By NorthX Biologics

At Innovation for Health 2026, this reality was front and centre. The conversation was candid: the frameworks are not yet fully established, expectations are shifting, and the organisations navigating this space need both scientific depth and considerable adaptability.

Scale-up: where complexity stops hiding

Moving from a laboratory bench to GMP-grade manufacturing rarely goes as expected but with vesicle-based products, the gap can be particularly stark. Variables that seem manageable at small scale, such as media composition, temperature, and purification approach, can produce meaningful shifts in vesicle size, composition, and biological activity when conditions change.

Purification is where many teams first encounter the hard reality of this complexity. Vesicles do not separate cleanly from biological debris, lipoproteins, or aggregates; impurities that can share similar physical properties. Standard lab methods rarely scale without redesign, often requiring a shift to approaches such as tangential flow filtration or chromatography. And when the process changes, the analytics must follow.

Analytics: building the instrument as you play it

Perhaps the most striking message from the conference was the degree to which analytical science for vesicle-based therapeutics is still being written. There is no universally accepted framework. Standardisation across organisations, let alone across borders, remains limited.

For anyone responsible for quality and compliance, this demands a different mindset. A single metric (protein concentration, for instance) tells only a fraction of the story. Building a meaningful picture requires layering particle characterisation, compositional analysis, and functional assays. In practice, this means moving towards multi-attribute quality fingerprints, and accepting that in early development phases, demonstrating consistency may be a more honest and valuable goal than chasing absolute numbers.

"In early phases, consistency is the target"

For vesicle-based products, defining quality is not a one-time exercise. Critical quality attributes are rarely fully understood at the start of development, and analytical methods must evolve in parallel with the product. Multi-attribute fingerprinting; combining particle characterisation, compositional analysis, and functional assays, is emerging as the standard approach. In early phases, demonstrating batch-to-batch consistency is often a more meaningful and achievable goal than pursuing absolute quantitative targets. The framework is still being built, but it is being built deliberately.

Development as a learning loop

In most development programmes, manufacturing is a downstream concern: something you build once the science is settled. In the vesicle space, this distinction largely disappears. As products move into clinical stages, new data continuously reshapes how processes are designed, controlled, and interpreted.

For a GMP manufacturing partner, this calls for more than technical capability. It requires the kind of close collaboration where emerging scientific insight can be translated quickly into process decisions and where the control strategy, the analytical methods, and even the product definition can all adapt as understanding grows. Early alignment on an analytical target profile, combined with open regulatory dialogue, is proving to be one of the most effective ways to reduce risk further down the line.

A field being built in real time

The clearest takeaway from Innovation for Health 2026 was also the most human one: nobody has all the answers yet, and the organisations making the most progress are the ones who have accepted that, and built accordingly. This is a field where manufacturing, analytics, and biology must genuinely co-evolve. That demands flexibility, cross-disciplinary collaboration, and a readiness to keep learning.