2026年RNA治疗中dsRNA控制的监管期望

Originally published by Crystal NAX. Republished with company attribution. Original technical perspectives, references and source links are retained. The original website also credits A C.

This article covers what that argument must contain at each stage of development, how FDA and EMA expectations compare, what analytical validation requires, how to set a defensible specification, and what comparability obligations follow from process changes. We also have a page covering the mechanistic and biological details of dsRNA: what is it, how is it formed and the immunogenic risk associated with it.

Does the FDA or EMA Set a Limit for dsRNA in RNA Therapeutics?

FDA and EMA do not set a fixed numeric dsRNA limit because both agencies treat dsRNA as a process-related impurity with product-specific risk.

They instead require risk-based impurity characterization, validated analytical methods, and a documented control strategy justified by product context and development stage.

The absence of a harmonized numeric dsRNA limit reflects a product-specific impurity philosophy running through FDA’s CMC framework for gene therapy and RNA products. Regulators treat dsRNA as an impurity whose acceptable level depends on demonstrated biological impact within a defined therapeutic context.

This creates a more demanding task than meeting a table value. A sponsor must generate analytical data from development batches, correlate dsRNA levels with functional outcomes where feasible, define an acceptance criterion, and justify that criterion relative to the clinical population, dose, route of administration, and dosing frequency. dsRNA content also needs to be consistent across process changes and development stages.

How to Set a dsRNA Specification for an IND or CTA Submission

A dsRNA specification becomes defensible when the acceptance criterion connects analytical data, functional evidence, and clinical rationale. Reviewers expect a proposed limit with reasoning at the IND or CTA submission, not a locked number.

Reviewers look for three things:

• dsRNA levels measured across representative development batches
• Evidence that those levels are controllable through the manufacturing process
• A correlation with a functional readout such as translation efficiency or an interferon signaling assay

The proposed limit is set relative to the highest level observed in batches with acceptable biological performance.

CQA designation depends on demonstrated impact on potency or safety in the specific product context. A sponsor who designates dsRNA as a non-CQA carries the burden of showing two things: levels are controlled, and the biological impact at those concentrations is negligible.

The acceptance criterion set at this stage becomes the comparability anchor for every process change that follows. This makes the number itself a strategic choice. A limit that reflects actual, demonstrated process capability gives the program room to breathe during scale-up. A limit set at the widest range a reviewer might accept looks generous now but leaves no margin later, since any drift toward that ceiling becomes a comparability problem down the line.

dsRNA Control Expectations by Development Stage

Regulatory scrutiny of dsRNA increases with development stage, and the sharpest jump happens between IND or CTA and Phase II and III. The research and RUO stage carries no formal filing requirement, but sponsors that skip dsRNA tracking during early research often lack the historical batch data needed to justify their specification later.

At NAX, dsRNA characterization is integrated from early synthesis through lot release, using validated J2 ELISA with internally qualified dsRNA reference standards spanning 500 to 1500 base pairs, alongside IP-RP HPLC for orthogonal confirmation. If your program is approaching an IND-enabling or IND submission milestone, contact the NAX team to discuss your dsRNA control strategy.

What Happens to a dsRNA Specification After a Manufacturing Process Change

Every manufacturing change triggers a comparability obligation for dsRNA under ICH Q5E, and reviewers expect the impurity profile itself to stay consistent across the change, not only the total reported value. A process change that lowers total dsRNA while shifting the size distribution toward longer duplex species can raise immunogenic risk even when the reported number improves.

For dsRNA, reviewers expect evidence that the impurity profile, meaning the size distribution and structural composition of dsRNA species present, stays consistent across the change. Orthogonal analytical data, including size-resolved methods, are part of a credible comparability package for this reason. If the analytical method changes alongside the process, both require separate justification.

FDA’s July 2023 draft guidance on manufacturing changes and comparability applies specifically to cellular and gene therapy products, and its risk-tiered approach to comparability studies is often used by industry as a directional reference for other complex biologics, including RNA therapeutics, even though RNA products fall outside its formal scope. The guidance signals that the depth of the analytical comparability package should scale with the potential impact of the manufacturing change on product quality. That same logic applies to dsRNA. Every specification set during development becomes a comparability anchor, so its precision at the IND stage has direct consequences once the process moves to scale-up.

How Route of Administration and Indication Affect Acceptable dsRNA Levels

Acceptable dsRNA levels shift with route of administration, indication, delivery vehicle, and dosing frequency. A specification built for one program context should never be carried unadjusted into another.

• Route of administration determine the systemic dsRNA burden per administration. A level tolerated in a single intramuscular dose may produce measurable interferon responses in a repeat intravenous regimen.
• Indication shapes risk tolerance directly, as oncology vaccine programs where innate immune activation is pharmacologically desirable accept higher dsRNA levels than chronic protein replacement therapies where it is entirely unintended. A limit carried from a vaccine program into a protein replacement program without adjustment is a regulatory vulnerability.
• Delivery vehicle. LNP encapsulation delays endosomal exposure of RNA cargo and changes when pattern recognition receptors encounter dsRNA impurities, so the specification rationale must account for the formulated product, not just the drug substance.
• Dosing frequency amplifies risk in repeat-dose regimens, where subclinical innate activation at a single dose may produce cumulative immune priming over time. Specifications for repeat-dose programs should be justified using dosing-interval data where available.

FDA vs EMA: How Their dsRNA Requirements Compare

FDA and EMA follow the same underlying logic. Characterize the impurity, validate the method, justify the limit. The difference lies in emphasis, since FDA reviewers focus on the analytical validation package and the specification rationale as discrete documentary elements, while EMA reviewers expect one coherent mechanistic narrative connecting impurity formation, analytical strategy, and specification.

How to Reduce dsRNA Impurities During mRNA Manufacturing

Regulatory defensibility starts at process design. A process generating low and consistent dsRNA levels from early development forward is easier to file, scale, and defend at comparability than one where control gets retrofitted through purification alone.

• Template engineering reduces duplex formation by minimizing self-complementary motifs in the sequence design.
• Polymerase reaction optimization, including magnesium concentration, NTP ratios, and reaction duration, controls the kinetics of antisense transcription that generate duplex byproducts upstream of purification.
• Chromatographic options including anion-exchange, ion-pair, and cellulose-based purification separate duplex species from single-stranded transcript through structural and charge differences, with method selection depending on scale, equipment, and the dsRNA profile of the specific construct.
• In-process monitoring at post-IVT and post-purification stages, using the same validated method applied to lot release, provides the longitudinal data needed to demonstrate process control and detect drift before it reaches the drug substance.

How Is dsRNA Detected and Measured in mRNA Products

A dsRNA specification only holds up in a submission if the method behind it is validated for specificity, limit of detection (LOD), limit of quantification(LOQ), linearity, accuracy, and precision. Reference standard choice and modification status both affect how reliably a method detects dsRNA. Per ICH Q2(R2), these are the parameters that apply:

Specificity requires that the method discriminates dsRNA from single-stranded RNA, referred to as ssRNA, abortive transcripts, and other IVT-derived impurities. For J2 ELISA-based methods, specificity depends on antibody selectivity and the structural characteristics of the reference material used. J2 monoclonal antibody recognizes double-stranded RNA in a length-dependent manner, with reduced sensitivity below approximately 40 base pairs, which means specificity data must include performance against dsRNA species across the biologically relevant size range rather than against a single synthetic duplex.

LOD and LOQ require a calibrated dsRNA reference standard of defined length and sequence. Because no pharmacopeial dsRNA reference material currently exists for RNA therapeutics, sponsors qualify in-house standards using orthogonal methods and document the characterization in the analytical validation package. For example, some of our clients use our dsRNA reference standards spanning 500 to 1500 base pairs, with or without m1Ψ-modified RNA, to construct calibration curves and establish LOD and LOQ values specific to each assay format. Using a size-matched reference standard matters because dsRNA detection sensitivity varies with duplex length, and a calibration standard that does not reflect the expected impurity size range will produce LOD and LOQ values that misrepresent the method’s actual detection capability.

One consideration specific to RNA therapeutics is the interaction between nucleoside modification and dsRNA detection. m1Ψ-modified dsRNA is not recognized identically to unmodified dsRNA by all antibody-based assays. If the drug substance contains m1Ψ-modified RNA, validation should confirm that the detection method performs consistently across both modified and unmodified dsRNA species, or clearly define the boundary of its applicability.

Accuracy and precision data must cover the concentration range relevant to the specification, typically from the LOQ to at least twice the proposed acceptance limit. Inter-day and inter-analyst precision are both required. For methods that will support lot release at clinical stage, robustness data demonstrating performance under deliberate variations in key method parameters should also be included.

Frequently Asked Questions

Does FDA require a specific numeric dsRNA limit?

FDA expects impurity characterization and a scientifically justified specification, but does not define a universal numeric dsRNA threshold. The acceptance criterion must be data-driven and clinically justified within the product context.

Does EMA publish a fixed dsRNA specification?

No. EMA requires impurity control and validated methods but does not prescribe a global numeric dsRNA limit. Acceptability is product-specific and must be argued from mechanistic and analytical data.

When must dsRNA be designated a CQA?

dsRNA should be designated a Critical Quality Attribute when data demonstrate it affects potency, safety, or batch consistency within the defined product context, following EMA’s ATMP guideline and ICH Q5E. The designation is not automatic and requires evidence. A non-CQA designation equally requires a documented justification that dsRNA impact is negligible at controlled levels.

Is dsRNA relevant for all RNA modalities?

dsRNA is an intrinsic impurity for all IVT-derived RNA modalities including mRNA, saRNA, and dsRNA-based constructs. For circRNA, structural differences in the synthesis pathway mean the dsRNA impurity profile differs from linear IVT products, and the analytical strategy should reflect that distinction.

Can LNP encapsulation eliminate dsRNA risk?

No. Encapsulation modifies the kinetics of immune sensing by delaying endosomal exposure but does not remove duplex RNA impurities from the drug substance. The dsRNA specification applies to the drug substance before formulation, and LNP encapsulation efficiency should be considered separately when justifying the specification for the final drug product.

Is J2 ELISA sufficient for regulatory submission?

J2-based ELISA may be sufficient if it is fully validated to ICH Q2(R2) standards with defined LOD, LOQ, specificity across the relevant size range, and demonstrated accuracy and precision. Orthogonal confirmation using a structural or chromatographic method strengthens the submission, particularly at Phase II and beyond.

What happens to a dsRNA specification after a process change?

A process change triggers a comparability obligation under ICH Q5E. The comparability package must demonstrate consistency in the dsRNA impurity profile, not only the total reported level. If the change shifts dsRNA species distribution, that shift requires analytical characterization and a risk assessment even when the total value remains within specification.

Are dsRNA regulatory expectations changing in 2026?

Yes. As more RNA therapeutics advance toward licensure, regulatory scrutiny of impurity control and comparability is increasing. EMA’s July 2025 ATMP guideline and FDA’s 2023 draft comparability guidance both signal a direction toward more detailed process-level justification of impurity control strategies.

References

[1] European Medicines Agency. Guideline on quality, non-clinical and clinical requirements for investigational advanced therapy medicinal products in clinical trials. EMA/CAT/22473/2025. Effective July 1, 2025. https://www.ema.europa.eu/en/documents/scientific-guideline/guideline-quality-non-clinical-clinical-requirements-investigational-advanced-therapy-medicinal-products-clinical-trials\_en.pdf

[2] U.S. Food and Drug Administration. Chemistry, Manufacturing, and Control Information for Human Gene Therapy Investigational New Drug Applications. Final Guidance for Industry, January 2020. https://www.fda.gov/media/113760/download

[3] International Council for Harmonisation. Q5E, Comparability of Biotechnological or Biological Products Subject to Changes in Their Manufacturing Process. Effective June 2005. https://www.ich.org/page/quality-guidelines

[4] International Council for Harmonisation. Q2(R2), Validation of Analytical Procedures. Adopted November 2023, effective June 2024. https://www.ich.org/page/quality-guidelines

[5] U.S. Food and Drug Administration. Manufacturing Changes and Comparability for Human Cellular and Gene Therapy Products. Draft Guidance for Industry, July 2023. https://www.fda.gov/media/170731/download

Read the original article and source figures