Oligonucleotide manufacturing: are we ready for what’s next?

Oligonucleotide therapeutics are advancing, but with that momentum comes new complexity in how these molecules are designed, manufactured, and scaled. Oligonucleotide manufacturing challenges vary whether you’re developing pipelines, optimizing production, or evaluating new technologies.

In this panel discussion, industry and academic experts share practical insights on what’s changing from diversification in chemistry and delivery strategies (including enzymatic synthesis) to the realities of cost, scaling production, and process control.

Panelists

Timothy Tan, PhD, Associate Professor, School of Chemistry, Chemical Engineering and Biotechnology, Nanyang Technological University Singapore and Chief Scientific and Strategy Advisor at RNAscence Biotechnology, a biotech spin-off from the University and Singapore National Skin Centre. Tan and his team commercialized the world’s first gene silencing technology for the skin through RNAscence Biotechnology. The flagship product is a scar prevention and treatment patch, with pipeline products including modified siRNA for enhanced scar treatment and multitargeting siRNA for eczema. Tan’s broader work in oligonucleotide therapeutics includes leading an oligo foundry project funded by the Nucleic Acid Therapeutics Initiative (NATi) to advance synthesis capabilities for Singapore’s research community.

Barrie Cassey, PhD, Technology Director, Medicines Manufacturing Innovation Centre at CPI. After working at GSK, Cassey moved to CPI to lead oligonucleotide manufacturing innovation and develop a broad program around all parts of the manufacturing process.

Carina Strandh, Senior Product Manager, Oligonucleotide Synthesis Hardware at Cytiva. Carina went from a development organization, Astra Zeneca, to Cytiva, a supplier situation. Working with oligos and customers that produce active pharmaceutical ingredients (APIs) brings her close to the patient.

Q: What is changing in the oligonucleotide market? Is diversity playing a role?

The panelists concurred that diversification is in every part of the process—the chemistry, the therapeutic area, the administration routes, and personalized and broad therapeutic approaches. But the manufacturing expense mandates that new modifications introduced into a therapeutic molecule add significant value.

The large market oligos such as Inclisiran have shown the promise in terms of scalability, manufacturing, and investment. In addition, for very precise targeting and to impact treatment efficiency, different delivery strategies in terms of conjugations like GalNAc peptide antibodies, exosomes, and lipid nanoparticles extend diversity.

Q: Are there gaps or investment needed?

Carina: There are so many initiatives in R&D and question marks around different ways of producing the drugs. You need to really understand the unit operation, the complexity, the cost of raw material, and the process mass intensity (PMI) balance calculation (input and output in kg). With the currently available information, it is difficult to get a good overview and understand where the gaps are.

Barrie: One of the interesting things about genetic medicines is personalization. CRISPR is a bit of a hybrid. We talk about rare and nano-rare diseases, which affect a single person or a handful of people. There’s such opportunity to dramatically impact someone’s life. But manufacturing is a big sledgehammer for a small nail. A fundamental mind shift is needed in terms of manufacturing along with the regulations. It’s not economically viable if it’s going to cost $500K to make a therapeutic dose. The health economics need to be resolved. It’s a difficult argument and reducing costs would help.

Q: What about alternative evolving technologies like enzymatic synthesis?

Carina: One of the main claims and drivers for enzymatic synthesis is sustainability. Completely enzymatic is still moving from base-to-base coupling toward shortmer coupling. The coupling of those shortmers depends on who you speak to. Although sustainability is a popular word, there are questions marks. It is water-based, and you may need highly engineered enzymes and other unit operations and process components, which need to be included in a comparison. These different alternative manufacturing technologies give producers a choice. Ultimately, cost will drive the decision.

Barrie: Solid-phase oligo manufacturing, phosphoramidite chemistry, has been optimized over 40 to 50 years, but just like in therapeutics, we see diversification in synthesis. People who need large quantities or diverse types of molecules such as gRNA have to fundamentally change their approach. It is difficult to have a high yield when making a 100-mer with solid synthesis. Bringing in process analytical technology (PAT) for better control and platform approaches specific to a molecule and the extension process makes liquid phase synthesis (a templated process) exciting. Yet, you can’t just improve synthesis without considering the impacts on the complete process.

Tim: I only trust solid-phase synthesis. It’s been so optimized.

Q: How is the supply of raw materials as scale goes up (or in some cases comes down)?

Tim: Materials are crucial. Different suppliers have different grades, purities, and impurities. More predictability in terms of partnering with raw materials vendors would solve a lot of problems both upstream and downstream. Sticking to one or a few trustworthy raw material suppliers is at the top of my wish list.

Barrie: I agree. Your raw material is a complex molecule and your registered starter material. You’re putting an enormous amount of trust in manufacturers, and you can’t move because the impurity profile of your product is intrinsically linked to that raw material. You spend more time worrying about your raw material quality than your finished product.

So, an enzymatic approach is quite interesting. There’s some commonality in synthesis roots but making a nucleoside triphosphate is quite advanced. Chemically making a pyrophosphate linkage is part of the process. The first phosphate or pyrophosphorus is your raw material. It’s a challenging change because to be a registered start material, the stereochemistry must be done in the process. I’m not sure whether a nucleoside triphosphate or a type of file version would be an allowable starter material. The convergence is a serious matter. Census fragment assembly techniques with the fragments being the registered starting material. People think it’s viable with quality control.

Carina: I heard the same, and I’m just as surprised about using short 7-, 8-, or 9-mers as starting material. It comes down to how you define your manufacturing process. How is regulatory going to view it?

Q: What does the industry have to do to address and manage this?

Carina: It needs to be defined but not necessarily with new regulations. Where do these materials fit: as raw materials or intermediate materials? Where do you start your process? The difference partly depends on whether you are sourcing your material from a third party as a raw material, or if you are producing the shortmers in-house as intermediates. The material could then be classified differently depending on the regulations or the inspecting authority.

Barrie: Carina is right. The regulations are there for the control of starting materials. It’s not really changed.

Q: What would it mean if you made the starting shortmers in house?

Tim: It just becomes part of the regulations of the therapeutic. It should be the end product.

Barrie: If you can characterize effectively and control and understand the relationship between the impurities in your starting material, then what you’re claiming is your starting material is your product. In theory, it’s possible for a very sophisticated, technologically advanced company with huge resources. It could be an advantage because you can take the good manufacturing practices (GMP) line further forward. It doesn’t make any difference to the effort and quality in the manufacture of that starting material and reduces cost and the effort to maintain an inspected facility. It’s an interesting commercial decision. But particularly, stereochemistry is a huge question in raw materials.

Q: Is there anything that you want to see specifically from your suppliers?

Tim: Coming from a young company, I wish for predictability in terms of yield: the impurity profile as well as supply. In terms of synthesis and purification, I would like more predictability, a playbook if possible. At the start, making all oligos is an empirical art; trial and error. We play with parameters, and if it doesn’t work, we start again. It’s quite primitive. If my vendor can build predictive qualities into the system to predict parameter adjustment, raw materials, and expectations of the impurity profile, that would help guide the synthesis process and reduce downtime. We wasted a lot of solvents and time not knowing where to start. If I need to buy new equipment, it would be something I would love to get from my supplier.

Barrie: We just finished a proof-of-concept, machine learning-driven, predictive tool. I agree with Tim, it’s such a complicated thing that you almost want to just press go and have the system optimize itself. That’s what we were aiming for. Combining automation with predictive set points and going a step further by integrating more complex PAT to give live feedback for optimization is so valuable. Oligo chemists are hard to find and are expensive. The biggest barrier to scaling up is the cost. Manufacturing is never going to get sanctioned unless you have superb process control, which will come from integrated PAT and model-driven process control tools that rely on well-supported and understood models.

Carina: I dream of implementing this on a synthesizer: to be able to press start and for the system to sense what step to take next and how to adjust live, especially for process development. When you scale up and enter the regulatory field, how do you handle the chemistry, manufacturing, and controls (CMC)? Would you use ranges of volumes, concentrations, or time?

Barrie: We are looking at continuous tablet manufacturing with a series of sensors, etc. throughout the process for feedback into an adaptive process control system. It is a challenge because regulators are used to seeing operating ranges from X to Y. But they have been educating themselves. Some people use models to constrain where they can operate and end up with a very narrow range because it gives a perfect product. A model for process control should control and not be constrained.

Then it becomes a GMP issue instead of a regulatory issue. During inspection, the goal will be to show how the model works. It is a shift; a difficult concept for some people to get their heads around. Not all regulatory authorities are as sophisticated as the US Food and Drug Administration (FDA) or the European Medicines Agency (EMA). So, you end up with a two-tier system depending on where you’re supplying. But it’ll come because it’s so much better process control rather than relying on rigid parameters. Good oligo chemists automatically do certain things. Just imagine if you could get twenty of the best oligo chemists, take the information in their brains and put it into a tool. That’s what we tried to do, to get some pre-emptive information, test certain things and get the machine learning (ML) process to adapt. A tool will come where you put in your sequence, and the process will run.

Q: What are the challenges of scaling up, especially when scaling up is expensive? Is there advice that can save time later?

Barrie: Think about the entire process, from solvent supply, solvent waste, etc. The actual synthesis process is this little thing in the middle. There is other activity, even just cycling columns, cleaning things, prepping analytes, and the amount of analytical support you need for the process. Not everything scales linearly.

Q: Would going modular and synthesizing in parallel derisk your scaling?

Carina: That would make the software complex. If you’re going to use the same pumps for different reagents in different columns, you need to ensure that the software is recording the right chemical, reagent, or solvent into the right column to the right volume and at the right time. Scaling is easier and more controlled by using a larger column.

Barrie: It does not simplify the process because the mistakes don’t happen on the synthesizer. Humans make the mistakes. Plenty of people go from 1 to 1200 mmol in one step because they’ve got confidence in the synthesis and the reproducibility. The core process isn’t the risk. It is all the peripheral activities and the analytics. Every time you run a different batch, you have analytical tests and in-process checks. You’ll become an analysis factory instead of an oligo factory.

Q: Do you have any advice for someone thinking of moving a process in house?

Tim: If I were to give advice, the vendor should ease them into the process by having a more definite synthesis pathway to reduce the trial and error. I did not have a playbook, so I had to problem solve with limited resources. Solid-phase synthesis is the way to go, at least for the next 10 to 20 years. I think it’s worth investing in optimizing the process and applying AI and ML. Enzymatic synthesis still needs R&D. It is coming, but it will take years.

Barrie: Do not underestimate how complicated it is to get a good oligo and yield. Really understand what you’re getting into and why and have everything you need lined up. It looks simple, but it has a huge platform nature, and within that platform there’s a lot of variability. Do not expect your CDMO to give you all the secrets; that’s their business. They can’t make you an expert in 2 weeks. It is a steep learning curve and might take a couple of years. You need to be really committed. We’re launching training to help reduce the learning curve not just for scientists or engineers but also for procurement and finance. Enzymatic synthesis will happen quicker than you think, but only for products where it makes sense

Carina: I recommend that anyone who wants to invest in developing oligo therapeutics consider buying their own synthesis equipment first to learn the process and understand the details, such as defining parameters, important variables, setup requirements, and the unit operations. There are many new areas being investigated leading to more complicated chemistries, longer sequences, different conjugates to improve deliveries, etc., all driving diversification and development of manufacturing technologies.

Conclusion

As this discussion highlights, there is no single path forward in oligonucleotide manufacturing—only a rapidly evolving landscape shaped by innovation, trade-offs, and careful decision-making. Success will depend on balancing proven approaches with emerging technologies, building strong supplier partnerships, and investing in process understanding from the start. For organizations looking to stay competitive, the key is to plan early, stay flexible, and leverage the growing ecosystem of expertise and tools available. With the right strategy in place, you can navigate complexity with confidence and turn today’s challenges into tomorrow’s breakthroughs.

CY59287-24Jul26-AR