Emerging GT Modalities — mRNA, CRISPR, Base Editing: CMC at the Next Frontier

The CMC frameworks developed for AAV and lentiviral vectors are being applied to mRNA-based gene therapies, CRISPR/Cas9 delivery systems, and engineered virus-like particles — sometimes correctly, often not.

The gene therapy CMC discipline spent two decades building a rigorous analytical and regulatory architecture around viral vectors — capsid characterization, genome titer, potency, adventitious agent testing, replication competence assays. That architecture is now being borrowed, selectively and sometimes carelessly, by sponsors developing mRNA-based gene replacement products, CRISPR-Cas9 editing systems, and engineered nucleic acid delivery platforms. The regulatory consequence of that borrowing — when it is done without understanding where the analogy holds and where it breaks — is an IND CMC package that arrives at CBER Office of Therapeutic Products (OTP) with critical characterization gaps that trigger information requests, clinical holds, or, at minimum, review cycles that delay first-in-human dosing by twelve to eighteen months. The thesis here is not that existing GT frameworks are wrong — it is that next-generation modalities require a deliberately tailored CMC strategy that draws selectively from established GT, oligonucleotide, and biologics precedents while squarely acknowledging the modality-specific analytical and regulatory gaps that no existing guidance document has yet fully resolved.

mRNA Therapeutics CMC: Where mRNA Drug Products Share Architecture With AAV and Where They Diverge

The most useful regulatory anchor for mRNA-based gene therapy CMC is not actually a gene therapy guidance document — it is the FDA Guidance for Industry: Development and Licensure of Vaccines to Prevent COVID-19, which established the first substantive FDA-published analytical framework for mRNA drug substance characterization. Sponsors building IND CMC packages for mRNA-based gene replacement programs (distinct from vaccines) have used that precedent, appropriately, to justify analytical approaches for capping efficiency, poly-A tail length distribution, and dsRNA impurity quantification. The critical distinction, however, is that COVID-19 mRNA vaccines are prophylactic biologics reviewed under a vaccine framework, while mRNA-based gene therapies — particularly those encoding therapeutic proteins for rare disease — are reviewed by CBER Office of Therapeutic Products (OTP) under the CMC Information for Human Gene Therapy Investigational New Drug Applications (INDs) (2020), which imposes gene therapy-specific requirements including long-term follow-up considerations, potency assay development tied to the therapeutic mechanism of action, and product characterization depth that exceeds the vaccine precedent.

The most consequential mRNA-specific CMC gap I have seen in early IND packages is the failure to characterize double-stranded RNA (dsRNA) as a process-related impurity. In vitro transcription reactions — the standard manufacturing route for mRNA drug substance — generate dsRNA as a byproduct of the IVT polymerase, and dsRNA is a potent innate immune activator via Toll-like receptors and MDA5. The specification criterion is dsRNA content by dot blot assay of no more than 5% — a threshold drawn from the vaccine precedent and increasingly expected by CBER reviewers for therapeutic mRNA programs — and when that characterization is absent from the IND drug substance section, the reviewer has no basis for assessing the immunogenicity risk profile of the molecule, independent of any nonclinical data package. ICH Q6B, which governs specifications for biotechnology-derived proteins, provides the analytical framework for impurity identification and qualification, but it does not anticipate RNA-specific process impurities; sponsors must bridge that gap explicitly in the CMC narrative rather than relying on the reviewer to infer it.

A second divergence from AAV precedent involves potency assay design. For an AAV gene therapy, the potency strategy is well-established: a cell-based assay measuring the biological effect of the transgene product, calibrated to a reference standard, with acceptance criteria tied to the mechanism of action. For an mRNA gene therapy, the potency assay must capture not only protein expression but also translational efficiency and duration of expression — parameters that are fundamentally time-dependent in a way that AAV potency assays are not. CBER Office of Therapeutic Products (OTP) presentations on novel GT modalities have flagged the absence of validated potency assays as a recurring deficiency in mRNA gene therapy INDs, and ICH Q2(R2)/Q14, the updated guideline on analytical procedure validation and development, now provides the lifecycle framework within which early-phase potency assays should be developed and justified even when full validation is not yet required at IND.

CRISPR-Based Therapies: The CMC Complexity of a Two-Component (Guide RNA + Cas9) Drug Substance

CRISPR-Cas9 gene editing programs present a CMC complexity that is categorically different from single-component GT products because the drug substance — whether delivered as AAV dual-vector, LNP-formulated mRNA plus guide RNA, or pre-assembled ribonucleoprotein complex — consists of at minimum two functionally distinct molecular entities that must each be individually specified and jointly characterized for potency. The delivery modality choice drives the entire CMC strategy. AAV dual-vector CRISPR delivery, in which Cas9 is packaged in one AAV capsid and the guide RNA in a second, is constrained by the 4.7 kb packaging capacity of standard AAV serotypes — a physical limitation that has driven the use of split-intein Cas9 constructs or the smaller SaCas9, approximately 3.2 kb, to fit within a single vector. That architectural choice has direct CMC implications: a dual-vector program requires two independent drug substance CMC packages, two potency assay strategies, and a defined drug product combining ratio specification that CBER will scrutinize for clinical dose-response justification.

For LNP-delivered CRISPR programs — where Cas9 mRNA and the synthetic guide RNA are co-encapsulated or separately formulated — the drug substance characterization requirements expand to include Cas9 mRNA capping efficiency of no less than 90% by m7G cap analysis, poly-A tail length of 100 to 150 nucleotides, and dsRNA content of no more than 5%, alongside guide RNA specifications including sequence identity confirmed by oligonucleotide sequencing, purity of no less than 90% by HPLC or denaturing capillary gel electrophoresis, and endotoxin below 0.5 EU per microgram. These are not aspirational targets — they are the analytically grounded thresholds that CBER reviewers are applying to IND submissions in this space, informed by the vaccine mRNA precedent and the emerging CRISPR-specific guidance framework. The FDA Guidance for Human Gene Therapy for Rare Diseases (2020) applies to in vivo CRISPR editing programs, and it explicitly encompasses gene editing products under the gene therapy IND framework, meaning the full analytical rigor of CMC Information for Human Gene Therapy INDs (2020) applies.

The most operationally dangerous CMC gap in CRISPR IND packages is the absence of a documented off-target editing characterization strategy. CBER-specific CRISPR CMC expectations — articulated in CBER Office of Therapeutic Products (OTP) presentations and reinforced in EMA/CAT’s guideline on quality, non-clinical, and clinical aspects of medicinal products containing genetically modified cells, which addresses CRISPR-edited cell products as ATMPs — require genome-wide off-target analysis by GUIDE-seq, CIRCLE-seq, or SITE-seq conducted in cell types relevant to the clinical indication, and that analysis must be documented in the IND CMC package as part of the product characterization strategy, not deferred entirely to the nonclinical pharmacology section. The editing efficiency assay — allele editing frequency measured by NGS at the target locus, including quantification of HDR, NHEJ, and deletion outcomes — must be developed and presented as the potency release method, and CBER has been explicit that a potency assay not validated or at minimum qualified per ICH Q2(R2)/Q14 principles at IND will generate a deficiency. For ex vivo CRISPR applications, such as CAR-T manufacturing using TRAC locus insertion for CAR gene integration, editing efficiency of no less than 70% of T cell alleles at the target locus has been the analytical benchmark, with residual RNP characterization by ELISA required after the final wash step to confirm transient Cas9 exposure and mitigate CBER’s stated concern about persistent Cas9 protein immunogenicity.

Base Editing and Prime Editing: The Next-Generation GT Modalities Approaching IND Submission

Base editing and prime editing represent the leading edge of the CRISPR derivative landscape, and both modalities are now at or approaching IND submission readiness for rare genetic diseases — which means CMC teams are currently being asked to build regulatory packages in an environment where CBER has not yet issued modality-specific guidance and EMA/CAT has addressed base editing only in the context of its broader genetically-modified-cells guideline. Base editors — cytosine base editors and adenine base editors — consist of a Cas9 nickase fused to a deaminase domain, delivering single-nucleotide changes without a double-strand break; prime editors incorporate a reverse transcriptase domain and a prime editing guide RNA, capable of all twelve possible transition and transversion mutations plus small insertions and deletions. Both architectures are substantially larger than standard SpCas9, which directly affects the AAV packaging strategy and forces sponsors toward split-intein delivery approaches or LNP-mRNA platforms.

The regulatory classification question — whether a base editor or prime editor delivered by LNP qualifies as a gene therapy product under CBER Office of Therapeutic Products (OTP) jurisdiction or as a novel biologic under another framework — is not settled, and CBER has not issued written classification guidance. That ambiguity is a CMC liability: sponsors who proceed to IND submission without a written classification confirmation from CBER risk building a CMC package to the wrong regulatory standard, potentially requiring substantive restructuring after the IND is filed. The practical resolution is a pre-IND meeting — specifically a Type B pre-IND meeting — at which the classification question is placed explicitly on the agenda and CBER’s written response, captured in the meeting minutes, becomes the regulatory anchor for the CMC package development strategy.

The analytical gaps for base and prime editing are analogous to but not identical to CRISPR nuclease programs: off-target base conversion assessment requires bystander editing analysis (unintended base conversions within the editing window) in addition to genome-wide off-target locus analysis, and no standardized method has been validated for this in a regulatory submission context. CBER reviewers participating in DBRR-adjacent meetings on novel modalities have indicated that sponsors should propose the analytical strategy for bystander editing characterization in the IND and commit to a development timeline — but the absence of a defined acceptance criterion means the strategy must be scientifically justified rather than simply cited against a guideline.

XGene Emerging GT Modality CMC Readiness Framework

Building CMC Capabilities for Emerging GT Modalities Before the Regulatory Framework Is Fully Defined

The XGene Emerging GT Modality CMC Readiness Framework is a structured pre-IND assessment that maps known analytical requirements, guidance gaps, and CBER/EMA precedents for mRNA, CRISPR, base editing, and prime editing programs — building the IND CMC package with the highest probability of deferral-free review in an environment where authoritative guidance often lags clinical development timelines.

Step 1: Regulatory Classification and Jurisdictional Anchoring. Before a single analytical method is selected, the modality must be classified — gene therapy product vs. biologic vs. ATMP — and that classification must be confirmed in writing with CBER (and EMA/CAT for parallel EU development) through a pre-IND or scientific advice meeting. The classification determines which guidance documents govern the CMC package, which CBER division reviews it, and what the potency assay standard of evidence must be. Skipping this step is the single most common root cause of IND CMC structural failures in novel modality programs.

Step 2: Modality-Specific Analytical Gap Analysis Against Published Precedent. Each molecular component of the drug substance — mRNA, guide RNA, Cas9 protein, LNP formulation, viral vector — is mapped against the analytical requirements established in the COVID-19 mRNA vaccine guidance, CMC Information for Human Gene Therapy INDs (2020), ICH Q6B, and ICH Q2(R2)/Q14, with explicit identification of where those frameworks apply directly, where they apply by analogy, and where genuine regulatory gaps exist that require prospective scientific justification in the IND.

Step 3: Off-Target and Bystander Editing Characterization Strategy Documentation. For all CRISPR, base editing, and prime editing programs, the IND CMC package must include a documented genome-wide off-target analysis strategy specifying the assay platform (GUIDE-seq, CIRCLE-seq, or SITE-seq), the relevant cell type, the analytical sensitivity, and the follow-up plan for any off-target sites identified. For base editors, bystander editing within the editing window must be separately characterized. This documentation belongs in the drug substance characterization section of Module 3, not solely in the nonclinical pharmacology module.

Step 4: Potency Assay Development Roadmap Aligned to Mechanism of Action. An editing efficiency assay by NGS at the target locus — quantifying HDR, NHEJ, and deletion frequencies — must be positioned as the primary potency release method and presented with a development and validation roadmap consistent with ICH Q2(R2)/Q14. CBER will not accept the absence of a potency strategy at IND for gene editing products, and a characterization-only potency narrative without a path to release specification will generate a deficiency in first review.

Sponsors who engage XGene’s Emerging GT Modality CMC Readiness Framework before IND submission leave the pre-IND process with a fully documented analytical strategy, confirmed regulatory classification, and a CMC package architecture that CBER reviewers can evaluate without requesting fundamental restructuring.

The regulatory frameworks for mRNA gene therapies, CRISPR-based editing products, and base and prime editing systems are being written in real time — but the analytical requirements CBER is applying at IND review are not waiting for finalized guidance. Sponsors who treat the absence of a modality-specific FDA guidance document as permission to defer dsRNA characterization, off-target editing analysis, or potency assay development will encounter those deficiencies as information requests after IND filing, not before. The cost is not merely a delay in review — it is the erosion of the first-mover advantage that early IND submission is designed to capture. The CMC teams building these packages now have the opportunity to define the analytical precedent for their modality, or to discover at review that CBER’s expectations were higher than their package anticipated.

For your mRNA-based gene therapy or CRISPR delivery program, can you identify today whether your IND CMC package includes dsRNA impurity characterization (for mRNA) or off-target editing analysis strategy (for CRISPR), and whether the regulatory classification of your product as a gene therapy product vs. biologic has been confirmed in writing with CBER before IND submission?