De-risking gene therapy development through efficient stress testing of viral vectors

Despite its importance, stress testing viral vectors at speed and scale remains difficult due to multiple factors. Viral vectors are sensitive and heterogeneous, workflows can be time-intensive, and many traditional analytics require complex optimisation or high sample input, and have limited throughput. This can slow down development decisions when time matters most. Here, we explore how two relatively new technologies, mass photometry (MP) and macro MP, compare to other analytical tools for efficient stress testing of viral vectors.

MP works by measuring the mass of individual particles in solution. It uses interferometric scattering, where particles produce a contrast signal when they land on a surface, and that contrast is proportional to mass. By converting contrast to mass using a known calibrant, MP resolves distinct populations – such as empty, partially filled, full and over-filled capsids – based on cargo mass differences. MP is best suited to smaller viral vectors, particularly adeno-associated viruses (AAV), which fall within its mass measurement range of ~30kDa to 6MDa (the exact mass range depends on the instrument).

Macro MP extends MP by measuring both the contrast signal (a proxy for particle mass) and particle diameter through a vertical sweep. This dual-parameter approach improves resolution of heterogeneous samples and reduces the bias towards larger particles that can affect motion-based methods like dynamic light scattering (DLS) and nanoparticle tracking analysis (NTA). Macro MP is designed for larger viral vectors, including adenoviral vectors and lentiviral vectors, which exceed the upper mass limit of standard MP.

Both MP and macro MP share several practical advantages relevant to biologics analysis. Measurements are performed directly in solution, with no labelling or extensive sample preparation required. Sample consumption is low (10–20µL of sample at 1,011 particles/mL for MP; 5µL of sample at 108 to 109 particles/mL for macro MP) and measurements take only minutes.1,2 Because both techniques measure individual particles rather than ensemble averages, they provide population-level resolution that bulk methods cannot match. These properties make MP and macro MP well suited to workflows where sample is precious, throughput matters and analytical clarity is essential.

For AAV, a central concern in stress testing is the distribution of capsid populations – empty, partially filled, full and overfilled particles – and how that distribution shifts under stress conditions such as heat, freeze–thaw cycling or pH variation. Changes in fill ratio can directly affect therapeutic potency and immunogenicity, making rapid, reliable quantification essential throughout development. A range of analytical techniques are used for AAV characterisation, each with different strengths and practical trade-offs that become particularly significant in the context of stress testing, which involves iterative testing of multiple samples and conditions.

Analytical ultracentrifugation (AUC) is widely considered a benchmark method for capsid fill analysis, capable of resolving empty, partial, full and overfull populations by sedimentation coefficient. A single AUC run can take several hours, and meaningful stress studies require measurements across multiple conditions and timepoints – making the total analytical burden substantial. The technique also requires specialist instrumentation and highly trained operators, and is frequently outsourced rather than run in-house, which adds further delays to turnaround. For these reasons, AUC is typically reserved for later-stage characterisation rather than iterative screening.

Electron microscopy (EM) offers direct structural visualisation, and cryo-EM in particular can provide high-resolution information about capsid integrity. However, sample preparation is complex and time-consuming, throughput is low and resolving differently loaded capsids based on electron density differences alone is technically challenging – particularly for partially filled particles where the contrast difference relative to full capsids is subtle. Charge detection mass spectrometry (CDMS) offers high mass resolution at the single-particle level and can resolve capsid sub-populations with accuracy comparable to AUC.

However, it is dependent on advanced native mass spectrometry instrumentation, is not widely available outside specialist facilities, and requires considerable expertise for calibration, experimental set-up and data interpretation. Like AUC and EM, it is better suited to in-depth characterisation studies than to the rapid, high-throughput screening that routine stress testing demands. Size-exclusion chromatography with multi-angle light scattering and quantitative polymerase chain reaction/enzyme-linked immunosorbent assay are faster and relatively easy to deploy, and both have strong good manufacturing practice (GMP) support. However, as ensemble techniques, they measure average properties across the whole sample rather than resolving individual particle populations. Neither can distinguish partially filled or overfull capsids from full ones – a critical limitation when monitoring fill ratio changes is central to the stress study objective.

MP, by comparison, resolves empty, partial, full and overfilled AAV capsids within a single rapid measurement. As a result, it can deliver information comparable in scope to AUC or CDMS, but with much lower time and expertise requirements, minimal sample consumption and low running costs. Individual measurements take under five minutes, making MP well suited to the iterative screening across conditions and timepoints that occur in stress testing workflows. Beyond shifts in fill ratio, MP can detect other stress-induced species, including single-stranded DNA released from degraded capsids and degraded or aggregated capsids. This provides a more complete picture of the sample’s stress response – all with a single measurement.

Macro MP does not have these limitations; its dual-parameter readout is better able to resolve populations within heterogeneous samples. By measuring the mass of individual particles, it is not subject to the biases that affect DLS and NTA. For adenoviral and lentiviral vectors, where stress-induced changes may produce a range of species – intact particles, aggregates and degradation products – the ability to distinguish populations across two dimensions provides a more complete picture of sample quality than single-parameter methods alone.

Larger viral vectors, including adenoviral vectors and lentiviral vectors, present a different characterisation challenge to AAV. Adenoviral vectors are significantly larger than AAV – typically 70 to 90nm in diameter compared to 20 to 25nm for AAVs – and structurally more complex.3,4,5 Lentiviral vectors, meanwhile, are enveloped particles with an inherently variable morphology. Their lipid envelope is derived from the host cell membrane during budding, introducing size and shape heterogeneity that is present even in well-manufactured material. This baseline heterogeneity means that interpreting stress-induced changes requires distinguishing genuine degradation from natural sample variation, which places greater demands on the analytical technique.

The comparator landscape for larger vectors differs from AAV, reflecting both the size of the particles involved and the greater heterogeneity of these preparations. DLS remains widely used for aggregation detection in larger vector preparations, and its speed and simplicity make it a practical first-line tool. However, its ensemble-averaging approach and bias toward larger, more strongly scattering particles mean that minor sub-populations – including aggregates or degradation products – can be masked by dominant species. It also provides no information about mass or identity.

NTA offers better resolution of individual particle size distributions and can detect sub-populations that DLS would miss. However, tracking accuracy decreases in samples with wide size ranges, as fast-moving smaller particles and slow-moving larger ones are difficult to capture reliably in the same acquisition. Like DLS, it provides no direct mass information. AUC can in principle be applied to larger vectors, but the practical constraints – long run times, low throughput, specialist expertise – are amplified for more heterogeneous samples, and it sees limited use in routine larger-vector stress testing workflows. EM remains an option for structural confirmation but is similarly low-throughput and resource-intensive.

Stress studies often generate many samples across multiple conditions (temperature, pH, freeze–thaw, agitation, etc). The bottleneck is getting clear, actionable readouts fast enough to guide decisions without high sample consumption or complex workflows. Insights from stress testing also enable teams to design more robust formulations, optimise processes and build the analytical evidence base needed for regulatory submissions. As vector modalities diversify and development timelines compress, the demand for analytical tools that deliver resolution, speed and flexibility will only grow. MP and macro MP characterise distinct vector populations rapidly at the single-particle level – providing the kind of precise, population-resolved data that teams can act on, whether by making a formulation change, a process adjustment or a go/no-go decision.

Jacintha Victoria is a strategic marketing and analytical leader with over a decade of experience in the biotechnology and gene therapy sectors. As market development manager at Refeyn in Oxford, UK, she leads initiatives to drive the adoption of MP across critical applications in cell and gene therapy, including AAV, LVV, adenovirus and RNA-based platforms. With a deep understanding of the emerging analytical technologies, Jacintha advocates for MP as a transformative tool to enhance viral vector characterisation, accelerate product development and strengthen regulatory preparedness. She holds a Master of Science in Cellular and Molecular Biology from the University of New Haven, CT, US, and is certified as a technologist in Molecular Biology by the ASCP. Passionate about bridging science and strategy, Jacintha is dedicated to advancing innovative therapeutic solutions through data-driven marketing, cross-functional collaboration and thought leadership.

De-risking gene therapy development through efficient stress testing of viral vectors

Stress testing AAV: characterising capsid populations at speed

Stress testing larger vectors: resolution across greater heterogeneity

Conclusion

References: