New ideas and progress in targeting "undruggable" transcription factors

Transcription factors regulate almost the entire genome
by binding to specific DNA-binding sequences to help turn genes on or off.
All transcription factors (TFs) contain two generic protein domains: a DNA-binding domain (DBD) that binds to a specific DNA sequence, and an effector domain
that recruits coactivators.
Many transcription factors also contain one or more regulatory domains that are commonly used to regulate transcription factor localization and function, such as STAT3, which enables dimerization and regulates the expression
of target genes through its SH2 domain.

It is estimated that there are at least 1,600 TFs in the human genome, of which about 19% are associated
with disease phenotypes.
Such as p53, Myc, androgen receptor (AR), estrogen receptor (ER), hypoxia-inducible factor (HIF), NF-κB and other transcription factors, they are involved in various cancer-promoting events, p53 is the most studied tumor suppressor, in more than 50% of human cancers occur functional acquisition mutation inactivation or become carcinogens
.
Nuclear receptor AR and ER are closely related
to the pathogenesis of endocrine-driven prostate cancer and female cancers such as breast, ovarian, and endometrial cancers, respectively.
HIF-1 is overexpressed
in most solid tumors.
NF-κB is a key transcription factor that classically activates macrophages and mediates the production of a large number of inflammatory factors
.

Transcription factors play a fundamental role in selective gene regulation, so they have a higher specific disease regulatory capacity than upstream signaling proteins such as kinases, and because gene targets are specifically controlled by DNA binding, a single TF usually regulates only a limited set of gene targets, so this inhibitor is not prone to compensatory resistance mechanisms
common to other drugs.
Therefore, targeting abnormal transcription factors is an effective strategy
for treating diseases.

However, due to the structural heterogeneity of transcription factors and their lack of active sites, they have traditionally been considered difficult to druggable
.
Studies have shown that transcription factors have highly dynamic protein-DNA and protein-protein interactions, with protein-DNA interaction interfaces usually convex and positively charged, while protein-protein interaction interfaces are usually flatter and lack binding bags, making it difficult for small molecules to target directly
.

Among them, nuclear receptors (NRs) and the bHLH-PAS family are special TF families with a DNA-binding domain (DBD) and a ligand-binding domain (LBD), and the bHLH-PAS family is formed by dimerization with its family members
.
When bound to small molecule ligands, conformational changes in NRs are induced, regulating their interaction with various transcriptional co-regulators such as co-activators or co-inhibitors, thereby controlling the expression of downstream target genes, and the small molecule inhibitors currently developed mainly target these two classes of TFs
.

Nuclear receptors include androgen AR, estrogen ER, etc
.
The binding of AR to AR LBD leads to nuclear translocation and initiates target gene expression, which is a key driver of prostate cancer, so AR is an important therapeutic target
.
AR antagonists bind to LBD to block the AR signaling pathway, and enzalutamide, aparutamide, and darolutamide are second-generation AR antagonists that have recently been approved by the FDA for the treatment of human prostate cancer with greater efficacy and fewer
side effects than the first generation.
In addition, some AR antagonists are undergoing clinical trials, such as proxalutamide and EPI-7386
.
To avoid undesirable side effects such as the metabolic effects of AR signaling pathways on muscle and bone synthesis, selective AR modulators (SARMs) such as GTx-024, AR or SARM have been synthesized and developed over the past two decades to bind to AR LBD and induce conformational changes, thereby promoting the recruitment of coactivators and activating gene expression
.

Approximately 75% of breast cancers are ER-positive, and ER signaling is a key driver of
breast cancer.
The selective ER modulator SERM (e.
g.
, tamoxifen) binds to LBD and blocks the conformational changes required to recruit coactivators, thereby reducing target gene expression
.

Although inhibitors or selective regulators of AR or ER can reduce target gene expression, the emergence of intrinsic and acquired resistance limits its application
.
PROTAC technology that induces degradation of target proteins in an event-driven manner shows great promise in overcoming this problem and is described
below.

Finally, although NR is the most druggable due to LBD, targeting all NR family members remains a very challenging task
.
Examples include NRs known as orphan receptors that do not have a designated endogenous ligand or distinct ligand-binding pocket
.
In addition, in some diseases, the presence of NR splicing variants can render traditional drugs ineffective, such as AR-V7, which is a splicing variant
of AR.
Traditional NR-targeted drugs are bound to
their LBD pockets.
More recently, NR coactivator binding inhibitors have also been developed to modulate NR activity
.

Protein-protein interactions (PPIs) play an important role in many physiological processes and are often dysregulated
in disease.
Therefore, PPIs are also considered potential therapeutic targets
.
Examples include p53-MDM2 inhibitors and menin-MLL inhibitors
.
In cancer, PPIs between TF p53 ("guardians of the genome") and ubiquitin E3 ligase MDM2 are often used as a mechanism
by which cancer cells downregulate p53 levels through the ubiquitin-proteasome system to escape apoptosis.
However, p53 is a highly disordered TF, so developing a small molecule that binds to p53 and protects it from proteasome degradation is challenging
.
Fortunately, MDM2's p53 binding site is a relatively small and well-defined hydrophobic pocket, so small molecule drugs that inhibit their interaction, such as HDM201 and RG7388, can be designed to inhibit the ubiquitination
of p53.

The interaction of the coactivator factor menin with MLL fusion proteins is essential for MLL fusion-positive leukemia, so the development of small molecule inhibitors is effective in the treatment of leukemia, such as MI-538 and MI-1481, which have been developed, to inhibit leukemia
by blocking their interaction.

However, protein-protein interfaces are often hydrophobic, making it difficult to develop small molecule compounds
with desirable druggable properties.
Compared with the active site of the enzyme or the binding site of the receptor, the protein-protein interaction interface is flatter and more stable, which is difficult to target.

In addition, protein-protein interfaces are relatively featureless and lack a hydrophobic cavity
that can fully accommodate small molecule ligands.
Therefore, the future direction of protein-protein interaction inhibitors should be to prioritize the design of new chemical, biological, and computational tools, and expand the chemical library to screen effective small molecules
.

In cells, unnecessary or misfolded proteins must be destroyed to ensure cellular protein homeostasis, a process tightly controlled
by intracellular ubiquitin proteasome mechanisms.
Proteins of interest (PROTACs) targeted proteolysis were developed by hijacking endogenous E3 ligase degradation of proteins of interest (POIs)
by the ubiquitin-proteasome system.
PROTAC is a heterobifunctional molecule containing a POI-bound ligand and an E3 ubiquitin ligase covalently linked ligand, resulting in degradation
of protein targets via the ubiquitin-proteasome system.

Unlike the occupancy-driven mode of operation of traditional SMIs, PROTAC degrades POIs in an event-driven manner, eliminating the need to isolate ligands to identify targeted POIs, providing a new strategy to inhibit previously undruggable targets with strong target selectivity and the potential to overcome resistance common to traditional SMIs
.
ROTACs eliminate all functions of the target protein, not just those targeted by specific small molecule inhibitors
.
And this method is also catalytic, allowing a single molecule to mediate the destruction of multiple target proteins through repeated binding and degradation cycles, which can reduce the frequency
of dosing.
However, there are some limitations, such as poor pharmacokinetic profiles, heavy reliance on E3 ubiquitin ligase, and potential off-target toxicity
.

There are currently many
studies on PROTACs targeting AR.
In 2008, Crews Laboratories reported the first PROTAC AR degrader, which used MDM2 inhibitors as E3 ligase ligands and bicarutamide as AR antagonists
.
Although compound 131 only hydrolyzes AR proteins in cells at micromolar concentrations, it provides an important proof of
concept.

Compounds 135-137 designed PROTAC with highly potent AR antagonists and VHL ligands, despite both have excellent degradation capacity, have low oral bioavailability in animals, which limits their further development
.
ARD69 and ARD266 show that PROTACs with high rigid joints increase activity
.

AR PROTAC, DESIGNED BY HIGHLY EFFICIENT AR LIGANDS AND CRBN LIGANDS WITH RIGID LINKERS, HAS SHOWN EXCELLENT ARR DEGRADATION POTENCY, POTENT CELL GROWTH INHIBITORY ACTIVITY, GOOD PK CURVE, AND POTENT ANTI-TUMOR EFFICACY
.
At present, ARV-110 has entered the second phase of clinical trials, it is a potent inhibitor, CRBN ligand linked by rigid linkers, has a very excellent AR degradation effect, in wild type and prostate cancer patients with AR T878A and H875Y mutants
.
Avinas also has an oral PROTAC drug, ARV-766, in Phase I clinical trials
.

Although a large number of agonists and antagonists have been developed, only a few AR inhibitors and their derivatives have been reported to be used in AR PROTAC, including bicalutamide, enzalutamide, and compound 72 from Pfizer, and only CRBN ligands provide good oral bioavailability
.

Signal transduction and transcriptional activator 3 (STAT3), a member of the STAT family, is activated in response to a variety of cytokine, growth factor, and oncogene signals, and its activation is often associated with
poor prognosis.
STAT3 is dimerized by the interaction of one Tyr705 phosphorylated monomer with the binding pocket of the SH2 binding domain of another monomer, with subsequent activation of the target gene
.
Therefore, inhibitors of the STAT3 SH2 domain can inhibit STAT3 activity
.
However, STAT3 and other STAT family members share a highly structurally homologous SH2 domain, making it difficult to obtain highly selective STAT3 inhibitors
.
And the monomeric STAT3 protein also has transcriptional activity, so STAT3 SH2 domain inhibitors that block STAT3 dimerization only partially inhibit the gene transcriptional activity
of STAT3.
The researchers developed a highly potent STAT3 degrader, SD-36
, which is permeable to cells, by optimizing the previous STAT3 SH2 domain inhibitor CJ-887.

Forkhead box M1 (FOXM1), a transcription factor essential for cell division, has been linked to
poor outcomes in cancer patients.
A computer modeling of the DNA-binding domain of FOXM1 identified a suitable small molecule that inhibits the FOXM1-DNA interaction, and designed a CRBN-based PROTAC that significantly degrades the FOXM1 protein and has antitumor activity
.

These small molecule-based PROTACs offer improved oral bioavailability, low toxicity, and superior efficacy compared to classical transcriptionally active inhibitors, and these PROTACs targeting transcription factors are currently in
phase I and II clinical trials.

As mentioned above, SMIs can target some transcription factors such as AR, ER, and STAT3, allowing PROTACs
to be developed based on these SMI designs.
However, most transcription factors remain undruggable or difficult to treat
.
Transcription factors recognize cis-regulatory elements in promoter and enhancer DNA in a sequence-specific manner, but these key functional sites are also difficult to design SMI
.
Several studies have used these oligonucleotide sequences as ligands to develop PROTACs for targeted degradation of transcription factors
.

Crews has developed a TRAFTAC (chimera targeting transcription factor) that uses a heterobifunctional double-stranded DNA/CRISPR-RNA chimera to bind both the transcription factor of interest (TOI) and the dCas9-HaloTag7 fusion protein (dCas9-HT7, used as an intermediate protein).

CRISPR-RNA binds to ectopically expressed dCas9-HT7, which in turn recruits the VHL E3 ligase complex in the presence of haloPROTAC
.
The double-stranded DNA portion of the chimera then recognizes TOI and brings the TOI and E3 complexes close to induce TOI degradation
.
As a proof of concept, they first selected the typical oncogene NF-κB as the target
.
NF-κB-TRAFTAC binds to dCas9-HT7 fusion protein and NF-κB subunit p50 and induces significant degradation
of p50 after haloPROTAC treatment.

To develop a more direct way to degrade transcription factors, the researchers selectively degraded TOIs, called TF-PROTAC,
by linking DNA oligonucleotide strands to E3 ligase ligands through a click reaction.

Commercially
available azide-modified DNA oligomers (N3-ODN) are bound to bicyclooctyne (BCN)-modified VHL ligands with various linkers (VHLL-X-BCN) by metal-free copper-responsive azide-alkyne cycloaddition (SPAAC) reactions.
After a simple purification process to remove excess ligand VHLL-X-BCN, TF-PROTAC is ready to be transfected into cells to be degraded
by selective recognition of DNA oligonucleotide strands and binding to specific TFs.

Two VHL-based TF-PROTAC, NF-κB-PROTAC (dNF-κB) and E2F-PROTAC (dE2F), have been developed, which effectively degrade endogenous p65 and E2F1 proteins in cells, respectively, and show excellent antiproliferative activity
in cells.
Overall, TF-PROTACs provide a versatile platform for selective degradation of TFs, but their clinical application remains to be determined
due to the lack of in vivo data.

Although small molecule-based PROTACs have shown great promise in the clinic, finding the right ligand to selectively target transcription factors remains challenging
.
Since oligonucleotide DNA is a natural ligand for transcription factors, oligonucleotide-based PROTACs are universal and flexible, and can target other transcription factors by replacing the DNA sequence of chimeric oligonucleotides, paving the way
for targeting traditionally undruggable transcription factors.
In addition, in view of the stability, tissue permeability, immunogenicity and low toxicity of oligonucleotides, the oligonucleotide-based PROTAC platform offers broad prospects
for the clinical development of therapeutic PROTACs.

Nevertheless, there are numerous challenges to their preclinical and clinical development, such as the PROTACs of oligonucleotides always carry a large negative charge, and the second-best pharmacokinetics and relatively low efficacy may limit their future development
.

As a future extension of this technology, RNA-binding proteins can form RNA-protein interactions through RNA recognition sequences to participate in a variety of biological processes, including RNA transcription and mRNA translation, and given the similarities between transcription factors and RNA-binding proteins, oligonucleotide-based PROTACs provide ideas
for the targeted degradation of such proteins.
As the technology matures, it is worth looking forward to
identifying the clinical application prospects of DNA and RNA-based PROTACs.

References

       1.
Nat.
Rev.
Drug Discovery 2021, 20, 669?688.

       2.
Microbiol.
Immunol.
2010, 349, 25?60.

       3.
Molecules.
2020; 25: 2448.

       4.
Cell 95, 927–937 (1998).

       5.
Nat Rev Cancer.
2015; 15: 701–11.

       6.
Proc.
Natl.
Acad.
Sci.
U.
S.
A.
2001, 98, 8554?8559.

       7.
J.
Med.
Chem.
2022, 65, 8772?8797

       8.
Cancer Cell 36, 2019 498–511.

       9.
J.
Med.
Chem.
2021, 64, 17098?17114

       10.
Cell Chemical Biology 2021.
28, 648–661

       11.
J.
Am.
Chem.
Soc.
2021, 143, 8902?8910

       12.
J.
Med.
Chem.
2022, 65, 10183-10194