Erythromycin
Antiinfectives for treatment of acne · Macrolides · given by ophthalmic, oral, topical · on the market since 1982
What it's made of
Each functional group on the molecule does a specific, predictable job.
Basic amine
What it does. Neutral enough to slip through membranes, then becomes charged inside the acidic compartments of the cell and gets trapped there. Concentrations inside cells can reach many times the blood level.
What it opens up. Diseases of the cell's internal recycling compartments, and a common reason a drug accumulates in lung, liver, and eye tissue far beyond what the blood level suggests.
Sugar ring
What it does. Extremely water-loving. Adding one to a molecule usually destroys its ability to cross the gut wall, which is why many such drugs must be injected.
What it opens up. If given by mouth it stays in the gut and acts there only — a targeting advantage for intestinal disease.
Ester
What it does. Blood and tissue enzymes cut this bond within minutes. Whatever was attached falls off.
What it opens up. Deliberate short action, or a delivery trick — attach an ester to sneak a molecule somewhere, then let the body cut it loose at the destination.
Where it can get to
Predicted from molecular weight, lipophilicity, polar surface area and hydrogen-bond donor count. A drug can only act where it distributes, so this constrains which conditions are addressable before any biology is considered. derived
What it binds — and whether we can reach it · 1 of 5 reachable
Binding affinity is only meaningful against the plasma concentration reached at the approved dose. A target requiring higher concentrations cannot be engaged in a patient, whatever the in vitro potency. Most repurposing hypotheses fail at this step, and it is rarely checked before a trial is proposed. measured
| Target | Kind | Action | Affinity (pAct) | Can it be engaged? |
|---|---|---|---|---|
| KCNH2 | Ion channel | — | 7.41 | Engaged |
| ABCB11 | Transporter | — | 5.39 | Out of reach |
| AOX1 | Enzyme | — | 4.82 | Not reachable |
| MLNR | GPCR | — | 4.43 | Not reachable |
| CYP1B1 | Enzyme | — | 4.4 | Not reachable |
This molecule builds up inside cells far above blood levels. Local concentration therefore exceeds plasma, so targets that look unreachable on a plasma reading may still be engaged. assumed
New uses, and who is already trying them
One card per candidate indication: the reasoning link by link, the evidence tier that governs how strongly it can be stated, and any registered trials.
5 of the 70 conditions linked to this drug’s targets have been checked against trial records so far. The rest are not ruled out — they are unchecked.
Romano-Ward syndrome
Wide openIt acts on a target linked to this disease; untested, and the direction of effect is unverified.
progressive familial intrahepatic cholestasis type 2
Wide openIt acts on a target linked to this disease; untested, and the direction of effect is unverified.
Juvenile glaucoma
Wide openIt acts on a target linked to this disease; untested, and the direction of effect is unverified.
benign recurrent intrahepatic cholestasis type 2
Wide openIt acts on a target linked to this disease; untested, and the direction of effect is unverified.
Familial short QT syndrome
Wide openIt acts on a target linked to this disease; untested, and the direction of effect is unverified.
Why nobody will fund it
15 companies make this across 34 approved products. With that many manufacturers the price has fallen to production cost. A company funding a trial could not recover it — competitors could market the same molecule for the new indication immediately, without contributing.
A new route or dosage form qualifies for a 505(b)(2) application, which carries three years of exclusivity for that formulation. It is the only remaining route to exclusivity on a generic molecule.
What it would cost to find out
Derived, not quoted. Events needed is 4(zα+zβ)²/ln(HR)²; enrolment is that divided by the event rate and follow-up. The 1/ln(HR)² term dominates: halving the detectable effect size roughly quadruples enrolment. derived
Cost per person reached is the relevant metric for a non-commercial funder, since there is no revenue to set against the outlay.
Ask for ideas nobody has listed
Reasons from the functional groups and distribution above, excluding conditions already listed. Output is hypotheses for falsification, not findings. Not medical advice.
Who might pay for the trial
Trials of off-patent drugs are rarely funded by venture capital or pharma. They are paid for by government grants and disease foundations. Here is who has funded work on this molecule before, what federal mechanisms are open now, and which foundations fund trials in the disease areas it touches. measured
Open notices with an explicit mandate to fund work on approved or off-patent molecules.
Grants.gov lists US federal money. Many off-patent drug trials are supported by disease foundations and nonprofits. These are matched to the body systems this drug touches. assumed
Who manufactures it
Companies holding an approved abbreviated application for this molecule in the United States, from the FDA Orange Book. Supply already exists at scale, so a trial would not need a manufacturing programme — but any of these could also be a partner, or an objector.
Nothing on this page is medical advice. Candidate uses are unproven hypotheses derived from public data; established uses are listed for verification only. Figures carry their provenance inline — anything marked assumed or derived is not a measurement.
← Back to the full ledger