In vivo therapeutic engineering is often framed as the next logical evolution of ex vivo cell and gene therapy. By eliminating complex manufacturing steps and enabling direct delivery of therapeutic payloads into the body, these approaches promise broader access, faster treatment timelines, and potentially lower cost.
However, this framing overlooks a fundamental shift that occurs when engineering moves from the laboratory to the patient. In vivo approaches are not simply streamlined versions of established modalities. They represent a fundamentally different biological paradigm in which control shifts from the developer to the patient’s biology.
From a genomic integrity standpoint, this shift introduces one of the most important challenges facing emerging in vivo therapeutic platforms today. Developers lose direct visibility into what is actually being created inside the patient.
At KROMATID, we believe the long-term success of in vivo therapeutic engineering will depend not only on delivery efficiency or editing performance, but also on the ability to measure, understand, and manage the genomic consequences of engineering cells within the body.
The Most Important Opportunity to Study Risk Occurs Before the Patient
The genomic outcomes of in vivo engineering ultimately unfold in the patient. Preclinical model systems cannot fully predict those outcomes, but they provide the only opportunity to investigate and begin to de-risk potential genomic consequences before a therapy reaches the clinic.
During development, delivery systems and engineering strategies are optimized in model systems including cell lines, primary cells, organoids, and animal models. Yet the structural consequences of these interventions, the genomic architecture that may be created during engineering, are often evaluated only indirectly.
This gap is becoming increasingly important as genome engineering technologies move into the clinic. Clinical studies have already demonstrated the feasibility of performing CRISPR-based editing directly in patients using systemic delivery approaches such as lipid nanoparticles.¹
At the same time, research has shown that genome editing can generate complex structural genomic outcomes including large deletions, chromosomal rearrangements, and chromothripsis. These events extend well beyond the small insertions or deletions typically measured by sequencing assays.²–⁴
These findings highlight an emerging reality. Genome engineering outcomes are often more complex than originally anticipated, and understanding those outcomes requires analytical tools capable of observing structural genomic changes at the cellular level.
Control Was Never Just Operational, It Was Genomic
Ex vivo cell and gene therapies advanced rapidly in part because developers retained control over critical genomic variables prior to administration. Engineered cells could be characterized before infusion, allowing investigators to determine:
• which cells were modified
• whether genetic cargo was integrated or edits were introduced
• how many integrations or edits occurred per cell
• where integrations or edits occurred within the genome
• how heterogeneous the engineered cell population was
• whether chromosome-scale genomic integrity was preserved
Over time, it became clear that these genomic features can directly influence therapeutic durability, toxicity, and long-term safety.
In vivo therapeutic approaches relinquish many of these controls by design. Once a delivery system is administered systemically, developers no longer determine:
• which cell types or subsets are modified
• whether unintended cell populations are affected
• how many edits or integrations occur per cell
• how genomic changes are distributed across the target population
• whether structural genomic alterations arise during the engineering process
This is not a failure of the approach. It is an inherent feature of in vivo engineering. However, it means genomic uncertainty is no longer an edge case. It becomes a central risk that must be measured, understood, and managed.
Delivery Systems Can Become Genomic Cargo
Delivery systems are often discussed in terms of efficiency, tissue targeting, or durability of expression. However, they also represent genomic variables that can influence the structural outcomes of engineering.
Integrating vectors introduce long-term expression but also create variable copy numbers and insertion sites across cells. Non-integrating delivery systems avoid intentional genomic integration but can still interact with the genome in complex ways.
For example, studies have demonstrated that adeno-associated virus vectors used for delivery can integrate into genomic DNA, particularly at sites of CRISPR-induced double-strand breaks. In some systems, fragments of the AAV genome, including inverted terminal repeats, have been incorporated directly into edited loci. Experimental studies have observed AAV vector sequences inserted into CRISPR-induced DNA breaks in up to approximately 47 percent of edited alleles, demonstrating that delivery vectors themselves can become genomic cargo during editing.⁵
AAV integrations have also been identified near oncogenic regions in human hepatocellular carcinoma, suggesting that vector integration events can have biological consequences under certain conditions.⁶
These findings illustrate an important principle. Delivery systems and engineering systems cannot be evaluated independently. Both contribute to the genomic architecture ultimately created in engineered cells.
Target Cell Populations Are Not Uniform
Another common assumption in in vivo therapeutic engineering is that target cell populations can be treated as uniform. In reality, cellular populations within tissues are highly heterogeneous and dynamic.
Cells differ in differentiation state, replication capacity, prior environmental exposure, and DNA repair activity. These biological differences can influence how cells respond to delivery systems and how they repair engineered DNA breaks.
Because of this complexity, bulk sequencing approaches are often insufficient to capture the full spectrum of genomic outcomes. Rare events such as structural rearrangements, high-copy integrations, or aberrant editing outcomes may occur in small subpopulations that remain invisible in population-averaged measurements.
Single-cell studies of genome editing outcomes have demonstrated mosaic genomic changes and chromosome-scale alterations that would not be detectable using bulk assays alone.⁷
For in vivo therapeutic strategies, where engineered cells cannot be isolated and characterized prior to infusion, understanding this cellular heterogeneity becomes even more important.
The Active Ingredient Problem
In vivo therapeutic engineering challenges traditional regulatory frameworks because the administered delivery system is not the therapeutic product. The actual therapeutic agent is the engineered cell that emerges after delivery.
This creates a fundamental analytical gap.
Batch release testing evaluates the delivery vector.
Clinical outcomes depend on the cells created inside the patient.
Bridging this gap will require analytical strategies capable of demonstrating reproducibility of genomic outcomes, detecting rare high-risk events, and monitoring genomic stability over time.
Regulatory agencies are already signaling increased attention to genomic outcomes in engineered cell therapies. In 2023, the U.S. Food and Drug Administration initiated an investigation into reports of secondary T-cell malignancies occurring after CAR-T therapy. In 2024, the agency required class-wide boxed warnings for all approved CAR-T products to highlight the potential risk of treatment-related T-cell cancers and recommended lifelong monitoring of treated patients for secondary malignancies.⁸
These actions reflect a broader regulatory emphasis on understanding the genomic consequences of engineered cell therapies. As gene and cell therapy platforms expand toward in vivo applications, developers will increasingly need analytical strategies that characterize the genomic outcomes their therapies create.
Building Genomic Visibility During Development
The most effective place to establish genomic understanding is during method development and preclinical evaluation, where delivery systems and engineering strategies can be studied across multiple model systems.
Sequencing technologies have dramatically improved the ability to detect nucleotide-level changes. However, they are not always optimized to resolve the full structural architecture of engineered genomes. Structural variants such as chromosomal rearrangements, large inversions, translocations, and complex multi-breakpoint events may span large genomic distances or involve repetitive genomic regions that are difficult to reconstruct using short-read sequencing alone.
Cytogenetic approaches that directly visualize chromosome structure provide a complementary perspective. These methods allow structural alterations to be observed at the single-cell level and enable detection of rare genomic configurations that may be diluted in population-based sequencing measurements.
KROMATID’s KROMASURE platform was developed to provide this level of genomic visibility across development stages.
KROMASURE PinPoint enables high-throughput analysis of cargo integration and edit-site integrity, including in non-dividing cells, allowing thousands of individual cells to be evaluated for integration outcomes. KROMASURE InSite provides deeper structural characterization in dividing cell models, enabling mapping of rearrangements and structural changes surrounding engineered loci. KROMASURE SCREEN and K-BAND assays provide genome-wide cytogenetic evaluation of structural abnormalities and chromosome-scale genomic stability.
Together, these approaches allow developers to understand genomic outcomes during the stages where interventions can still be optimized, before therapies reach patients.
Lower Cost Alone Will Not Offset Genomic Blind Spots
In vivo therapeutic engineering is often promoted as a path toward lower cost and broader access. While direct delivery approaches may reduce manufacturing complexity, cost alone does not determine adoption in advanced therapies.
Clinical durability, safety confidence, and regulatory clarity remain the primary drivers of value. Without robust genomic characterization, unexplained variability or unexpected safety signals can slow development, increase monitoring requirements, and ultimately negate operational advantages.
Access will improve not simply when therapies are easier to deliver, but when they are predictable, interpretable, and trusted.
The KROMATID View: In Vivo Engineering Must Become Measurable
In vivo, therapeutic engineering represents a powerful expansion of the cell and gene therapy toolkit. However, it cannot be evaluated or regulated using assumptions inherited from ex vivo models.
Progress in this field will ultimately depend on answering a single question.
Can we see, at single-cell resolution, what biology we are creating inside the patient?
With the right analytical tools, developers can characterize integration outcomes, detect rare genomic events, and monitor chromosome-scale genome integrity across thousands of engineered cells before therapies ever reach patients.
Without that visibility, developers are forced to trade manufacturing simplicity for biological opacity. With it, in vivo engineering becomes a controllable and characterizable modality capable of delivering on its promise.
References
- Gillmore JD et al. CRISPR-Cas9 In Vivo Gene Editing for Transthyretin Amyloidosis. NEJM, 2021.
- Kosicki M et al. Repair of double-strand breaks induced by CRISPR–Cas9 leads to large deletions and complex rearrangements. Nature Biotechnology, 2018.
- Leibowitz ML et al. Chromothripsis as an on-target consequence of CRISPR-Cas9 genome editing. Nature, 2021.
- Adikusuma F et al. Large deletions induced by Cas9 cleavage. Nature Methods, 2018.
- Hanlon KS et al. High levels of AAV vector integration into CRISPR-induced DNA breaks. Nature Communications, 2019.
- Nault JC et al. Recurrent AAV insertions in human hepatocellular carcinoma. Nature Genetics, 2015.
- Zuccaro MV et al. Allele-specific chromosome removal after CRISPR-Cas9 cleavage. Nature Biotechnology, 2020.
- U.S. Food and Drug Administration. FDA Requires Boxed Warning for T-Cell Malignancies Following Treatment with BCMA-Directed or CD19-Directed CAR-T Cell Therapy. 2024.
- Rurik JG et al. CAR T cells produced in vivo to treat cardiac fibrosis. Science, 2022.