Making Waves in Peptide Synthesis

Synthesis sat down with Matt Bio, Chief Scientific Officer, Cambrex & President, Snapdragon Chemistry, for a forward‑looking perspective on the forces reshaping peptide manufacturing…

Earlier this month at TIDES 2026 in Boston, there was a workshop entitled: “The Next Wave of Innovation in Peptide Manufacturing.” What does this next wave look like to you?

I’d say we’re very much in the middle of the current wave of transitioning from solid-phase peptide synthesis (SPPS) to liquid-phase peptide synthesis (LPPS). It’s an ongoing process that will take 5–10 years, I believe.

What do I see beyond that? Well, peptide crystallization is becoming very important. Thinking more specifically in terms of synthesis, if I look further out and consider what’s happening in academia, cell-free omniligase enzymatic approaches for putting together peptides are rapidly emerging. There are also innovations on the synthetic side; Merck published a fascinating paper on unprotected amino acid couplings.

What I like about both of these approaches is that they rely on traditional equipment, so you’re not locked into a specialty facility. I believe that both areas will continue to evolve and, hopefully, spin up into viable manufacturing.

What are your thoughts on microbial recombinant production?

 To answer that question, we need to look at what’s happening inside the discovery teams hunting for new peptides. They typically begin with a natural sequence to identify a target or binder before systematically replacing natural amino acids with unnatural ones to resist some of the metabolic challenges faced by canonical peptides. As we increasingly rely on the incorporation of these unnatural amino acids, I believe the utility of microbial recombinant production will be limited.

For decades, there have been attempts to expand the genetic code to use unnatural amino acids in cell-based or cell-free biosynthetic manufacturing with ribosomes, but success is somewhat elusive. Admittedly, there have been some small expansions of the genetic code to gain access to a few non-canonical amino acids, but they really limit the medicinal chemist’s toolbox in terms of modifications that can extend half-life, increase bioavailability or enhance binding.

If you look at the first generation GLP-1s and long-acting amylins, you’d likely find that one or two of the 40 amino acid chain were replaced with unnatural amino acids. In the third generation, you might find 10 unnatural amino acids. And in the case of the exciting macrocyclic peptides being discovered using mRNA display technology, unnatural amino acids dominate. Take Merck’s enlicitide (formerly known as MK-0616), details of which were published in the Journal of the American Chemical Society in 2025; there’s not a single natural amino acid to be found despite the discovery platform being all natural.

In short, enzymes certainly have their place, but I do not foresee a significant number of peptides of interest being amenable to full ribosomal manufacture.

You used the phrase “exciting macrocyclic peptides.” Could you please elaborate?

Many of these macrocycles are being developed against targets that were deemed undruggable by synthetic molecules and only accessible via monoclonal antibodies (mAbs) or large expressed proteins. But now we’re seeing macrocyclic peptides that are capable of disrupting or engaging protein–protein interactions – often with the added advantage of being able to enter cells. I recently read that around 35 percent of these macrocycles are turning out to be bioavailable orally. Simply amazing.

Two macrocycle cases in point: J&J’s recently FDA-approved drug targeting IL23 for autoimmune diseases (icotrokinra, formerly known as JNJ-77242113) and the aforementioned enlicitide from Merck are both oral drugs for targets that previously had only been tackled by mAbs. Also amazing.

Where do (common) peptide fragments fit into the current or next wave of innovation in manufacturing?

Given that many peptide therapeutics are being taken forward by synthetic process development groups that are used to doing small molecule manufacture, there has been a great deal of debate over what counts as a regulatory starting material (RSM).

Successful crystallization of small fragment peptides (tetramers, pentamers, or even hexamers) has yielded well-defined, fully characterized, and highly stable material, raising a fundamental question: Which fragment peptides fit into the RSM “box” that was developed around traditional small molecule synthesis? After all, the answer determines where GMP begins – an important consideration given the many downstream steps found in today’s convergent LPPS approaches.

Notably, assembling a longer peptide from high-purity crystalline fragments takes pressure off purity control at a downstream step or the API itself. Indeed, from an analytical chemistry perspective, it is much easier to find a single enantiomer impurity in a tetramer than a 40mer!

The ongoing discussion has resulted in (at least) tetramers being characterized as RSMs, which means more attention has focused on securing a reliable and/or local supply. To that end, we’ve launched a program at our Karlskoga (Sweden) site to manufacture common tetramers found in incretin molecules, where there is a great deal of sequence homology. By the end of 2026, we expect to be able to manufacture at the hundreds of kilogram scale, and we’re already speaking with customers about becoming part of their supply chain.

As we continue to work on LPPS as an organization and also as an industry, we’re identifying larger fragments that are crystalline, stable, isolable, and in high purity. I think there will be continued pressure on the definition of “regulatory starting material.” And Cambrex will be well-positioned to make those fragments for folks looking to de-risk their pipeline.

You briefly mentioned analytical chemistry challenges. What innovations are you seeing in this space?

Peptide synthesis results in several impurities that are analytically challenging, which makes HPLC method development complex and time-consuming. We are actively working on automation and algorithmic optimization technologies to ease the burden. In brief, we take a large design space and, after initial automatic column screening, allow Bayesian optimization to “go to town” on all the other parameters needed for a validatable method.

In THE METHOD directly below, you’ll find a link to a presentation by Adrian Amador that explores this and other areas that are ripe for automation and algorithms. I don’t want to steal his thunder, but reducing method development from four weeks to two days is a big deal.

We can’t discuss peptide manufacture without covering the environmental angle…

It’s a big topic. And one big number that you may see commonly referenced is the average process mass intensity (PMI) of SPPS: 13,000. For those unfamiliar, that equates to 13,000 kg of waste per 1 kg of peptide. The number comes from an excellent paper by a consortium of companies with peptide-based therapeutics in their portfolio (all members of the American Chemical Society Green Chemistry Institute Pharmaceutical Roundtable). Notably, that PMI of ~13,000 is for Commercial/Phase 3. If you look at the data for Phase 1/Preclinical, the PMI rises to ~33,000. In state-of-the-art LPPS, PMIs are two orders of magnitude lower, approaching the PMI ranges seen with small molecule processes.

For large volume products, reducing PMI is not only a question of environmental sustainability but also financial sustainability – especially when it comes to the use of expensive starting materials, such as the unnatural amino acids I mentioned earlier.

On the green chemistry front, we are also exploring how we can avoid solvents that are undesirable in the environment; for example, completely removing N,N-dimethylformamide (DMF) and instead using greener alternatives like tetrahydrofuran (THF) or, even better, 2-MeTHF. We are also interested in how we can recycle solvents that are essential to the process, whether through distillation or organic solvent filtration. This latter innovation also promises a greener and faster alternative to chromatography (another big solvent sink) when it comes to purification.

Finally, any thoughts on the evolving regulatory landscape?

The BLA/NDA divide at 40 amino acids is a really interesting topic from a CMC perspective. To do synthetic manufacture, it stands to reason that you need a plant designed around synthetic manufacture, which raises a few questions. Where in a convergent synthesis sequence do you start to move away from NDA-like CMC towards BLA-like CMC? And what risk assessments need to go into such decision making?

And though it’s certainly not my area of expertise, I also wonder what the 40 amino acid divide looks like from a drug discovery perspective. I’d be very interested to hear what readers think. What are the thought processes involved? Do you stop at 40 amino acids, or do you consider adding on a few as a design tool to embrace BLA as a way to avoid generic competition down the road?! Let me know!