The problem: a dark kinase with no tools and no structure
HIPK4 sits in the unlit corner of the kinome. Knockout work implicates it in spermiogenesis (F-actin/keratin-scaffolded sperm-head shaping) and in cutaneous squamous cell carcinoma, where it looks like a negative regulator of keratinocyte differentiation. But there has been no high-quality selective inhibitor and no deposited HIPK4 structure — the protein resists efficient expression and crystallization — so nobody could test whether its genetic phenotypes are pharmacologically addressable.
The strategy: borrow bosutinib’s pharmacophore, delete its promiscuity
Bosutinib is a potent but famously broad-spectrum 3-cyanoquinoline kinase inhibitor. Across deposited co-crystals its bioactive conformation is highly conserved, with a flexible solvent-exposed tail doing little for binding. The Frankfurt/SGC team (Zerva, Raig, Knapp, Hanke and co-workers, J. Med. Chem. 2026) made the elegant bet that macrocyclization — tethering the quinoline 6-O to the back-pocket aryl — would pre-organize the scaffold into fewer conformations and buy selectivity, following the lorlatinib/pacritinib playbook.

The chemistry: build the core, then close the ring
The route assembles a 4-chloro-6-hydroxy-7-methoxy-3-cyanoquinoline core (16), installs the back-pocket aniline by SNAr with a 2-aminophenol to give a linear bis-phenol, then performs the decisive step: an intramolecular double Williamson etherification with a short bis-electrophilic linker that stitches the two phenols into a macrocycle. Systematically walking the linker length, the solvent-exposed R¹, and the back-pocket R² delivered AZ137 (28e) — the Ar⁶ / diethylene-glycol-tether macrocycle.

The structural read-out: a preorganized macrocycle in a surrogate pocket
With no HIPK4 structure available, the team crystallized their macrocycles in CLK3, the closest tractable surrogate (newly deposited as PDB 29MP/29MO), and confirmed a preorganized binding conformation consistent with an AlphaFold HIPK4 model. Shown here is a macrocyclic inhibitor bound in CLK3 (PDB 6Z51) — the same kinase used as the structural proxy — threading its ether-linked ring past the hinge region (near Leu290, one of the few residues that differ from HIPK4’s Met143).

The lead: AZ137 (28e)
- HIPK4 IC50 = 11 nM; cellular EC50 = 76 nM
- Exceptional selectivity across three orthogonal panels — in the Eurofins scanMAX 468-kinase screen at 100 nM, HIPK4 (PoC 7.3%) was essentially the only genuine hit; 30–80× over the nearest CLK isoforms
- High solubility, no detectable cytotoxicity
- Cellular target engagement confirmed via HIPK4-dependent F-actin remodeling
- Shipped as a probe set with a structurally matched negative-control compound (>100× less active)
Why it matters
- First high-quality, selective HIPK4 chemical probe — the reagent needed to de-orphanize a dark kinase and test its SCC/reproductive biology pharmacologically.
- A clean demonstration that macrocyclization is a selectivity tool, not just a PK trick, for the tightly conserved CMGC group.
- Delivering a probe without a target crystal structure — leaning on a surrogate kinase plus an AlphaFold model — is an increasingly common, and underrated, medicinal-chemistry capability.
Reference & image credit: Zerva, A.; Raig, N. D. et al. “Macrocyclization of Broad-Spectrum Kinase Inhibitor Bosutinib Leads to Potent and Selective Quinoline-Based HIPK4 Inhibitor AZ137.” J. Med. Chem. 2026, 69, 20616–20638. DOI: 10.1021/acs.jmedchem.6c01320. Design figure reproduced under a CC-BY 4.0 license © 2026 The Authors, published by the American Chemical Society. Synthesis scheme redrawn and binding-site figure (CLK3 surrogate, PDB 6Z51) rendered by Astinova Bio Labs.
Macrocyclization, conformational restriction, original route design, structure-based rationale — this is the work Astinova Bio Labs does for discovery partners. Talk to us about your scaffold.