What does it take to develop and manufacture a Protein / Enzyme (Recombinant) drug?
A recombinant protein or enzyme is not synthesized on a bench the way a small molecule is. It is grown. You insert the gene for your protein into a host cell, then coax that cell to express, fold, and (for many proteins) glycosylate the molecule correctly, and the choice of expression system shapes everything downstream. Mammalian systems, almost always CHO, are the workhorse when you need human-like glycosylation and complex folding, which covers most therapeutic enzymes, fusion proteins, hormones, and cytokines. Microbial systems, E. coli or yeast such as Pichia, are faster and cheaper and fit non-glycosylated proteins, smaller constructs, and many industrial-style enzymes. Picking the wrong host early is one of the more expensive mistakes in this modality, because a switch late in development can mean re-running cell line, process, and a chunk of your analytical work.
Development runs as a chain of linked workstreams, and most programs outsource several of them. Cell line development creates and screens clones for titer, growth, and product quality, then banks a stable master cell bank under GMP. Process development splits into upstream (the bioreactor: media, feed strategy, fed-batch or perfusion, scale-up from shake flask to production scale) and downstream (the purification train: capture chromatography, polishing, viral clearance for mammalian product, and ultrafiltration/diafiltration). Analytical development is unusually heavy for this modality, because a protein has to be characterized for identity, purity, charge and size variants, aggregation, host cell protein and DNA residuals, and for enzymes a validated potency or activity assay that actually tracks the biology. Formulation keeps the protein folded and stable in the vial, and then GMP drug substance manufacturing, fill-finish, and QC release turn the process into clinical and commercial supply.
Where a specialist recombinant protein CDMO pulls away from a generalist is in the parts that punish inexperience. A group that has run many CHO and microbial programs already knows the platform media, the capture resins, the common host cell protein assays, and the viral clearance expectations, so they design the process once rather than learning on your molecule. They have seen aggregation and clipping problems before and have a playbook. For enzymes specifically, a real potency assay (often a kinetic activity assay tied to the mechanism, not just a protein concentration readout) is a capability some generalists simply do not have running. A generalist CDMO can make protein, but the depth in glycan analysis, comparability after a process change, and immunogenicity risk assessment is what separates a clean program from one that stalls at a regulatory review.
How do you choose a CRO or CDMO for Protein / Enzyme (Recombinant)?
Match the supplier to the molecule and the expression system first, then weigh everything else. A site that is excellent at CHO monoclonal-style work is not automatically the right home for a microbial enzyme with refolding and inclusion-body recovery, and the reverse holds too. Score two or three candidates against the same written scope so the quotes measure the same thing, and weight relevant platform experience and analytical depth above headline price, because a comparability failure or a re-run process costs far more than the difference between bids.
- Relevant platform and track record: confirmed experience in your expression system (CHO, E. coli, or yeast) and your protein class (enzyme, fusion protein, hormone, cytokine), with named programs that reached the clinic or market, not just generic capability claims
- GxP and analytical capability for this modality: GMP for drug substance and fill-finish, plus the analytical bench this modality demands (identity and sequence confirmation, SEC and aggregation, charge variants, glycan profiling for mammalian product, host cell protein and residual DNA assays, and a validated enzyme potency/activity assay)
- Capacity and scale: bioreactor scale that fits your phase and a credible path from clinical to commercial volume, current queue and slot availability (often the binding constraint, not the science), and whether single-use or stainless-steel suits your titer and batch size
- Regulatory experience: a clean FDA and EMA inspection history, comparability protocols for the inevitable process change, viral safety and clearance for mammalian product, and immunogenicity risk assessment, since the immune response to a protein is itself a safety question
- IP and tech transfer: who owns the cell line and process know-how, whether any expression platform or host carries licensing or royalty strings, and how cleanly the cell bank, process, and analytical methods transfer out if you move sites or scale up elsewhere