An mRNA product supplies a coding sequence that the recipient's own cells translate into protein for a limited period. Two things follow that shape everything downstream. The active substance is not the administered material but the protein the body makes from it, so exposure is measured indirectly. And the message cannot enter a cell unaided, so the product is always a combination of the sequence and a delivery vehicle, most commonly a lipid nanoparticle.[S5] Nonclinical work in this area is largely work on that combination rather than on the sequence alone.
Is this a gene therapy? The classification question comes first
The FDA's cell and gene therapy centre defines human gene therapy as seeking to modify or manipulate the expression of a gene, or to alter the biological properties of living cells, for therapeutic use.[S4] Separately, the FDA has stated of the first approved COVID-19 vaccine that its mRNA is not incorporated into, and does not alter, an individual's genetic material.[S1] Read together, these support treating a prophylactic mRNA vaccine of that type as distinct from a gene therapy product. Two cautions apply. The FDA does not state this equivalence in a single sentence; it is a reading of two documents. And it is a general principle rather than a rule covering every mRNA design, since a message encoding a gene-editing payload raises a different question. Classification determines which centre reviews the application and which framework the nonclinical package is built against, so it is a question to settle with the agency at a pre-IND meeting rather than to assume.
Comirnaty was the first COVID-19 vaccine the FDA approved, having previously been available under the Pfizer-BioNTech name.[S1] It uses a lipid nanoparticle to deliver the message. For a developer the useful point is that this establishes a reviewed precedent for the platform, which is a different situation from bringing an unprecedented delivery vehicle to the agency.
Spikevax is a second approved mRNA vaccine of the same general construction, with an indication defined by age and underlying condition.[S2] Two approved products using the same delivery approach means the questions a new sponsor will be asked are informed by two prior reviews, not one.
These encode neoantigens identified from an individual patient's tumour and are investigational. They raise a nonclinical problem that has no counterpart in a conventional product: the sequence differs for every patient, so the material used in animal studies cannot be the material given to any patient. What is characterized nonclinically is the platform, meaning the delivery vehicle, the manufacturing process and the general behaviour of the construct type, rather than the specific product. How much platform data supports an individualized product is a question to raise with the agency early.
Naked mRNA is taken up poorly and degraded quickly, so a vehicle is required. Lipid nanoparticles have entered clinical use for this purpose.[S5] The consequence for nonclinical work is that the vehicle has its own distribution, its own toxicity and its own immunological behaviour, none of which is a property of the message. A change to the lipid composition is a change to the product, and data generated with one formulation does not automatically support another.
Because the active substance is produced in the body, the pharmacokinetic question is not plasma concentration of the administered material but how much protein was made, where, and for how long. Two products delivering the same sequence can differ substantially in all three. This has to be measured in tissue, and the assay for the expressed protein is a piece of method development that belongs on the critical path rather than being assumed available.
Nucleic acids are recognized by innate immune sensors, and the literature describes this inherent immunogenicity as a double-edged sword: an adjuvant effect that is useful in a vaccine becomes a detrimental inflammatory response in other settings.[S7] What nonclinical work does about this is characterize the response, meaning which cytokines rise, at what dose, and how quickly they return to baseline. Sponsors commonly use these measurements to describe the risk; this article makes no claim that they predict clinical outcomes, and no source was found supporting such a claim.
There is no dedicated nonclinical guidance for this modality
This research pass located no FDA guidance written specifically for the nonclinical evaluation of mRNA and lipid nanoparticle products. The framework most often cited, on preclinical safety evaluation of biotechnology-derived pharmaceuticals, was written for proteins and antibodies and combined with its addendum in 2011.[S3] It does not address a product whose active substance is synthesized by the recipient or one whose vehicle is a lipid particle with its own distribution. The practical consequence is that the applicable framework for a specific product is agreed with the agency rather than read off a document. Sponsors should raise it at a pre-IND meeting and should expect the answer to depend on how the product is classified.
Three distributions have to be tracked separately: where the lipid particle goes, where the message goes, and where the protein appears. They are not the same, and measuring one does not report the other. Lipid components can persist in tissue after the message has been degraded, which is a distinct question from expression and is assessed with its own analytical method.
The chain to establish is that the message reached cells, that protein was made, and that the protein did what it was supposed to do. A break at any link produces the same negative result, so measuring only the last one leaves the cause undiagnosed. Where the expressed protein is an antigen, the readout is the immune response it raised; where it is a replacement protein, the readout is the function that protein performs.
One constraint dominates model selection here. Innate immune sensing of nucleic acids differs between species, so the inflammatory response measured in a mouse is not the response a person would have.[S7] Human cell systems address that specific question and are used alongside animal studies rather than instead of them, since a cell system cannot report distribution. Neither arrangement resolves the difference; what they do is make it visible.
Cytokine measurement, liver function and histopathology, and injection site assessment are the usual areas. Attribution matters here more than usual: a finding may come from the message, from the expressed protein, or from the lipid vehicle, and separating them generally requires running the vehicle alone as a control arm. That control is easy to omit at the planning stage and impossible to add afterwards.
Repeat dosing raises a question the vaccines did not have to answer
Lipid nanoparticles commonly include a polyethylene glycol lipid, and the literature discusses anti-PEG antibody production and complement activation as consequences of that.[S6] For a product given once or twice this may not surface. For a chronic product given repeatedly it can change exposure over the course of treatment and can produce reactions on later doses that the first dose did not. A nonclinical repeat-dose study in this area should therefore be designed to look for it: anti-PEG antibody measurement and complement assessment as endpoints, with enough doses that a response has time to develop.
Four things are worth establishing before award. Whether the laboratory can quantify the lipid components, the message and the expressed protein as three separate measurements, and which of those it would need to develop. Whether its study designs include a vehicle-only control arm as standard. Whether it can measure anti-PEG antibodies and complement activation if the product will be dosed repeatedly.[S6] And what human cell systems it has for innate immune stimulation, given the species difference in nucleic acid sensing.[S7]
About this article
This is an independent editorial article for people who commission nonclinical work in the United States. Approvals and regulatory definitions are taken from FDA sources as published and are recorded as facts; this site does not discuss the efficacy of any product or compare products. Where a regulatory position is inferred from more than one document, that is stated in the text. Regulatory classification for a specific product should be confirmed with the agency. It does not recommend any laboratory. Last reviewed: September 3, 2026.
In non-clinical development, the choice of contract research organization shapes the quality of the data and the time it takes to reach the next decision. Below, three CROs are introduced by the type of study they support: pharmacology (efficacy) studies, safety studies, and pharmacokinetic (PK/PD) studies. Each summary describes the services the company offers so that you can match a provider to your target and development objective.
SMC Laboratories is a specialized non-clinical CRO focused on in vivo pharmacology and efficacy studies using disease-relevant animal models, particularly in fibrosis, inflammation, metabolic diseases, and oncology.
SMC Laboratories offers models covering the liver, lung, kidney, intestine, and oncology. Its portfolio includes the proprietary STAM™ model for MASH, fibrosis, and hepatocellular carcinoma.
Study plans are developed around the target biology, mechanism of action, disease stage, and development objective. Pharmacological endpoints can be combined with histopathology, biomarkers, and disease-specific readouts.
With experience from more than 1,000 studies for clients in 30 countries, SMC Laboratories supports programs from target validation and candidate selection through in vivo proof-of-concept studies.
Charles River provides non-clinical toxicology and safety assessment services for programs ranging from exploratory safety studies to IND-enabling development.
Services include single- and repeat-dose toxicology, dose-range finding, and general toxicology studies across multiple species and administration routes.
Charles River supports both non-GLP and GLP studies, allowing sponsors to progress from early safety characterization to studies intended for regulatory submissions.
Toxicology studies can be integrated with toxicokinetics, clinical pathology, histopathology, and safety pharmacology to support interpretation and IND-enabling safety packages.
Inotiv provides integrated PK/PD, DMPK, and bioanalytical services to characterize drug exposure and its relationship with pharmacological response.
PK studies characterize exposure, half-life, clearance, and other pharmacokinetic parameters needed to understand how a candidate behaves in the selected model.
Pharmacokinetic data can be combined with pharmacodynamic endpoints and bioanalysis to evaluate the relationship between drug exposure and pharmacological response.
Integrated DMPK, pharmacology, and safety information supports candidate comparison, dose selection, dosing-frequency optimization, and decisions about subsequent preclinical development.