Co-founder, CSO, Junevity
Transcription factors are the most powerful genes in our body. They are master regulators and have been implicated in our biggest diseases, even aging itself. Yet, many in the pharmaceutical industry call these genes “undruggable” and “too risky” due to historical challenges in harnessing their power.
New technology is changing this. Specifically, siRNA, omics, and AI/ML are reaching maturity. Together, they are allowing scientists to finally understand and unlock the potential of transcription factors.
Co-founder, COO, Junevity
These technologies could open a new drug category and lead to a wave of FDA approvals. By reprogramming cells back to a healthy state, it could even change how we treat disease.
From novel to Nobel
The 2006 paper by Shinya Yamanaka, MD, PhD, set the biology world on fire by showing that activating a combination of just four transcription factors (Oct3/4, Sox2, c-Myc, and Klf4) could revert differentiated cells to pluripotency.1 If transcription factors can conduct this feat, what else could they do? His work led to the Nobel Prize in Physiology or Medicine in 2012.
Broad transcriptional change characterizes many of our toughest diseases, including cardiometabolic, immunology, inflammation, oncology, musculoskeletal, neurodegeneration, and aging more broadly. For example, roughly 3,000 genes in the aged adult brain (60-79 years old) were expressed differentially versus the younger adult brain (20-39 years old).2 Other patient datasets show similar levels of dysregulation in aging and disease, including in the brain, liver, heart, adipose, muscle, kidney, eye, and beyond.
Modulating a single downstream gene target likely would not be enough to fix all the dysregulation. Modulating a transcription factor can reset cells towards health by altering hundreds of genes—as evidenced in stem cell biology.
Undruggable and too risky?
Biotech and pharma have shied away from transcription factors. Less than four percent of FDA-approved drugs in the last five years target transcription factors, despite making up 10% of genes and their enormous potential. What is holding them back?
First, traditional small molecule drugs struggle to safely target transcription factors. These proteins lack defined binding pockets and are intrinsically disordered. They have large, challenging interaction surfaces. Targeting the DNA-binding location of a transcription factor is similarly challenging. It is hard to hit the right gene, and hard to avoid off-target effects, putting safety at risk.
Second, even when you can drug them, it is hard to identify them as disease drivers, given the difficulty of establishing genetic linkages to disease. Transcription factors have broad effects that can be unique and pleiotropic by cell type, development stage, and disease state. Because of the varying effects, clear genetic linkages to disease with transcription factors are challenging. For example, a transcription factor may have a role in normal embryonic brain development, yet later in life, cause liver inflammation and exacerbate type 2 diabetes. As another example, resmetirom, the first FDA-approved compound for metabolic dysfunction-associated steatohepatitis (MASH), targets THR-b, a transcription factor. MASH-associated genetic studies have not identified THR-b gene variants as established risk factors. Meanwhile, the drug is effective, safe, and on track to cross $1B in sales in 2026.
Third, understanding the downstream effects of transcription factors has been difficult, making it harder to predict efficacy and safety. Previously, biologists might have noticed that a certain transcription factor was differentially expressed in disease. This sometimes gave a helpful clue, but they were still in the dark about the full effects. Transcription factors can regulate dozens or even hundreds of genes. These regulatory effects differ by cell type, development stage, disease state, and more. Academic labs have been working on predicting regulatory networks for a long time, but the lack of large-scale data and advanced machine learning algorithms limited their efforts until recently.
Together, these barriers have increased the perceived risk of drugs targeting transcription factors and thus limited the funding and development.
Technology tailwind
How can we overcome these barriers? Three new technologies are reaching maturity at exactly the right time: siRNA, omics, and AI/ML.
First, siRNA. How do you solve the hard-to-drug nature of transcription factor proteins? Simple: skip the protein and silence the mRNA before it can become a protein. That’s what siRNA enables. The foundation for siRNA therapeutics was laid in 1998, when Andrew Fire, PhD, and Craig Mello, PhD, described the mechanism of RNA interference (RNAi)—a discovery that earned them the 2006 Nobel Prize in Physiology or Medicine. In 2002, Alnylam Pharmaceuticals launched to translate RNAi into medicines. Alnylam’s original pitch deck reportedly argued that a key use case of siRNA is for transcription factors. After 16 years of development, Alnylam achieved the first FDA approval for an siRNA therapeutic in 2018. Since then, seven siRNA drugs have been approved—averaging roughly one per year (Table 1).
siRNA can target specific tissues, which is especially important for ensuring safety with the broad effects of transcription factors. Chemically conjugating siRNA with N-acetylgalactosamine (GalNAc) enables selective delivery to hepatocytes, reducing off-tissue effects (all currently approved siRNA therapeutics are liver-targeted). However, extrahepatic delivery—encompassing targets in the central nervous system (CNS), muscle, and other tissues—is an area of intense preclinical and early clinical exploration. As delivery technologies expand the tissue addressable space, siRNA will continue to open new therapeutic opportunities, particularly for transcription factors.
Second, omics. How can scientists predict whether a transcription factor will be a good drug target, especially where genetic linkages aren’t obvious? Large amounts of data are needed to model the downstream effects. We are fortunate to be living through an exponential explosion of large-scale human genetic, transcriptomic, and proteomic datasets. For example, 10 years ago, roughly 100,000 cells of scRNA-seq data were available in public databases. Today, a single database, CELLXGENE, has over 100M cells, approximately a 1000-times increase, and they are promising a billion soon.3 Human datasets can provide insight into every gene in the cell, in a variety of situations (tissue, disease, age, etc.). The history of technology says that when you see long-term exponential growth in a critical area, like biological data, you pay attention.
Third, AI. Omics data is already far beyond the ability of humans to comprehend and analyze. AI/ML allows unpacking the complex dynamics from the data. Ten years ago, machine learning and bioinformatics algorithms helped analyze the early omics databases. While useful, the progress in AI/ML in the last decade has been breathtaking. The scale for AI compute (measured by the number of operations per second of all installed GPUs) has recently been doubling every three to four months. The maximum model parameter size has been doubling every 12 months. AlphaFold launched in 2021, ChatGPT in 2022, and Nobel Prizes were awarded to AI pioneers in 2024. Cutting-edge AI/ML are actively being applied to omics and predicting regulatory networks.
When two complementary technologies go exponential (in this case, biological data and AI), you stop whatever you’re doing and go work in that field.
Once dismissed as “undruggable” and “too risky,” transcription factors are now within reach. siRNA delivers the precision to drug them, and omics supplies the data for AI to pinpoint safe, effective targets.
100 new FDA-approved drugs by 2045?
Based on these technological breakthroughs, how many new transcription factor-targeting drugs could earn FDA approval in the next 20 years?
Even without the benefit of these technologies, there were nine FDA approvals for drugs targeting transcription factors in the last five years, which is roughly four percent of all approvals (Table 2). This suggests a pace of 36 FDA approvals across 20 years.
targeting therapeutics since 2021
Now, consider if transcription factors could earn 10% of approvals, which is roughly their share of total human genes. This share could be considered the minimum expected amount, given the importance of transcription factors. The FDA has been approving about 50 new drugs per year. If transcription factors earned a 10% share, that would mean five new drugs per year and 100 across the next 20 years.
The potential may be greater. Transcription factors are relevant to many of our biggest diseases, including most aging-related indications. With an aging global population, there is an acute and growing interest in treating these diseases, and funding will flow. siRNA delivery technology will continue to open up new tissues, with muscle, adipose, and CNS next in line, and kidney, eye, skin, and others to follow. With new tissues available for targeting, new drug targets will become feasible. Further, each new target can lead to multiple approvals. For example, 12 FDA approvals now target the JAK-STAT axis, a key transcriptional pathway.
While it is hard to know with precision, 100 new FDA approvals by 2045 for drugs targeting transcription factors is feasible.
Looking at the bigger picture, imagine the therapies for our toughest diseases by leveraging our most powerful genes. Advances in siRNA, omics, and AI unlock this opportunity. The momentum will build every year. Undruggable no more, biotech and pharma could end up investing hundreds of billions of dollars in this category. Billions of patients around the world would benefit, with less disease and healthier, longer lives.
References
- Takahashi K, Yamanaka S. Induction of pluripotent stem cells from mouse embryonic and adult fibroblast cultures by defined factors. Cell. 2006;126(4):663-676. doi:10.1016/j.cell.2006.07.024
- The GTEx Consortium. The GTEx Consortium atlas of genetic regulatory effects across human tissues. Science. 2020; 369:1318-1330. doi:10.1126/science.aaz1776
- CZI Single-Cell Biology, et al. CZ CELLxGENE Discover: A single-cell data platform for scalable exploration, analysis and modeling of aggregated data. Preprint on
bioRxiv. 2023.10.30; doi: 10.1101/2023.10.30.563174
Janine Sengstack, PhD, is co-founder and CSO of Junevity. Rob Cahill is co-founder and COO of Junevity.
