Table 2.

Considerations for the safety assessment of viral vectors

Topic and Guidance Ref.	Considerations for the Safety Assessment of Viral Vectors
Replication competency 2,50, Table 1—Guidance L, N, Y	If vector is replication competent, what is the therapeutic rationale and implications to the patient? What are risks to close contacts/health care providers from accidental exposure or shed virus? Does the shed or released virus pose an environmental risk? Does patient treatment or immune status affect risk of local/systemic viremia? If vector is not intended to be replication competent, are replication-competent viral particles present as an impurity during manufacture? Have nonessential accessory genes been removed or other engineering/manufacture steps been taken to minimize risk that recombination could lead to replication competency (e.g., low sequence homology, use of self-inactivating vectors, separation of vector sequences onto multiple plasmids)? For oncolytic viruses, demonstrate tumor selectivity.
Viral strain/serotype2	What is the derivation of the parental viral strain/serotype used for the vector? Is it a passaged strain or novel isolate? How does this affect potency/viremia or other viral properties? Are there known liabilities from infection by the (wild-type) viral vector (including risks to pregnancy and immunocompromised patients)?
Engineering/attenuation 2, Table 1—Guidance L, Y	What genetic elements comprise the therapy and what sequence modifications have been made to the parental viral strain? How do these affect replication competency, relative potency, biodistribution/tropism, cell-specific replication, or interactions with host immunity? Does vector transgene lead to expression of a nontolerated or immunogenic protein/oligonucleotide? Is expression local/systemic, acute/chronic?
Recombination 2	If viral vector is replication competent, what are possible risks from novel strains arising from recombination with wild-type viruses (or itself)? If vector is not intended to be replication competent, what modifications to virus or product manufacture are taken to minimize risk of recombination (e.g., separation of viral components onto distinct plasmids, incorporation of self-inactivating designs)? Are there homologous sequences in the genome that generate ‘hot spots' for recombination?
Complementation	If virus has been attenuated, can other viruses complement for the deficiency to restore virulence?
Tropism (biodistribution) 2,50, Table 1—Guidance N, Y	What is the natural tissue tropism of the parental viral strain? Does viral engineering alter tropism (i.e., pseudotyping) or mechanisms of infection? If so, how does this potentially impact biodistribution/safety? Is there evidence of vector in germ cells or of transgenerational gene transfer (or integrated vector in germinal DNA)? How is the virus cleared and what are the routes secretion/excretion that could give rise to secondary exposure?
Host range 2,50, Table 1—Guidance L, Y	What species can be infected with the virus? Does this pose an environmental risk from viral shedding/accidental release? Does species host range support use of selected species for pharmacology, biodistribution, or safety assessment? Is this supported by evidence of transfection, transduction, infectivity, and/or replication in cells/tissues of the selected species? Do genetic elements or expressed proteins of the transduced vector function in that species (i.e., promoter or regulatory elements, expressed protein/oligonucleotide)? Are there significant differences in transduction efficiency between species and how does this impact considerations for translation of the therapeutic dose–response and safety margin to humans.
Vector copy number 2	How many copies of the transgene are incorporated per target cell? Does the safety program support the copy number range? Note: Standards for acceptable vector copies per cell have not been formalized, but for retroviral vector levels, <5 are generally accepted by regulatory agencies for ex vivo engineered cell products.17
Genomic integration, genomic risk, insertional mutagenesis 2, Table 1—Guidance Y, Z	Does the virus recombine or integrate with the host genome? [Note that episomal vectors, such as AAV, have low-frequency integration events that are prompting regulatory expectations for these assessments.] If so, is the observed integration profile similar to that reported for the wild-type viral strain? Do the sites of integration disrupt cell/gene regulatory mechanisms? Could these integration events promote mutagenic transformation (e.g., insertional disruption, insertion of promoter sequences) or cause other forms of genomic risk (e.g., breaks, recombination)? Provide detailed analytical and bioinformatic methods for regulatory review.
Availability of antiviral therapy	For replicating virus, are antiviral therapies available in case of a virus-related adverse effect or accidental exposure? Has viral engineering altered viral response to antiviral therapy?
Viral load, persistence, and shedding 2,50, Table 1—Guidance L, N	For replicating viruses, what is the viral burden over time? Is viral replication dependent on patient-specific factors (e.g., tumor load, immune status, comedication, age)? What is the persistence of the virus in the patient (or nonclinical model)? For replication-competent viruses, what are the routes for viral shedding? What precautions are required to minimize environmental release or exposure to close contacts? How long are precautions required to protect from exposure to shed virus?
Latency/reactivation 2,50, Table 1—Guidance L, N	Is there evidence that the virus can undergo a latent infection, with subsequent reactivation/viremia? Does viral engineering, including attenuation, alter this risk? How does latency/reactivation affect shedding risk? Are nonclinical models available to assess risk of latency/reactivation?
Special populations 2	For attenuated or replication-competent viruses, does disease state or host immune status confer differential risk? Should special precautions be warranted for coadministration of agents that suppress or alter immunity? For both replication-competent and defective viruses, are there risks to pregnancy or of maternal/fetal transfer?
Translational systems 2, Table 1—Guidance L, Q	Are patient-derived or engineered animal cells, or disease animal models (natural or genetically altered) available that can inform pharmacology proof of concept, safety, and/or dose selection? Can alternative/surrogate viral serotypes vector constructs facilitate pivotal nonclinical pharmacology or safety studies as compared with the intended human clinical product?

AAV, adeno-associated virus.