Drug Discovery Pro

🧪 A 7-Year Lesson in Keap1–Nrf2 Drug Discovery: When Potency Misleads Medicinal Chemistry

The story of Keap1–Nrf2 signaling pathway inhibitors offers a powerful reminder for medicinal chemists working in modern Design–Make–Test–Analyze (DMTA) cycles.

It begins with a promising discovery.

🔬 2013 — A powerful hit appears https://www.sciencedirect.com/science/article/abs/pii/S0968089613003453?via%3Dihub

A high-throughput screening campaign from the Evotec Lead Discovery Library identified a naphthalene-based hit as an inhibitor of Kelch-like ECH-associated protein 1, disrupting the interaction with Nuclear factor erythroid 2–related factor 2.

The molecule looked excellent on paper:

But the real story had only just begun.

⚙️ 2014 — Structure-guided optimization

Researchers led by Jiang et al. modified the scaffold into a dicarboxylate derivative, achieving nanomolar potency through structure-based design.

• IC₅₀ = 25 nM
• Strong binding
• Clear biochemical activity
https://pubs.acs.org/doi/10.1021/jm5000529
Encouraged by the activity, the group continued exploring the scaffold for diagnostic probes.

The chemistry looked promising.

Yet a deeper issue remained hidden.

🧬 2018–2020 — The optimization struggle
https://pubs.acs.org/doi/10.1021/acs.jmedchem.8b01133
The group of Terry Moor attempted to turn the hit into a drug-like molecule.

The objectives were clear:

• Improve metabolic stability
• Enhance solubility
• Achieve cellular activity

By 2020, these goals were partially achieved.

But there was a price.

The potency dropped to IC₅₀ = 73 nM.
https://pubs.acs.org/doi/abs/10.1021/acs.jmedchem.9b01074
Despite seven years of work across two groups, the naphthalene chemotype proved extremely difficult to develop without sacrificing activity.

🧭 2022 — A different strategy
https://pubs.acs.org/doi/10.1021/acs.jmedchem.2c00830

Instead of further forcing the original scaffold, Bach et al. explored chemical space through virtual screening.

They identified a fluorene-based hit — a completely different chemotype.

Using fragment growing, they optimized the structure to achieve:

• Ki = 0.28 μM
• Strong binding to Keap1

And more importantly:

A developable scaffold.

🚀 2024 — The breakthrough
https://pubs.acs.org/doi/10.1021/acs.jmedchem.4c01221

The fluorene scaffold was fully optimized.

The result:

• 85-fold potency improvement
• High metabolic stability
• Strong cellular activity
• No compromise in potency

After years of struggle, the problem was not optimization.

The problem was the starting chemotype.

💡 The Medicinal Chemistry Lesson

In drug discovery, potency can be seductive.

But potency alone does not define a good hit.

The Keap1–Nrf2 story shows that:

✔ A highly potent hit may still be chemically non-developable
✔ Forcing optimization on the wrong scaffold can waste years
✔ Exploring chemical space can unlock entirely new optimization trajectories

Sometimes the fastest DMTA cycle is not optimizing the hit.

It is replacing it.

🧠 For medicinal chemists:

Before committing to a hit, always ask:

Is this scaffold truly developable?

Because in modern drug discovery, the quality of the starting chemotype often determines the success of the entire program.