Science Wiki

A viral vector is a modified virus that is used as a vehicle to deliver genetic material into cells. The design and production of such vectors is called viral vector engineering. Viral vectors are used in gene therapy, in vaccine development, in oncolytic therapy, in genome editing, and in basic research in molecular biology. Viruses are obligate intracellular parasites that have evolved efficient ways to bind to cells, enter them, evade some intracellular defenses, and deliver their nucleic acids to the cytoplasm or nucleus. In vector engineering, genes needed for replication and for pathogenesis are deleted or inactivated, and a therapeutic or antigen-encoding sequence (the transgene) is inserted in their place. This separates the delivery function of the virus from its disease-causing life cycle. The vector is not necessarily harmless. It can still cause immune responses, toxicity, and, in rare cases, cancer, and safety is a main subject of vector design and of clinical practice.

Non-viral methods, such as lipid nanoparticles, electroporation, and naked DNA, are also used for delivery, and viral vectors are chosen when efficient delivery, long-term expression, or specific tissue targeting is required. In 2017, the first gene therapy that uses an AAV vector for an inherited disease was approved in the United States, and by the mid-2020s, vectors based on adeno-associated virus, lentiviruses, gammaretroviruses, adenoviruses, herpes simplex virus, and poxviruses were in approved products or in advanced clinical trials.

The idea of using viruses to carry genes dates from the 1960s and 1970s, when Paul Berg and others constructed recombinant DNA from the virus SV40, and when the possibility of correcting genetic disease by gene transfer was first discussed. Retroviral vectors, which were derived from murine leukemia virus, were developed in the early 1980s by Richard Mulligan, Inder Verma, and others. The first clinical gene transfer trials, in 1989 and 1990, used retroviral vectors. A trial for adenosine deaminase deficiency (a form of severe combined immunodeficiency, SCID) began in 1990. In 1999, the death of Jesse Gelsinger, an 18-year-old participant in a trial of an adenoviral vector for ornithine transcarbamylase deficiency, who developed a severe inflammatory response, led to major changes in the oversight of trials. In the early 2000s, trials of a gammaretroviral vector for X-linked SCID (SCID-X1) cured many children but also caused T-cell leukemia in several of them, because the vector integrated near the proto-oncogene LMO2, and this led to the development of safer designs. Lentiviral vectors, which were developed in the late 1990s from HIV-1, and AAV vectors became the main platforms in the 2010s and 2020s.

Cis- and trans-acting elements

A vector system divides the functions of the virus into elements that must be present on the vector genome (the cis-acting elements) and functions that can be supplied from another source (the trans-acting functions). The vector genome contains the transgene with a promoter and the regulatory elements, flanked by the signals that are needed for packaging and, where relevant, for reverse transcription and integration. These are the long terminal repeats (LTRs) in retroviral and lentiviral vectors, and the inverted terminal repeats (ITRs) in AAV vectors. The viral proteins that are needed to make the particle are made from separate constructs.

Production

Vectors are produced in packaging or producer cells. The usual method is the transient transfection of human embryonic kidney 293 cells (HEK293 or 293T, which express the SV40 large T antigen) with several plasmids, although stable producer cell lines and baculovirus-insect cell systems (for AAV) are also used. The transfer plasmid carries the vector genome, which is the only nucleic acid that is packaged. Packaging plasmids carry the genes for the structural proteins and enzymes (such as gag, pol, and rev for lentiviruses, or rep and cap for AAV), and an envelope or capsid plasmid provides the surface proteins. For AAV, adenoviral helper genes (E1A, E1B, E2A, E4, and VA RNA) are supplied, and the 293 cells provide E1. Splitting the viral genes between several plasmids, with no overlapping homology, reduces the risk of recombination to a replication-competent virus (RCV), and the vectors are tested for RCV before use. The vector particles can transduce a cell once but cannot make new infectious particles. The virions are purified by ultracentrifugation, chromatography, and tangential-flow filtration. The titer is measured as vector genomes per milliliter (by quantitative PCR or digital PCR), as transducing units, or as infectious units. The transduction efficiency depends on the multiplicity of infection (MOI), which is the ratio of vector particles to target cells. If particles attach to cells independently at random, the fraction of cells that receive at least one particle follows a Poisson distribution:

so that an MOI of 1 transduces about 63 percent of cells, and an MOI of 3 transduces about 95 percent. In practice, the efficiency is lower than this ideal, because only a part of the particles are infectious and because cells differ in their susceptibility.

Tropism is the range of cell types that a virus or vector infects. It depends on the interaction between surface proteins and receptors on the cell, and on the ability of the cell to support later steps. For vectors, tropism can be changed in several ways:

• Pseudotyping is the incorporation of an envelope protein from another virus into the particle. The most common pseudotype is the lentiviral vector with the glycoprotein G of vesicular stomatitis virus (VSV-G). VSV-G binds the low-density lipoprotein receptor (LDLR) and other family members, which are widely expressed, so the vectors transduce a wide range of cells, from many species. It is not true that VSV-G changes the lentivirus from CD4-targeting to a vector that infects "almost all" cells without exception. Some cells are resistant, for example because they have few receptors, and quiescent T cells and hematopoietic stem cells are not efficiently transduced without special conditions. VSV-G gives particles that are sufficiently stable to be concentrated by ultracentrifugation. A disadvantage is its cytotoxicity, which makes stable packaging cell lines difficult, and its sensitivity to complement in human serum. Other envelopes include the glycoproteins of amphotropic murine leukemia virus, gibbon ape leukemia virus, measles virus, and baboon endogenous retrovirus (used to transduce hematopoietic stem cells and T cells), and targeted envelopes with antibody-derived or ligand domains.
• Capsid engineering. For AAV, tropism is determined by the capsid. Naturally occurring serotypes and variants differ in their tropism. AAV2 binds heparan sulfate proteoglycans and transduces many tissues, and it has been widely used in the eye and the brain. AAV8 and AAV5 transduce hepatocytes efficiently in many species. AAV9 transduces the heart, muscle, and the central nervous system after intravenous delivery, and it crosses the blood-brain barrier in neonates and in some animal models, to a lower degree in adults. Capsids are also engineered by rational design, by directed evolution (for example by the selection of random peptide insertions in the capsid, as for the AAV-PHP.B family, which crosses the blood-brain barrier efficiently in some strains of mice but not in other mouse strains or in primates), and by machine-learning approaches.
• Transcriptional and post-transcriptional targeting. Tissue-specific promoters and enhancers, and microRNA target sites that suppress expression in unwanted cells (for example in the liver or in dorsal root ganglia), reduce off-target expression.

Gammaretroviral vectors

Gammaretroviral vectors, from murine leukemia virus, integrate their DNA into the host genome and so give stable expression that is inherited by the daughter cells. They transduce only dividing cells, because the preintegration complex needs the breakdown of the nuclear envelope in mitosis to reach the chromosomes. They prefer to integrate near transcription start sites and enhancers, which raises the risk of insertional mutagenesis. Gammaretroviral vectors have been used in many clinical trials and in approved ex vivo therapies, such as some CAR-T cell products and some therapies for SCID.

Lentiviral vectors

Lentiviruses are a genus in the subfamily Orthoretrovirinae of the family Retroviridae. Their genomes are two copies of single-stranded, positive-sense RNA. The vectors are derived mainly from HIV-1. The mechanism of transduction is that, after entry and uncoating, reverse transcriptase makes a double-stranded DNA copy of the vector RNA, as the core moves through the cytoplasm. The DNA is part of a complex with viral proteins, and it enters the nucleus through the nuclear pore, which allows transduction of non-dividing cells, such as neurons, hepatocytes, and macrophages. The integrase then inserts the DNA into the host genome. Lentiviral integration favors the bodies of actively transcribed genes, and this pattern, in contrast to the preference of gammaretroviruses for promoters and enhancers, is thought to carry a lower risk of oncogenic activation, but not zero risk.

Lentiviral vectors are used for stable genetic modification of cells ex vivo, in particular hematopoietic stem cells (for example in therapies for beta thalassemia, cerebral adrenoleukodystrophy, metachromatic leukodystrophy, and Wiskott-Aldrich syndrome) and T cells (for the expression of chimeric antigen receptors in CAR-T cell products). It is not accurate to say that they are limited to "rapidly dividing" cells, and they are also used in vivo in some research and in trials. The transfer of the genetic information to the daughter cells during mitosis is the reason for their use in dividing cell populations, such as stem cells. Cases of leukemia and myelodysplastic syndrome have been reported after some lentiviral therapies, and they have required long-term follow-up.

Safety design

The development of lentiviral vector systems has progressed from first generation (which kept most accessory genes of HIV-1) to second generation (accessory genes vif, vpr, vpu, and nef removed) and third generation systems. In third generation systems, the tat gene is removed (the vector does not depend on Tat for transcription), the rev gene is on a separate plasmid, and the envelope is on another plasmid, so that four plasmids are used. Self-inactivating (SIN) vectors have a deletion in the U3 region of the 3′ LTR, which is copied to the 5′ LTR during reverse transcription, so that the viral promoter and enhancer are inactive in the integrated provirus, and the transgene is expressed from an internal promoter. This reduces the risk of activation of neighboring genes, and it is a key factor in their safety, although it does not eliminate the risk, since the internal promoter, and the integration pattern, still matter. Insulators, and the choice of cellular promoters, are also used.

Adeno-associated virus (AAV) is a small, non-enveloped virus of the family Parvoviridae, genus Dependoparvovirus, with a single-stranded DNA genome of about 4.7 kb. Wild-type AAV is not known to cause disease in humans. It depends on a helper virus (such as adenovirus or herpesvirus) for replication, and in the absence of a helper, it can establish latency, in which wild-type AAV2 integrates preferentially at a specific site on human chromosome 19 (AAVS1) through the activity of the Rep protein. Vectors lack rep and so do not integrate site-specifically.

Mechanism and persistence

AAV vectors bind to cells by receptors that depend on the serotype (such as AAVR, a common receptor for most serotypes), are taken up by endocytosis, escape from the endosome, and travel to the nucleus, where the capsid uncoats. The single-stranded genome is converted to double-stranded DNA, by second-strand synthesis or by annealing of plus and minus strands, which is a rate-limiting step. Self-complementary AAV (scAAV) vectors carry a genome that is a hairpin of half the size, and they skip this step, so that expression is faster, but they carry only about 2.3 kb. Inside the nucleus, the vector genomes mostly form circular and concatemeric episomes, which persist as nonreplicating chromatin. Most of the AAV vector DNA in cells remains extrachromosomal, and integration occurs at a low frequency (of the order of 0.1 to a few percent of the vector genomes, with a preference for transcribed regions and sites of DNA breaks). The risk of insertional mutagenesis is lower than for integrating vectors, but it is not zero, and hepatocellular carcinoma and AAV integration have been studied in mice and in a few human cases. In dividing cells, the episomes are diluted by cell division, so that AAV is most suitable for tissues with slowly dividing or post-mitotic cells, such as neurons, muscle, the retina, and hepatocytes (though the liver of children grows, which dilutes the vector). The duration of expression in humans has been followed for years, and in some cases for more than a decade, but the durability depends on the tissue and the dose.

Capacity and approaches for large genes

The packaging capacity of AAV is limited to about 4.7 kb of genome including the ITRs, which means a coding sequence of about 4.4 kb with the promoter and regulatory sequences. Slight overpacking can occur but gives truncated, fragmented genomes. For larger genes, such as DMD (dystrophin, with a coding sequence of about 11 kb), the strategies include shortened "micro-" or "mini-" versions that retain function (as in the approved micro-dystrophin therapy for Duchenne muscular dystrophy), and the splitting of the gene between two or three vectors, which recombine, or whose protein products join (trans-splicing, overlapping, or intein-mediated methods). The efficiency of dual vectors is lower, since both vectors have to transduce the same cell.

Immunity and safety

Pre-existing neutralizing antibodies against AAV capsids, which are found in a large fraction of the population (about 30 to 60 percent for AAV2, depending on the population and serotype, and lower for some others) reduce or block transduction after systemic administration, so patients are screened, and many trials exclude those with a high titer. Delivery to immune-privileged sites, such as the eye (subretinal or intravitreal) and the brain, is less affected. After vector administration, antibody and T cell responses against the capsid can occur and can eliminate transduced cells, which is why immunosuppression (such as corticosteroids) is used in many protocols, and the antibodies prevent re-administration. At high systemic doses (of the order of 10¹⁴ vector genomes per kilogram), serious toxicity, including liver failure, thrombotic microangiopathy, myocarditis, and deaths, has been reported in clinical trials and in treated patients, and so dose and immune responses are the focus of current research. Examples of approved AAV therapies include voretigene neparvovec (Luxturna, RPE65-related retinal dystrophy, subretinal), onasemnogene abeparvovec (Zolgensma, spinal muscular atrophy, intravenous), etranacogene dezaparvovec (hemophilia B), valoctocogene roxaparvovec (hemophilia A), and delandistrogene moxeparvovec (Elevidys, Duchenne muscular dystrophy, whose use has been subject to safety reviews and restrictions in some jurisdictions). The retina is a common target for AAV therapy, as the cells are post-mitotic, the space is small, and the eye is an immune-privileged organ. The statement that AAV is used "exclusively" in post-mitotic tissues is not correct. It is used in many tissues, for example the liver, which contains cells that divide slowly. The statement that it presents "near-zero risk" of insertional mutagenesis is not supported, since integration happens at a low frequency.

Adenoviruses are non-enveloped viruses with a linear double-stranded DNA genome of about 36 kb and an icosahedral capsid of about 90 nm. They infect dividing and non-dividing cells, and the genome remains episomal in the nucleus. Vectors have been developed from human serotypes (especially Ad5, and also Ad26 and Ad35) and from chimpanzee and other nonhuman adenoviruses. In first-generation vectors, the E1 (and often E3) genes are deleted, which makes them replication-defective (they are grown in cells that supply E1, such as 293 cells) and allows inserts of about 5 to 8 kb. Later generations delete more of the viral genes (E2 and E4). "Gutless" or helper-dependent vectors lack all viral coding sequences, keeping only the ITRs and the packaging signal, and they can carry up to about 36 kb, but they must be produced with a helper virus.

Adenoviral vectors induce strong innate and adaptive immune responses. The capsid is detected by Toll-like receptors, and by other sensors, and it activates the complement system and cytokine release. After systemic administration of a high dose, the inflammatory response can be fatal, as in the case of Jesse Gelsinger in 1999. In addition, most people have pre-existing immunity to common serotypes such as Ad5, since the viruses cause common colds, and they lack the capacity for long-term expression, since the cells that express the transgene are cleared by T cells. For these reasons, adenoviral vectors are used when transient expression and immunogenicity are advantageous, mainly in vaccines and in oncolytic and some cancer gene therapies. Adenoviral vector vaccines have been used against Ebola virus (Ad26.ZEBOV, with MVA-BN-Filo, and Ad5-EBOV), against SARS-CoV-2 (ChAdOx1 nCoV-19, Ad26.COV2.S, Ad5-nCoV, and Gam-COVID-Vac, which combines Ad26 and Ad5), and against other pathogens. Chimpanzee adenoviruses, such as ChAdOx1, are used to avoid pre-existing human immunity. Rare cases of vaccine-induced immune thrombotic thrombocytopenia (VITT), a clotting disorder, were identified after the use of the ChAdOx1 and Ad26 COVID-19 vaccines, at a rate of a few cases per 100,000 to a million doses, which led to restrictions on their use in some countries. The first adenoviral vector gene therapy for cancer, an adenovirus with TP53 (Gendicine), was approved in China in 2003, and nadofaragene firadenovec, which delivers interferon alfa-2b to the bladder, was approved in the United States in 2022 for bladder cancer.

• Herpes simplex virus (HSV). HSV-1 has a large double-stranded DNA genome of about 152 kb and can carry large inserts (up to about 30 to 50 kb, depending on the design). It infects neurons and establishes latency, so that it has been developed for gene delivery to the nervous system and as an oncolytic virus. Talimogene laherparepvec (T-VEC), an HSV-1 that lacks the genes ICP34.5 and ICP47 and carries the gene for GM-CSF, is approved for melanoma, and beremagene geperpavec (a replication-defective HSV-1 vector that delivers collagen VII for dystrophic epidermolysis bullosa) is approved for topical use.
• Poxviruses (vaccinia and modified vaccinia Ankara, MVA). They replicate in the cytoplasm, have large capacity (over 25 kb), and are used in vaccines and as oncolytic vectors.
• Other vectors. These include vesicular stomatitis virus (the rVSV-ZEBOV vaccine against Ebola virus), Sendai virus, alphavirus replicons, measles virus, and baculovirus (for the production of AAV and for transduction of mammalian cells).

Safety issues include insertional mutagenesis (for integrating vectors), immune responses to the vector and to the transgene, toxicity at high doses, the presence of replication-competent viruses, off-target expression, and the shedding of the vector, which is relevant for the environment and for transmission to others. Regulatory authorities, such as the United States Food and Drug Administration and the European Medicines Agency, require long-term follow-up (up to 15 years for integrating vectors), testing for RCV, and characterization of the product. Manufacturing is complex and costly, which affects the price of the therapies, some of which cost millions of dollars per patient, and the distribution of them. Another problem is the "empty" capsids, which lack the genome and which contribute to the immune response, so that the ratio of full to empty capsids is a quality measure.

What is the difference between lentiviral vectors and AAV vectors in gene therapy?

Lentiviral vectors (derived from HIV-1, capacity ~8–10 kb) reverse-transcribe their RNA genome and permanently integrate into host chromosomes, making them ideal for *ex vivo* modification of dividing hematopoietic stem cells and CAR-T cells. Adeno-associated virus (AAV) vectors (ssDNA, capacity ~4.7 kb) persist primarily as non-integrating nuclear episomes, making them the preferred platform for *in vivo* gene delivery to long-lived, post-mitotic tissues such as neurons, retina, skeletal muscle, and liver.

How does a self-inactivating (SIN) lentiviral vector prevent insertional oncogenesis?

In third-generation self-inactivating (SIN) lentiviral vectors, the viral transcriptional enhancer and promoter elements within the U3 region of the 3' Long Terminal Repeat (LTR) are deleted. During reverse transcription, this deletion is copied to the 5' LTR of the integrated provirus, transcriptionally silencing both viral LTRs so they cannot accidentally activate adjacent host proto-oncogenes (such as *LMO2*).

Why can't systemic AAV gene therapies easily be redosed in the same patient?

Following a single intravenous infusion of an AAV vector, the host adaptive immune system mounts a high-titer neutralizing antibody response against the viral capsid proteins that persists for years. These circulating antibodies immediately opsonize and block any subsequent dose of the same AAV serotype from transducing target tissues, driving research into capsid switching, IgG-cleaving proteases (imlifidase), and transient B-cell depletion.

• Gene therapy
• Adeno-associated virus
• Lentiviral vector in gene therapy
• Retrovirus
• Adenovirus
• Viral vector vaccine
• CAR-T cell therapy