell-specific response. But they are recalcitrant targets for small molecule inhibitors as, with few exceptions, they lack classic SM-friendly pockets. Instead, TFs do their business largely through protein-DNA and protein-protein interactions, mediated by relatively featureless surfaces or helices with side chains that reach into DNA grooves to make base-specific contacts.

MYC:  A prototypical TF in cancer that’s been out of therapeutic reach

TFs are particularly compelling targets in cancer since several of them are clear drivers of tumor growth. Consider, for example, MYC, one of the earliest-identified cellular oncogenes. MYC is a TF that affects the expression of numerous genes, with critical roles in cell cycle progression, biomass increase, apoptosis and tumorigenesis. MYC is dysregulated in roughly 70% of human cancers, often as a consequence of gene amplification. While its relevance to cancer was first discovered in Burkitt lymphoma, MYC is associated with cancers of the colon, breast, ovary, prostate, lung, liver, stomach and cervix.

MYC shows high cancer selectivity. Its expression is very low in normal cells. Tumors over-expressing MYC typically have other driver mutations. One notable example is ARF, which represses MYC activity in normal cells. Many cancers have both MYC over-expression and loss of ARF, leading to lack of control over MYC expression.

Despite the deep and broad association of MYC with cancer – its role as an oncogene was discovered 35 years ago – attempts to target MYC activity have repeatedly failed. The MYC  protein structure offers virtually nothing in the way of concave pockets, and so has proved refractory to the development of SM inhibitors. Attempts to target MYC expression with oligonucleotide drugs have been frustrated by poor delivery of these agents to tumors.

Numerous TFs in Cancer

MYC is not the only oncogenic TF. Alterations in ETS family members, especially through chromosomal translations as well as gene amplification, can drive tumor initiation, progression and metastasis in a variety of cancer types.  Additional cancer target TFs include STAT3 and other members of the signal transducer and activator of transcription (STAT) family, b-catenin, FOS, JUN, MYB, SOX2, RUNX1 and SMADs. Beyond primary tumor growth, TFs have been shown to be critical in epithelial-mesenchymal transition, tumor cell invasion and metastasis. These include TWIST1, Snail, and ZEB.

Another target whose inhibition could lead to broad anticancer activity is hypoxia-inducible factor 1-a (HIF1a). HIF1a is a subunit of a heterodimeric TF. It is the master transcriptional regulator of cellular response to hypoxia, a state of low oxygen levels found in solid tumors.  The dysregulation and overexpression of HIF1a by either hypoxia or genetic alternations have been strongly implicated in cancer biology, including tumor survival, metabolism, angiogenesis and chemoresistance.

TFs in Broad Disease Biology

TFs can act as master regulators of cell phenotype and lineage, with potentially powerful leverage on disease-relevant biology.  An excellent example is in the differentiation and activity of distinct T lymphocyte subsets. The TF TBX21 (encodes T-bet) drives Th1 cells, while GATA3 directs Th2 cells, and RORγT drives Th17 cells. FOXP3 drives the differentiation and activity of regulatory T cells (Tregs). The modulation of these TFs could be an effective treatment for inflammatory diseases or for promoting an anti-tumor immune response in an immuno-oncology application. Beyond T cells, a compelling TF target is IRF5.  IRF5 upregulation induces the differentiation of tumor stimulating M2 macrophages to anti-tumor M1 macrophages.

Bringing TFs within Therapeutic Reach

Various nucleic acid-based therapeutic strategies are being deployed to overcome the challenges in addressing targets that have not been druggable by SMs and antibodies.  Oligonucleotide therapeutics such as antisense or siRNA have been developed to reduce deleterious gene expression. More recently, CRISPR technology has been a focus as a means of correcting genetic lesions. But these modalities all face challenges in drug delivery in vivo. Successful delivery has been limited primarily to the liver and to local compartments (e.g., intrathecal injection into the spinal cord for the treatment of spinal muscular atrophy). In addition, the long-term safety of these therapeutics in chronic use is not yet clear.

Arrakis’s approach to modulating proteins that themselves are difficult to antagonize or agonize with a SM is to instead target the RNA encoding or regulating the protein. mRNA possesses secondary and tertiary structures that have numerous druggable pockets for SMs. Via selective binding to certain mRNA structures in functionally important regions, we intend to modulate the expression level of proteins whose activity we cannot attack directly. RNA molecules not only have druggable pockets, but our bioinformatics analysis suggests there is vast structural diversity enabling the identification of potent and selective binders. We believe that targeting RNA with small-molecule medicines can access a wealth of previously undruggable drivers of human disease, opening up new classes of targets, such as transcription factors. Stay tuned, as we share more of our thinking on new biology enabled by the technology we are developing here at Arrakis.