Most drug developers already know how to select a CDMO, often working from a due diligence checklist for Phase I selection. The standard CDMO selection criteria cover quality history, capacity, cost, timelines, project management and regulatory support, and these still apply to complex molecules. What changes is that the usual questions stop being enough.
When the molecule is a PROTAC, a molecular glue or a drug-linker conjugate, three additional factors often decide the choice: whether the partner can contain potent chemistry, purify what crystallization cannot, and keep the analytics connected to the molecule through conjugation.
What makes a molecule complex
These three modalities differ in size and mechanism. PROTACs are potent, relatively large molecules and difficult to purify by conventional crystallization. Molecular glues are small molecules, but mechanism-driven and scientifically demanding to produce. Drug-linker conjugates bring strict containment requirements and significant analytical complexity.
A partner offering drug substance manufacturing for complex molecules should be able to show how it handles each of these, from route scouting and stereochemistry control through to CMC regulatory submission.
PROTACs: ask how they purify what will not crystallize
PROTACs, the first of the three complex chemistry modalities, are bivalent molecules of typically 700 to 1,000 Da. They connect an E3 ligase ligand and a target warhead via a linker, and they fall outside Lipinski’s Rule of Five.
That profile points to three questions for a partner:
- Purification. Where crystallization falls short, preparative HPLC has to take over. Ask what chromatography capacity is available under GMP.
- Stereochemistry. PROTAC chemistry needs a strong focus on stereochemical control. Ask how it is built into the route.
- Solubility. Poorly soluble PROTACs with BCS Class II and IV profiles may need spray drying or hot melt extrusion. Ask whether that sits with the same partner.
Molecular glues: small does not mean simple
Molecular glues are monovalent targeted protein degraders, typically under 500 Da and within Lipinski’s Rule of Five. Their mode of action is anything but simple. They work through weak binding affinities to both the E3 ligase and the target protein, acting as event-driven catalytic agents that redirect the cell’s own protein degradation machinery. Their targets are often unpredictable.
Development therefore requires scientific judgment, not just protocol. The capabilities to probe are physicochemical characterization (LogP, pKa, PSD, HRXRPD, DSC, TGA-MS, NMR), amorphous solid dispersion screening (or other enabling technologies) for solubility challenges that standard formulation cannot solve, and phase-appropriate CMC support.
Ask, too, whether the partner has taken a glue into the clinic. For CYRS1542, a GSPT1 molecular glue degrader developed by Cyrus Therapeutics, Ardena was responsible for the formulation development of oral dosage forms and the GMP manufacturing of clinical trial material. Cyrus received IND approvals from the US FDA and the Korean MFDS in 2025.
Drug-linker conjugates: follow the molecule through conjugation
Drug-linker conjugates, including ADCs and XDCs, combine the selectivity of a carrier with the potency of a payload, connected by a linker that must be precisely engineered and controlled.
Here the selection turns on continuity. Ask whether one partner can cover ADC and XDC development from drug-linker and payload synthesis through conjugation. Ask how drug-to-antibody ratio (DAR) and conjugation heterogeneity are controlled, and how free payloads and catalysts are removed. Finally, ask whether bioanalysis can measure total antibody, total ADC, conjugated payload and free payload.
Containment: ask for evidence
PROTACs are potent, and drug-linker conjugates carry the potency of their payload, so containment belongs in every assessment. Ask for evidence rather than a claim: occupational exposure measurements with a representative model compound, segregated manufacturing suites, and a product changeover based on a full MACO calculation using health-based exposure limits from a certified toxicologist.
Ardena’s high potency containment facility has manufactured highly potent APIs at OEB-5 level since 2012 and is routinely inspected by the Dutch Health Authorities.
What to probe during a CDMO audit
Use these questions during a CDMO audit or qualification visit.
| Modality | What to probe | Question to ask |
|---|---|---|
| PROTACs | Purification without crystallization, stereochemical control, solubility | What preparative HPLC capacity runs under GMP, and who handles solubility enhancement? |
| Molecular glues | Characterization depth, ASD screening (and other enabling technologies), clinical track record | Which glue programs have you supported into the clinic, and what did you deliver? |
| Drug-linker conjugates | Continuity from synthesis to conjugation, DAR control, free payload removal | Can one team make the drug-linker, conjugate it, and measure conjugated and free payload? |
| High-potency containment | Measured operator exposure, changeover cleaning | Can you show exposure data with a model compound, and how is MACO calculated? |
The bottom line
For complex molecules, the strongest CDMO partnership is with an integrated CDMO that keeps chemistry, purification, containment and analytics in one program. At Ardena, drug substance chemistry, drug product formulation, bioanalysis and CMC regulatory support are coordinated within one program, so the scientists working on linker chemistry talk to the scientists designing the bioanalytical assays.
Related reading
- How to Evaluate HPAPI Contract Manufacturing Companies in 2026
- ADC and XDC Drug Development: What CDMOs Need to Deliver
- Quality Agreements in Pharmaceutical Outsourcing: What They Must Cover