Preclinical oncology
Making chemotherapy work again in colorectal cancer.
We design isoform-selective PHD1 (EGLN2) inhibitors to restore chemotherapy sensitivity in microsatellite-stable colorectal cancer. The program is preclinical, and enzyme-level selectivity is the next result we are generating.
How it is designed to work
Prime, then activate.
Resistant tumor cells lean on PHD1 to adapt. The design logic is to block PHD1 so a latent p53 death program is primed, then let standard chemotherapy activate it. Click through the steps.
This is the proposed mechanism. Enzyme and combination data come next.
The problem we target
Microsatellite-stable (MSS) colorectal cancer responds poorly to immunotherapy and depends on chemotherapy that tumors eventually resist. No approved agent restores that lost chemosensitivity.
Every clinical HIF-prolyl-hydroxylase inhibitor to date is pan-PHD and was developed for anemia. Evolvia is designing for selectivity toward PHD1, the isoform tied to the p53/NF-κB apoptotic axis, by targeting regions outside the conserved catalytic core.
Sources: Colorectal Cancer Alliance; Front. Oncol. 2022.
Prime
Selective PHD1 inhibition is designed to prime a latent p53 death program inside resistant tumor cells.
Activate
Standard chemotherapy (oxaliplatin, SN-38) then activates that primed program.
Spare
The design goal is selectivity that leaves PHD2 and PHD3 alone, to avoid HIF-driven toxicity. Enzyme assays will test this.
The differentiator
One enzyme, not three.
Today's PHD drugs block all three PHD enzymes. PHD2 is the main oxygen sensor behind red-blood-cell production, which is why those drugs were built for anemia and carry HIF-driven liabilities. We are designing for PHD1 alone.
Developed for anemia. Hitting PHD2 brings cardiovascular and thrombotic liabilities.
Block PHD1 and leave PHD2 and PHD3 alone. This is a design goal. Enzyme assays will test it.
Updated October 2026
Where the evidence stands
Preclinical, with computational and early cell data. Each step says what has been shown and what has not.
- Computational screenComplete
- Compounds madeComplete
- Cell studiesEarly data
- Enzyme selectivityNext
- CombinationNext
- ClinicLater
| Computational screen | 88 compounds assessed across PHD1, PHD2, PHD3 and CP4H. | Complete |
|---|---|---|
| Compound supply | EBP-4 and one analogue synthesized for biochemical testing. | Complete |
| p53-axis engagement | Western blot and qPCR in HCT-116 colorectal cancer cells. | Early data |
| Enzyme selectivity | IC50 across PHD1, PHD2 and PHD3 at an outside lab. | Next |
| Chemotherapy combination | Oxaliplatin and SN-38 in p53 wild-type and mutant models. | Next |
| Evolvia AI candidate ranking | Held until pose recovery, enrichment, counter-screens and biochemical calibration are met. | Locked |
No clinical data. Early cell results come from single experiments and need replication.
Evolvia AI platform
Candidates earn a ranking. They do not start with one.
Evolvia AI is the computational engine behind our programs. It checks protein structures, metal chemistry, docking poses and controls in one traceable workflow, and it holds back any ranking until the chemistry holds up.
Updates
- Evolvia AI redocking now models crystal contacts
The receptor-preparation step adds crystal-symmetry neighbours, with automated tests. Status page updated.
- HDAC structural qualification published
The validation ledger lists what passed, the method limit found, and what remains locked.
- Preprint: HIF-P4H-1/EGLN2 stabilizes IKKα
bioRxiv, not peer-reviewed. It supports the biology behind the PHD1 program.
Published science
Built on published biology.
Our PHD1 program builds on peer-reviewed work and a recent preprint from our founder.
Questions, answered plainly
Is EBP-4 a drug I could take?
No. EBP-4 is a preclinical tool compound. It has not been tested in people, and its enzyme selectivity has not yet been measured.
What does "prime, then activate" mean?
Selective PHD1 inhibition is designed to prime a latent p53 cell-death program. Standard chemotherapy, such as oxaliplatin or SN-38, is then meant to activate it. The idea is a combination effect, not a stand-alone cytotoxic drug.
Why PHD1 and not PHD2?
PHD2 is the dominant oxygen sensor behind red-blood-cell production. Drugs that block it were built for anemia and carry HIF-driven liabilities. PHD1 is the isoform tied to the p53 and NF-κB cell-death axis, so that is the one we are designing against.
What does Evolvia AI do, and does it pick the drugs?
It is a computational engine that checks structures, metal chemistry and docking poses, and logs every decision. It does not replace experiments. Candidate ranking stays locked until its validation gates pass and biochemical data back it up.
How can we work with Evolvia?
There are three routes: co-discovery on a partner-defined target, access to the Evolvia AI platform after technical review, and advancement of our own molecules once lab results justify it. Use the contact page to start a conversation.
Working on a hard target or a chemotherapy-combination idea?
We work with pharma and academic partners on traceable, experimentally grounded discovery.